Suppression of cholesterol 7α-hydroxylase transcription and bile acid synthesis by an α1-antitrypsin peptide via interaction with α1-fetoprotein transcription factor

α1-Antitrypsin (α1-AT) is a serum protease inhibitor that is synthesized mainly in the liver, and its rate of synthesis markedly increases in response to inflammation. This increase in α1-AT synthesis results in an increase in peptides, like its carboxyl-terminal C-36 peptide (C-36), resulting from α1-AT cleavage by proteases. Atherosclerosis is a form of chronic inflammation, and one of the risk factors is elevated plasma cholesterol levels. Because of the correlation between atherosclerosis, plasma cholesterol content, inflammation, and α1-AT rate of synthesis, we investigated the effect of the C-36 serpin peptide on hepatic bile acid biosynthesis. We discovered that C-36 is a powerful and specific transcriptional down-regulator of bile acid synthesis in primary rat hepatocytes, through inhibition of the cholesterol 7α-hydroxylase/CYP7A1 (7α-hydroxylase) promoter. Mice injected with the C-36 peptide also showed a decrease in 7α-hydroxylase mRNA. A mutated but very similar peptide did not have any effect on 7α-hydroxylase mRNA or its promoter. The sterol 12α-hydroxylase/CYP8B1 (12α-hydroxylase) promoter is also down-regulated by the C-36 peptide in HepG2 cells but not by the mutated peptide. The DNA element involved in the C-36-mediated regulation of 7α- and 12α-hydroxylase promoters mapped to the α1-fetoprotein transcription factor (FTF) site in both promoters. The C-36 peptide prevented binding of FTF to its target DNA recognition site by direct interaction with FTF. We hypothesize that the C-36 peptide specifically interacts with FTF and induces a conformational change that results in loss of its DNA binding ability, which results in suppression of 7α- and 12α-hydroxylase transcription. These results suggest that peptides derived from specific serum proteins may alter hepatic gene expression in a highly specific manner.

␣ 1 -Antitrypsin (␣ 1 -AT) is a serum protease inhibitor that is synthesized mainly in the liver, and its rate of synthesis markedly increases in response to inflammation. This increase in ␣ 1 -AT synthesis results in an increase in peptides, like its carboxyl-terminal C-36 peptide (C-36), resulting from ␣ 1 -AT cleavage by proteases. Atherosclerosis is a form of chronic inflammation, and one of the risk factors is elevated plasma cholesterol levels. Because of the correlation between atherosclerosis, plasma cholesterol content, inflammation, and ␣ 1 -AT rate of synthesis, we investigated the effect of the C-36 serpin peptide on hepatic bile acid biosynthesis. We discovered that C-36 is a powerful and specific transcriptional down-regulator of bile acid synthesis in primary rat hepatocytes, through inhibition of the cholesterol 7␣-hydroxylase/CYP7A1 (7␣-hydroxylase) promoter. Mice injected with the C-36 peptide also showed a decrease in 7␣-hydroxylase mRNA. A mutated but very similar peptide did not have any effect on 7␣-hydroxylase mRNA or its promoter. The sterol 12␣-hydroxylase/ CYP8B1 (12␣-hydroxylase) promoter is also downregulated by the C-36 peptide in HepG2 cells but not by the mutated peptide. The DNA element involved in the C-36-mediated regulation of 7␣-and 12␣-hydroxylase promoters mapped to the ␣ 1 -fetoprotein transcription factor (FTF) site in both promoters. The C-36 peptide prevented binding of FTF to its target DNA recognition site by direct interaction with FTF. We hypothesize that the C-36 peptide specifically interacts with FTF and induces a conformational change that results in loss of its DNA binding ability, which results in suppression of 7␣and 12␣-hydroxylase transcription. These results suggest that peptides derived from specific serum proteins may alter hepatic gene expression in a highly specific manner.
