Original Contributions |
From the Department of Medicine and The John P. Robarts Research Institute, University of Western Ontario, London, Ontario, Canada (J.R.B., D.B.M., D.E.T, S.J.K., M.W.H.), and the Departments of Bioengineering and Medicine, University of Washington, Seattle (P.H.R.B., P.V.).
Correspondence to Murray W. Huff, The John P. Robarts Research Institute, 4-16, University of Western Ontario, 100 Perth Dr, London, Ontario N6A 5K8, Canada. E-mail mhuff{at}julian.uwo.ca
| Abstract |
|---|
|
|
|---|
Key Words: HMG-CoA reductase inhibitor atorvastatin triglyceride-rich lipoproteins tracer kinetics compartmental model
| Introduction |
|---|
|
|
|---|
There is increasing evidence from cell culture studies in vitro8 9 and human studies in vivo10 11 12 13 14 15 that postprandial, triglyceride-rich, apoB-containing lipoproteins are atherogenic. Furthermore, these postprandial lipoproteins may be more closely related to atherogenic risk than those lipoproteins in the fasting state. Consequently, the measurement of postprandial lipids and lipoproteins may be superior to fasting lipid analyses in predicting the existence and/or progression of coronary heart disease.
The most commonly used approach for studying chylomicron
metabolism is an indirect one, involving the oral
administration of vitamin A (retinol) to endogenously label
intestinally derived triglyceride-rich lipoproteins
(TRLs).16 The rationale is that retinol is taken
up by enterocytes, esterified to retinyl esters [predominantly retinyl
palmitate (RP)], incorporated along with other neutral lipids into the
core of the chylomicron,17 and secreted into the
lymphatics. The assumption that the retinyl esters remain associated
with the chylomicron remnant, however, may not be entirely correct.
Krasinski et al18 found in humans that during the
late postprandial period (
9 hours), up to one third of the remaining
plasma RP was associated with HDL and LDL. Nevertheless, RP data
provide a close approximation of plasma chylomicron and chylomicron
remnant metabolism during the early postprandial period.
The plasma concentrations of retinyl esters with time reflect the
balance between (1) the assembly and secretion of chylomicrons into
plasma and (2) the formation and clearance of chylomicron remnants.
Most studies to date have used the areas under the retinyl ester curves
to determine differences in the kinetics of intestinally derived
retinyl esters in the postprandial state. However, this approach does
not allow one to distinguish the secretion rate from the clearance
rate.
3-Hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, an endoplasmic reticulum (ER) protein, catalyzes the rate-determining reductive deacylation of HMG-CoA to mevalonate7 19 in cholesterol biosynthesis. A class of lipid-lowering compounds, the HMG-CoA reductase inhibitors, have been shown to interfere with the rate-limiting step in cholesterol biosynthesis.20 These compounds are potent LDL cholesterollowering agents; however, relative to the LDL response, their ability to reduce triglyceride concentrations is modest. Moreover, the effects of HMG-CoA reductase inhibitors on hepatic apoB secretion in humans are inconsistent and have been the focus of a recent review by Huff and Burnett.21
As reviewed by Cohn,22 many studies have examined the effect of HMG-CoA reductase inhibitors on postprandial lipoprotein metabolism.23 24 25 26 27 28 29 30 Although the HMG-CoA reductase inhibitors as a class tend to decrease postprandial triglyceride concentrations, the effects of treatment are inconsistent but appear to correlate with the degree of fasting plasma triglyceride reduction.22 No multicompartmental modeling of any retinyl ester kinetic data, however, was performed in any of these studies.
Atorvastatin is a new synthetic, chiral, tissue-selective inhibitor of HMG-CoA reductase.31 32 Clinical trials of atorvastatin in humans have demonstrated marked plasma LDL cholesterol and triglyceride reductions.33 34 35 We have previously shown in vivo that inhibition of HMG-CoA reductase by atorvastatin in miniature pigs significantly decreases hepatic apoB-containing lipoprotein secretion into plasma in the fasting state,36 with minimal effect on rate of clearance.
