Effects of lipoproteins on plasma viscosity

89

Atherosclerosis, 38 (1981) 89-95 @ Elsevier/North-Holland Scientific Publishers, Ltd.

EFFECTS

OF LIPOPROTEINS

ALAN H. SEPLOWITZ,

ON PLASMA VISCOSITY

SHU CHIEN and FRANK REES SMITH

Arteriosclerosis Research Center and Departments of Medicine and Physiology, Columbia University, College of Physicians and Surgeons, New York, NY (U.S.A.) (Received 9 April, 1980) (Revised, received 25 June, 1980) (Accepted 7 July, 1980)

Summary Patients with hypertriglyceridemia and mixed hyperlipidemia have been found to have mean plasma viscosities significantly higher than controls (P < 0.005). In a group of 70 hyperlipidemic patients and controls, plasma viscosity was correlated with plasma triglyceride concentration (r = 0.56, P< 0.01) and to a lesser extent with the concentration of plasma cholesterol (r = 0.29, P< 0.05). When isolated lipoprotein fractions were added to lipoprotein-free plasma in increasing concentration over a physiological range, a highly significant linear relationship between plasma viscosity and chylomicron concentrations (r = 0.98, P< 0.001) was apparent. Furthermore, when chylomicrons were removed by ultracentrifugation, viscosity returned to baseline levels. Added VLDL produced a lesser effect (r = 0.70, P< 0.001) and added LDL, over the range of cholesterol concentration studied, had no influence on viscosity. These studies indicate that chylomicrons in particular can increase plasma viscosity. Viscosity increases of the magnitude demonstrated may in turn alter blood flow and thus contribute to symptoms such as intermittent claudication. Chylomicron-induced increases in plasma viscosity and subsequent decreases in local pancreatic blood flow may be one of the factors involved in the known relationship between severe chylomicronemia and acute pancreatitis. Key words:

Chylomicrons - Hypercholesterolemia - Hyperlipidemia eridemia -Lipoproteins - Triglycerides - Viscosity

- Hypertriglyc-

This research was sumorted in part by arants HL-21006 (SCOR) and HL-19464 from the National Heart. Lung and Blood Institute and ENG 75-19243 from the National Science Foundation. Address reprint requests to: Shu Chien. M.D., Ph.D., Columbia University, College of Physicians and Surgeons, 630 West 168th Street, New York, NY 10032. U.S.A.

90

Introduction Increased concentrations of plasma cholesterol and triglycerides are associated with an increased risk of atherosclerotic cardiovascular disease [l-7]. Hyperlipidemia may also alter peripheral blood flow; patients with type III hyperlipoproteinemia, in particular, have impaired peripheral blood flow which improves after the hyperlipidemia is controlled [ 81. Additionally, microcirculatory changes in the eyes have been linked to hyperlipidemia [9,10]. Finally, hypertriglyceridemia, long known to be clinically associated with pancreatitis, induces pancreatitis in the isolated dog pancreas [ 111. Increased blood viscosity may be one mechanism whereby hyperlipidemia may mediate some of these effects. Elevated blood viscosity has been found in hypertriglyceridemic patients [ 121, and elevated plasma viscosity has been described in patients with hypertriglyceridemia and/or hypercholesterolemia [ 131. Ingestion of a fatty meal may raise both whole blood [ 14,153 and plasma [ 141 viscosity. McMillan found increased serum viscosities in patients with diabetes, especially those with microanglopathy, but did not report lipid levels [16]. Dormandy et al. reported that intermittent claudication is inversely correlated with initial blood viscosity, but not with initial serum lipid concentrations [ 171. In the studies reported here we confirm the presence of increased viscosity in untreated hypertriglyceridemic patients. Using isolated lipoprotein fractions we further demonstrate that chylomicrons in increasing amounts produce a more marked linear increase in plasma viscosity than do either VLDL or LDL over a range of lipid concentrations likely to be encountered in clinical practice. Materials and Methods Forty-five patients with hyperlipidemia and 25 control subjects (see Table 1) kere studied after written informed consent was obtained. Of the 70 volunteers, 49 were male and 21 were female. No subject was taking lipid-lowering drugs at the time of study and none of the females was using oral contraceptives or homone replacement therapy. Diagnoses of hyperlipidemia were based on an arbitrary working definition of levels of cholesterol (at the time of the study) in excess of 275 mg/dl and or levels of triglyceride in excess of 200 mg/ dl in plasma obtained after a 14 h fast. Patients were classified as having hypercholesterolemia, hypertriglyceridemia or both (mixed hyperlipidemia). In terms of the Fredrickson classification of the hyperlipoproteinemias, all patients in the “hypercholesteremia” group had the IIa phenotype. The 19 patients in the “hypertriglyceridemia” group had phenotype IV at least some of the time; some of these patients sometimes manifested phenotype V and others sometimes showed type IIb. The 11 patients with “mixed hyperlipidemla” consistently fell into the type V category. Venous blood was obtained from subjects after 14 h of fasting and 24 h of abstinence from alcohol. Cholesterol and triglycerides were measured in plasma with a Technicon AutoAnalyzer I (Technicon Instruments Corporation, Tarrytown, NY) using the method N-24a for cholesterol [18] and a modification [ 191 of the Kessler and Lederer technique [ 201 for triglycerides. Specific lipo-

