- Email: [email protected]
Retinyl ester retention in chronic renal failure Further evidence for a defect in chylomicron remnant metabolism
Atherosclerosis,
189
57 (1985) 189-197
Elsevier
ATH 03666
Retinyl Ester Retention in Chronic Renal Failure Further Evidence for a Defect in Chylomicron Remnant Metabolism Dana E. Wilson *, Ing-Fong Chan, Alfred K. Cheung, Wolfgang and Kenneth N. Buchi
Dutz **
Veterans Administration Hospital (11 I E), 500 Foothill Drive, Salt Lake City, UT 84148 and Department Internal Medicine, Divisions of Endocrinology and Metabolism, Gastroenterology, and Nephrology, The University of Utah School of Medicine, UT (U.S.A.)
of
(Received 27 November, 1984) (Revised, received 8 March, 1985) (Accepted 11 March, 1985)
Summary
The metabolic remnants of triglyceride-rich lipoproteins are atherogenic in man and experimental animals. Particles resembling lipoprotein remnants have been found in plasma from patients with chronic renal failure (CRF). In this study we took advantage of the observation that retinyl esters are transported only by lipoproteins that originate in the intestine, that is, by chylomicrons (CM) and their remnants. To investigate further remnant metabolism in CRF, plasma RE were measured by reverse-phase high performance liquid chromatography in 20 non-diabetic hemodialyzed patients with CRF and 20 hospitalized non-diabetic control subjects 12-15 h after the administration of retinyl ester, 25 000 IU orally. Total plasma RE were increased 3-fold in the CRF patients (P < 0.001). Quantitative analysis of retinoids and lipids in fractions separated by unit-gravity flotation and flocculation in 3% polyvinylpyrrolidone indicated that the plasma RE were not
Supported in part by NIH Grant HL23535 and the Veterans Administration. A grant-in-aid from the Nora Eccles Treadwell Foundation supported non-clinical portions of the work. * To whom reprint requests should be addressed. ** Medizinische Klinik, Ernst-Moritz-Amdt-Universitat, 22 Greifswald, G.D.R. Abbreviations: CM = chylomicrons, CMr = chylomicron remnants, CRF = chronic renal failure, HDL = high density lipoproteins, IDL = intermediate density lipoproteins, LDL = low density lipoproteins. RE = retinyl esters, TG = triglycerides, VLDL = very low density lipoproteins.
0021-9150/85/$03.30
0 1985 Elsevier Scientific
Publishers
Ireland,
Ltd
190
contained among intact CM. Mean plasma retinol in CRF was also elevated consistent with previous observations and the known role of the kidney in retinolbinding protein metabolism. Although postabsorptive RE concentration was correlated positively and significantly with plasma triglyceride concentration in both groups, RE were higher in CRF patients at comparable plasma triglyceride concentrations. These data support the proposal that atherogenic lipoprotein remnants accumulate in the plasma of patients with CRF. Key words: Atherosclerosis - Chylomicron remnants - Chylomicrons - Hemodialysis - Lipoproteins - Renal failure - Retinoids - Retinol - Retinyl esters
Introduction
Patients with chronic renal failure (CRF) are predisposed to premature atherosclerosis [l]. As a group they have mild hypertriglyceridemia, increased concentrations of triglyceride-rich lipoproteins and decreased concentrations of high density lipoprotein (HDL). However, the progression of atherosclerosis in CRF may be accelerated disproportionately relative to the modest lipid abnormalities that have been described. This discrepancy could be explained by additive or synergistic effects of known risk factors in CRF or by the presence of atherogenic lipoprotein subpopulations. Lipoprotein remnants, cholesterol-enriched catabolic products of chylomicrons (CM) and very low density lipoproteins (VLDL), may represent such a subpopulation. They are atherogenic in patients with Type III hyperlipoproteinemia and remnant-like lipoproteins are found in cholesterol-fed animals that develop florid atherosclerosis [2]. Moreover, it has been proposed that remnant particles arising after high-fat, high-cholesterol meals contribute to atherogenesis in humans without inherited lipid transport abnormalities [3]. Several reports now indicate that particles with compositional properties of lipoprotein remnants are present in the VLDL and intermediate density lipoproteins (IDL) of patients with CRF [4-S]. We investigated further the hypothesis that CM remnants (CMr) accumulate in CRF using fed retinyl esters (RE) as markers of lipoproteins that originate in the intestine [9]. The results of this study show that there is disproportionate retention of RE in CRF providing additional, independent evidence for increased plasma concentrations of CM remnants. Materials and Methods
Subjects Twenty non-diabetic patients (10 males and 10 females, mean age 53.1 years, range 20-69 years) with chronic renal failure undergoing maintenance hemodialysis 3 times weekly were selected at random for study. Renal failure was due to: amyloidosis (1); glomerulonephritis (8); interstitial nephritis (2); medullary cystic
191
disease (1); nephrosclerosis (1); oxalate nephropathy (2); polycystic kidney disease (2); chronic pyelonephritis (1); lupus nephritis (1) and; unknown (1). Two-thirds of the patients had been dialyzed with bicarbonate-containing solutions and one-third with acetate. Average predialysis serum urea nitrogen, creatinine and albumin concentrations were 73 mg/dl, 10.9 mg/dl and 3.7 g/dl, respectively. Two patients were receiving beta-adrenergic blocking agents, 4 androgenic/ anabolic steroids. and 1 mysoline and phenytoin. None was receiving glucocorticoid therapy. Twenty hospitalized non-diabetic patients whose plasma triglyceride (TG) concentrations were less than 318 mg/dl were selected as controls (19 males, 1 female, mean age 55.5 years, range 32-73 years). They had a variety of medical illnesses including premature coronary heart disease. Several of them were receiving drugs known to affect plasma triglycerides including thiazide diuretics (5) beta-adrenergic blocking agents (3) and prednisone, 60 mg daily, for pulmonary disease (1). None of the 20 had serum urea nitrogen concentrations > 20 mg/dl or creatinine values > 1.5 mg/dl. Retinyl ester challenge