Serine protease inhibitors (serpins) are the most common protease inhibitors in mammals and are part of the acute phase response. At sites of inflammation, proteolytic enzymes are released by neutrophils, platelets, mast cells, macrophages, or by any invading microorganisms. Because an uncontrolled proteolytic activity would result in serious unwanted destruction of surrounding tissues, the synthesis of serpins (i.e. ␣ 1 -antichymotrypsin, ␣ 1 -antitrypsin (␣ 1 -AT), 1 and plasminogen activator inhibitor I) is markedly increased (1) to restore homeostasis. Inhibitory serpins interact with their target proteases at a reactive site located within a loop structure of 30 -40 amino acid residues from the carboxyl-terminal end (2). Formation of a stable complex between ␣ 1 -AT and human leukocyte elastase results from the cleavage of the P 1 -PЈ 1 bond in the reactive site loop, generating a 4-kDa carboxyl-terminal fragment of 36 amino acids corresponding to residues 359 -394 (C-36), that remains non-covalently bound to the cleaved ␣ 1 -AT in complex with the protease (3). This reactive site loop is also susceptible to proteolysis by non-target proteases (4), which can give rise to the carboxyl-terminal peptide fragments of the serpin. Both of these reactions cause a loss of inhibitor activity and, in the case of ␣ 1 -AT, it results in a ␣ 1 -AT molecule composed of two associated chains, a long amino-terminal fragment and a 36-residue carboxyl-terminal fragment, C-36 (5). This cleaved form of ␣ 1 -AT has been shown to have biological activities, such as chemotatic activity (6) and regulation of native ␣ 1 -AT synthesis (7). These observations support the notion that ␣ 1 -AT may also act as a reservoir of these biological activities that are released upon cleavage of the native form. ␣ 1 -AT is the archetype of the serpin superfamily and the most abundant serpin in plasma (20 -53 M) and is mainly synthesized by hepatocytes. Its hepatic synthesis increases during acute phase response due to stimulation by the release of interleukin 6 (8). ␣ 1 -AT concentration in serum can increase up to 8-fold during inflammatory events (9). The ␣ 1 -AT-elastase complexes are cleared from the circulation by the low density lipoprotein (LDL)-related protein receptor or the serpin-enzyme complex receptor, located mainly on the surface of the hepatocytes (7, 10). A 5-amino acid peptide within C-36 is actually recognized by the receptor (7). It should be mentioned that free C-36 carboxyl-terminal fragment and other cleavage products have been isolated from plasma, bile, and the spleen (11).
The acute phase response that is triggered by infection, inflammation, and trauma is associated with changes in plasma lipid and lipoprotein levels (12) as well as in a variety of proteins, such as ␣ 1 -AT (13). Plasma cholesterol also increases albeit more moderately than plasma triglycerides (14). Interestingly, however, in experimentally induced acute phase in rabbits, the lipoprotein profile distribution shifts dramatically toward a more atherogenetic profile (14). These changes in cholesterol metabolism can also be induced by the administration of lipopolysaccharide, which mimics Gram-negative infections (12). Administration of lipopolysaccharide to rodents increases plasma cholesterol levels resulting in an increase in LDL (15).
Atherosclerosis is a form of chronic inflammation resulting from interaction between modified lipoproteins, macrophages, and other cells (for a review see Ref. 16). High plasma cholesterol levels are unique in being sufficient to drive the development of atherosclerosis in humans and experimental animals, even in the absence of other known risk factors, and thus cholesterol homeostasis is important. One of the pathways that play a critical role in maintaining cholesterol homeostasis is the bile acid biosynthetic pathway, because nearly 50% of the body cholesterol is catabolized to bile acids. Cholesterol conversion to bile acids occurs via the "classic" (neutral) or the "alternative" (acidic) bile acid biosynthesis pathways (17). Cholic acid and chenodeoxycholic acid are the end products of these pathways and the major primary bile acids found in most vertebrates. Cholic acid is hydroxylated at position 12␣, whereas chenodeoxycholic acid is not. There are two enzymes that play major regulatory roles in these two pathways. Cholesterol 7␣-hydroxylase/CYP7A1 (7␣-hydroxylase) is the ratelimiting enzyme in the classic pathway. Sterol 12␣-hydroxylase/CYP8B1 (12␣-hydroxylase) is the specific enzyme for cholic acid synthesis and determines the ratio of cholic acid to chenodeoxycholic acid and thus the hydrophobicity of the circulating pool. Because chenodeoxycholic acid is a more potent suppressor of HMG-CoA reductase and 7␣-hydroxylase than cholic acid (18,19), the relative activity of 12␣-hydroxylase may play an important role in the regulation of hepatic cholesterol homeostasis. The alteration of cholic/chenodeoxycholic acid ratio affects biliary cholesterol and phospholipid secretion, thus altering intestinal cholesterol absorption and receptor-mediated lipoprotein uptake by the hepatocyte (19).