91

proteins were separated and isolated from plasma by means of a Beckman LZ65 ultracentrifuge (Beckman Instruments, Inc., Palo Alto, CA). After 30 min of centrifugation at 25,000 rpm in a swinging bucket rotor (Beckman SW 25.1), chylomicrons were separated with a spatula and resuspended in isotonic saline. VLDL and LDL were isolated from chylomicron-free plasma by ultracentrifugation in a Beckman 40.3 rotor according to the method of Have1 et al. [ 211. VLDL were taken as the fraction of density less than 1.006 g/ml from nonchylomicronemic plasma. LDL were taken as the fraction with density 1.006 to 1.063 g/ml. Lipoprotein-free plasma for baseline measurements of viscosity was obtained by adjusting plasma to a density of 1.21 g/ml with potassium bromide (KBr), followed by ultracentrifugation for 48 h at 45,000 rpm in a Beckman 50 Ti rotor. In all isolated fractions, KBr was removed by means of dialysis for 24 h at 4°C against 0.05 M sodium phosphate buffer, pH 7.6. In some experiments, human albumin 9.0 g/d1 in 0.9% NaCl, was used in lieu of lipoproteinfree plasma. In all experiments, chylomicrons were obtained from the blood of fasting chylomicronemic subjects. Plasma viscosity was measured in a modified GDM coaxial cylinder viscometer (22) at 37’C as previously described (23). In this system the shear rate is proportional to the rotational speed; viscosity was calculated from the ratio of the torque to the rotational speed by means of a geometric conversion factor [ 231. Viscometric measurements were made at shear rates of 0.5, 5, and 52 set-‘; plasma viscosity showed no shear rate dependence so that values at those shear rates were averaged to give the plasma viscosity [ 241. Protein concentration was measured by the method of Lowry et al. [25]. Analysis of the difference between mean values was performed with a one-way analysis of variance; linear regression, correlation and multiple co-variance analysis were performed on a Wang programming calculator using equations [26] adapted by Dr. J.E. Smith. Results When plasma viscosities in 70 hyperlipidemic and control subjects were analyzed as a function of plasma lipid concentration, the correlation with triglycerides (r = 0.56, P < 0.01) or log (triglycerides) (r = 0.59, P< 0.01) was stronger than that with cholesterol (r = 0.29, P< 0.05). Multivariate analysis indicated that the cholesterol effect was independent but accounted for only 4% of the observed variation in viscosity. As shown in Table 1, the mean plasma viscosity

TABLE 1 CHARACTERISTICS Group

Control Hypercholesterolemia Hypertriglyceridemia Mixed hyperlipidemia

OF CONTROL Number

25 15 19 11

SUBJECTS AND HYPERLIPIDEMIC

PATIENTS

Age a

Cholesterol

Triglycerides

Viscocity

(years)

(ml3/dl)

(muw

(eP)

41 f 48 f 50 f 51*

205 371 229 341

127 f 50 124i 08 458 f 294 847 f 696

1.3aO 1.380 1.420 1.500

10 16 11 08

a Meanf SD. b Probability that value differs from control, P < 0.005.

f f i f

37 75 26 64

f f * f

0.08 0.10 0.08 b 0.16 b

92

b E 6 z

1.4 1.2

t I.0 I .g 0.8 x 0.6 ' ._ = 0.4 c 2 0.2 5 w I ki Y -

0 TRIGLYCERIDE CONCENTRATION (mg /dl x IO-‘)

I

TRIGLYCERIDE OR CHOLESTEROL CONCENTRATION (mg/dl x IO-‘) Fig. 1. Effects of chylomicrons (upper panel), VLDL and LDL (lower panel) on lipoprotein-free plasma viscosity. In the upper panel the equation for the chylomicron vs. viscosity regression line is Y = 0.025+ 0.00011x. r = 0.98, P < 0.001. Dark triangles connected with light triangles indicate the plasma viscosities after removal of chylomicrons by ultracentrifugation. In the lower panel, the equation for the regression of VLDL (squares) on viscosity is y = 0.014 + 0.000049x. r = 0.10. P < 0.001 and is plotted as a broken line. For LDL (circles), the slope of the regression line is not significantly different from zero.