We have now developed a multicompartmental model of TRL metabolism and have used kinetic analysis to determine the metabolic parameters of postprandial exogenous TRL in plasma with atorvastatin treatment. These studies were designed to test the hypothesis that the inhibition of cholesterol biosynthesis by atorvastatin modulates exogenous TRL metabolism in vivo.
| Methods |
|---|
|
|
|---|
Six animals received the HMG-CoA reductase inhibitor
atorvastatin (Parke-Davis) at a dose of
3 mg · kg body
wt-1 · d-1 (80
mg/d) for 28 days before postprandial studies. This dose is equivalent
to the maximum therapeutic dose used in humans but is
2.6-fold
greater than that used in humans when body weight is taken into
consideration. Atorvastatin is a synthetic inhibitor of
HMG-CoA reductase with an in vitro IC50 of 7.5
nmol/L for rat liver microsomal HMG-CoA reductase
activity.37 The chemical structure of
atorvastatin has been illustrated previously.38
Atorvastatin was placed in gelatin capsules and, to ensure ingestion,
was administered by hand before the daily feeding. The 6 control
animals received a placebo capsule. The atorvastatin was given at 9
AM each day after a 24-hour fast.
Two weeks before the postprandial studies, an indwelling silicone elastomer (Silastic) catheter (1.96-mm ID) was surgically implanted in an external jugular vein.39 Isoflurane USP (Abbott Laboratories Ltd) was used as the anesthetic and ketamine USP (Vetrepharm Canada Inc) as the preanesthetic. Catheters that were kept patent by being filled with 7% EDTA-Na2 allowed blood sampling throughout each postprandial study in unrestrained, unanesthetized animals. The Animal Care Committee of the University of Western Ontario approved the experimental protocol.
Oral Fat Tolerance Test
After a 24-hour fast, pigs were fed the diet described above in
an amount calculated to provide 2 g fat/kg body wt and either
placebo or atorvastatin. This test meal was supplemented with 50 000
IU of retinol (vitamin A capsules USP, Novopharm Ltd) and
consumed within 10 minutes. The animals were not fed for the 12 hours
of the study but had free access to drinking water. Venous blood
samples (20 mL) were drawn at 0 (before the test meal), 0.5, 1, 1.5, 2,
2.5, 3, 4, 5, 6, 8, and 12 hours and collected into tubes containing
EDTA-Na2. Samples were kept on ice before
isolation of plasma lipoproteins and protected from light during
processing. Plasma was obtained by centrifugation at
1000g for 25 minutes at 4°C. The isolated plasma underwent
preparative ultracentrifugation at d=1.006
g/mL in a Beckman 50.4 Ti rotor at 35 500 rpm at 12°C for 16 hours.
TRL fractions (d<1.006 g/mL; Sf>20)
were isolated by tube slicing, and each plasma sample and TRL fraction
was analyzed for cholesterol,
triglyceride, and retinyl ester concentrations.
Retinyl Ester Analysis
Retinyl ester (RP and retinyl stearate) concentrations were
determined in total plasma and in the TRL fraction by a modification of
the high-performance liquid chromatography
(HPLC) method of Weintraub et al.40 Extractions
and analyses were carried out with HPLC-grade solvents and
under subdued light. Retinyl acetate was added to the samples as an
internal standard, and retinyl esters were extracted by a mixture of
ethanol, hexane, and water (1:5:0.5). The hexane layer was evaporated
under nitrogen. The samples were redissolved in ethanol and separated
on a 5-µm Hypersil C18 (7.5x0.32 cm) column. Pump A consisted of
88% methanol in water, whereas pump B consisted of 75% methanol and
25% isopropanol. Pump A was used at 100% for the first 4 minutes,
reduced to 20% pump A to 9 minutes, and further reduced to 0% pump A
to 12 minutes, at a flow rate of 0.6 mL/min. The absorbance of the
eluent was measured at 325 nm, and the retinyl ester concentrations
were quantified by ratio of peak height to that of the internal
standard. Areas under curves (AUCs) for the retinyl ester were
calculated. Within-assay and between-assay coefficients of variation
for TRL RP were <4.0% and <8.6%, respectively.
