INTRODUCTION

Conversion of cholesterol to bile salts is the principal regulated pathway whereby cholesterol is eliminated from the body. The first and rate-limiting step in this pathway is catalyzed by hepatic cholesterol 7-hydroxylase (1). Hepatic 7-hydroxylase activity appears to be regulated at the transcriptional level by the quantity of bile salts fluxing through the enterohepatic circulation (2, 3). Thus, interventions that accelerate bile salt loss from the enterohepatic circulation (thereby reducing bile salt pool size) lead to derepression of hepatic 7-hydroxylase expression and enhanced bile salt synthesis. Conversely, expansion of the enterohepatic pool of bile salts leads to suppression of 7-hydroxylase activity and reduced rates of bile salt formation.

How bile salts fluxing through the enterohepatic circulation regulate the expression of hepatic 7-hydroxylase remains unclear. In the bile-diverted rat, depletion of the enterohepatic pool of bile salts leads to derepression of hepatic 7-hydroxylase activity and an increase in the rate of bile salt synthesis, effects that can be prevented by the intraduodenal administration of bile salts (4, 5). The conventional interpretation of these experiments is that the flux of bile salts through the hepatocyte (or, more likely, the concentration of bile salts within the hepatocyte) is being sensed, resulting in feedback repression of the rate-limiting enzyme in the bile salt biosynthetic pathway. However, early studies using hepatocyte suspensions or cultures failed to demonstrate any direct regulatory effect of exogenous bile salts on the secretion of bile salts into the media (6, 7). Moreover, additional studies in the rat failed to support a direct role for bile salts in the regulation of hepatic 7-hydroxylase expression. Thus, obstruction of the common bile duct, which leads to an accumulation of bile salts within the liver, paradoxically increases hepatic 7-hydroxylase activity and bile salt synthesis in the rat (8, 9, 10). In addition, the creation of a thoracic duct lymph fistula in rats with an intact enterohepatic circulation leads to derepression of hepatic 7-hydroxylase, suggesting the presence of an intestinally derived inhibitory factor carried in lymph (11). Recently, Pandak et al. (12) administered equimolar amounts of taurocholate intravenously or intraduodenally to bile-diverted rats and found that intraduodenal, but not intravenously administered, taurocholate exerted negative feedback inhibition on hepatic 7-hydroxylase expression. Together, these data suggest that bile salts may have no direct effect on hepatic 7-hydroxylase expression. Rather, these studies suggest that an intestinal factor, released or absorbed in the presence of bile salts, may play a role in the regulation of bile salt synthesis.

Most of the work cited above has been carried out in the rat, an animal model that differs significantly from humans with respect to sterol and lipoprotein metabolism. In contrast to the rat, the hamster more closely resembles humans with respect to basal rates of cholesterol and bile salt synthesis (13), composition of the bile salt pool (14, 15), various aspects of lipoprotein transport, and response to dietary and pharmacologic interventions (13, 16, 17). Using adenovirus-mediated gene transfer, we recently noted that primary overexpression of an exogenous 7-hydroxylase gene in the hamster resulted in reciprocal down-regulation of the endogenous gene, suggesting that bile salts may directly regulate expression of the 7-hydroxylase gene in this species (18). In these studies, however, primary overexpression of the exogenous 7-hydroxylase gene was accompanied by expansion of the enterohepatic pool of bile salts, raising the possibility that suppression of the endogenous gene may have resulted from an increased flux of bile salts in the expanded enterohepatic pool.

The purpose of the present work was to determine if bile salts directly regulate hepatic 7-hydroxylase expression in the bile-diverted hamster, an animal model that more closely resembles the human situation than does the rat (13, 14, 15). Three independent approaches were used to increase bile salt availability within the liver: 1) overexpression of an exogenous 7-hydroxylase gene, 2) complete biliary obstruction, and 3) intravenous infusion of taurocholate. All three approaches resulted in suppression of hepatic 7-hydroxylase expression, indicating that bile salts suppress their own synthesis at the level of the hepatocyte through feedback repression of 7-hydroxylase expression.

MATERIALS AND METHODS

Male Golden Syrian hamsters (Sasco, Inc., Omaha, NE) were housed in colony cages and subjected to light cycling for at least 3 weeks prior to use in specific experiments. During this time, animals were maintained on commercial rodent diet (Wayne Lab Blox, Allied Mills, Chicago, IL). Under brief diethyl ether anesthesia, gallbladder bile was aspirated, the cystic duct was tied, and a polyurethane cannula (0.012-inch inner diameter, Braintree Scientific, Braintree, MA) was placed in the common bile duct. In addition, each animal was fitted with a femoral vein catheter. The animals were placed in individual cages and received a glucose-electrolyte infusion at the rate of 1.2-2 ml/h. The infusate contained glucose (10%), NaCl (80 mM), KCl (20 mM), K₂HPO₄ (20 mM), and NaHCO₃ (30 mM). The enterohepatic pool was allowed to drain out for 12-14 h. After an additional 24 h of biliary diversion, animals were sacrificed, and aliquots of plasma and liver were taken for specific determinations. Timed bile collections were obtained throughout the 36-h period of biliary diversion.