Recent studies have delineated many of the factors involved in the expression and regulation of these two bile acid biosynthetic enzymes. Thus, liver receptor homolog-1, also known as CYP7A promoter-binding factor, NR5A2 (20), or ␣ 1 -fetoprotein transcription factor (FTF) (Genome Data base Nomenclature Committee) (21), has been proposed to be required for the transcription of the 7␣-hydroxylase gene (22,23). Bile acids activate transcription of the small heterodimer partner 1 (SHP) via binding of the hormone receptor farnesoid X receptor to its binding site in the SHP promoter. In turn, it has been proposed that SHP dimerizes with FTF and diminishes its activity on the 7␣-hydroxylase promoter by mechanisms not well understood yet (24,25).
FTF also plays a key role in the expression and in the regulation of 12␣-hydroxylase (26). FTF binds to two sites within the rat 12␣-hydroxylase promoter and both sites are required for both promoter activity and bile acid-mediated regulation. We have also shown that SHP is involved in the downregulation of the 12␣-hydroxylase promoter (27). Overexpression of SHP in HepG2 cells suppresses 12␣-hydroxylase promoter activity. We also showed that SHP prevents binding of FTF to its binding sites within the 12␣-hydroxylase promoter providing a mechanism of action for the SHP-mediated suppression of 12␣-hydroxylase transcription (27). FTF is also the target for a specific sterol regulatory binding protein 2-mediated suppression of the 12␣-hydroxylase promoter (28).
Because of the well documented dramatic increase in ␣ 1 -AT levels in response to inflammation, which in turns gives rise to an increase in peptides resulting from ␣ 1 -AT cleavage by proteases and the alteration in cholesterol metabolism during the acute phase response, we have investigated whether there is a connection between the inflammatory response and cholesterol homeostasis through a specific suppression of bile acid biosynthesis by the ␣ 1 -AT-derived peptide C-36. We report here that the ␣ 1 -AT C-36 peptide is a powerful inhibitor of the 7␣-and 12␣-hydroxylase promoters by specifically interacting with FTF, a positive-acting transcription factor.
Primary Cultures of Rat Hepatocytes-Hepatocytes from adult male Sprague-Dawley rats (200 -300 g) were isolated according to the method of Bissell and Guzelian (30). Parenchymal hepatocytes (3.5 ϫ 10 6 ) were plated onto 60-mm Falcon culture dishes coated with rat tail collagen (Vitrogen 100, Cohesion, Palo Alto, CA). Prior to plating, cells were judged to be Ͼ90% viable using trypan blue exclusion (0.04%). Cells were incubated in 3 ml of serum-free William's E medium supplemented with 1.0 M thyroxine, 0.1 M dexamethasone (31), 0.25 unit/ml insulin, and 100 units/ml penicillin in a 5% CO 2 atmosphere at 37°C. Culture medium was changed daily. Peptides were added 48 h after plating, unless otherwise indicated. Hepatocytes were harvested at the indicated times.
Bile Acid Synthesis by Primary Hepatocytes-The conversion of [ 14 C]cholesterol into methanol-water-soluble material was determined as previously described (31) according to the Folch technique.
Peptide Injections into Mice-Adult male C57BBL mice (22-25 g) were injected through the tail vain twice, at 9:00 a.m. and 4:00 p.m., with 2 mg of the C-36 or the C-35 (F 3 A) peptides dissolved in 150 l of phosphate-buffered saline (PBS) solution. Control mice were injected with PBS alone. Mice were killed 24 h after the first injection, and liver RNA was extracted by standard procedures (32).
