modestly decreased 7a-hydroxylase activity (mean 29% suppression). ... rate of cholesterol synthesis that can be regulated over a wide range in response to ...
THEJOURNAL OF BIOLOG~CAL CHEMISTRY 0 1992 by The AmericanSociety for Biochemistry and Molecular Biology, Inc.
Vol. 267, No. 8, Issue of March 15, pp. 55846591,1992 Printed in U.S.A.
Regulation of Hepatic Sterol Metabolism in The Rat PARALLEL REGULATION OF ACTIVITY AND mRNA FOR 7a-HYDROXYLASE BUTNOT 3-HYDROXY-3METHYLGLUTARYL-COENZYME A REDUCTASE ORLOW DENSITYLIPOPROTEINRECEPTOR* (Received for publication, October 3, 1991)
David K. Spady andJennifer A. Cuthbert From the Department of Internal Medicine, The University of Texas Southwestern Medical Center, Dallas, Texas 75235-8887
In vivo regulation of hepatic sterol metabolismwas examined in the rat. Sodium cholate markedly suppressed hepatic 7a-hydroxylase mRNA levels and activity when fed torats on a low cholesterol diet. Sterol balance was maintained solely by decreasing hepatic cholesterol synthesis. Compensatory mechanisms were inadequate when cholate was fed to rats on a high cholesterol diet and massive amounts of cholesterol accumulated in the liver and plasma. Suppression of bile salt synthesiswas not responsiblesince cholate did not suppress 7a-hydroxylaseactivity when fed torats on a high cholesterol diet. Moreover, total hepatic low densitylipoproteinreceptor activity was notsuppressed even though liver cholesteryl ester levels were increased more than360-fold. Changes in 7a-hydroxylase activity were always accompanied by parallel changes in mRNA, whereas mRNA levels for 3-hydroxy-3-methylglutaryl-coenzymeA (HMG-CoA) reductase were reducedby 60%or less, even when cholesterol synthesis was suppressed by 98%.HMG-CoA reductase and low density lipoprotein receptor activities were regulated independently although mRNA levels for these two proteins were coordinately regulated. These findings indicate that 7a-hydroxylase is controlled bymRNA levels, whereas in vivo cholesterol synthesis is predominantlycontrolledbyposttranscriptional regulation of HMG-CoA reductase activity.
The liver plays a central role both in the maintenance of whole body sterol balance and in the regulation of plasma lipoprotein concentrations (1).The bulk of dietary cholesterol is transported into the liver by way of the chylomicron remnant receptor pathway, and biliary secretion of sterols, either as cholesterol or after conversion to bile salts, provides the only significant mechanism for cholesterol removal from the body. In response to changes in cholesterol influx or efflux, sterol balance across the hepatocyte is maintained by altering the flux of cholesterol through two sterol-regulated pathways, endogenous cholesterol synthesis,and receptor-dependent LDL’ uptake (2, 3). In cultured cells, the activity of 3-hydroxy-3-methylglutaryl-coenzymeA (HMG-CoA) reductase, the rate-limiting *This research was supported by National Institutes of Health Grants A117653, HL38409, and a grant-in-aid from the American Heart Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. The abbreviations used are: LDL, low density lipoprotein; FABP, fatty acid binding protein; GAPDH, glyceraldehyde-3-phosphate deA; hydrogenase; HMG-CoA, 3-hydroxy-3-methylglutaryl-coenzyme HPLC, high-performance liquid chromatography; bp, base pair(s).
enzyme in the cholesterol biosynthetic pathway, andthe number of cell surface LDL receptors are coordinately regulated inresponse to changes in sterolavailability (2-5). HMGCoA reductase and LDL receptor mRNA levels are also regulated in parallel in cultured cells, suggesting that a common mechanism may be responsible for controlling the expression of these two genes in vitro (5). Whether the expression of these two genes is coordinately regulated in vivo is notknown; however, regulation of de nouo cholesterol synthesis appears to be the primary, and often the only, response to experimental manipulations that alter cholesterol balance across the liver in animals (6, 7). Up to 80% of total LDL turnover is normally mediated by LDL receptors located in the liver (810). If endogenous cholesterol synthesis was, or couldbe, regulated independently of the LDL receptor pathway, the changes in circulating LDL levels that normally occur in response to dietary cholesterol loads might be moderated. Conversion of cholesterol to bile salts provides the major route for cholesterol elimination from the body. The initial and rate-limiting step in this pathway is catalyzed by 7ahydroxylase (11-13). ‘la-Hydroxylase activity is subject to end-product suppression by lipophilic bile salts in the enterohepatic circulation (14) and, at least in some speciessuch as the rat, tosubstrate induction (15). Recently, the mRNAs for rat andhuman 7a-hydroxylase have been isolated and cloned (16-19). Subsequent studiesin therat have shown that suppression of 7a-hydroxylase activity by lipophilic bile salts and stimulation by dietary cholesterol are due to changes in gene transcription (20). The rat is extremely resistant to the effects of dietary cholesterol but can be made hypercholesterolemic when cholesterol is fed in combination with lipophilic bile salts (21). The mechanisms involved in the resistance of the rat to dietary cholesterol and how dietary cholesterol and bile salts interact to regulate plasma LDL levels is notwell understood. The present studies were undertaken to examine the mechanism by which dietary bile salts and cholesterol regulate 7ahydroxylase activity, cholesterol synthesis, and receptor-dependent LDL transport in the rat and determine to how these sterol-regulated pathways interact to control plasma LDL concentrations. MATERIALS ANDMETHODS
Animals and Diets-All studies were performed in female SpragueDawley rats (SASCO Inc., Omaha, NE). The animals were obtained in the weight range of 125-150 g and housed in an isolation room with 12-h light cycling and allowed free access to water and standard rodent diet for at least 2 weeks before beginning the experimental diets. The experimental diets were prepared by mixing various amounts of cholesterol or bile salts with ground Wayne Lab Blox (Allied Mills, Chicago, IL) using a commercial food mixer. Unconjugated bile salts were purchased from Sigma and were >98% pure by HPLC. The experimental diets were fed for 6 weeks prior to specific
5584
5585