Generation of the recombinant adenoviruses AdCMV7 (carrying a gene encoding rat 7-hydroxylase), AdCMVLuc (carrying a gene encoding firefly luciferase), and AdCMVgal (carrying a gene encoding a nuclear localizing variant of Escherichia coli -galactosidase) have been previously described (18, 19). Large scale production of recombinant adenovirus was carried out by infecting confluent monolayers of 293 cells grown in 15-cm tissue culture plates with primary stock at a multiplicity of infection of 0.1-1.0 (20). Infected monolayers were lysed with Nonidet P-40 (final concentration 0.1%) when >90% of the cells showed cytopathic changes. Recombinant virus was purified by precipitation with polyethylene glycol 8000, centrifugation on a discontinuous d = 1.3-1.4 g/ml density gradient, and desalting by chromatography on Sepharose CL4B in an isotonic saline buffer (10 mM Tris-HCl, pH 7.4, 137 mM NaCl, 5 mM KCl, 1 mM MgCl₂). Purified virus eluting in the void volume was collected and, after the addition of sterile bovine serum albumin to a final concentration of 0.1 mg/ml, snap-frozen in liquid nitrogen and stored at 80° until used. The titer of infectious viral particles in purified stocks was determined by plaque assay in monolayers of 293 cells (21) and was routinely >10¹⁰ plaque-forming units (pfu)¹/ml.

Determination of Hepatic 7-Hydroxylase Activity

Hepatic 7-hydroxylase activity was measured using an HPLC-spectrophotometric assay that quantifies the mass of 7-hydroxycholesterol formed from endogenous microsomal cholesterol after enzymatic conversion to 7-hydroxy-4-cholesten-3-one using cholesterol oxidase (22).

One-half g aliquots of liver, along with tracer amounts of sodium tauro[¹⁴C]cholate as an internal standard, were refluxed at 80 °C in absolute ethanol containing 0.1% NH₄OH for 2 h. After evaporation of the ethanol, the residue was resuspended in 9 ml of 0.1 N NaOH and applied to a prewetted C18 Bond Elut^® column (Varian, Harbor City, CA). The Bond Elut^® column was washed with 5 ml of 0.1 N NaOH, 10 ml of H₂O, 10 ml of Hexane, and 10 ml of H₂O. Bile salts were eluted from the column with 20 ml of methanol/H₂0, 75:25. After evaporation, the dried residue was resuspended in 500 µl of buffered saline (pH 7) containing 50 mg/ml bovine serum albumin. Fifty-µl aliquots were taken for the determination of total bile salts using a 3-hydroxysteroid dehydrogenase/diaphorase enzymatic kit (Sigma). The same kit was used to quantify total bile salts in plasma and in diluted aliquots of bile. The bile salt composition of bile was determined by HPLC. Individual bile salts were separated on a 4.6 × 250-mm Hypersil ODS (C18) column (Altech, Deerfield, IL) and detected using a Waters 410 differential refractometer.

Rates of hepatic cholesterol synthesis were measured in vivo using [³H]water. As described previously (23), the animals were administered ~100 mCi of [³H]water intravenously and then returned to individual cages under a fume hood. One h after the injection 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 nanomoles of [³H]water incorporated into digitonin-precipitable sterols per hour per gram of liver (nmol/h/g).

Plasma was obtained from normocholesterolemic hamster and human donors. The LDL was isolated from plasma by preparative ultracentrifugation in the density range of 1.020-1.055 g/ml and labeled with ¹²⁵I- or ¹³¹I-tyramine cellobiose as described previously (24). The human LDL was also reductively methylated to completely eliminate its recognition by the LDL receptor (25). Rates of hepatic LDL uptake were measured using primed infusions of ¹²⁵I-tyramine cellobiose-labeled LDL (26). The infusions of ¹²⁵I-tyramine cellobiose-labeled LDL were continued for 4 h, at which time each animal was administered a bolus of ¹³¹I-tyramine cellobiose-labeled LDL as a volume marker and killed 10 min later by exsanguination through the abdominal aorta. Tissue samples along with aliquots of plasma were assayed for radioactivity in a -counter (Packard Instrument Co., Inc., Downers Grove, IL). The amount of labeled LDL in the various tissues at 10 min (¹³¹I disintegrations per minute per gram of tissue divided by the specific activity of ¹³¹I in plasma) and at 4 h (¹²⁵I disintegrations per minute per gram of tissue divided by the specific activity of ¹²⁵I in plasma) was then calculated. The increase in the tissue content of LDL cholesterol or LDL protein with time represents the rate of LDL uptake in micrograms of LDL cholesterol or LDL protein taken up per hour per gram of tissue.