Each subject was fed a capsule containing 25 000 IU of retinyl ester along with supper on the evening preceding dialysis. Venous blood samples for lipid and retinoid analyses were obtained the following morning after a 12-15 h fast, 2-3 days following the last dialysis and prior to that day’s dialysis session. At the time of blood sampling, patients had not received any heparin for 4468 h. Hospitalized controls were studied under identical conditions. Blood was collected on ice in tubes containing disodium EDTA, 1 mg/ml, final concentration. Plasma samples were kept refrigerated at 4°C in metal-foil-covered tubes and handled in subdued light. Chylomicron
flotation
and flocculation
The presence of intact CM was excluded by quantitative analysis of retinoids and lipids using 2 different methods. When whole plasma is allowed to stand in a test tube overnight at 4°C CM float to the top (unit-gravity flotation-or the “icebox test”). Accordingly, unit-gravity flotation was carried out and the presence or absence of floating CM was noted. The upper l/5 of the sample was removed by gentle aspiration (the unit-gravity “top”). This fraction and the bottom 4/5 (“bottom”) were then analyzed for cholesterol, triglyceride, retinol and retinyl ester concentrations. Recovery values were calculated as: [(top + bottom)/total] x 100. Recoveries for these 4 components averaged 99.2, 98.8, 100.2 and 98.7%, respectively. The 2nd method, polyvinylpyrrolidone flocculation, was performed as described by O’Hara and others [lo]. A l-ml portion of plasma was layered under 10 ml of 3% polyvinylpyrrolidone (PVP) and allowed to stand for 18 h at 37°C. This technique separates lymph CM (“primary particles”) from “secondary particles” of less certain derivation. The tubes were examined for turbidity and the top fractions (containing primary particles) and the bottom fractions (containing secondary particles and lipoproteins with density > 1.006 g/ml) were removed for lipid and retinoid analyses. Retinyl ester recoveries for the 11 CRF patients with total plasma retinyl
192
ester concentrations of 15 pg/dl or more were 70.6 percent. Cholesterol, triglyceride, and retinol recoveries (calculated as described above) were 88.2, 94.5 and 98.5%, respectively. Lipoprotein retinoid analysis Plasma retinol and retinyl ester concentrations were measured on extracts of unfrozen plasma by high performance liquid chromatography within a few days of collection [11,12]. The coefficient of variation for replicate analyses of retinol and retinyl ester standards was less than 5%. The variability of post-challenge plasma RE concentrations was examined by replicate clinical studies in 3 subjects. The 1st subject, a woman with severe Type V hyperlipoproteinemia (TG > 2000 mg/dl, increased pre-beta lipoproteins and chylomicronemia by lipoprotein electrophoresis) had normal renal function. Replicate challenges on 7 separate occasions showed a mean value of 209 pg/dl (SD 39.8, range 161-286 pg/dl, n = 7). Patient #2, a CRF patient whose initial RE value was 15 pg/dl, had a mean concentration of 11.1 pg/dl (SD 4.3, range 5.9-17.0 pg/dl, n = 6). Patient #3 was a CRF patient who had no measurable retinyl esters on initial examination. With repeated study on 6 occasions, mean retinyl ester concentration was 0.5 pg/dl (SD 1.3, range O-3.2 pg/dl). Thus, between-day reproducibility was satisfactory. Other methods Triglyceride and cholesterol were measured in coupled enzymatic assays using microbial lipase or cholesterol esterase/cholesterol oxidase, respectively (Beckman Instruments, Carlsbad, CA). Agarose lipoprotein electrophoresis was performed on agarose plates (Corning, Palo Alto, CA). Electropherograms were first interpreted blindly without reference to plasma lipid values. A phenotype was assigned when an abnormal pattern was noted and fasting plasma cholesterol and/or triglyceride values exceed the age- and sex-specific upper decile cutpoints for Visit 2 in the Lipid Research Clinics database [13]. Data analysis To minimize potentially confounding effects of inherited hyperlipidemias, only patients whose plasma triglyceride concentrations (TG) fell below an arbitrarilyselected value, 318 mg/dl (the upper 5th percentile cutpoint for males, aged 40-44 [13]), were compared with control subjects. One patient, a 61-year-old woman, had suffered from polycystic kidney disease and hypertension. Her plasma cholesterol ranged from 223-251 mg/dl and triglycerides 437-498 mg/dl. Twelve hours after oral retinyl ester her total retinyl ester concentration was 74.4 pg/dl. On agarose lipoprotein electrophoresis she proved to have beta-migrating VLDL, a characteristic finding in dysbetalipoproteinemia [2]. For this reason, her data were excluded from the final analysis. Post-challenge retinoid data were skewed upward and so statistical analyses were carried out with non-parametric methods [14]. The Mann-Whitney U-test was used to determine significance between groups, Spearman’s rank correlation coefficient to test the significance of associations between paired variables and Fisher’s Exact Test was used for nominal data.
193
Results
Fasting lipids and lipoprotein electrophoresis Mean fasting plasma triglyceride concentrations were 158 mg/dl (13.6 SEM) in the 20 hospitalized control subjects and 182 mg/dl (14.4 SEM) in the 19 CRF patients (P > 0.2). Mean plasma cholesterol was 197 mg/dl (16.6 SEM) in the control and 168 (7.6 SEM) in the CRF patients (P > 0.2). Thus, the differences in plasma lipid values between CRF patients and hospitalized patients without CRF were not statistically significant. Lipoprotein electropherograms in the controls showed Type IIA in 2 and Type IV hyperlipoproteinemia in 3 patients. Eight of the 19 CRF patients had Type IV hyperlipoproteinemia defined by elevated pre-beta lipoproteins and fasting triglycerides that fell within the age- and sex-specific upper decile [13].
Plasma retinoids Postabsorptive plasma retinol and retinyl ester concentrations in control and CRF patients are shown in Fig. 1. Both distributions were skewed upwards. Plasma retinol concentration in the CRF patients (mean 177 pg/dl, SEM 15.2) was nearly twice that of controls (mean 95.5 pg/dl, SEM 7.7, P < O.OOl), an expected finding given the role of the kidney in retinol-binding protein excretion [15]. Plasma retinyl
3.