Regulation of Hepatic Sterol Metabolismin Vivo experiments, and all experiments were carried out during the middark phase of the light cycle. Determination of Cholesterol Synthesis Rates and HMG-CoA Reductase Actiuity-For measurements of absolute rates of cholesterol synthesis in uiuo, animals were administered -50 mCi of [3H]water intravenously as previously described (22, 23), and then returned to individual cages under a fume hood. One h after theinjection of [3H] water the animals were anesthetized and exsanguinated through the abdominal aorta. Aliquots of plasma were taken for the determination of body water specific activity, and samples of liver were taken for the isolation of digitonin-precipitable sterols. Rates of sterol synthesis are expressed as the nanomoles of [3H]waterincorporated into digitonin-precipitable sterols per h per ofg tissue (nanomoles/h per 9). HMG-CoA reductase activity was measured as previously described (24). Briefly, hepatic microsomes were prepared in the presence of 50 mM sodium fluoride. Following a preincubation period, during which microsomes were incubated in the presence or absence of Escherichia coli alkaline phosphatase, HMG-CoA reductase activity was measured using ["CIHMG-CoA (Du Pont-New England Nuclear) as substrate. The ["C]mevalonate formed was converted to thelactone and isolated by thin-layer chromatography using [3H]mevalonate (Du Pont-New England Nuclear) to correct for recovery. Determination of Tissue LDL Uptake Rates in Viuo-Plasma was obtained from normocholesterolemic rat and human donors. The LDL fraction was isolated in the density range of 1.020-1.055 g/ml, washed and concentrated, and labeled with '"1- or I3'I-tyramine cellobiose as previously described (25). The human LDL was also subjected to reductive methylation to eliminate its recognition by the LDL receptor (26). Rates of tissue LDL uptake were determined using a primed infusion of '"I-tyramine cellobiose-labeledLDL (27). cellobiose-labeledLDL were continued The infusions of 1Z51-tyramine for 6 h at which time each animal was administered a bolus of I3'Ityramine cellobiose-labeled LDL as a volume marker and killed 10 min later by exsanguination through the abdominal aorta. Samples of the liver along with aliquots of plasma were assayed for radioactivity ina gamma counter(Packard Instrument Co., Inc., Downers Grove, IL). The amount of labeled LDL in the liver at 10 min ( I 3 l I disintegrations per min per gof tissue divided by the specific activity of I3'I in plasma) and at 6 h ('"1 disintegrations per min per g of tissue divided by the specific activity ofIz5I in plasma) was then calculated and has the units of micrograms of LDL-cholesterol or LDL-protein per g of tissue. The increase in the tissue content of LDL-cholesterol or LDL-protein with time represents the rate of LDL uptake inmicrograms of LDL-cholesterol or LDL-protein taken up per h per g of tissue (micrograms/h per g) or per whole organ (micrograms/h per organ). Since receptor-dependent LDL uptakeby the liver is saturable and since plasma LDL concentrations varied among the different experimental groups, changes in receptor-dependent LDL uptake could not be directly equated with changes in LDL receptor activity (8, 9). In order to relate changes in receptor-dependent LDL uptake to changes in LDL receptor activity, the experimentally determined uptake rates were superimposed on kinetic curves defining the relationship between hepatic LDL uptake and circulating LDL concentrations in normal animals. These kinetic curves were established in control rats by measuring rates of total andreceptor-independent LDL transport under conditions where plasma LDL concentrations were acutely varied over a wide range by infusing mass amounts of unlabeled LDL (9). Therelationship between total LDL uptake (JJ and the concentration of LDL in plasma (C,) can be described by the equation Jt = (*PG)/(*K, + C1) + *PCl, where *P equals the apparent maximal uptake velocity, *K, equals the concentration of LDL in plasma necessary to achieve one-half of this maximal uptake rate and *P equals the apparent uptake constant for receptor-independent transport. The kinetic curves for normal LDL uptake used in thesestudies were constructed using previously published values for *P,*K,, and *P (9). By relating the rates of receptor-dependent and receptorindependent LDL uptake in theexperimental animals to these normal kinetic curves, it was possible to determine how the various dietary manipulations affected LDL receptor activity (defined as the rateof receptor-dependent LDL uptake in an experimental animal relative to the rate of receptor-dependent LDL uptake that would be seen in control animals at thesame LDL concentration). Determination of Hepatic 7a-Hydroxyhe Actiuity-Hepatic 701hydroxylase activity was determined using an HPLC-spectrophotometric assay which quantifies the amount of 7a-hydroxycholesterol formed from endogenous microsomal cholesterol after enzymatic
conversion to 7a-hydroxy-4-cholesten-3-one using cholesterol oxidase (28). Determination of mRNA Leuels-Levels of mRNA for 7a-hydroxylase, HMG-CoA reductase, LDL receptor, fatty acid-binding protein (FABP), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were determined by nuclease protection as previously described (29, 30). Plasmids containingcDNA's encoding rat 7a-hydroxylase (Sac7) (16), HMG-CoA reductase (pGEM-HMGR), LDL receptor (pLDL1)(31), FABP (pJG418) (32), and GAPDH (pRcGAP123) (33), were generously provided by Drs. David Russell and Diane Jelinek (University of Texas Southwestern Medical Center, Dallas, TX), Sohaib Khan (University of Cincinnati, OH), Richard Tanaka (BristolMeyers Squibb Research Institute, Princeton, NJ), Jeffrey Gordon (Washington University, S t Louis, MO), and Ray Wu (Cornel1 University, Ithaca, NY), respectively. Single-stranded 32P-labeledprobes of varying specific activity were synthesized after subcloning the cDNA's into the bacteriophage vector M13. Briefly, the following fragments were subcloned after digesting the cDNAs with the indicated restriction endonucleases as follows. Rat 7a-hydroxylase, 246-bp BamHI-Xbal fragment (nucleotides 1188-1434); rat HMG-CoA reductase, 420-bp Pstl fragment (nucleotides 104-524); rat LDL receptor, 186-bp HincII-Sac1 fragment (nucleotides 2300-2486), rat FABP, 335-bp PuuII-EcoRI fragment (nucleotides 11-346); rat GAPDH, 447-bp Sau3A-TqI fragment (nucleotides 754-1201). Probes were synthesized as detailed previously (30), using 0.5 p~ [32P]dCTP(3000 Ci/mmol, Amersham Radiochemicals, Arlington Heights, IL) and varying concentrations of unlabeled dCTP (1 p~ for HMG-CoA reductase, 2 p~ for LDL receptor and 7a-hydroxylase, 75 p M for GAPDH and 300pM for FABP). Synthesis of excess radiolabeled cDNA probe for hybridization with the abundant mRNAs (GAPDH and FABP) was ensured by increasing the amount of single-stranded template. The sizes of the full-length, undigested single-stranded cDNA probes, including M13 sequences, were 7a-hydroxylase = 314 nucleotides; HMG-CoA reductase = 495 nucleotides; LDL receptor = 241 nucleotides; FABP = 395 nucleotides; and GAPDH = 530 nucleotides. Samples of rat liver were homogenized in guanidinium