Retinyl Ester Kinetic Analysis
A compartmental model of TRL metabolism was
developed by use of the multicompartmental modeling program SAAM II
(SAAM Institute Inc, Seattle, WA) running on a Pentium-based personal
computer. The model was developed assuming that the fractional rate
constants (k) were time invariant and first order. The model
chosen to describe the data is shown in Figure 1
. This model is locally
identifiable41 42 in that it has, a priori,
1 solution for k(2,1) and k(0,3) but 2 distinct
solutions for k(0,4) and k(4,3). One of the 2
solutions for k(4,3) is consistently negative and
therefore must be rejected (all transfer constants must be positive).
Thus, the model is uniquely identifiable, with the constraint that all
of the transfer rates must be positive. Dietary retinol was used to
endogenously label intestinally derived TRL with RP.
|
Figure 1
shows the compartments of the model and the pathways that
connect the compartments. Compartment 1 represents the dosing
compartment. On the basis of earlier studies,43
it was assumed that all of the ingested retinol appeared in
chylomicrons as RP. Compartment 2, the delay compartment, contains 5
compartments in series. It was assumed that the residence time of
material in compartment 1 and in the delay compartments (compartment 2)
were equal. The functions of these compartments were to provide a delay
that corresponded to the time required for the synthesis of
chylomicrons and their secretion into plasma. From the delay
compartment, material enters compartment 3, which is assumed to
represent a rapid-turnover chylomicron pool. Chylomicrons in
compartment 3 are either removed from plasma or converted to a
slower-turnover plasma pool, compartment 4. Chylomicrons and
chylomicron remnants in compartment 4 are subsequently removed from
plasma. The sum of compartments 3 and 4 represents total plasma
TRL RP.
The compartmental model was fit to each individual data set. Simplifying the model to 1 plasma chylomicron pool clearly demonstrated the necessity for the more complex model structure to describe the tail component of the data. We were unable, however, to obtain acceptable coefficients of variation for any of the kinetic parameters reported for animal 3 in the atorvastatin-treated group. The experimental data for this animal did not fit the proposed model, presumably for technical reasons, and this pig represents an outlier relative to the whole population and was thus excluded from subsequent analysis.
Relationship Between TRL Fractional Clearance Rate (FCR) and
Fasting Hepatic VLDL ApoB Synthetic Rates
To examine the relationship between the kinetics of postprandial
retinyl esters and those of fasting hepatic apoB synthesis, 11 pigs (6
control and 5 atorvastatin-treated) participated in both an oral fat
load study and an apoB kinetic study using simultaneous
triple-isotope (131I-labeled
VLDL,125I-labeled LDL, and
3H-labeled leucine) labeling. These pigs
represent a subset of those reported
previously.36 The 2 protocols were performed 1
week apart.
Plasma Lipids and Lipoproteins
Cholesterol and triglyceride
concentrations were determined in the plasma and the TRL fraction by
enzymatic, colorimetric assays using reagents obtained
from Boehringer Mannheim GmbH. Areas under the
triglyceride curves were calculated. Lipoprotein protein
was determined by the method of Markwell et al.44
Tests for statistical significance of differences in lipid and RP
concentrations and kinetic parameters were compared by
unpaired t test. A value of P<0.05 was
considered significant.
| Results |
|---|
|
|
|---|
|
During the course of the oral fat tolerance test, plasma
cholesterol did not change significantly from baseline, at
any time point, in either the control or the treatment group (Figure 2
). TRL cholesterol was
maximally lowered by 37% in the atorvastatin-treated animals at 2
hours, but this was not statistically significant
(P=0.143).
|
Atorvastatin treatment resulted in a significant decrease in both
plasma and TRL triglycerides at 2 and 2.5 hours. The 0- to
12-hour AUC for plasma and TRL triglycerides decreased by
20% (Figure 3
; Table 2
) and 24%, respectively, in the
atorvastatin-treated animals; however, these reductions were not
statistically significant. Moreover, the peak TRL
triglyceride concentration was decreased 29%, and the time
to achieve this peak was delayed (5.2 versus 2.3 hours;
P<0.01). Consistent with the decrease in peak TRL
triglyceride, a decrease in the ratio of TRL
triglyceride to protein at the 2-hour time point was
observed in all atorvastatin-treated animals. However, this 23%
decrease was not statistically significant (P=0.081; Table 3
).
|
|
|
The 0- to 12-hour AUC for plasma and TRL RP decreased by 16%
(P=0.047; Figure 4
) and 15%
(P=0.193), respectively. When total plasma retinyl esters
were compared, similar differences were observed. No significant
differences were seen in either plasma or TRL retinyl esters for any
single time point between the control and treatment groups.
|
A fit of the model to the TRL RP data using the parameters
derived from the kinetic analysis for a
representative atorvastatin-treated animal is shown in
Figure 5
. A fit of the model to the mean
TRL RP data using the parameters derived from the kinetic
analysis is shown in Figure 6
.