Hepatic 7-hydroxylase, HMG-CoA reductase, LDL receptor, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, used as an invariant control) mRNA levels were determined by nuclease protection as described previously (18). Probes were synthesized using 0.5 µM [³²P]dCTP and 1 µM (7-hydroxylase, HMG-CoA reductase, and LDL receptor) or 300 µM (GAPDH) unlabeled dCTP.

Samples of hamster liver were homogenized in guanidinium thiocyanate, and the RNA was isolated by the method of Chomczynski and Sacchi (27). Total RNA (40 µg) was hybridized with the ³²P-labeled cDNA probes simultaneously at 48 °C overnight. Unhybridized probe, present in excess relative to the amount of specific mRNA, was then digested with 40 units of mung bean nuclease (Life Technologies, Inc.). The mRNA-protected ³²P-labeled probes were separated on 7 M urea, 6% polyacrylamide gels together with ³²P-labeled MspI-digested pBR322 size standards. The radioactivity in each band, as well as background radioactivity, was quantified using an isotopic imaging system (Ambis, Inc., San Diego, CA). The level of GAPDH mRNA did not vary among the various experimental groups and was used to correct for any procedural losses.

Liver cholesterol was quantified by capillary gas-liquid chromatography. The cholesterol distribution in plasma was determined by gel filtration chromatography using a Superose 6 column (Sigma). Two-ml aliquots were collected and assayed for cholesterol using an enzymatic kit (Boehringer Diagnostics, Indianapolis, IN).

The data are presented as means ± S.D. To test for differences among experimental groups, one-way analysis of variance was performed. Significant results were further analyzed using the Tukey multiple comparison procedure.

RESULTS

Under in vivo conditions, >95% of the bile salts secreted into bile are primary and secondary bile salts fluxing through the liver in the enterohepatic circulation (28). Consequently, the potential role of specific bile salts in the regulation of 7-hydroxylase expression is necessarily investigated following removal of the enterohepatic pool of bile salts by biliary diversion. Using adenovirus-mediated gene transfer, we recently noted that primary overexpression of an exogenous 7-hydroxylase gene in hamsters with an intact enterohepatic circulation resulted in reciprocal down-regulation of the endogenous gene (18). However, it was not possible to determine from these studies whether suppression of 7-hydroxylase was due to a direct effect of the newly synthesized bile salts (or their precursors) on 7-hydroxylase at the level of the hepatocyte, or due to an increased transhepatic flux of bile salts (primary or secondary) in the expanded enterohepatic pool.

Our initial studies were therefore undertaken to characterize the effect of biliary diversion on bile salt synthesis and 7-hydroxylase expression in the hamster and to determine if primary overexpression of an exogenous 7-hydroxylase gene in bile-diverted animals leads to inhibition of endogenous gene expression. Biliary catheters were placed, and the enterohepatic pool of bile salts was allowed to drain out over the next 12-14 h. Following drainage of the enterohepatic pool of bile salts, animals were administered 10¹⁰ pfu AdCMV7 or control virus intravenously, after which biliary diversion was continued for an additional 24 h. As illustrated in Fig. 1, rates of bile salt secretion during the first hour after insertion of the biliary catheter equaled 2-2.5 µmol/h/100 g of body weight. Rates of bile salt secretion fell rapidly with depletion of the bile salt pool and reached a nadir of ~0.2 µmol/h/100 g of body weight 12-14 h after initiation of biliary diversion. Over the ensuing 24 h, the rate of bile salt secretion increased ~3-fold in animals administered control virus (or vehicle alone) and 4-5-fold in animals administered AdCMV7. In animals administered control virus, cholate and chenodeoxycholate conjugates accounted for 85 and 11%, respectively, of total secreted bile salts. In animals administered AdCMV7, these values equaled 72 and 23%. Hepatic 7-hydroxylase activity equaled 61 pmol/h/mg of protein shortly after drainage of the enterohepatic pool and, over the subsequent 24 h, increased ~10-fold in animals administered AdCMV7 and 4-5-fold in animals administered control virus or vehicle alone (Fig. 1, bottom panel). Only small amounts of 7-hydroxycholesterol (the product of the 7-hydroxylase reaction) were present in hepatic microsomes prior to incubation, even in animals overexpressing 7-hydroxylase. In animals with the highest level of 7-hydroxylase activity (bile-diverted animals administered AdCMV7), hepatic microsomes contained 5-10 pmol of 7-hydroxycholesterol/mg of protein, an amount equaling 1-2% of the 7-hydroxycholesterol formed per hour per milligram of microsomal protein.

Effect of biliary diversion and AdCMV7 administration on bile salt secretion and 7-hydroxylase activity in the hamster.