A
RETINYL
AETINOL
.
ESTER .
. 60
300. > zoo:.
G \ ?
.
. . : 4 . :
40
r .
.
1001
. 2 .
o.. CONTROL
20
: I
!
.
+-CRF
CONTROL
CRF
0
Fig. 1. Postabsorptive plasma retinal and retinyl ester concentrations in 20 non-azotemic control and 19 CRF patients (pg/dI). Plasma samples were obtained after a 12-15-h fast following supper along with 25000 IU of retinyl ester. Differences between control and CRF patients were significantly different (P < 0.001 for retinol and < 0.01 for retinyl ester) by Mann-Whitney U.
194 TABLE
1
FRACTIONAL DISTRIBUTIONS OF CHOLESTEROL, TRIGLYCERIDES, RETINOL AND RETINYL ESTERS WITH OVERNIGHT UNIT-GRAVITY FLOTATION AND 3% POLYVINYLPYRROLIDONE FLOCCULATION [lo] OF WHOLE PLASMA FROM CRF PATIENTS flotation; The upper one-fifth of the total volume was taken for the “top” fraction in the unit-gravity thus, if no redistribution of CM had taken place, the fractional distribution of a component would be 0.200 in the top and 0.800 in the bottom fraction. Means and standard errors of the means are given. Cholesterol (n = 19)
Triglyceride (n = 19)
Retinol (n =19)
Retinyl ester (n =18)
0.199 (0.001) 0.800 (0.001)
0.195 (0.002) 0.805 (0.002)
0.196 (0.002) 0.804 (0.002)
0.209 (0.004) 0.192 (0.004)
flocculation 0.011 (0.010) 0.989 (0.002)
0.015 (0.028) 0.989 (0.003)
0.009 (0.010) 0.991 (0.002)
0.000
Unit gravity flotation Top Bottom
Polyvinylpyrrolidone Top Bottom
PLASMA
Fig. 2. Best fit linear 20 control subjects correlation) existed intercepts are given
TRIGLYCERIDES
1.000
(mg/dl)
regression plots for retinyl esters (ordinate) versus plasma triglycerides (abscissa) for (0) and 19 CRF patients (0). Significant positive correlations (Spearman’s rank for both the CRF patients (P < 0.05) and the controls (P -C 0.01). Slopes and in the text.
195
esters for the 20 controls were 8.1 pg/dl (SEM 3.5) and for the 19 CRF patients. 23.9 ,ug/dl (SEM 4.2, P < 0.01). The difference between retinyl ester concentrations in controls and CRF patients was significant using cutoff values of either 10 or 20 pg/dl (P < 0.0005) by Fisher’s exact probability [14]. None of the control or CRF plasma samples showed fasting chylomicronemia by inspection. The absence of lipids and retinoids among particles with the properties of unaltered CM was verified after unit-gravity flotation and polyvinylpyrrolidone flocculation of CRF plasma (Table 1). During unit-gravity flotation the distribution of all components closely approximated that expected in the absence of chylomicronemia. All of the measured retinyl ester was accounted for in the lower portion of the tube and none among the “primary particles” during PVP flocculation. Figure 2 shows the relationship between fasting plasma retinyl ester and TG concentrations for the 2 groups. The least squares best fit line for control patients had a slope of 0.125 and a y-intercept of - 11.6 (y, = 0.565, P -c 0.01). The least squares regression for the CRF patients had a slope of 0.121 and an intercept of 1.7 (r, = 0.515, P < 0.05). There was a weak positive correlation between TG and retinol values in the CRF group (slope 0.404, y-intercept 103.7, r, = 0.455, 0.1 > P > 0.05) which was not statistically significant. As expected, correlations between plasma retinol and retinyl ester concentrations were not significant either for the CRF patients (r, = 0.311, P > 0.2) or the controls (r, = 0.230, P > 0.2). Other variables
Data were examined for other differences that might account for the skewed and variable distributions observed in measured retinoid concentrations. There were no significant correlations between serum creatinine and either retinol or retinyl ester concentrations indicating that the severity of renal failure was not an important determinant of retinoid retention in hemodialyzed patients with established CRF. By chance there was a disproportionate number of male patients in the control sample. However, there were no statistically significant differences in age, total cholesterol, triglycerides or retinyl esters between males and females in the CRF group. Retinol levels were higher in CRF males than females (206 pg/dl vs 146 pgg/dl, respectively. 0.05 > P > 0.02 by Mann-Whitney U-test). Retinyl ester and retinol concentrations were significantly greater for either gender in the CRF group than they were in controls. Neither beta-blocking agents, thiazide diuretics, glucocorticoids or anabolic/androgenic steroid therapy could be related to retinoid retention. Thus, there appeared to be unidentified variables, either acquired or genetic, that affected RE disappearance in the subjects. Discussion In this study as well as previous ones [1,6], CRF patients have had only modest hypertriglyceridemia accompanied by normal total plasma cholesterol concentrations. The unimpressive abnormalities in circulating lipids and lipoproteins appear inadequate to account for the increased risk of atherosclerosis in patients with CRF.