isothiocyanate, and RNA was isolated by centrifugation on cesium chloride. Total RNA(20-40pg)was hybridized with the 32P-labeledcDNA probes simultaneously at 48 'C overnight. Unhybridized probe, present in excess relative to the amount of specific mRNA, was then digested with 20-40 units of mung bean nuclease as described previously (30). The mRNA-protected 32P-labeledprobes were separated on 7 M urea, 6% polyacrylamide gels together with 32P-labeledMspldigested pBR322 size standards and identified by autoradiography. The radiolabeled bands were excised and assayed for radioactivity by liquid scintillation spectroscopy. Identically sized bands from samples containing no RNA were also counted as a measure of nonspecific background radioactivity. The levels of FABP and GAPDH mRNA did not vary with dietary changes and were used to correct for any procedural losses. Determination of Liver and Plasma Cholesterol Distribution-Hepatic esterified and unesterified cholesterol were separated using silicic acid/celite columns and quantified by capillary gas-liquid chromatography (34). The cholesterol distribution in plasma was determined by simultaneously centrifuging plasma at densities of1.020 and 1.063 g/ml.The cholesterol in the top one-third and bottom twothirds of each tube was determined colorimetrically (34). Determination of Bile Salt Pool Size and Composition-The size and composition of the enterohepatic pool of bile salts was determined as previously described (35). Individual bile salts were separated and quantified by HPLC using a 2 5 0 - m ~018reversed-phase column (Hypersil ODs, Alltech) and a Waters 410 differential refractometer.
RESULTS Effect of Bile Salts on Responses to Dietary Cholesterol-
Bile salts greatly influenced the effect of dietary cholesterol on liver and plasma cholesterol concentrations. As shown in Fig. 1, dietary cholesterol alone had relatively little effect on liver or plasma cholesterol levels in the rat, even when fed in large quantities (2% by weight). Similarly, bile salts had essentially no effect on liver or plasma cholesterol levels when fed to animals on a low cholesterol diet. However,when administered to animals on a high cholesterol diet, cholate (a relatively lipophilic bile salt) greatly increased liver and
in Vivo
Regulation of Metabolism Hepatic Sterol
5586
plasma cholesterol levels, whereas ursodeoxycholate (a hydrophilic bile salt) had the opposite effect. T o investigate the mechanistic basis for this striking interaction, experiments were carried out to examine the effects of dietary bile salts and cholesterol on the major pathways that determine net sterol balance across the liver and on the processes that determine the concentration of LDL in plasma. Groups of animals were fed 0.5% cholate or ursodeoxycholate in the presence or absence of 2% cholesterol for 6 weeks. Weight gain was identical in all groups of animals. Table I summarizes the effect of the experimental diets on liver weight and on the distribution of cholesterol in liver and plasma. Final liver weight wassimilar in all groups with the exception of the cholate plus cholesterol-fed group where liver weights were 40% higher. Bile salts had no effect on hepatic cholesteryl esters when added to a low cholesterol diet. However, when added to a high cholesterol diet, cholate greatly increased hepatic cholesteryl esters, whereas ursodeoxycholate had the opposite effect. None of the experimental diets significantly altered the unesterified cholesterol content of the liver. Plasma cholesterol was significantly elevated only in
300
200
100
0.0
0.5
1.0
1.5
I 2.0
Dietary Cholesterol (% by wt)
FIG.1. Effect of dietary bile salts and cholesterol on liver and plasma cholesterol. Animals were feddiets supplemented with varying amounts of cholesterol (0-2%) in the presence or absence of 0.5% cholate or ursodeoxycholate for 6 weeks. Each value represents the mean f 1 S.D. for data obtained in six animals.
the group fed cholate plus cholesterol and this increase was due to increases in both LDL and 6-very low density lipoprotein. Parallel Regulation of 7a-Hydroxylase Activity and mRNAIn the rat,conversion of cholesterol to bile salts accounts for about 80% of all cholesterol eliminated from the body. Thus, changes in the rate of bile salt synthesis will markedly alter sterol balance across the liver. The initial and rate-limiting step in this pathway is catalyzed by 7a-hydroxylase. The first group of experiments was carried out to determine the effect of dietary bile salts andcholesterol on 7a-hydroxylase activity and mRNA levels. ?a-hydroxylase mRNA levels were determined using a nuclease protection assay. An autoradiogram depicting changes in 7a-hydroxylase mRNA levels is shown in Fig. 2 and the mean changes in 7a-hydroxylase activity and mRNA levels are summarized in Fig. 3. In the absence of dietary cholesterol, 0.5% dietary cholate suppressed 7a-hydroxylase activity by >go%, whereas ursodeoxycholate only modestly decreased 7a-hydroxylase activity (mean 29% suppression). In contrast, dietary cholesterol (2%) increased 7a-hydroxylase activity by approximately 2-fold. Of note, dietary cholesterol completely abolished the effect of cholate on ?a-hydroxylase activity. Indeed, when added to a high cholesterol diet, cholate tended to further increase 7a-hydroxylase activity. Thus, the massive accumulation of liver and plasma cholesterol inanimals fed cholate and cholesterol could not be attributedto suppression of 7a-hydroxylase activity. As shown by the solid barsin Fig. 3, changes in7ahydroxylase activity were accompanied by similar changes in 7a-hydroxylase mRNA, suggesting that regulation of gene transcription was the major mechanism modulating enzyme activity under the conditions studied. The changes in 7ahydroxylase activity were highly correlated with the changes in mRNA levels ( r 2= 0.85). Only in animals fed cholate plus cholesterol was the change in 7a-hydroxylase activity not entirely accounted for by a parallel change in 7a-hydroxylase mRNA. It is possible, therefore, that posttranscriptional events or alterations in substrate availability may have contributed to the regulation of 7a-hydroxylase activity under these conditions. Regulation of Endogenous Sterol Synthesis by Dietary Bile Salts and Cholesterol-The rat has an exceptionally high basal rate of cholesterol synthesis that can be regulated over a wide range in response to changes in sterol balance across the liver. T o determine the magnitude of compensatory changes, absoluterates of hepatic cholesterol synthesisand HMG-CoA reductase mRNA levels were measured in animals fed bile salts and cholesterol. A representative autoradiogram demonstrating changes in HMG-CoA reductase mRNA levels is