The fractional rate constants, delay times, and FCRs derived from the
model analyses are shown in Table 4
. The TRL RP k(4,3) between
the rapid-turnover and slow-turnover plasma compartments was decreased
by 78% (0.306 versus 1.074 pools/h; P=NS) with atorvastatin
treatment. The TRL RP FCR was increased by 1.4-fold in the
atorvastatin-treated pigs (3.093 versus 2.276 pools/h;
P=0.012). The percent conversion of TRL RP from the
rapid-turnover to the slow-turnover compartment decreased markedly, by
47% (P=NS), with atorvastatin treatment. The TRL RP FCR was
inversely correlated with the change in VLDL apoB production
rate (r=-0.49) for the 11 animals for which fasting apoB
kinetic data were also determined (Figure 7
). However, the observed relationship
was not statistically significant.
|
|
|
|
Because the chylomicron pool size is changing with respect to time, it
is not possible to determine absolute secretion rates. Each pig in the
study received the same dose of retinol, and we have made the
assumption that the amount of retinol per TRL particle secreted was the
same for the control and treatment animals. However, 3 lines of
evidence suggest no significant change in the number of intestinal
particles assembled and secreted into plasma with atorvastatin
treatment. First, the TRL RP curves overlapped during the first 1.5
hours (Figure 4
). Second, the TRL RP peak heights and times to peak
coincided in both the control and treatment animals. Third, the
integral of the mass in compartment 2 provides a time course for the
accumulation of the dose in the plasma. We can predict from the model
that the time required for 50% of the retinol to appear in the plasma
was unaffected by atorvastatin treatment (1.94 versus 1.98 hours;
P=NS).
| Discussion |
|---|
|
|
|---|
We have demonstrated previously in miniature pigs that atorvastatin treatment decreases fasting cholesterol, triglyceride, and VLDL triglyceride concentrations in plasma36 ; similar responses have been reported in human subjects.33 34 35 In the present study, decreases were seen in postprandial TRL cholesterol, triglyceride, and RP concentrations. The shapes of the TRL cholesterol and triglyceride curves were similar, with a reduced AUC during the first 4 hours of the study in atorvastatin-treated animals but overlapping curves during the final 8 hours of the study. In contrast, the TRL RP curves were overlapping during the first 1.5 hours after the fat meal, with the atorvastatin-treated pigs having lower values than control animals at the latter time points. The rate at which plasma triglycerides increase after a fat meal is a function of gastric emptying, intestinal absorption, chylomicron assembly and secretion, and LPL-mediated catabolism. However, the increase in postprandial triglycerides may not be entirely of intestinal origin, because recent studies have shown that hepatogenous VLDL triglycerides contribute appreciably to the TRL triglyceride response, particularly after the peak.11 45 46 47 48 As discussed below, this may be due to enhanced competition for VLDL triglyceride clearance by postprandial chylomicrons.
Our findings of decreased postprandial RP concentrations, most noticeable at the latter time points, are consistent with those found by Castro Cabezas et al25 in familial combined hyperlipidemic subjects treated with simvastatin (mean dose, 24.5 mg/d), Weintraub et al23 in type IIb subjects treated with lovastatin (80 mg/d), Gylling et al28 in type III (apoE2/2) hyperlipoproteinemic subjects treated with lovastatin (40 mg/d), and Simo et al26 in subjects with mild to moderate hypertriglyceridemia (1.75 to 7.5 mmol/L) and low HDL (<0.9 mmol/L) treated with lovastatin (40 mg/d). These reductions were associated with a mean reduction of 23% for the plasma RP AUC in these 4 studies. Statistically significant reductions of RP were achieved only in the studies by Weintraub at al23 for plasma, Sf>1000, and Sf<1000 0- to 24-hour AUCs and Castro Cabezas et al25 for the Sf<1000 0- to 24-hour AUC. However, the mechanism for the decreased AUCs could not be determined, because multicompartmental modeling was not carried out.