Fig. 2 illustrates the changes in expression of mRNAs encoding the endogenous and transduced 7-hydroxylase genes in bile-diverted hamsters. Total liver RNA was prepared from bile-diverted hamsters 24 h after administration of AdCMV7 or control virus and assayed for 7-hydroxylase mRNA by nuclease protection using ³²P-labeled cDNA probes specific for hamster or rat 7-hydroxylase. In this assay, the RNA-protected rat 7-hydroxylase probe is 388 nucleotides, whereas the hamster 7-hydroxylase probe is 240 nucleotides. RNA isolated from hamsters infected with control virus protected the hamster-specific 7-hydroxylase probe but did not protect the rat probe. In these animals, biliary diversion significantly increased the abundance of mRNA encoding hamster 7-hydroxylase (lane 2 compared with lane 1). Hamsters infected with recombinant virus encoding rat 7-hydroxylase expressed transgene (rat)-specific mRNA, whereas expression of mRNA from the endogenous (hamster) gene was suppressed as summarized in Fig. 3. Based on the length and specific activity of the hamster and rat probes, it could be estimated that animals administered 10¹⁰ pfu of AdCMV7 accumulated ~2-fold higher overall levels of 7-hydroxylase mRNA in the liver than did uninfected animals or animals infected with the same dose of control virus. Infection of animals with smaller quantities of AdCMV7 produced less expression of transgene-specific mRNA and, in a reciprocal manner, less suppression of the endogenous hamster gene. In contrast to AdCMV7, neither control virus (AdCMVLuc or AdCMVgal at 10¹⁰ pfu) altered hepatic 7-hydroxylase mRNA or activity in the hamster. Thus, overexpression of an exogenous 7-hydroxylase gene resulted in suppression of endogenous 7-hydroxylase gene expression even in the absence of an enterohepatic circulation of bile salts, suggesting that primary bile salts suppress their own synthesis in the liver by feedback inhibition of 7-hydroxylase expression.

Nuclease protection analysis of cholesterol 7-hydroxylase mRNA levels in hamsters administered AdCMV7 or control virus.

Hepatic 7-hydroxylase expression in hamsters administered AdCMV7.

If bile salts synthesized in the liver exert feedback inhibition on 7-hydroxylase expression, then ligation of the common bile duct, which is known to increase bile salt levels in the liver, should also result in down-regulation of hepatic 7-hydroxylase expression. As illustrated in Fig. 4 (top panel) biliary catheters were placed, and the enterohepatic pool of bile salts was allowed to drain out over a 12-14-h period of time. Following depletion of the enterohepatic pool of bile salts, the biliary catheters were tied, and total biliary obstruction was continued for 24 h. As shown in Fig. 4 (bottom panel), hepatic 7-hydroxylase activity in control bile-diverted animals increased ~4-fold but remained essentially unchanged in bile duct-ligated animals. These data suggest that bile salts synthesized in the liver suppress their own synthesis by exerting feedback inhibition on hepatic 7-hydroxylase expression.

Effect of biliary obstruction on hepatic 7-hydroxylase activity in the hamster.

Studies were next undertaken to determine if intravenously administered bile salts exert negative feedback control on hepatic 7-hydroxylase expression in bile-diverted hamsters. As illustrated in Fig. 5 (top panel), biliary catheters were placed, and after 12-14 h (to allow the enterohepatic pool to drain out) animals were infused intravenously with taurocholate at a rate of ~10 µmol/h/100 g of body weight for 24 h. Trace amounts of tauro[¹⁴C]cholate were added to the infusions to assess recovery of the infused bile salt in bile. Under the conditions employed in these studies, recovery of infused tauro[¹⁴C]cholate in bile was essentially complete (87-96%). In addition, the mass of taurocholate secreted into bile equaled the rate of bile salt infusion (93-102%). As shown in Fig. 5 (bottom panel), intravenously administered taurocholate at rates of ~10 µmol/h/100 g of body weight prevented the 4-fold derepression of hepatic 7-hydroxylase activity observed in control bile-diverted animals. Although not shown, infusion of taurocholate at rates of ~10 µmol/h/100 g of body weight also prevented the derepression of hepatic 7-hydroxylase mRNA observed in control bile-diverted animals.

Effect of intravenously administered taurocholate on hepatic 7-hydroxylase activity in the hamster.

The dose-response relationship between the rate of taurocholate infusion and the activity of hepatic 7-hydroxylase was next investigated using a protocol identical to that shown in Fig. 5, and these data are summarized in Fig. 6. Intravenously administered taurocholate at infusion rates of up to 6 µmol/h/100 g of body weight had little effect on hepatic 7-hydroxylase activity. At an infusion rate of ~8 µmol/h/100 g of body weight, intravenously administered taurocholate partially prevented the derepression of 7-hydroxylase activity associated with biliary diversion, and at rates of 10 µmol/h/100 g of body weight, derepression of 7-hydroxylase activity was largely prevented.