196
For this reason, investigators have examined the possibility that there is a subpopulation of potentially atherogenic lipoprotein remnants in the plasma of patients with CRF [1,4,5,8]. The major new finding in the present report was that non-chylomicronemic CRF patients had elevated plasma levels of chylomicron remnants identified by their retinyl ester content. This approach capitalizes on the observation that retinyl esters are transported mainly if not exclusively by CM and CM remnants 191.The validity of the approach is supported by evidence that retinyl esters are conserved among CM and CM remnants [9,11,16] and that retinyl esters are not secreted from the liver with VLDL [16]. Plasma RE concentrations were positively and significantly correlated with plasma TG concentration in both control subjects and CRF patients (Fig. 2). This relationship is expected since CM and endogenous VLDL share a common removal pathway [17] and since endogenous (overproduction) hypertriglyceridemia leads to a secondary decrease in the removal of CM [17] and CM-derived particles. However, in the present study plasma retinyl ester concentration was greater for any given plasma triglyceride concentration in patients with CRF than it was in control subjects. A number of other reported observations are consistent with abnormal persistence of enterogenous lipoprotein remnants in CRF: (1) kinetic studies have demonstrated delayed removal of TG-rich lipoproteins in uremic patients and those undergoing dialysis [18]; (2) the lipoproteins that accumulate in CRF are denser than typical VLDL and relatively enriched in cholesterol, properties of lipoprotein remnants [2,4,5,8]; (3) these lipoproteins contain apolipoprotein B-48 which, in man, appears to be a specific marker of CM and CM catabolic products [5]; (4) the hepatic postheparin plasma triglyceride hydrolase activity is decreased in CRF [6,19,20]; this enzyme plays a role in the latter stages of TG-rich lipoprotein catabolism [21]; and, (5) elevated concentrations of sialylated C apoproteins have been demonstrated in VLDL obtained from patients with CRF [22]. Since the relative concentrations of the C and E apolipoproteins carried by TG-rich lipoproteins affect remnant removal from plasma [23] it is also possible that abnormal apolipoprotein composition in CRF patients contributes to remnant retention. The persistence of fed retinyl esters in the plasma after a 12-15-h fast is additional, independent evidence for abnormal lipoprotein remnant metabolism in CRF. Acknowledgements
We thank Asta Frischat, Tran Thu-Ha, Tran Due-Duy and Martha Wolfe for their valuable technical assistance. We are grateful to Dr. Martin Gregory for helping with some of the clinical studies. References 1 Editorial, Uraemia, lipoproteins and atherosclerosis, Lance& ii (1981) 1151. 2 Brown, MS., Goldstein, J.L. and Fredrickson, D.S., Familial type 3 hyperlipoproteinemia poproteinemia). In: J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein
(dysbetaliand M.S.
197 Brown (Eds.), The Metabolic Ch. 32, p. 655. 3 Zilversmit, D.B., Atherogenesis 4 Ron, D., Oren, I., Aviram, 5 6 7 8
9 10 11 12 13 14 15 16 17 18 19 20
21 22 23
Basis of Inherited -
A postprandial
Disease,
5th edition,
phenomenon,
McGraw-Hill,
Circulation,
M., Better, OS. and Brook, J.G., Accumulation
New York, 1983,
60 (1979) 473. of lipoprotein
remnants
in
patients with chronic renal failure, Atherosclerosis, 46 (1983) 67. Nestel. P.J., Fidge, N.H. and Tan, M.H., Increased lipoprotein-remnant formation in chronic renal failure, N. Engl. J. Med., 307 (1982) 329. Bolanzo, K., Krempler, F. and Sandhofer, F., Hepatic and extrahepatic triglyceride lipase activity in uraemic patients on chronic hemodialysis, Europ. J. Clin. Invest., 8 (1978) 289. Norbeck, H.E. and Carlson, L.A., Increased frequency of late pre-beta-lipoproteins (LPbeta) in isolated serum very low density lipoproteins in uraemia, Europ. J. Clin. Invest., 10 (1980) 423. Minamisono, T., Wada, W., Akamatsu, A., Okabe, M., Handa, Y., Morita, T., Asagami, C.. Naito. H.D., Nakamoto, S., Lewis, L.A. and Mise, J., Dyslipoproteinemia, a remnant lipoprotein disease in uremic patients on hemodialysis, Clin. Chim. Acta, 84 (1978) 163. Goodman, D.S., Vitamin A and retinoids in health and disease, N. Engl. J. Med., 310 (1984) 1023. O’Hara, D.O., Porte, Jr., D. and Williams, R.H.. Use of constant composition polyvinylpyrrolidone columns to study the interaction of fat particles with plasma, J. Lipid Res., 7 (1966) 264. Wilson, D.E. and Chan, L-F., In vitro transfer of chylomicron retinol and retinyl esters, Biochem. Biophys. Res. Comm., 115 (1983) 958. Wilson, D.E., Chan, L-F., Buchi, K.N. and Horton, S.C., Postchallenge plasma lipoprotein retinoids - Chylomicron remnants in endogenous hypertriglyceridemia, Metabolism, 34 (1985) 551. The Lipid Research Clinics Population Studies Data Book, Volume 1 (The Prevalence Study), (DHEW Publication No. (NIH) 79-1527), National Institutes of Health, 1979, p. 58. Siegel, S., Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill, New York, 1958, pp. 104, 116, 202. Smith, F.R. and Goodman, D.S., The effects of diseases of the liver, thyroid, and kidneys on the transport of vitamin A in human plasma, J. Clin. Invest., 50 (1971) 2426. Thompson, K.H., Hughes, L.B. and Zilversmit, D.B., Lack of secretion of retinyl ester by livers of normal and cholesterol-fed rabbits, J. Nutr., 113 (1983) 1995. Grundy, SM. and Mok, H.Y.I., Chylomicron clearance in normal and hyperlipidemic man, Metabolism, 25 (1976) 1225. Heuck, C.-C. and Ritz, E., Hyperlipoproteinemia in renal insufficiency, Nephron, 25 (1980) 1. Mordasini R., Frey, F., Flury, W., Klose, G. and Greten, H., Selective deficiency of hepatic triglyceride lipase in uremic patients, N. Engl. J. Med., 297 (1977) 1362. Applebaum-Bowden, D., Goldberg, A.P., Hazard, W.R., Sherrard, D.J., Brunzell. J.D., Huttunen, J.K., Nikkila, E.A. and Ehnholm, C., Postheparin plasma triglyceride lipases in chronic hemodialysis - Evidence for a role for hepatic lipase in lipoprotein metabolism, Metabolism, 28 (1979) 917. Jackson, R.L., Lipoprotein lipase and hepatic lipase. In: P.D. Boyer (Ed.), The Enzymes, 3rd edition. Academic Press, New York, 1983, p. 141. Holdsworth, G., Stocks, J., Dodson, P. and Galton, D.J., An abnormal triglyceride-rich lipoprotein containing excess sialylated apolipoprotein C-III, J. Clin. Invest., 69 (1982) 932. Windler, E., Chao, Y. and Havel, R.J., Determinants of hepatic uptake of triglyceride-rich lipoproteins and their remnants in the rat, J. Biol. Chem., 255 (1980) 5475.