TABLE I Effect of dietary bile salts and cholesterol on liver and plasma cholesterol Animals were fed diets supplemented with 0.5% cholate or ursodeoxycholate in the presence or absence of 2% cholesterol for 6 weeks. Each value represents the mean f 1 S.D. for data obtained in six animals. cholesterol distribution cholesterol Plasma Liver weight
Liver
Diet
Unesterified Esterified g/100 g body wtmg/g
Control Cholate Ursodeoxycholate Cholesterol Cholesterol + cholate Cholesterol ursodeoxycholate
+
3.5 f 0.3 3.7 f 0.4 3.5 f 0.4 3.9 f 0.3 5.1 f 0.4" 3.8 f 0.3
d < 1.020, g/ml
wldl
mg/g
0.2 2.1 f 0.2 0.3 2.1 f 0.4 2.2 0.2 f 0.4 2.36.1f 0.3 2.7 f75.2 0.6 2.3 f 0.4
f 0.1 f 0.1 f 0.1 f 1.3" f 8.7" 2.1 f 0.5"
3 f 1 9 5f f1 1 3 f 1 7 f 1" 265 f49 41" 3 f 1
Significantly different from the corresponding control value, p < 0.05.
d
=
1.020-1.063, g/ml
d > 1.063, g/ml
mddl W/dl
9f1 9f1 53 11 f 1 f 3" 10 f 1
45 f 5 48 f 7 47 f 7 f8 40 & 6 48 f 6
Regulation of Hepatic Sterol Metabolism in Vivo
5587
shown in Fig. 4 and the mean changes in rates of cholesterol synthesis, HMG-CoA reductase activity and mRNA levelsare summarized in Fig. 5. In the absence of dietary cholesterol, feeding 0.5% cholate suppressed hepatic cholesterol synthesis by 76%, whereas ursodeoxycholate had no effect. Dietary 622 527 cholesterol suppressed hepatic cholesterol synthesis by >95% GAPDH and the furtheraddition of cholate or ursodeoxycholate had 404 noadditional effect. Hepatic HMG-CoA reductase mRNA FABP was only modestly reduced (by -40%) in animals fed cholate 309 or cholesterol despite the marked suppression of cholesterol 7a-hydroxylase synthesis. Even in animals fed both cholate and cholesterol, FIG. 2. Measurement of hepatic 7a-hydroxylasemRNA lev- where liver cholesteryl esters were increased by >350-foldand els. Autoradiogram of undigested 32P-labeledcDNA probes (lane 1 ) hepatic cholesterol synthesis was suppressed by >98%, HMGand cDNA probes protected from mung bean nuclease digestion by CoA reductase mRNA was reduced by only 50%. rat liver mRNA (lanes 3-5). Total RNA was isolated from livers of The activity of HMG-CoA reductase can be modulated over rats fed diets supplemented with 0.5% cholate or ursodeoxycholate or a wide range by phosphorylation/dephosphorylation. To de2% cholesterol for 6 weeks, 20 pg was hybridized with 32P-labeled probes and GAPDH, FABP, and 7a-hydroxylase bands resistant to termine if the changes in cholesterol synthesis rates in animung bean nuclease digestion were identified, after separationby gel mals fed cholate or cholesterol were due to changes in the electrophoresis, as described under "Materials and Methods." amount or in the activity of HMG-CoA reductase, HMG-CoA reductase activity was measured in hepatic microsomes prepared in the presence of 50 mM fluoride (to prevent enzyme activation) or after incubation with alkaline phosphatase No Dietary Cholesterol 2% Dietary Cholesterol (fully active enzyme). Changes in HMG-CoA reductase activity paralleled the changes in rates of cholesterol synthesis in all experimental groups. In addition, the ratio of active to inactive forms of the enzyme was not affected by any of the experimental diets (data not shown). These findings suggest that regulation of cholesterol synthesis occurred at the level of HMG-CoA reductase and that the amount of reductase enzyme, rather than its specific activity, accounted for the observed changes in ratesof cholesterol synthesis. Luck of Regulation of LDL Receptor Activityby Dietary Bile Saltsand Cholesterol-Intracellular sterols exert negative feedback control onHMG-CoA reductase and theLDL receptor in vitro and in most systems these two pathways are coordinately regulated to maintain cholesterol homeostasis in the cell. In rat liver, however, cholesterol synthesis can be FIG. 3. Parallel regulation of hepatic 7a-hydroxylase activ- regulated over a wide range with no change in receptority and mRNA by dietary bile salts and cholesterol. Animals dependent LDL transport. To examine the effects of dietary were fed diets supplemented with 0.5% cholate or ursodeoxycholate bile salts andcholesterol on the LDL receptor pathway, rates in thepresence or absence of 2% cholesterol for 6 weeks. 7a-Hydrox- of receptor-dependentLDL transportand LDL receptor ylase activity was measured using endogenous microsomal cholesterol mRNA levels were measured. Absolute rates of total and as substrate. 7a-Hydroxylase activity equalled 1.58 f 0.22 nmol/h per mg of microsomal protein in the control group and activities in receptor-independent LDLuptake were measured in vivo the various experimental groups are expressed as apercentage of this using homologous and methylated human LDL, respectively. value. 7cr-Hydroxylase mRNA levels were measured in the same livers In order to relate the changes in absolute rates of LDLand are expressed as a percentage of the control levels. Each value cholesterol uptake tochanges in receptor activity, the values represents the mean & 1S.D. for data obtained in eight to tenanimals. for total and receptor-independent LDL uptake determined in theexperimental animalswere superimposed on the kinetic curves that define the relationship between LDL-cholesterol uptakeand circulating LDL-cholesterol concentrationsin control animals fed standard rodent diet. Fig. 6 shows the normal kinetic curves for hepatic LDL uptake in rats fed standard rodent diet. The shaded areas representthe relationship between total (stippled) and receptor-independent 622 527 IAPOH (hatched) LDL-cholesterol uptake and circulating LDL-choMG-CoA reductase 404 lesterol concentrations over the range of LDL levels observed ABP in these studies. Superimposed on thesenormal kinetic curves 309 are the actual rates of total and receptor-independent LDL FIG. 4. Measurement of hepatic HMG-CoA reductase uptake for animals fed cholate or ursodeoxycholate in the mRNA levels. Autoradiogram of undigested 32P-labeled cDNA probes (lane 1) and cDNA probes protected from mung bean nuclease presence or absence of cholesterol. As shown by the open digestion by rat liver mRNA (lanes 3-5). Total RNA was isolated circles, rates of total and receptor-independent LDL uptake from livers of rats fed diets supplemented with 0.5% cholate or in control animals (no added bile salts or cholesterol) fell on ursodeoxycholate or 2% cholesterol for 6 weeks, 40 pg of RNA was the normal kinetic curves for hepatic LDL transport in the hybridized with 32P-labeledprobes, and GAPDH, HMG-CoA reductase, and FABP bands resistant to mung bean nuclease digestion rat. Similarly, rates of total and receptor-independent LDL were identified after separation by gel electrophoresis as described uptake in animals fed cholate (open squares), ursodeoxycholate (open triangles), cholesterol alone (solid circles), and under "Materials and Methods."