Dose-response relationship between rates of taurocholate infusion and hepatic 7-hydroxylase activity in the hamster.

Fig. 7 shows the effect of AdCMV7 administration, taurocholate infusion, or bile duct ligation on the concentration of bile salts in liver and plasma. Hamsters were fitted with biliary and femoral vein catheters. After 12-14 h (to allow for drainage of the enterohepatic pool of bile salts), groups of animals were administered 10¹⁰ pfu of AdCMV7, infused intravenously with taurocholate at a rate of 10 µmol/h/100 g of body weight, or subjected to bile duct ligation. Animals were sacrificed 24 h later, and samples of liver and plasma were taken for the determination of total bile salt content. The hepatic bile salt concentration equaled 55 ± 15 nmol/g in control bile-diverted hamsters and was significantly elevated in animals administered AdCMV7 (~3-fold), infused with taurocholate (~4-fold), or subjected to bile duct ligation (~6-fold). The hepatic bile salt concentration in control animals with an intact enterohepatic circulation equaled 167 ± 35 nmol/g (data not shown). The plasma bile salt concentration in control bile-diverted hamsters equaled 2 ± 1 nmol/ml. Administration of AdCMV7 had no effect on plasma bile salt concentrations. Plasma bile salt concentrations were significantly elevated in animals infused with taurocholate (~4-fold) or subjected to bile duct ligation (~9-fold). The plasma bile salt concentration in control hamsters with an intact enterohepatic circulation equaled 4 ± 1 nmol/ml (data not shown).

Hepatic and plasma bile salt concentrations in bile-diverted animals subjected to AdCMV7 administration, taurocholate infusion, or bile duct ligation.

The effect of biliary diversion (with or without AdCMV7 administration, bile duct ligation, or taurocholate infusion) on rates of hepatic cholesterol synthesis was next determined in parallel groups of animals subjected to the experimental protocols shown in Figs. 1, 4, and 5. Hamsters were fitted with biliary catheters and allowed 12-14 h of biliary drainage to eliminate the enterohepatic pool of bile salts. Animals were then administered AdCMV7 intravenously, subjected to bile duct ligation, or administered taurocholate as a continuous intravenous infusion at a rate of ~10 µmol/h/100 g of body weight for 24 h. Rates of hepatic cholesterol synthesis were quantified in vivo using [³H]water. As shown in Fig. 8, the rate of hepatic sterol synthesis equaled 0.6 µmol/h/100 g of body weight in hamsters with an intact enterohepatic circulation and increased to 2.2 µmol/h/100 g of body weight after 12 h of biliary diversion and to 4.7 µmol/h/100 g of body weight after an additional 24 h of biliary diversion. Neither bile duct ligation nor infusion of taurocholate prevented the marked derepression of hepatic cholesterol synthesis associated with biliary diversion despite the fact that both experimental conditions prevented derepression of hepatic 7-hydroxylase activity. Similarly, administration of AdCMV7 did not significantly alter the rate of hepatic sterol synthesis compared with control bile-diverted animals. HMG-CoA reductase mRNA was measured in separate but identically treated animals and was found to increase ~3-fold with biliary diversion, an effect that was not prevented by AdCMV7 administration, biliary obstruction, or infusion of taurocholate (data not shown). The marked increase in rates of hepatic cholesterol synthesis appeared to be related to biliary diversion (and not to stress or the glucose-rich infusate), since a parallel group of animals treated identically but not subjected to biliary diversion showed no increase in rates of hepatic sterol synthesis (data not shown).

Effect of AdCMV7 administration, taurocholate infusion, or bile duct ligation on the rate of hepatic cholesterol synthesis in bile-diverted animals.

Rates of receptor-dependent LDL transport were quantified in vivo in parallel groups of animals subjected to the same experimental protocol illustrated in Fig. 8. As shown in Fig. 9 (top panel), hepatic LDL receptor activity was not significantly derepressed in response to biliary diversion. Moreover, administration of AdCMV7, bile duct ligation, or infusion of taurocholate at a rate of ~10 µmol/h/100 g of body weight all failed to alter hepatic LDL receptor activity over the relatively short time frame of these studies. Although not shown, hepatic LDL receptor mRNA levels were measured in these same animals by nuclease protection and were similarly unaffected by short term biliary diversion with or without the administration of AdCMV7, taurocholate infusion, or bile duct ligation. As shown in Fig. 9 (bottom panel) plasma LDL-cholesterol concentrations were not significantly altered in bile-diverted animals compared with animals with an intact enterohepatic circulation. Similarly, administration of AdCMV7 and taurocholate infusion both failed to alter circulating LDL levels over the relatively short time course of these studies. Bile duct ligation greatly increased the amount of cholesterol carried in the LDL fraction as determined by Superose 6 chromatography. However, the increase in cholesterol within this fraction was almost entirely due to an increase in unesterified cholesterol, suggesting the accumulation of lipoprotein-X particles. The presence of lipoprotein-X particles within this fraction was confirmed by agarose gel electrophoresis.