Total RNA Isolation and Quantification-Total RNA was isolated from primary hepatocytes using the guanidine thiocyanate cesium chloride centrifugation method (32). 7␣-Hydroxylase, HMG-CoA reductase, CYP27, actin, glyceraldehyde-3-phosphate dehydrogenase, and cyclophilin mRNAs were quantified by RPA as already described (29) according to the manufacturer's protocol (Ambion, Austin, TX). Fatty acid synthase (FAS), HMG-CoA synthase, and LDL receptor mRNAs were analyzed by Northern blot as previously described (33). In brief, total RNA was size-fractionated by electrophoresis on 1% agarose gel con-taining 7% formaldehyde and then transferred to nitrocellulose membranes (Bio-Rad). To standardize loaded mRNAs, a cyclophilin cDNA fragment was labeled by the same method and used to probe the same membranes under modified hybridization conditions. Transient Transfections of Hepatocytes and Luciferase Assays-HepG2 cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). Rat primary hepatocytes and HepG2 cells were both transfected with Effectene Transfection Reagent from Qiagen (Valencia, CA) according to the manufacturer's instructions. Rat primary hepatocytes were plated in William's E medium containing 10% fetal calf serum, 1.0 M thyroxine, 0.1 M dexamethasone, and 100 units/ml penicillin in six-well Primaria tissue culture plates (Becton Dickinson, Franklin Lakes, NJ). Rat primary hepatocytes were transfected with 500 ng of test plasmid, 5 ng of pCMV-␤Galactosidase and pBluescript (used as a carrier DNA to adjust the total DNA amount to 1.5 g) in serum-free, penicillin and hormone-containing (dexamethasone, thyroxine) William's E medium. HepG2 cells were plated on 12-well plates in supplemented minimal essential medium containing 10% fetal calf serum, 0.125 unit/ml insulin, and 0.1 M dexamethasone. Twenty-four hours later, cells were transfected with 10 ng of test plasmid and 0.2 ng of pCMVsport␤Galactosidase and pBluescript (used as a carrier DNA to adjust the total DNA amount to 400 ng). After 16 h, the DNA complexed to Effectene reagent was removed and fresh medium was added with or without the C-36 peptide. Rat primary hepatocytes and HepG2 cells were harvested 24 and 48 h later, respectively, using the Dual-Light Kit (Tropix, Bedford, MA) according to the manufacturer's instructions. Luciferase activity was measured using a Monolight 2010 luminometer (Pharmingen, San Diego, CA), and background activity (pGL3-basic) was subtracted. Luciferase activity was normalized to ␤-galactosidase activity for transfection efficiency.
Electrophoretic Mobility Shift Assay-DNA oligonucleotides used as EMSA probes were labeled using [␥-32 P]ATP and T4 polynucleotide kinase (New England BioLabs) according to the manufacturer's instructions. In vitro translation/transcription of cDNAs encoding human FTF and human HNF-4␣, was performed using the TNT T7-coupled Rabbit Reticulocyte Lysate system from Promega (Madison, WI), according to the manufacturer's protocol. DNA binding reactions were performed in 50 mM KCl, 20 mM Tris-HCl (pH 8.0), 0.2 mM EDTA, 4% Ficoll, 1.0 g of poly(dI-dC), 1000-fold molar excess of an irrelevant single-stranded DNA, 4 l of the in vitro translated protein, and the indicated amounts of peptide in a final volume of 20 l. After 15-min incubation on ice the indicated 32 P-labeled DNA probe was added to the reaction. After 20min incubation on ice, samples were loaded onto a 4% polyacrylamide gel and subjected to electrophoresis at 4°C. Gels were dried, and exposures were made to XAR film (Kodak).
FTF-Peptide ELISA Binding Assay-Ninety-six well microtiter plates (Dynatech Laboratories) were coated with either C-36, C-35 (F 3 A) peptides, or bovine serum albumin (BSA) by incubating the plates overnight at 37°C with either peptide or BSA at 2.5 g/ml in phosphate-buffered saline solution, 0.05% NaN 3 . Plates were washed and treated with blocking buffer (0.017 M Na 2 B 4 O 7 , 0.12 NaCl, 0.05% Tween 20, 1 mM EDTA, 0.25% BSA, 0.05% NaN 3 ). FTF, in the form of a recombinant glutathione S-transferase (GST) fusion protein containing amino acids 141-501 of the human FTF, was then added. After 1-h incubation at 37°C, plates were washed, blocked, and incubated with goat anti-GST (Amersham Biosciences) for 1 h at 37°C. After washing and reblocking, a secondary rabbit anti-goat IgG conjugated to horseradish peroxidase was added and incubated for 1 h at 37°C. A Dako TMB One-Step Substrate system (Dako Corp., Carpenteria, CA) was used to develop the peroxidase activity, the reaction was terminated by adding 0.18 M H 2 SO 4 , and absorbance was measured at 450 nm. Specific binding was defined by subtracting the absorbance of the BSA wells from the corresponding peptide wells, and the number was multiplied by 10. For the competition experiments, the indicated molar excess of either C-36 or C-35 (F 3 A) peptides was included during the incubation with FTF.