in the 15 hypercholesterolemic subjects did not differ significantly from that of the controls. In contrast, the mean viscosity in the 19 patients with hypertriglyceridemia (1.42 cP) and in the 11 patients with mixed hyperlipidemia (1.50 cP) differed significantly (P < 0.005) from the mean control values (1.33 cP). The effect of adding increasing amounts of isolated lipoproteins to lipoprotein-free plasma is shown in Fig. 1. Chylomicrons added over a very wide range, to a maximum concentration of 12,000 mg of triglyceride per dl of plasma, produced a linear increase in plasma viscosity (r = 0.98, P< 0.0001). With chylomicron triglyceride concentration limited to less than 1,500 mg/dl (16 data points), the correlation with plasma viscosity was still highly significant (r = 0.72, P < 0.001). This latter, limited range of chylomicron concentration was comparable to the range of VLDL which we investigated (o-1337 mg/dl; 20 data points). Although VLDL also showed a correlation with viscosity (r = 0.70, P-C O.OOl), the effect was less pronounced, with the slope of the regression line being much flatter. The difference in slopes between the chylomicron and VLDL data over a comparable triglyceride range was highly significant by covariance analysis (P < 0.005). As indicated in Fig. 1, for samples in which the added chylomicrons were removed by ultracentrifugation, plasma viscosity fell to baseline levels. LDL were also added to lipoprotein-free plasma and viscosity was determined over a cholesterol concentration of O-653 mg/dl. The slope of the regression equation was not significantly different from zero.

93

Discussion Rheological properties of blood may constitute one of the parameters involved in the regulation of blood flow. The viscosity of blood at a given temperature is determined by plasma viscosity, red cell concentration and deformability, and the interaction between plasma and cells [ 271. Plasma viscosity has been correlated with plasma protein concentrations, especially fibrinogen, but there is less information on its relation to plasma lipid composition. Marked viscosity-related disturbances of blood and plasma have been observed in the hyperviscosity syndromes where elevated plasma concentrations of abnormal plasma proteins result in plasma or serum viscosity values as high as 10 CP with attendant manifestations of flow disturbances [ 281. Alterations in plasma viscosity of lesser magnitude have been measured in patients in the early stages of myocardial infarction where values of 1.46 + 0.04 CP were attained on the 4th day [ 291. This post-infarction increase in plasma viscosity has been correlated with elevations in plasma fibrinogen and cY,-globulin concentrations. In the post-infarction period, there are fluctuations of not only fibrinogen concentration and split products of fibrinogen [ 301, but also of free fatty acid and lipoprotein concentrations [ 311. It is not known to what extent these lipid changes contribute to the post-infarction elevation in plasma viscosity. Studies on stable, untreated hyperlipidemic patients provide a situation in which the relationships between serum lipids and viscosity may be examined. Fifteen years ago, investigators explored the effect of a fatty meal on rheological change in volunteers [14] and in 2 hyperlipidemic patients [ 151. While transient increases in whole blood viscosity were seen, no alterations in plasma viscosity were demonstrated. More recent studies by Bartoli et al. [12] suggest that hypertriglyceridemic patients have higher whole blood viscosities than do control subjects. Leonhardt et al. [ 131 studied 39 patients with hyperlipidemia and found mean plasma viscosity values increased over controls (mean 1.26 cP), with patients with mixed hyperlipidemia having the highest value (mean 1.38 cP). Plasma viscosities in patients with hypertriglyceridemia (mean 1.34 cP) and in those with hypercholesterolemia (mean 1.31 cP) were also higher than the mean plasma viscosity of control subjects [ 131. The present data support and extend the conclusions of Leonhardt et al. [13]. Our patients with mixed hyperlipidemia and those with hypertriglyceridemia had comparable elevations of mean plasma viscosity over controls. Our studies with isolated lipoproteins indicate that triglyceride-rich particles may significantly influence plasma viscosity. With such an artificially controlled system, we were able to provide data that complemented our observations on patients with diverse types of hyperlipoproteinemia. These studies suggest that chylomicrons, and to a lesser degree VLDL, by virtue of particle size and/or interparticle interaction, have effects on plasma viscosity and hence on the rheological properties of whole blood. In the microcirculation, where the hematocrit is lower than that in the venous blood [32], the role of plasma viscosity as a determinant of blood viscosity is enhanced. Therefore, the observed increase in plasma viscosity in the presence of chylomicronemia may have a greater influence in curtailing blood flow through the