Although a large number of studies have investigated postprandial lipoprotein metabolism, as reviewed by Cohn,22 few have undertaken multicompartmental modeling to analyze chylomicron metabolism.43 49 50 51 52 53 To further investigate the apparent enhanced clearance of TRL RP, we developed a multicompartmental model of TRL metabolism, 2 compartments of which represent the total plasma TRL RP. We found that the TRL RP FCR was increased by 1.4-fold with atorvastatin treatment. Furthermore, the percent conversion of RP from the rapid-turnover plasma compartment to the slow-turnover plasma compartment was decreased, by 26%, in atorvastatin-treated pigs. Although Martins et al,54 using lipid emulsions in rats, demonstrated that particle number is an important factor regulating the metabolism of chylomicrons and their remnants, we did not find any evidence for a difference in chylomicron secretion with atorvastatin treatment. The TRL RP curves overlapped during the first 1.5 hours, and TRL RP peak heights and times to peak coincided in both the control and treatment animals.
The mechanism whereby the liver rapidly and specifically removes chylomicron remnants is still unclear. The hepatic uptake of chylomicron remnants has been shown to be mediated by the LDL receptor and LDL receptorrelated protein (LRP),55 56 57 58 59 and turnover studies suggest that the liver is the main site of chylomicron metabolism.60 Various possible ligand proteins, including apoE, LPL, and hepatic lipase, have been proposed.61 Heparan sulfate proteoglycans (HSPGs) have been shown to enhance chylomicron clearance.62 63 The VLDL receptor has been implicated in chylomicron remnant removal by nonhepatic tissues.64 65 We have previously shown that inhibition of cholesterol synthesis by atorvastatin decreases both VLDL and LDL apoB production rates in the miniature pig.36 In these studies, neither the VLDL apoB FCR nor LDL apoB FCR was significantly affected by atorvastatin treatment. Furthermore, hepatic or intestinal LDL receptor mRNA abundances, as measured by RNase protection assay, were also unchanged. Our results in this animal model would suggest that upregulation of LDL receptors is not the reason for the enhanced TRL clearance. We could not exclude the possibility of an effect of atorvastatin treatment on LRP or VLDL receptor expression.
Our results differ from those of other postprandial studies using HMG-CoA reductase inhibitors, in that we demonstrated reductions in postprandial TRL triglycerides in the early postprandial period (0 to 4 hours). In contrast to the studies by Simo et al26 and Castro Cabezas et al,25 in which the reduced triglyceride concentrations observed at baseline were maintained throughout the duration of the study, the differences we observed in the postprandial period with atorvastatin treatment were greater than the differences observed at baseline and were confined to the first 4 hours in the postprandial period. This may represent a unique effect of atorvastatin treatment.
The decrease in TRL triglyceride seen in the early (0 to 4
hours) postprandial phase presumably represents a change in
lipoprotein composition of postprandial intestinal and/or hepatic
lipoproteins with atorvastatin treatment. Consistent with the
decrease in peak TRL triglyceride, a trend toward a
decrease (23%) in the ratio of TRL triglyceride to protein
at the 2-hour time point was observed with atorvastatin treatment
(Table 3
). It is possible that atorvastatin treatment results in the
secretion of a chylomicron particle with (1) a reduced triglyceride
content and/or (2) an altered apolipoprotein composition, thereby
allowing a more rapid rate of lipolysis and subsequent hepatic
clearance of triglyceride-depleted remnants.