Effect of biliary diversion, with or without AdCMV7 administration, taurocholate infusion, or biliary obstruction, on receptor-dependent LDL uptake by the liver and plasma LDL cholesterol concentrations.

DISCUSSION

Hepatic 7-hydroxylase catalyzes the initial and rate-limiting step in the bile salt biosynthetic pathway. The expression of hepatic 7-hydroxylase appears to be regulated at the transcriptional level by the quantity of bile salts fluxing through the enterohepatic circulation; however, the mechanisms responsible for bile salt-mediated regulation of 7-hydroxylase are poorly understood. Several lines of evidence suggest that bile salts may not directly inhibit their own synthesis at the level of the hepatocyte. Thus, there is not a simple inverse relationship between hepatic 7-hydroxylase activity and the concentration of bile salts in the liver or portal vein (29, 30, 31). Moreover, several groups of investigators were unable to demonstrate any direct inhibitory effect of bile salts on bile salt synthesis in hepatocyte suspensions or cultures (6, 7, 32). Although more recent studies in cultured hepatocytes suggest that hydrophobic bile salts may suppress 7-hydroxylase expression and regulate 7-hydroxylase promoter activity (33, 34, 35, 36, 37), this seems to occur only at hepatocyte bile salt concentrations greatly exceeding those found in vivo (38), raising questions about the physiologic relevance of these findings. Recently, Pandak et al. (12) reported that intraduodenally administered taurocholate exerted negative feedback inhibition on hepatic 7-hydroxylase expression in the rat, whereas equimolar quantities of intravenously administered taurocholate had no effect. Together, these studies suggest that bile salts may have no direct effect on hepatic 7-hydroxylase expression. Rather, these observations raise the possibility that an intestinal factor, released or absorbed in the presence of bile salts, may play a role in the regulation of bile salt synthesis. The concept of a gut-derived hormonal substance controlling hepatic bile salt synthesis is appealing due to the pronounced effects of various hormones on bile salt synthesis in cultured hepatocytes (39). However, cross-circulation studies in the dog failed to provide any evidence that a hormonal substance is involved in bile salt-dependent regulation of 7-hydroxylase expression in vivo (40).

The present work, carried out in the hamster, supports the view that bile salts inhibit their own synthesis at the level of the hepatocyte through feedback repression of 7-hydroxylase expression. In these studies, the transhepatic flux of bile salts was eliminated by biliary diversion. Biliary diversion was accompanied by a 67% decline in hepatic bile salt concentrations, a ~4-fold increase in hepatic 7-hydroxylase expression, and a ~3-fold increase in the rate of bile salt synthesis. Following drainage of the enterohepatic pool of bile salts, three independent approaches were employed to increase hepatic bile salt concentrations. In the first approach, an exogenous 7-hydroxylase gene was expressed in the liver using adenovirus-mediated gene transfer. Overexpression of an exogenous 7-hydroxylase gene in these animals significantly raised hepatic bile salt concentrations and reduced mRNA encoding the endogenous 7-hydroxylase gene. Administering smaller quantities of AdCMV7 produced less expression of transgene-specific mRNA and, in a reciprocal manner, less suppression of the endogenous gene. These data suggest that bile salts synthesized within the liver (primary bile salts) are capable of regulating their own synthesis by feedback inhibition of 7-hydroxylase expression.

Rates of bile salt secretion in bile-diverted animals expressing the 7-hydroxylase transgene were still significantly less than the normal transhepatic flux of bile salts in hamsters with an intact enterohepatic circulation. Despite this, levels of mRNA encoding the endogenous 7-hydroxylase gene were lower in bile-diverted animals expressing the 7-hydroxylase transgene than in control animals with an intact enterohepatic circulation. This observation raises the possibility that de novo synthesized bile salts may exert greater suppressive activity on 7-hydroxylase than bile salts passing through the hepatocyte in the enterohepatic circulation. It is possible that other enzymes in the bile salt biosynthetic pathway may become rate-limiting under conditions of 7-hydroxylase overexpression. If this were true, then a variety of intermediates might accumulate, some of which may suppress 7-hydroxylase expression. This seems unlikely in the present studies, since hepatic 7-hydroxylase activity in animals administered the largest dose of AdCMV7 exceeded enzyme activity in control bile-diverted animals by only 2-3-fold.