189
57 (1985) 189-197
Elsevier
ATH 03666
Retinyl Ester Retention in Chronic Renal Failure Further Evidence for a Defect in Chylomicron Remnant Metabolism Dana E. Wilson *, Ing-Fong Chan, Alfred K. Cheung, Wolfgang and Kenneth N. Buchi
Dutz **
Veterans Administration Hospital (11 I E), 500 Foothill Drive, Salt Lake City, UT 84148 and Department Internal Medicine, Divisions of Endocrinology and Metabolism, Gastroenterology, and Nephrology, The University of Utah School of Medicine, UT (U.S.A.)
of
(Received 27 November, 1984) (Revised, received 8 March, 1985) (Accepted 11 March, 1985)
Summary
The metabolic remnants of triglyceride-rich lipoproteins are atherogenic in man and experimental animals. Particles resembling lipoprotein remnants have been found in plasma from patients with chronic renal failure (CRF). In this study we took advantage of the observation that retinyl esters are transported only by lipoproteins that originate in the intestine, that is, by chylomicrons (CM) and their remnants. To investigate further remnant metabolism in CRF, plasma RE were measured by reverse-phase high performance liquid chromatography in 20 non-diabetic hemodialyzed patients with CRF and 20 hospitalized non-diabetic control subjects 12-15 h after the administration of retinyl ester, 25 000 IU orally. Total plasma RE were increased 3-fold in the CRF patients (P < 0.001). Quantitative analysis of retinoids and lipids in fractions separated by unit-gravity flotation and flocculation in 3% polyvinylpyrrolidone indicated that the plasma RE were not
Supported in part by NIH Grant HL23535 and the Veterans Administration. A grant-in-aid from the Nora Eccles Treadwell Foundation supported non-clinical portions of the work. * To whom reprint requests should be addressed. ** Medizinische Klinik, Ernst-Moritz-Amdt-Universitat, 22 Greifswald, G.D.R. Abbreviations: CM = chylomicrons, CMr = chylomicron remnants, CRF = chronic renal failure, HDL = high density lipoproteins, IDL = intermediate density lipoproteins, LDL = low density lipoproteins. RE = retinyl esters, TG = triglycerides, VLDL = very low density lipoproteins.
0021-9150/85/$03.30
0 1985 Elsevier Scientific
Publishers
Ireland,
Ltd
190
contained among intact CM. Mean plasma retinol in CRF was also elevated consistent with previous observations and the known role of the kidney in retinolbinding protein metabolism. Although postabsorptive RE concentration was correlated positively and significantly with plasma triglyceride concentration in both groups, RE were higher in CRF patients at comparable plasma triglyceride concentrations. These data support the proposal that atherogenic lipoprotein remnants accumulate in the plasma of patients with CRF. Key words: Atherosclerosis - Chylomicron remnants - Chylomicrons - Hemodialysis - Lipoproteins - Renal failure - Retinoids - Retinol - Retinyl esters
Introduction
Patients with chronic renal failure (CRF) are predisposed to premature atherosclerosis [l]. As a group they have mild hypertriglyceridemia, increased concentrations of triglyceride-rich lipoproteins and decreased concentrations of high density lipoprotein (HDL). However, the progression of atherosclerosis in CRF may be accelerated disproportionately relative to the modest lipid abnormalities that have been described. This discrepancy could be explained by additive or synergistic effects of known risk factors in CRF or by the presence of atherogenic lipoprotein subpopulations. Lipoprotein remnants, cholesterol-enriched catabolic products of chylomicrons (CM) and very low density lipoproteins (VLDL), may represent such a subpopulation. They are atherogenic in patients with Type III hyperlipoproteinemia and remnant-like lipoproteins are found in cholesterol-fed animals that develop florid atherosclerosis [2]. Moreover, it has been proposed that remnant particles arising after high-fat, high-cholesterol meals contribute to atherogenesis in humans without inherited lipid transport abnormalities [3]. Several reports now indicate that particles with compositional properties of lipoprotein remnants are present in the VLDL and intermediate density lipoproteins (IDL) of patients with CRF [4-S]. We investigated further the hypothesis that CM remnants (CMr) accumulate in CRF using fed retinyl esters (RE) as markers of lipoproteins that originate in the intestine [9]. The results of this study show that there is disproportionate retention of RE in CRF providing additional, independent evidence for increased plasma concentrations of CM remnants. Materials and Methods
Subjects Twenty non-diabetic patients (10 males and 10 females, mean age 53.1 years, range 20-69 years) with chronic renal failure undergoing maintenance hemodialysis 3 times weekly were selected at random for study. Renal failure was due to: amyloidosis (1); glomerulonephritis (8); interstitial nephritis (2); medullary cystic
191
disease (1); nephrosclerosis (1); oxalate nephropathy (2); polycystic kidney disease (2); chronic pyelonephritis (1); lupus nephritis (1) and; unknown (1). Two-thirds of the patients had been dialyzed with bicarbonate-containing solutions and one-third with acetate. Average predialysis serum urea nitrogen, creatinine and albumin concentrations were 73 mg/dl, 10.9 mg/dl and 3.7 g/dl, respectively. Two patients were receiving beta-adrenergic blocking agents, 4 androgenic/ anabolic steroids. and 1 mysoline and phenytoin. None was receiving glucocorticoid therapy. Twenty hospitalized non-diabetic patients whose plasma triglyceride (TG) concentrations were less than 318 mg/dl were selected as controls (19 males, 1 female, mean age 55.5 years, range 32-73 years). They had a variety of medical illnesses including premature coronary heart disease. Several of them were receiving drugs known to affect plasma triglycerides including thiazide diuretics (5) beta-adrenergic blocking agents (3) and prednisone, 60 mg daily, for pulmonary disease (1). None of the 20 had serum urea nitrogen concentrations > 20 mg/dl or creatinine values > 1.5 mg/dl. Retinyl ester challenge