Statistics-Results are expressed as the means Ϯ S.E. Statistical significance between two groups was determined by using the Student's t test. A p value of Ͻ0.05 was considered significant.

The C-36 Peptide Decreases Bile Acid Biosynthesis Rates by
Lowering 7␣-Hydroxylase mRNA Levels-We first investigated whether the C-36 peptide, resulting from cleavage of ␣ 1 -AT by neutrophil elastase, could alter the rates of cholesterol conversion into bile acids in primary rat hepatocytes. Rates of bile acid biosynthesis were measured by the conversion of [ 14 C]cho-lesterol into methanol-water-soluble material ( 14 C-bile acids). As shown in Fig. 1, the C-36 peptide significantly decreased bile acid biosynthesis rates by about 3-fold. No evidence of C-36 peptide toxicity was apparent from lactate dehydrogenase assay (data not shown) or from microscopic examination of the treated hepatocytes.
Cholesterol 7␣-hydroxylase is the rate-limiting step in the conversion of cholesterol into bile acids via the neutral pathway, so we next determined if the C-36 peptide could regulate 7␣-hydroxylase mRNA levels. Forty-eight hours after plating, the C-36 peptide was added to primary rat hepatocytes and total RNA was isolated after 24 h. Fig. 2 shows that the C-36 peptide down-regulated 7␣-hydroxylase mRNA levels in a concentrationdependent manner up to 10-fold. Fig. 3 shows that the C-36-mediated suppression of 7␣-hydroxylase mRNA is time-dependent, and it reaches its peak at 24 h. C-36 had no effect on 7␣-hydroxylase mRNA half-life based on actinomycin D experiments (data not shown).
To demonstrate the specificity of the C-36 peptide on 7␣-hydroxylase mRNA levels, we used a mutated peptide, C-35 (F 3 A), that has the same amino acid sequence as the C-36 peptide, except it has a mutation at residue 372 (F 3 A) and a deletion at residue 384 (Fig. 4). This mutation should prevent interaction of C-36 with the serpin-enzyme complex receptor, because the penta-amino acid peptide FVYLI is required for the binding (34). This peptide was tested for its ability to downregulate CYP7A1 mRNA levels in primary rat hepatocytes as compared with C-36. As shown in Fig. 4, the mutated peptide C-35 (F 3 A) had no significant effect on 7␣-hydroxylase mRNA levels, as compared with the C-36 peptide.
The C-36 Peptide Down-regulates mRNA Levels of Other Genes Involved in Cholesterol Metabolism-Because the regulation of a number of genes involved in cholesterol metabolism and, to a larger extent, in lipid metabolism depends on the size of the free cholesterol pool, we measured mRNA levels of HMG-CoA reductase, HMG-CoA synthase, and FAS by RPA or Northern blot analyses. Following the addition of 10 M C-36 peptide to primary rat hepatocytes, HMG-CoA reductase, HMG-CoA synthase, and FAS mRNA were suppressed between 2-and 3-fold (Fig. 5). Moreover, there was a smaller effect of the C-36 peptide on LDL receptor mRNA levels. Expression of housekeeping genes, such as actin and glyceraldehyde-3-phosphate dehydrogenase was not regulated (Fig. 5).
The C-36 Peptide Specifically Down-regulates 7␣-Hydroxylase in Vivo-To investigate whether the C-36-mediated suppression of 7␣-hydroxylase mRNA that we observed in tissue culture cells also occurs in vivo, mice were injected twice with 2 mg of the C-36 peptide each time. As controls, two other sets of mice were injected with either vehicle alone (PBS) or the C-35 (F 3 A) control peptide. Fig. 6 shows that injection of the C-36 peptide reduced 7␣-hydroxylase expression by more than 2-fold. Control C-35 (F 3 A) peptide had no significant effect.