In the rat, obstruction of the common bile duct, which leads to an accumulation of bile salts in the liver, paradoxically leads to derepression of 7-hydroxylase expression (8, 9, 10). This observation is commonly used to support the hypothesis that bile salts exert no direct effect on 7-hydroxylase expression. In contrast to previous observations in the rat, bile duct ligation suppressed 7-hydroxylase expression in the hamster. This was true even when the enterohepatic pool was first depleted, resulting in the elimination of all secondary bile salts. These data provide further support for the view that bile salts synthesized within the liver (primary bile salts) are capable of suppressing their own synthesis through feedback repression of 7-hydroxylase in the hamster. These data are consistent with observations in the rhesus monkey suggesting that obstruction of the common bile duct also leads to repression of bile salt synthesis in primates (41). These findings are also consistent with observations in patients with obstructive jaundice where 7-hydroxylation of cholesterol (42) and bile salt synthesis (42, 43, 44) appear to be reduced. Why bile duct obstruction is accompanied by a paradoxical increase in 7-hydroxylase expression in the rat but by suppression of enzyme expression in the hamster and other species is not entirely clear. Of note, the rat responds to bile duct ligation by synthesizing predominantly hydrophilic, weakly detergent bile salts that are known not to suppress 7-hydroxylase expression (8, 9, 45). In contrast, hamsters (46) and humans (42, 43, 44) are apparently unable to significantly alter the pattern of bile salts synthesized by the liver and continue to synthesize mainly cholate and chenodeoxycholate derivatives during biliary obstruction.

If primary bile salts inhibit their own synthesis through feedback repression of hepatic 7-hydroxylase, then intravenous administration of bile salts should also suppress enzyme activity. However, whereas intraduodenally administered bile salts suppress hepatic 7-hydroxylase activity, several investigators have reported no effect of intravenously administered taurocholate on enzyme expression in the bile-diverted rat (12, 47, 48, 49). In the present studies, intravenously administered taurocholate clearly suppressed hepatic 7-hydroxylase expression, or at least prevented derepression of enzyme expression, in bile-diverted hamsters. The transhepatic flux of taurocholate necessary to prevent derepression of enzyme expression was roughly twice the normal transhepatic flux of bile salts in hamsters with an intact enterohepatic circulation. However, in the intact enterohepatic circulation, chenodeoxycholate and deoxycholate compose ~40% of total bile salts, and it is likely that these relatively hydrophobic bile salts are more active than taurocholate at suppressing 7-hydroxylase expression (50). Why intravenous taurocholate suppresses hepatic 7-hydroxylase expression in the hamster but not in the rat is not entirely clear. From a technical standpoint, the normal transhepatic flux of bile salts would appear to be higher in the rat than in the hamster. As a consequence, higher infusion rates (or concentrations) of bile salts must be administered to rats to achieve physiologic transhepatic flux rates. Of the bile salts known to suppress 7-hydroxylase expression when administered intraduodenally, all except taurocholate are strong detergents, precluding their intravenous administration due to severe hemolysis. Even taurocholate causes hemolysis when infused intravenously in concentrations exceeding ~20 mM, necessitating the administration of large volumes of infusate to approach physiologic transhepatic flux rates in the rat (12).

Given that an intravenous infusion of bile salts suppresses hepatic 7-hydroxylase expression in the hamster, the question follows as to how bile salts fluxing through the liver are sensed and how this signal leads to changes in 7-hydroxylase gene transcription. Presumably, it is the concentration of bile salts within the hepatocyte that is important for feedback inhibition of 7-hydroxylase expression, and the current data are consistent with this supposition. It is well-recognized, however, that hepatic (or portal vein) bile salt concentrations frequently do not change in response to less extreme experimental interventions. Thus, hepatic 7-hydroxylase activity can be varied over a 10-20-fold range (by feeding agents that expand or contract the enterohepatic pool of bile salts), with relatively little, if any, change in hepatic (or portal vein) bile salt concentrations (29, 30, 31). The lack of a simple inverse relationship between hepatic 7-hydroxylase activity and hepatic bile salt concentrations has been used to support the view that bile salts do not directly regulate 7-hydroxylase expression at the level of the hepatocyte. It could be argued, however, that compensatory changes in 7-hydroxylase activity tend to normalize hepatic bile salt concentrations and that only when this compensatory mechanism is overwhelmed are hepatic bile salt concentrations significantly altered. This would be somewhat analogous to hepatic HMG-CoA reductase activity, which can also be regulated over a 10-20-fold range (by interventions that alter net hepatic sterol balance) with little change in liver cholesterol levels (16). Only when the capacity of HMG-CoA reductase (and the cholesterol biosynthetic pathway) to compensate for changes in hepatic sterol balance is exceeded, does the cholesterol content of the liver significantly change.