Each subject was fed a capsule containing 25 000 IU of retinyl ester along with supper on the evening preceding dialysis. Venous blood samples for lipid and retinoid analyses were obtained the following morning after a 12-15 h fast, 2-3 days following the last dialysis and prior to that day’s dialysis session. At the time of blood sampling, patients had not received any heparin for 4468 h. Hospitalized controls were studied under identical conditions. Blood was collected on ice in tubes containing disodium EDTA, 1 mg/ml, final concentration. Plasma samples were kept refrigerated at 4°C in metal-foil-covered tubes and handled in subdued light. Chylomicron
flotation
and flocculation
The presence of intact CM was excluded by quantitative analysis of retinoids and lipids using 2 different methods. When whole plasma is allowed to stand in a test tube overnight at 4°C CM float to the top (unit-gravity flotation-or the “icebox test”). Accordingly, unit-gravity flotation was carried out and the presence or absence of floating CM was noted. The upper l/5 of the sample was removed by gentle aspiration (the unit-gravity “top”). This fraction and the bottom 4/5 (“bottom”) were then analyzed for cholesterol, triglyceride, retinol and retinyl ester concentrations. Recovery values were calculated as: [(top + bottom)/total] x 100. Recoveries for these 4 components averaged 99.2, 98.8, 100.2 and 98.7%, respectively. The 2nd method, polyvinylpyrrolidone flocculation, was performed as described by O’Hara and others [lo]. A l-ml portion of plasma was layered under 10 ml of 3% polyvinylpyrrolidone (PVP) and allowed to stand for 18 h at 37°C. This technique separates lymph CM (“primary particles”) from “secondary particles” of less certain derivation. The tubes were examined for turbidity and the top fractions (containing primary particles) and the bottom fractions (containing secondary particles and lipoproteins with density > 1.006 g/ml) were removed for lipid and retinoid analyses. Retinyl ester recoveries for the 11 CRF patients with total plasma retinyl
192
ester concentrations of 15 pg/dl or more were 70.6 percent. Cholesterol, triglyceride, and retinol recoveries (calculated as described above) were 88.2, 94.5 and 98.5%, respectively. Lipoprotein retinoid analysis Plasma retinol and retinyl ester concentrations were measured on extracts of unfrozen plasma by high performance liquid chromatography within a few days of collection [11,12]. The coefficient of variation for replicate analyses of retinol and retinyl ester standards was less than 5%. The variability of post-challenge plasma RE concentrations was examined by replicate clinical studies in 3 subjects. The 1st subject, a woman with severe Type V hyperlipoproteinemia (TG > 2000 mg/dl, increased pre-beta lipoproteins and chylomicronemia by lipoprotein electrophoresis) had normal renal function. Replicate challenges on 7 separate occasions showed a mean value of 209 pg/dl (SD 39.8, range 161-286 pg/dl, n = 7). Patient #2, a CRF patient whose initial RE value was 15 pg/dl, had a mean concentration of 11.1 pg/dl (SD 4.3, range 5.9-17.0 pg/dl, n = 6). Patient #3 was a CRF patient who had no measurable retinyl esters on initial examination. With repeated study on 6 occasions, mean retinyl ester concentration was 0.5 pg/dl (SD 1.3, range O-3.2 pg/dl). Thus, between-day reproducibility was satisfactory. Other methods Triglyceride and cholesterol were measured in coupled enzymatic assays using microbial lipase or cholesterol esterase/cholesterol oxidase, respectively (Beckman Instruments, Carlsbad, CA). Agarose lipoprotein electrophoresis was performed on agarose plates (Corning, Palo Alto, CA). Electropherograms were first interpreted blindly without reference to plasma lipid values. A phenotype was assigned when an abnormal pattern was noted and fasting plasma cholesterol and/or triglyceride values exceed the age- and sex-specific upper decile cutpoints for Visit 2 in the Lipid Research Clinics database [13]. Data analysis To minimize potentially confounding effects of inherited hyperlipidemias, only patients whose plasma triglyceride concentrations (TG) fell below an arbitrarilyselected value, 318 mg/dl (the upper 5th percentile cutpoint for males, aged 40-44 [13]), were compared with control subjects. One patient, a 61-year-old woman, had suffered from polycystic kidney disease and hypertension. Her plasma cholesterol ranged from 223-251 mg/dl and triglycerides 437-498 mg/dl. Twelve hours after oral retinyl ester her total retinyl ester concentration was 74.4 pg/dl. On agarose lipoprotein electrophoresis she proved to have beta-migrating VLDL, a characteristic finding in dysbetalipoproteinemia [2]. For this reason, her data were excluded from the final analysis. Post-challenge retinoid data were skewed upward and so statistical analyses were carried out with non-parametric methods [14]. The Mann-Whitney U-test was used to determine significance between groups, Spearman’s rank correlation coefficient to test the significance of associations between paired variables and Fisher’s Exact Test was used for nominal data.
193
Results
Fasting lipids and lipoprotein electrophoresis Mean fasting plasma triglyceride concentrations were 158 mg/dl (13.6 SEM) in the 20 hospitalized control subjects and 182 mg/dl (14.4 SEM) in the 19 CRF patients (P > 0.2). Mean plasma cholesterol was 197 mg/dl (16.6 SEM) in the control and 168 (7.6 SEM) in the CRF patients (P > 0.2). Thus, the differences in plasma lipid values between CRF patients and hospitalized patients without CRF were not statistically significant. Lipoprotein electropherograms in the controls showed Type IIA in 2 and Type IV hyperlipoproteinemia in 3 patients. Eight of the 19 CRF patients had Type IV hyperlipoproteinemia defined by elevated pre-beta lipoproteins and fasting triglycerides that fell within the age- and sex-specific upper decile [13].
Plasma retinoids Postabsorptive plasma retinol and retinyl ester concentrations in control and CRF patients are shown in Fig. 1. Both distributions were skewed upwards. Plasma retinol concentration in the CRF patients (mean 177 pg/dl, SEM 15.2) was nearly twice that of controls (mean 95.5 pg/dl, SEM 7.7, P < O.OOl), an expected finding given the role of the kidney in retinol-binding protein excretion [15]. Plasma retinyl
3.
A
RETINYL
AETINOL
.
ESTER .