Regulation of Hepatic Sterol Metabolismin Vivo
5588
No Dietary Cholesterol
2% Dietary Cholesterol "
I
Cholesterol Synthesis
63
125
Reductase mRNA Reductase activity
100
75 50
25
0
"
FIG. 5. Regulation of hepatic cholesterol synthesis and HMG-CoA reductase activity and mRNA by dietary bilesalts and cholesterol.Animals were fed diets supplemented with 0.5% cholate or ursodeoxycholate in the presence or absence of 2% cholesterol for 6 weeks. Rates of hepatic cholesterol synthesis were quantified i n vivo using ["]water. The rate of cholesterol synthesis equalled 2,413 f 188 nmol/h per g of liver in the control group and synthesis ratesin the various experimental groups are expressed as a percentage of this value. HMGCoA reductase activity and mRNA levels were measured in identically treated parallel groups of animals. HMGCoA reductase measured in thepresence of alkaline phosphatase equalled 58.8 f 19 nmol/h per mg of microsomal protein in the control group, and activities in the experimental groups are expressed as a percentage of this value. Each value represents the mean f 1 S.D. for data obtained in eight to ten animals.
Cholesterol None 2% 60 - Bile Salt None 0 . 0.5% Cholate 0 50 - 0.5% Urso A A
Total Uptake
~
t
40 -
-
30 -
622 527 404
20 -
309
-
10.
242
-
n "-
0
-
iAPDH ABP
.DL receptor
10
20
30 50
40
Plasma LDL Cholesterol Concentration (mgW
FIG.6. Measurement of hepatic receptor-dependent LDL uptake. Animals were fed diets supplemented with 0.5% cholate or ursodeoxycholate in the presence or absence of 2% cholesterol for 6 weeks. The shaded areas represent the kinetic curves for total (stippled) and receptor-independent (hatched) LDL-cholesterol uptake determined in control animals fed standard rodent diet as described under"Materials and Methods." Superimposed on these normal kinetic curves are the absolute rates of total andreceptor-independent LDL-cholesterol uptake in the experimental animals plotted as a function of the plasma LDL-cholesterol concentration in the same animals. Each point represents the mean f 1 S.D. for data obtained in 12 animals.
FIG. 7. Measurement of hepatic LDL receptor mRNA levels. Representative autoradiogram of undigested "P-labeled cDNA probes (lanes I ) and cDNA probes protected from mung bean nuclease digestion by rat liver mRNA (lanes 3-5). Total RNA was isolated from livers of rats fed dietssupplemented with 0.5% cholate or ursodeoxycholate or 2%cholesterol for 6 weeks, 20 pg was hybridized with 32P-labeledprobes and GAPDH, FABP,and LDL receptor bands resistantto mung bean nuclease digestion were identified, after separation by gel electrophoresis, as described under "Materials and Methods."
ArepresentativeexperimentmeasuringLDL receptor mRNA levels is shown in Fig. 7 and the mean changes in hepatic LDL receptor activity and LDLreceptor mRNA levels are summarized in Fig. 8. Receptor-dependent LDLtransport cholesterol plus ursodeoxycholate (solid squares) were not was not significantly altered by any of the dietary manipuladisplaced from the normal kineticcurves indicating that these tions. Even in the livers of animals fed both cholate and experimental conditions hadno significant effect on receptor- cholesterol, where cholesteryl ester levels were elevated >350dependent or receptor-independent LDL transport by the fold and cholesterol synthesis was suppressed by >98%, recepliver. In animals fed cholate plus cholesterol (solid triungks), tor-dependent LDL transport was reduced by only 25-30% total LDL-cholesterol uptake by the liver equalled 35 pg/h when expressed per g liver. Since liver weight increased by per g at a plasma LDL-cholesterol level of 49 mg/dl, whereas nearly 40% inanimals fed cholate with cholesterol, total normal animals would transport -47 pg/h per g at this LDL hepatic LDL receptor activity was essentially unchanged in concentration.Sincereceptor-independentLDL transport these animalsand the5-fold increase in circulating LDL levels was normal in these animals, the reduction in total LDL- was the result of a nearly &fold increase in the rate of LDL cholesterol uptake was due entirelyto a reduction in receptor- formation, ratherthan suppression of receptor-dependent dependent transport, which was suppressed -27% relative to LDL clearance by the liver (data not shown). Hepatic LDL receptor mRNA was modestly reduced (by 30-40%) in animals normal animals at thesame LDL-cholesterol concentration.