Postprandial apolipoproteins, particularly apoC and apoE, and lipid compositional changes in chylomicrons, their remnants, and VLDL, have been demonstrated in humans by immunoaffinity chromatography.66 67 ApoC-III, a component of chylomicrons and VLDL, has been shown to play an important role in TRL metabolism. In vitro, apoC-III can inhibit the hydrolysis of triglyceride68 69 and reduce TRL clearance.70 71 In vivo, transgenic mice overexpressing human apoC-III develop a marked hypertriglyceridemia resulting from impaired clearance of TRL due to apoE insufficiency.72 73 In contrast, homozygous apoC-III knockout mice have hypotriglyceridemia and enhanced TRL clearance.74 ApoE-enriched subfractions of large VLDL from hypertriglyceridemic subjects show enhanced triglyceride hydrolysis by LPL.75 ApoE is important for the hepatic recognition of TRL remnants by the LDL receptor,76 the LRP,55 59 and HSPGs.62 HSPGs play a significant role in plasma clearance and hepatic uptake of TRL remnants in mice.63 It is possible that atorvastatin treatment decreases apoC-III and/or increases apoE concentrations of TRL. However, in the present study, TRL apoC-III and apoE concentrations were not determined.
Large TRLs, and in particular chylomicrons, are the favored substrates for LPL. Essentially all TRLs share the same lipolytic pathway, that is, the hydrolysis of the core triglyceride content by LPL. After a fat meal, chylomicrons and endogenous TRLs of hepatic origin (VLDL) are mixed in blood and can compete for LPL. It has been shown that endogenous TRLs accumulate in plasma after oral fat intake11 47 48 because of the failure of these lipoproteins to effectively compete with chylomicrons for lipolysis by LPL.2 3 4 5 We have established that in the fasting state, atorvastatin treatment of miniature pigs decreases the production rate of hepatic VLDL apoB.36 Assuming that atorvastatin continued to decrease hepatic VLDL apoB secretion in the postprandial state, the enhanced fractional clearance of intestinally derived TRL observed with atorvastatin treatment would be consistent with decreased competition for removal processes by hepatic VLDL. A negative correlation was observed between TRL (d<1.006 g/mL; Sf>20) RP FCR and VLDL apoB production rate (r=-0.49; P=NS) in the 11 animals in which both protocols had been carried out. Consistent with this observation, Le et al,51 in a group of normolipidemic controls and hypertriglyceridemic subjects, found a positive correlation between the production rate for VLDL apoB in the fasting state and the half-life (the inverse of FCR) of retinyl esters in both Sf 20 to 400 (mostly remnants) and Sf>400 lipoproteins.
In conclusion, we have demonstrated in a multicompartmental model of TRL metabolism that the inhibition of HMG-CoA reductase by atorvastatin has no significant effect on intestinal TRL assembly and secretion; however, plasma clearance of TRL was increased, perhaps related to decreased competition for removal processes by hepatic VLDL.
| Acknowledgments |
|---|
Received December 15, 1997; accepted May 26, 1998.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
F. Basso, L. A. Freeman, C. Ko, C. Joyce, M. J. Amar, R. D. Shamburek, T. Tansey, F. Thomas, J. Wu, B. Paigen, et al. Hepatic ABCG5/G8 overexpression reduces apoB-lipoproteins and atherosclerosis when cholesterol absorption is inhibited J. Lipid Res., January 1, 2007; 48(1): 114 - 126. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Bach-Ngohou, K. Ouguerram, R. Frenais, P. Maugere, B. Ripolles-Piquer, Y. Zair, M. Krempf, and J. M. Bard Influence of Atorvastatin on Apolipoprotein E and AI Kinetics in Patients with Type 2 Diabetes J. Pharmacol. Exp. Ther., October 1, 2005; 315(1): 363 - 369. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Castro Cabezas, C. Verseyden, S. Meijssen, H. Jansen, and D. W. Erkelens Effects of Atorvastatin on the Clearance of Triglyceride-Rich Lipoproteins in Familial Combined Hyperlipidemia J. Clin. Endocrinol. Metab., December 1, 2004; 89(12): 5972 - 5980. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Verseyden, S. Meijssen, H. van Dijk, H. Jansen, and M. C. Cabezas Effects of atorvastatin on fasting and postprandial complement component 3 response in familial combined hyperlipidemia J. Lipid Res., November 1, 2003; 44(11): 2100 - 2108. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. G. Parhofer, E. Laubach, and P. H. R. Barrett Effect of atorvastatin on postprandial lipoprotein metabolism in hypertriglyceridemic patients J. Lipid Res., June 1, 2003; 44(6): 1192 - 1198. [Abstract] [Full Text] [PDF] |
||||
|
|