94

microcirculations, such as those in the region of the small intestine or pancreas. In earlier investigations, Swank observed visceral hyperemia in hamsters with elevated blood viscosity following high fat feedings [ 331. The same author also showed that significant slowing of the circulation may occur in fat-fed rabbits even in the absence of an increase in blood viscosity [ 341. More recently, Saharia et al. showed that chylomicron infusions in the isolated perfused dog pancreas produce pancreatitis [ 111. These studies and ours suggest the following hypothesis: Very high concentrations of triglyceride-rich chylomicrons in the regional pancreatic circulation result in hyperviscosity, decreased capillary flow and even stasis. These flow derangements and the attendant metabolic changes may result in activation and release of pancreatic lipase, which then acts on the chylomicrons to release large quantities of free fatty acids. The released free fatty acids may act locally to initiate inflammatory changes and thus pancreatitis. In summary, our studies indicate that triglyceride-rich particles, and chylomicrons in particular, can increase plasma viscosity. Additional ‘effects on whole blood viscosity, not examined in this study, probably also occur and provide an area for possible future investigation. These viscosity changes may have pathophysiological significance with regard to blood flow abnormalities in patients with intermittent claudication, coronary heart disease or pancreatitis. Acknowledgements The authors thank Ms. Dagmara Igals, Ms. Minnie Myers and Ms. Maria PazAlvir for excellent technical assistance. The authors thank Dr. Dewitt S. Goodman for helpful criticism and support and Dr. John Edgar Smith for the statistical programs used for univariate analysis. We thank Drs. Ralph Dell, R. Ramakrishnan and R. Sciacca for performing the multivariate analysis of our data. References Kannel, W.B., Caste& W.P., Gordon, T. and McNamsra, P.M., Serum cholesterol, lipoproteins and the risk of coronary heart disease -The Framingham Study, Ann. ht. Med., ‘74 (1971) 1. Carlson, L.A. and Bottinger. L.E., Ischemic heart disease in relation to fasting values of plasma triglycerides and Cholesterol-Stockholm Prospective Study. Lancet. i (1972) 865. Castelli, W.P.. Doyle, J.T.. Gordon, T., Hames, C.G.. Hjortland. M.C.. Hulley. S.B., Kagan. A. and Zukel. W.J.. HDL cholesterol and other lipids in coronary heart disease -The Cooperative LiPoProtein Phenotyping Study, Circulation, 55 (1977) 767. Goldstein, J.L., Hazzard, W.R.. Schrott, H.G.. Bierman. E.L. and Motulsky. A.G.. Hyperlipidemia in coronary heart disease, Part 1 (Lipid levels in 500 survivors of myocsrdial infarction). J. Chn. Invest., 52 (1973) 1633. Mabuchi, H., Haba. T.. Ueda. K., Ueda. R.. Tatami, R., Ito. S., Kametani, T., Koisumi, J.. Miyamoto. S., Ohta. M., Takeda, R., Takegosbi. T. and Takeshita. H., Serum lipids and CoronarY heart disease in heterozygous familial hypercholesterolemia in the Hokuriku district of Japan, Atherosclerosis, 28 (1977) 417. Ross, R. and Harker. maintains lesions by

L., Hyperlipidemia and atherosclerosis - Chronic hyperlipidemia endothelial cell desquamation and lipid accumulation, Science,

initiates and 193 (1976)

1094. Vogelberg, K.H.. Berchtold. P., Berger. H., Gries. F.A., Klinger, H.. Kubler. W. and Stolze, T.. Primary hyperlipoproteinemias as risk factors in peripheral artery disease documented by arteriopraphy, Atherosclerosis, 22 (1975) 271. Zelis, R., Mason, D.T., Braunwald. E. and Levy. R.I.. Effects of hyperlipoproteinemias and their treatment on the peripheral circulation. J. CIin. Invest., 49 (1970) 1007. Friedman. M., Rosenman. R.H. and Byers, S.. Serum lipids and conjunctival CircUlation after fat ingestion in men exhibiting type A behavior pattern, Circulation, 29 (1964) 874.

95

10 11

Charan,

B.I.,

Lipemia

retinalis -

Ferguson,

Saharia,

P., Margolis.

13 14 15 16 17

18

19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

Cast&i,

S.. Zuidema.