While the present studies suggest that bile salts directly regulate their own synthesis at the level of the hepatocyte, it is clear that other substances also contribute to the regulation of 7-hydroxylase expression. Several hormones alter 7-hydroxylase expression in cultured hepatocytes (39), and it is likely that hormonal factors contribute to the diurnal changes in 7-hydroxylase expression that are observed in some species (51, 52). Circumstantial evidence suggests that an intestinal factor may contribute to the control of hepatic 7-hydroxylase expression (12). Lymphatic diversion leads to up-regulation of hepatic 7-hydroxylase activity, suggesting that the putative inhibitory factor is carried in lymph (11). We have found that dietary triglyceride significantly reduces hepatic 7-hydroxylase expression in rats and hamsters consuming a very low fat diet.² In the rat, suppression of hepatic 7-hydroxylase activity by dietary triglyceride was previously reported (53). Moreover, oleic acid was shown to reduce bile salt synthesis and secretion in isolated perfused rat liver (54). Thus, fatty acids may represent a gut-derived substance that is dependent on luminal bile salts for absorption, transported in intestinal lymph, and capable of exerting feedback repression on hepatic 7-hydroxylase expression. A major function of bile salts is to facilitate the digestion and absorption of dietary triglyceride. It would make a certain amount of mechanistic sense if bile salt synthesis were up-regulated in response to a severe reduction in postprandial triglyceride (or fatty acid) delivery to the liver.

Whereas regulation of hepatic 7-hydroxylase by fatty acids remains somewhat speculative, regulation by cholesterol has been much better characterized. In some species, most notably the rat and mouse (55, 56, 57), dietary cholesterol up-regulates hepatic 7-hydroxylase expression. Recent studies suggest that cholesterol (and not an oxygenated metabolite) directly up-regulates 7-hydroxylase expression in primary rat hepatocyte cultures (58). We previously found that hepatic 7-hydroxylase expression is not significantly up-regulated by physiologically relevant levels of dietary cholesterol in the hamster (16). Failure to up-regulate hepatic 7-hydroxylase expression in response to excess dietary cholesterol accounts, in part, for the greater sensitivity of the hamster (compared with the rat and mouse) to the cholesterolemic effects of dietary cholesterol (16).

The liver responded to the accelerated loss of sterol associated with biliary diversion by markedly increasing the rate of de novo cholesterol synthesis (~8-fold). The increase in HMG-CoA reductase mRNA levels was less striking (~3-fold), suggesting that posttranscriptional regulation of HMG-CoA reductase activity plays a major role in controlling the rate of hepatic cholesterol synthesis under these conditions. Derepression of hepatic cholesterol synthesis was apparently triggered by the increased rate of conversion of cholesterol to bile salts (and not to stress or the glucose-rich infusions), since sterol synthesis rates did not increase in identically treated sham-operated animals. Notably, neither biliary obstruction nor taurocholate infusion prevented derepression of hepatic cholesterol synthesis despite the fact that both experimental interventions prevented derepression of 7-hydroxylase expression. These data are consistent with previous studies in bile-diverted rats (59, 60) and support the view that bile salts have no direct inhibitory effect on hepatic sterol synthesis. It is less clear, however, why inhibition of 7-hydroxylase activity would not lead to increased cholesterol availability with secondary suppression of hepatic cholesterol synthesis. In the case of taurocholate infusion, decreased conversion of cholesterol to bile salts would be offset by the increased secretion of biliary cholesterol that accompanies bile salt infusions (15).

Changes in bile salt synthesis had little effect on hepatic LDL receptor expression during the relatively short time frame of these studies. AdCMV7 did not up-regulate hepatic LDL receptor expression in bile-diverted hamsters. Of note, all animals in the present studies were maintained on a low cholesterol, low fat diet. These findings in bile-diverted animals are consistent with previous studies performed in hamsters with an intact enterohepatic circulation, where AdCMV7 restored receptor activity in animals maintained on a Western-type (cholesterol- and fat-enriched) diet, but did not increase receptor activity in animals maintained on a low cholesterol, low fat control diet (18). In the previous study, plasma cholesterol levels progressively fell following AdCMV7 administration, reaching a minimum at 3-4 days. In the current studies, which were relatively short-term (animals sacrificed 24 h after virus administration), plasma cholesterol concentrations were not significantly reduced by AdCMV7 administration. Presumably, longer term studies would show a decline in plasma LDL levels in bile-diverted animals administered AdCMV7.

Biliary obstruction produced a rapid and marked increase in cholesterol carried in the LDL fraction as determined by Superose 6 chromatography. The excess cholesterol in this fraction was entirely unesterified, suggesting the accumulation of lipoprotein-X particles. Lipoprotein-X is a bilayer vesicle composed of equimolar amounts of unesterified cholesterol and phospholipid enclosing an aqueous compartment containing albumin (61). Accumulation of lipoprotein-X is a characteristic feature of cholestasis and biliary obstruction but may also occur in individuals with lecithin:cholesterol acyltransferase deficiency and during intravenous lipid emulsion administration. Lipoprotein-X is not transported by the LDL receptor pathway and is mainly cleared by cells of the reticuloendothelial system. As a consequence, lipoprotein-X exerts no feedback control on hepatic sterol synthesis or LDL receptor expression (62, 63), consistent with the current observations in the bile duct-ligated hamster.