Regulation ofVivo Metabolism Hepatic Sterol in fed cholate or ursodeoxycholate, either alone or in combination with cholesterol. Thus, mRNA levels for HMG-CoA reductase and LDL receptor changed in parallel and to the same extent under the dietary conditions studied, whereas actual rates of cholesterol synthesis and receptor-dependent LDL transport were not coordinately regulated. Changes in Bile Salt Pool with Dietary SupplemntatwnSince dietarybile salts may undergo biotransformation inthe intestine andliver, the effect of the experimental dietson the size and composition of the bile salt pool wasdetermined and these data are summarized in Table 11. In control animals, the total bile salt pool size equalled 402 pmol per kg of body weight and themajor bile salts were muricholate, cholate, and chenodeoxycholate. Feeding cholate or ursodeoxycholate increased the total bile salt pool size by -35% and markedly altered the composition of the pool with lipophilic bile salts (cholate and deoxycholate) predominating inanimals fed cholate and hydrophilic bile salts (muricholateand ursodeoxycholate) predominating inanimals fed ursodeoxycholate. Dietary cholesterol increased the size of the bile salt pool by -45% and, in addition, altered the composition of the pool toward more hydrophilic species. DISCUSSION
The ratis resistant to theeffects of dietary cholesterol but does become markedly hypercholesterolemic when fed cholesNo Dietary Cholesterol 2% Dietary Cholesterol
FIG. 8. Regulation of hepatic LDL receptor activity and mRNA by dietary bile salts and cholesterol. Animals were fed diets supplemented with 0.5% cholate or ursodeoxycholate in the presence or absence of 2% cholesterol for 6 weeks. Hepatic LDL receptor activity was determined as illustrated inFig. 6. LDL receptor mRNA levels were measured in identically treated parallel groups of animals and are expressed as a percentage of the control levels. Each value represents the mean f 1 S.D. for data obtained in eight to ten animals.
5589 terol in combination with lipophilic bile salts. These studies were undertaken to determine how dietary cholesterol and bile salts interact to regulate 7a-hydroxylase activity, cholesterol synthesis, and receptor-dependent LDL transport in the liver and to examine the mechanisms involved in the regulation of these pathways. It is generally accepted that bile salts returning to the liver in the enterohepatic circulation exert negative feedback control on 7a-hydroxylase activity (11, 12) and that lipophilic bile salts are much more active in this regard than are hydrophilic bile salts (14). In the present studies, cholate suppressed 7a-hydroxylase activity by more than 90%, whereas ursodeoxycholate had little effect. Since conversion of cholesterol to bile salts is the major route for cholesterol removal from the body, regulation of this pathway by bile salts markedly alters sterol balance across the liver and wholebody. When dietary cholate decreased bile salt synthesis, sterol balance was completely restored by appropriate changes in the rateof hepatic cholesterol synthesis and hepatic cholesteryl ester levels, hepatic LDL receptor activity, and plasma LDL concentrations all remained unchanged. Similarly, when faced with a large influx of dietary cholesterol, the liver compensated by completely suppressing de nouo cholesterol synthesis and, in addition, by increasing 7a-hydroxylase activity. Together,these two mechanisms were nearly fully compensatory since the liver accumulated only modest amounts of cholesterol and hepatic LDL receptor activity and plasma LDL concentrations remained unchanged. Whereas dietary cholesterol and cholate had relatively little effect on liver and plasma cholesterol levels when fed individually, the two in combination caused a massive accumulation of cholesterol in the liver (>350-fold increase in cholesteryl esters) and a marked increase in plasma cholesterol concentrations. Augmentation of 7a-hydroxylase is an important mechanism for dealing with increasing loads of dietary cholesterol in the rat and, at least in animals fed a low cholesterol diet, cholate almost totally suppressed this excretory pathway. A plausible explanation for the cholesterol accumulation in animals fed cholesterol plus cholate, therefore, was that cholate inhibitedthe major excretory pathway by which the liver compensates for increasing loads of dietary cholesterol. However, in contrast to the marked suppression (>go%) in 7ahydroxylase activity that occurred when cholate was fed to animals on a low cholesterol diet, cholate did not suppress 7a-hydroxylase activity when fed to animals on a high cholesterol diet. Thus, thestriking elevations in liver and plasma cholesterol levels when cholate was fed to animals on a high cholesterol diet could not be attributed to suppression of 7ahydroxylase. The changes may be due, at least in part, to the known ability of cholate to increase cholesterol absorption (21, 36).
TABLEI1 Effect of dietary bile salts and cholesterol on the size and compositionof the bile salt pool Animals were fed diets supplemented with 0.5% cholate or ursodeoxycholate in the presence or absence of 2% cholesterol for 6 weeks. Each value represents the mean f 1 S.D. for four animals. Diet
Total pool size rmollkg
%
Control 55 Cholate Ursodeoxycholate Cholesterol 31 Cholesterol cholate Cholesterol ursodeoxycholate
402 f 38 535 +. 43 540 f 49 612 f 72 796 f 88 807 f 99
33 8 22 57 26 23
+ +
Muricholate
Composition of bile salt pool Ursodeoxycholate Cholate Chenodeoxycholate %
%
39
81 27
53
59 14
Deoxycholate
%
%
11 2 11 12 9 9
1 9 1 6 1
5590
Regulation of Hepatic Sterol Metabolism in Vivo