Studies with an isolated 12

B.D.,

microcirculatory perfused

W.P.,

Touborg.

J.N.F.,

Balodimos,

changes and lipid studies in a family. G.D.

and Cameron,

canine pancreas,

J.L., Acute

Surgery,

M.C. and Rutstein, Metabolism,

pancreatitis

82 (1977)

D.D.,

18 (1969)

with hyperliaidemia

978. -

60.

Bsrtoli, V., Pasquini. G. and Dorigo. B.. Viscosita ematica e triglyceridemia, Minerva Medica, 68 (1977) 3081. Leonhardt, H., Arntz, H.R. and Klemens, U.H.. Studies of plasma viscosity in primary hyperli?oprw teinema, Atherosclerosis, 28 (1977) 29. Charm, S., McComis. W.. Tejada. C. and Kurland, G., Effect of a fatty meal on whole blood a?d olasma viscosity, J. Appl. Physiol., 19 (1963) 1217. MerrilI. E.W., Gilliland, E.R.. Margetts, W.G. and Hatch, F.T., Rheology of human blood and hyperlipidemia, J. Appl. Physiol., 19 (1964) 493. McMlllan, D.W., Disturbance of serum viscosity in diabetes mellitus, J. Clin. Invest., 53 (1974) 1071. Dormandy, J.A.. Hoare, E., Khattab, A.H., Arrowsmith. D.E. and Dormandy, T.L., Prognostic sisnificance of rheological and biochemical findings in patients with intermittent claudication. Rrit. Med. J.. 4 (1973) 581. Block. W.D., Jarret, K.J. and Levine, J.B., Use of a single color reagent to improve the automated determination of serum total cholesterol. In: L.T. Skeggs, Jr., (Ed). Automation in Analytical Chemistry, Mediad. New York, NY, 1965. pp. 345-347. Noble, R.P. and Campbell. F.M., Improved accuracy in automated fluorometric determination of olasma triglycerides, Clin. Chem.. 16 (1970) 166. Kessler, G. and Lederer, H., Fluorometric measurements of triglycerides. In: L.T. Skegqs, Jr. (Ed.), Automation in Analytical Chemistry, Mediad, New York, NY, 1965. PP. 341-344. Havel. R.J., Eder, H.A. and Bragdon, J.H.. The distribution and chemical composition of ultvacentrifugally separated lipoproteins in human serum, J. Clin. Invest., 34 (1955) 1345. Gilinson, P.J., Dauwalter, C.R. and Merrill, E.W., A rotational viscometer using an AC torque to balance loop and air bearing, Trans. Sot. Rheol., 7 (1963) 319. Chien, S.. Dellenback. R.J. and Bryant. C.A., Comparative hemorheology -Implications of species differences in biood viscosity. Biorheology. 9 (1971) 35. Chien. S., Usami, S., Dellenback, R.J. and Gregersen, MI., Shear-dependent interaction of plasma proteins with erythrocytes in blood rheology. Amer. J. Physiol., 219 (1970) 143. Lowry, O.H., Rosebrough, N.J.. Fan-, A.L. and Randall R.J., Protein measurement with the Folin phenol reagent. J. Biol. Chem., 193 (1951) 265. Snedecor. G.W. and Cochran, W.G.. Statistical Methods. 6th edition, Iowa State University Press, Ames, IA, 1967. Chien, S., Biophysical behavior of red cells in suspensions. In: D.M. Surgenor (Ed.), The Red Blood Cell, Vol. 2. Academic Press, New York, NY, 1975, pp. 1031-1133. Somer. T., The viscosity of blood. plasma and serum in dys- and para-proteinemias, Acta Med. &and., 180, Suppl. 458 (1966) 1. Jan, K.M., Bigger, Jr., J.T. and Chien, S., Observations on blood viscosity changes after acute myocardial infarction, Circulation, 51 (1975) 1079. Almer. L.O., Hedner, U. and Nilsson. I.M., Fibrin degradation products in the early differential diagnosis of acute myocardial infarction, Thromb. Res., 1 (1972) 59. Kirkeby, K.. Disturbances in serum lipids and in their fatty acid composition following acute mymardid infarction, Acta Med. Scand., 192 (1972) 523. Johnson, P.C., Red cell separation in the mesenteric capillary network, Amer. J. Physiol., 221 (1971) 99. Swank, R.L., Effect of high fat feedings on viscosity of the blood, Science. 120 (1954) 427. Swank, R.L., Changes in blood of dogs and rabbits by high fat intake, Amer. J. Physiol.. 196 (1959) 473.