. 60
300. > zoo:.
G \ ?
.
. . : 4 . :
40
r .
.
1001
. 2 .
o.. CONTROL
20
: I
!
.
+-CRF
CONTROL
CRF
0
Fig. 1. Postabsorptive plasma retinal and retinyl ester concentrations in 20 non-azotemic control and 19 CRF patients (pg/dI). Plasma samples were obtained after a 12-15-h fast following supper along with 25000 IU of retinyl ester. Differences between control and CRF patients were significantly different (P < 0.001 for retinol and < 0.01 for retinyl ester) by Mann-Whitney U.
194 TABLE
1
FRACTIONAL DISTRIBUTIONS OF CHOLESTEROL, TRIGLYCERIDES, RETINOL AND RETINYL ESTERS WITH OVERNIGHT UNIT-GRAVITY FLOTATION AND 3% POLYVINYLPYRROLIDONE FLOCCULATION [lo] OF WHOLE PLASMA FROM CRF PATIENTS flotation; The upper one-fifth of the total volume was taken for the “top” fraction in the unit-gravity thus, if no redistribution of CM had taken place, the fractional distribution of a component would be 0.200 in the top and 0.800 in the bottom fraction. Means and standard errors of the means are given. Cholesterol (n = 19)
Triglyceride (n = 19)
Retinol (n =19)
Retinyl ester (n =18)
0.199 (0.001) 0.800 (0.001)
0.195 (0.002) 0.805 (0.002)
0.196 (0.002) 0.804 (0.002)
0.209 (0.004) 0.192 (0.004)
flocculation 0.011 (0.010) 0.989 (0.002)
0.015 (0.028) 0.989 (0.003)
0.009 (0.010) 0.991 (0.002)
0.000
Unit gravity flotation Top Bottom
Polyvinylpyrrolidone Top Bottom
PLASMA
Fig. 2. Best fit linear 20 control subjects correlation) existed intercepts are given
TRIGLYCERIDES
1.000
(mg/dl)
regression plots for retinyl esters (ordinate) versus plasma triglycerides (abscissa) for (0) and 19 CRF patients (0). Significant positive correlations (Spearman’s rank for both the CRF patients (P < 0.05) and the controls (P -C 0.01). Slopes and in the text.
195
esters for the 20 controls were 8.1 pg/dl (SEM 3.5) and for the 19 CRF patients. 23.9 ,ug/dl (SEM 4.2, P < 0.01). The difference between retinyl ester concentrations in controls and CRF patients was significant using cutoff values of either 10 or 20 pg/dl (P < 0.0005) by Fisher’s exact probability [14]. None of the control or CRF plasma samples showed fasting chylomicronemia by inspection. The absence of lipids and retinoids among particles with the properties of unaltered CM was verified after unit-gravity flotation and polyvinylpyrrolidone flocculation of CRF plasma (Table 1). During unit-gravity flotation the distribution of all components closely approximated that expected in the absence of chylomicronemia. All of the measured retinyl ester was accounted for in the lower portion of the tube and none among the “primary particles” during PVP flocculation. Figure 2 shows the relationship between fasting plasma retinyl ester and TG concentrations for the 2 groups. The least squares best fit line for control patients had a slope of 0.125 and a y-intercept of - 11.6 (y, = 0.565, P -c 0.01). The least squares regression for the CRF patients had a slope of 0.121 and an intercept of 1.7 (r, = 0.515, P < 0.05). There was a weak positive correlation between TG and retinol values in the CRF group (slope 0.404, y-intercept 103.7, r, = 0.455, 0.1 > P > 0.05) which was not statistically significant. As expected, correlations between plasma retinol and retinyl ester concentrations were not significant either for the CRF patients (r, = 0.311, P > 0.2) or the controls (r, = 0.230, P > 0.2). Other variables
Data were examined for other differences that might account for the skewed and variable distributions observed in measured retinoid concentrations. There were no significant correlations between serum creatinine and either retinol or retinyl ester concentrations indicating that the severity of renal failure was not an important determinant of retinoid retention in hemodialyzed patients with established CRF. By chance there was a disproportionate number of male patients in the control sample. However, there were no statistically significant differences in age, total cholesterol, triglycerides or retinyl esters between males and females in the CRF group. Retinol levels were higher in CRF males than females (206 pg/dl vs 146 pgg/dl, respectively. 0.05 > P > 0.02 by Mann-Whitney U-test). Retinyl ester and retinol concentrations were significantly greater for either gender in the CRF group than they were in controls. Neither beta-blocking agents, thiazide diuretics, glucocorticoids or anabolic/androgenic steroid therapy could be related to retinoid retention. Thus, there appeared to be unidentified variables, either acquired or genetic, that affected RE disappearance in the subjects. Discussion In this study as well as previous ones [1,6], CRF patients have had only modest hypertriglyceridemia accompanied by normal total plasma cholesterol concentrations. The unimpressive abnormalities in circulating lipids and lipoproteins appear inadequate to account for the increased risk of atherosclerosis in patients with CRF.