The precise mechanism by which dietary bile salts and cholesterol regulate 7a-hydroxylase activity is not known. In the present studies, changes in 7a-hydroxylase activity were accompanied by parallel changes in 7a-hydroxylase mRNA levels under all experimental conditions that were examined. These results are consistentwith recent studiesshowing that suppression of 7a-hydroxylase activity by lipophilic bile salts and stimulation by dietary cholesterol are due largely to changes in gene transcription (20). How cholesterol and bile salts interact to regulate 7a-hydroxylase gene transcription is unknown. One possibility is that cholesterol and lipophilic bile salts independently exert positive and negative control, respectively, on gene expression. A second possibility is that cholesterol regulates gene transcription and that the effect of bile salts is secondary to theireffects on cholesterol availability. This has been suggested previously, based on the observation that cholesterol, but not bile salts, regulates 7a-hydroxylase activity in isolated hepatocytes (37-39). Clearly, the effect of cholesterol dominates in rat liver and overall, 7ahydroxylase activity and mRNA levels correlated better with the sterol content of the liver than with the size or composition of the enterohepatic pool of bile salts. On the otherhand, dietary cholesterol did shift the composition of the enterohepatic pool toward more hydrophilic bile salts (which do not suppress 7a-hydroxylase activity), anditis possible that regulation of 7a-hydroxylase by dietary cholesterol is in part secondary to alterations in the composition of the bile salt pool. The contribution of posttranscriptional mechanisms to the regulation of 7a-hydroxylase expression appears to be minimal in the current studies. Only in animals fed cholate plus cholesterol did the change in 7a-hydroxylase activity exceed the change in 7a-hydroxylase mRNA. In contrast, bile diversion for 48 h was recently reported to increase bile salt synthesis predominantly by increasing the catalytic activity of 7a-hydroxylase, with only a small increase in enzyme mass, although intravenous infusion of taurocholate decreased both enzyme activity and mass proportionally, consistent with transcriptional control (40). Although 7a-hydroxylase appears to be regulated largely by changes in mRNA levels, regulation of HMG-CoA reductase is known to occur at multiple levels (3). In rat liver, both diurnal changes in HMG-CoA reductase and induction of enzyme activity by cholestyramine and mevinolin are accompanied by parallel changes in mRNA levels (41, 42). In contrast, the marked reductions in rates of hepatic cholesterol synthesis in animals fed cholate (75% decrease) or cholesterol (95% decrease) were accompanied by only modest reductions (-40%) in HMG-CoA reductase mRNA levels. Even under circumstances where hepatic sterol synthesis was suppressed >98% and cholesteryl ester levels were increased >350-fold (animals fed cholesterol and cholate), HMG-CoA reductase mRNA was reduced by only 50%. Changes in HMG-CoA reductase activity measured in microsomes paralleled changes in rates of cholesterol synthesis measured in vivo using [3H] water; there were no significant differences in the ratio of activities measured in the presence of fluoride or in the presence of alkaline phosphatase, suggesting that regulation of synthesis occurred at thelevel of HMG-CoA reductase and that a reduction in enzyme protein rather than changes in enzyme specific activity was responsible for the suppression of cholesterol synthesis. Thus, regulation of HMG-CoA reductase by dietary cholesterol and lipophilic bile saltsis largely posttranscriptional in rat liver, as recently reported (43), although there is a modest contribution from inhibition of mRNA levels.
In cultured cells, HMG-CoA reductase activity and LDL receptor number are coordinately regulated so as tomaintain cholesterol homeostasis (2-5). In ratliver, however, lipophilic bile salts and cholesterol suppressed hepatic sterol synthesis but had noeffect on receptor-dependent LDL transport. Even in the case where hepatic cholesteryl ester levels were massively elevated and hepatic sterol synthesis totallysuppressed (animals fed cholate plus cholesterol), receptor-dependent LDL uptake was reduced by only 25-30% when expressed per g of liver and was virtually unchanged when expressed per whole liver. In contrast to rates of cholesterol synthesis and receptor-dependent LDL transport, which were completely dissociated under these experimental conditions, mRNA levels for HMG-CoA reductase and the LDL receptor were regulated in parallel. Sterol-mediated regulation of mRNA levels in rat liver in vivo thus approximates the regulatory mechanisms observed in cultured cells such as thefibroblast, although the magnitude differs greatly. The LDL receptor pathway in rat liver in vivo is resistant to dietary sterols even when hepatic and plasma cholesterol levels are markedly elevated. The explanation for this resistance is unknown. The basal activity of the LDL receptor pathway and sensitivity to cellular sterols may vary among different cell types. Indeed, the LDL receptor pathway does appear to be more resistant to sterols incultured hepatocytes than in cultured fibroblasts (4, 5, 44-46). However, hepatic LDL receptor activity is regulated by dietary sterols in other species such as the hamster (27). This may be due in part to less effective mechanisms for dealing with dietary cholesterol in this species. For example, the basal rate of hepatic cholesterol synthesis is extremely low in the hamster, so that only minimal amounts of dietary cholesterol can be compensated for by suppression of de novo synthesis (23). In addition, the hamster may be unable to increase rates of bile salt synthesis in response to dietary cholesterol. These differences not withstanding, the LDL receptor pathway in rat liver is not suppressed at liver cholesterol levels that are associated with marked suppression of LDL receptor activity in the hamster (27). This may reflect species differences in the promoter region of the LDL receptor gene or in other factors regulating mRNA levels. Rates of hepatic cholesterol synthesis, receptor-dependent LDL transport, and bile salt synthesis are regulated so as to maintain cholesterol balance across the liver. Suppression of hepatic cholesterol synthesis appears to be the primary response to increasing dietary cholesterol loads in the rat and probably in other species as well (6). If this mechanism is inadequate to restore cholesterol balance, additional compensatory mechanisms are called into play. In many species, down-regulation of the LDL receptor pathway occurs, whereas other species, including the rat, respond by increasing the rate of bile salt synthesis. Such differences in the regulation of these three pathways may contribute to themarked variability in response to dietary cholesterol seen in different species and even in different individuals within the same species. Acknowledgments-We thank Drs. David Russell, Diane Jelinek, Sohaib Khan, Richard Tanaka, Jeffrey Gordon, and Ray Wu for making cDNAs available and Russell Nelson, Jody Houston, and Chris White for excellent technical assistance. REFERENCES 1. Turley, S. D., and Dietschy, J. M. (1988)in The Liver: Biology and Pathobiology (Schacter, D., and Shafritz, D.A., e&) pp. 617-641,Raven Press, Ltd., New York 2. Goldstein, J. L., Brown, M. S., Anderson, R. G . W., Russell, D. W., and Schneider, W. J. (1985)Annu. Rev. Cell BioL 1,l-39
Regulation of Hepatic St:erol Metabolism in Vivo
5591
3. Goldstein, J. L., and Brown, M. S. (1990)Nature 343,425-430 T. L., and Holcombe, K. S. (1980)Proc. Nutl. Acad. Sci. U.S. A. 77,225-229 4. Goldstein, J. L., and Brown, M. S. (1977)Annu. Rev. Biochem. 27. Spady, D. K., and Dietschy, J. M. (1988)J. Clin. Invest. 81,30046,897-930 309 5. Molowa, D. T., and Cimis, G.M. (1989)Biochem. J. 260, 73128. Hylemon, P. B., Studer, E. J., Pandak, W. M., Heuman, D. M., 736 Vlahcevic, Z. R., and Chiang, J. Y. L. (1989)Anal. Biochem. 6. Spady, D. K., Turley, S. D., and Dietschy, J. M. (1985)J. Lipid 182,212-216 Res. 26,465-472 7. Spady, D. K., and Dietschy, J. M. (1989)J. Lipid Res. 30, 559- 29. Williams, D. L., Newman, T. C., Shelness, G. S., and Gordon, D. A., (1986)Methods Enzymol. 128,671-689 569 8. Spady, D. K., Bilheimer, D. W., and Dietschy, J. M. (1983)Proc. 30. Cuthbert, J. A., and Lipsky, P. E. (1990)J. Lipid Res. 31,20672078 Nutl. Acad. Sci. U.S. A. 80, 3499-3503 9. Spady, D. K., Meddings, J. B., and Dietachy, J. M. (1986)J. Clin. 31. Lee, L. Y., Mohler, W.A., Schafer, B. L., Freudenberger, J. S., Byme-Connolly, N., Eager, K. B., Mosley, S. T., Leighton, J. Invest. 77,1474-1481 K., Thrift, R. N., Davis, R.A., andTanaka, R. D. (1989) 10. Spady, D. K., Huettinger, M., Bilheimer, D. W., and Dietschy, J. Nucleic Acids Res. 17,1259-1260 M. (1987)J. Lipid Res. 28,32-41 32. Gordon, J. I., Alpers, D. H., Ockner, R. K., and Strauss, A. W. 11. Danielsson, H., Einarsson, K., and Johansson, G. (1967).Eur. J. (1982)J. Bwl. Chem. 268,3356-3363 Biochem. 2,44-49 33. Tso, J. Y.,Sun, W.-H., Kao, T.-h., Reece,K. S., and Wu, R. 12. Shefer, S., Hauser, J., Berkersky, J., and Mosbach, E. H. (1970) (1985)Nucleic Acids Res. 13,2485-2502 J. Lipid Res. 11,404-411 34. Andersen, J. M., and Dietschy, J. M. (1978)J. Bwl. Chem. 263, 13. Myant, N. B., and Mitropoulos, K. A. (1977)J. Lipid Res. 18, 9024-9032 135-153 35. Turley, S. D., Spady, D. K., and Dietschy, J. M. (1983)Gastro14. Heuman, D. M., Hylemon, P. B., and Vlahcevic, Z. R. (1989)J. enterology 84,253-264 Lipid Res. 30,1161-1171 36. Raicht, R. F., Cohen, B. I., and Mosbach, E. H. (1974)Gastroen15. Straka, M. S., Junker, L. H., Zacarro, L., Zogg, D. L., Dueland, terology 67, 1155-1161 S., Everson, G. T., and Davis, R. A. (1990)J. Biol. Chem. 265, 37. Arebalo, R. E., Tormanen, C. D., Hardgrave, J. E., Noland, B. J., 7145-7149 and Scallen, T. J. (1982)Proc. Nutl. Acad. Sci. U.S. A. 79,5116. Jelinek, D. F., Andersson, S., Slaughter, C.A., and Russell, D. 55 38. Davis, R. A., Hyde, P. M., Kuan, J.-C. W., Malone-McNeal, M., W. (1990)J. Biol. Chem. 265,8190-8197 and Archambault-Schexnayder,J. (1983)J. Biol. Chem. 258, 17. Noshiro, M., Nishimoto, M., Morohashi, K., and Okuda, K. (1989) 3661-3667 FEBS Lett. 257,97-100 39. Kubaska, W. M., Gurley, E. C., Hylemon, P. B., Guzelian, P. S., 18. Noshiro, M., and Okuda, K. (1990)FEBS Lett. 268, 137-140 and Vlahcevic, Z. R. (1985)J. Bwl. Chem. 260, 13459-13463 19. Li, Y. C., Wang, D. P., and Chiang, J. Y.L. (1990)J. Biol. Chem. 40. Shefer, S., Nguyen, L.B., Salen, G., Ness, G. C., Tint, G. S., 265, 12012-12019 Batta, A.K., Hauser, S., and Rani, I. (1991)J. Bwl.Chem. 20. Pandak, W. M., Li, Y. C., Chiang, J. Y. L., Studer, E. J., Gurley, 266,2693-2696 E. C., Heuman, D.M., Vlahcevic, Z. R., and Hylemon, P. B. 41. Clarke, C. F., Fogelman, A. M., and Edwards, P. A. (1984)J. Bwl. (1991)J. Bwl. Chem. 266,3416-3421 21. Cohen, B. I., Raicht, R. F., and Mosbach, E. H. (1977)J. Lipid 42. Chem. 259,10439-10447 Clarke, C.F., Edwards, P. A., Lan, S.-F., Tanaka, R.D., and Res. 18,223-231 Fogelman, A.M. (1983)Proc. Nutl. Acad. Sci. U.S. A. 80, 22. Jeske, D. J., and Dietschy, J. M. (1980)J. Lipid Res. 21, 3643305-3308 376 43. Ness, G.C., Keller, R. K., and Pendleton, L. C. (1991)J. Bwl. 23. Spady, D. K., and Dietschy, J. M. (1983)J. Lipid Res. 24, 303Chem. 266,14854-14857 315 44. Edge, S. L., Hoeg,J. M., Triche, T., Schneider, P. D., and Brewer, 24. Brown, M. S., Goldstein, J. L., and Dietschy, J. M. (1979)J. Bwl. H. B., Jr. (1986)J. Bwl. Chem. 261,3800-3806 Chem. 254,5144-5149 45. Salter, A. M., Bugaut, M., Saxton, J., Fisher, S. C., and Brindley, 25. Glass, C. K., Pittman, R.C., Keller, G.A., and Steinberg, D. D. N. (1987)Biochem. J. 247, 79-84 (1983)J. BwL Chem. 258, 7161-7167 46. Tamai, T., Patsch, W., and Schonfeld, G. (1988)Atherosclerosis 26. Mahley, R. W., Weisgraber, K. H., Melchior, G. W., Innerarity, 69,29-37