196
For this reason, investigators have examined the possibility that there is a subpopulation of potentially atherogenic lipoprotein remnants in the plasma of patients with CRF [1,4,5,8]. The major new finding in the present report was that non-chylomicronemic CRF patients had elevated plasma levels of chylomicron remnants identified by their retinyl ester content. This approach capitalizes on the observation that retinyl esters are transported mainly if not exclusively by CM and CM remnants 191.The validity of the approach is supported by evidence that retinyl esters are conserved among CM and CM remnants [9,11,16] and that retinyl esters are not secreted from the liver with VLDL [16]. Plasma RE concentrations were positively and significantly correlated with plasma TG concentration in both control subjects and CRF patients (Fig. 2). This relationship is expected since CM and endogenous VLDL share a common removal pathway [17] and since endogenous (overproduction) hypertriglyceridemia leads to a secondary decrease in the removal of CM [17] and CM-derived particles. However, in the present study plasma retinyl ester concentration was greater for any given plasma triglyceride concentration in patients with CRF than it was in control subjects. A number of other reported observations are consistent with abnormal persistence of enterogenous lipoprotein remnants in CRF: (1) kinetic studies have demonstrated delayed removal of TG-rich lipoproteins in uremic patients and those undergoing dialysis [18]; (2) the lipoproteins that accumulate in CRF are denser than typical VLDL and relatively enriched in cholesterol, properties of lipoprotein remnants [2,4,5,8]; (3) these lipoproteins contain apolipoprotein B-48 which, in man, appears to be a specific marker of CM and CM catabolic products [5]; (4) the hepatic postheparin plasma triglyceride hydrolase activity is decreased in CRF [6,19,20]; this enzyme plays a role in the latter stages of TG-rich lipoprotein catabolism [21]; and, (5) elevated concentrations of sialylated C apoproteins have been demonstrated in VLDL obtained from patients with CRF [22]. Since the relative concentrations of the C and E apolipoproteins carried by TG-rich lipoproteins affect remnant removal from plasma [23] it is also possible that abnormal apolipoprotein composition in CRF patients contributes to remnant retention. The persistence of fed retinyl esters in the plasma after a 12-15-h fast is additional, independent evidence for abnormal lipoprotein remnant metabolism in CRF. Acknowledgements
We thank Asta Frischat, Tran Thu-Ha, Tran Due-Duy and Martha Wolfe for their valuable technical assistance. We are grateful to Dr. Martin Gregory for helping with some of the clinical studies. References 1 Editorial, Uraemia, lipoproteins and atherosclerosis, Lance& ii (1981) 1151. 2 Brown, MS., Goldstein, J.L. and Fredrickson, D.S., Familial type 3 hyperlipoproteinemia poproteinemia). In: J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein
(dysbetaliand M.S.
197 Brown (Eds.), The Metabolic Ch. 32, p. 655. 3 Zilversmit, D.B., Atherogenesis 4 Ron, D., Oren, I., Aviram, 5 6 7 8
9 10 11 12 13 14 15 16 17 18 19 20
21 22 23
Basis of Inherited -
A postprandial
Disease,
5th edition,
phenomenon,
McGraw-Hill,
Circulation,
M., Better, OS. and Brook, J.G., Accumulation
New York, 1983,
60 (1979) 473. of lipoprotein
remnants
in
patients with chronic renal failure, Atherosclerosis, 46 (1983) 67. Nestel. P.J., Fidge, N.H. and Tan, M.H., Increased lipoprotein-remnant formation in chronic renal failure, N. Engl. J. Med., 307 (1982) 329. Bolanzo, K., Krempler, F. and Sandhofer, F., Hepatic and extrahepatic triglyceride lipase activity in uraemic patients on chronic hemodialysis, Europ. J. Clin. Invest., 8 (1978) 289. Norbeck, H.E. and Carlson, L.A., Increased frequency of late pre-beta-lipoproteins (LPbeta) in isolated serum very low density lipoproteins in uraemia, Europ. J. Clin. Invest., 10 (1980) 423. Minamisono, T., Wada, W., Akamatsu, A., Okabe, M., Handa, Y., Morita, T., Asagami, C.. Naito. H.D., Nakamoto, S., Lewis, L.A. and Mise, J., Dyslipoproteinemia, a remnant lipoprotein disease in uremic patients on hemodialysis, Clin. Chim. Acta, 84 (1978) 163. Goodman, D.S., Vitamin A and retinoids in health and disease, N. Engl. J. Med., 310 (1984) 1023. O’Hara, D.O., Porte, Jr., D. and Williams, R.H.. Use of constant composition polyvinylpyrrolidone columns to study the interaction of fat particles with plasma, J. Lipid Res., 7 (1966) 264. Wilson, D.E. and Chan, L-F., In vitro transfer of chylomicron retinol and retinyl esters, Biochem. Biophys. Res. Comm., 115 (1983) 958. Wilson, D.E., Chan, L-F., Buchi, K.N. and Horton, S.C., Postchallenge plasma lipoprotein retinoids - Chylomicron remnants in endogenous hypertriglyceridemia, Metabolism, 34 (1985) 551. The Lipid Research Clinics Population Studies Data Book, Volume 1 (The Prevalence Study), (DHEW Publication No. (NIH) 79-1527), National Institutes of Health, 1979, p. 58. Siegel, S., Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill, New York, 1958, pp. 104, 116, 202. Smith, F.R. and Goodman, D.S., The effects of diseases of the liver, thyroid, and kidneys on the transport of vitamin A in human plasma, J. Clin. Invest., 50 (1971) 2426. Thompson, K.H., Hughes, L.B. and Zilversmit, D.B., Lack of secretion of retinyl ester by livers of normal and cholesterol-fed rabbits, J. Nutr., 113 (1983) 1995. Grundy, SM. and Mok, H.Y.I., Chylomicron clearance in normal and hyperlipidemic man, Metabolism, 25 (1976) 1225. Heuck, C.-C. and Ritz, E., Hyperlipoproteinemia in renal insufficiency, Nephron, 25 (1980) 1. Mordasini R., Frey, F., Flury, W., Klose, G. and Greten, H., Selective deficiency of hepatic triglyceride lipase in uremic patients, N. Engl. J. Med., 297 (1977) 1362. Applebaum-Bowden, D., Goldberg, A.P., Hazard, W.R., Sherrard, D.J., Brunzell. J.D., Huttunen, J.K., Nikkila, E.A. and Ehnholm, C., Postheparin plasma triglyceride lipases in chronic hemodialysis - Evidence for a role for hepatic lipase in lipoprotein metabolism, Metabolism, 28 (1979) 917. Jackson, R.L., Lipoprotein lipase and hepatic lipase. In: P.D. Boyer (Ed.), The Enzymes, 3rd edition. Academic Press, New York, 1983, p. 141. Holdsworth, G., Stocks, J., Dodson, P. and Galton, D.J., An abnormal triglyceride-rich lipoprotein containing excess sialylated apolipoprotein C-III, J. Clin. Invest., 69 (1982) 932. Windler, E., Chao, Y. and Havel, R.J., Determinants of hepatic uptake of triglyceride-rich lipoproteins and their remnants in the rat, J. Biol. Chem., 255 (1980) 5475.