The C-36 Peptide Specifically Down-regulates FTF-dependent Promoters-To characterize the molecular mechanisms involved in the regulation of 7␣-hydroxylase expression by the C-36 peptide, we investigated whether the C-36 peptide could regulate 7␣-hydroxylase promoter activity. Primary rat hepatocytes were transfected with a chimeric gene containing 342 nucleotides of the rat 7␣-hydroxylase promoter in front of the luciferase gene (designated pGL3-R7␣-342, Fig. 7). This promoter fragment has been shown to contain all the necessary elements for transcription and regulation by bile acids (23). As shown in Fig. 7, the C-36 peptide suppressed 7␣-hydroxylase promoter activity by about 2.5-fold. The C-35 (F 3 A) mutant peptide did not have any effect (data not shown). Another similar construct that contained only 143 nucleotides of the 7␣-hydroxylase promoter was similarly regulated by the C-36 peptide (data non shown).
It has been reported that FTF is required factor for 7␣hydroxylase transcription, and expression of this factor alone supports 7␣-hydroxylase promoter activity even in non-liver cells (22,23). Given the extent of the C-36-mediated suppression of 7␣-hydroxylase expression (over 10-fold, Fig. 2), it is reasonable to hypothesize that C-36 might work by preventing FTF-mediated activation. FTF binds to its cis-element located at position Ϫ133 to Ϫ125, which overlaps a direct repeat element (DR-1), which is a binding site for HNF-4␣ (35). To determine whether the FTF site and/or the DR-1 element could be involved in the C-36-mediated regulation of 7␣-hydroxylase promoter activity, a mutation was introduced within each site (Fig. 7). The DR-1 mutant 7␣-hydroxylase promoter was regulated at least as much by C-36 as the wild type. However, mutation of the FTF site rendered an inactive promoter that could not be tested for regulation by C-36.
Expression of the 12␣-hydroxylase gene is also dependent on FTF (26). To test whether the C-36 peptide could also suppress 12␣-hydroxylase promoter activity, we transfected a 12␣-hydroxylase promoter construct containing 865 nucleotides of the 12␣-hydroxylase 5Ј-flanking region in front of the luciferase gene, and, after transfection, cells were treated with or without C-36. Fig. 8 shows that the C-36 peptide down-regulated 12␣hydroxylase promoter activity by about 2.6-fold. A 5Ј-deletion construct (pGL3-R12␣-106) that contained only 106 nucleotides of the 12␣-hydroxylase 5Ј-flanking region was also regulated by C-36 (Fig. 8), indicating that the DNA element that mediates regulation by the C-36 peptide is within the first 106 nucleotides of the 12␣-hydroxylase promoter. To better characterize this element, we mutated those 106 nucleotides in blocks of 10 -15 nucleotides to generate mutants A through D, and tested them as above for regulation by C-36. Fig. 8 shows that all the mutants A through D were still regulated by the C-36 peptide. Mutants of the two FTF sites were not tested, because we have previously shown that mutations at those sites cause a complete lost of promoter (26).
The experiments shown above point to the FTF site as the DNA element involved in the C-36-mediated regulation of 7␣and 12␣-hydroxylase promoters. To further test whether FTF sites are involved in the C-36-mediated suppression of promoter activity, we tested another promoter that also require FTF for its activity, ␣ 1 -fetoprotein (AFP) (21). The AFP promoter was placed in the reporter plasmid pGL3, transfected into HepG2 cells, and tested for regulation by the C-36 peptide. Expression of the AFP promoter was regulated 2.6-fold upon addition of the C-36 peptide (data not shown).
The C-36 Peptide, but Not the Mutated Peptide, Specifically Binds to the FTF Protein and Inhibits Its DNA Binding Activity to the 7␣-and 12␣-Hydroxylase Promoters-FTF is a required transcription factor for 7␣-and 12␣-hydroxylase promoter activities, and one way to prevent FTF from activating these promoters is to prevent FTF binding to its cis-element (27). To determine whether the C-36 peptide could directly alter FTF binding to its recognition sites within those promoters, we performed electrophoretic mobility shift assay using in vitro made FTF protein and the C-36 peptide. The data in Fig. 9 shows that the C-36 peptide inhibited FTF binding to both the 7␣-and 12␣-hydroxylase promoter sequences in a concentrationdependent manner. This inhibitory effect was specific for the C-36 serpin peptide, because the mutated peptide C-35 (F 3 A) had no significant effect on FTF binding to either the 7␣-or the 12␣-hydroxylase promoter sequences (Figs. 10, A and B) at concentrations up to 60 M.
HNF-4␣ has been shown to bind to a DR-1 element that overlaps the FTF sites within the 12␣-hydroxylase promoter (Fig. 8), and HNF-4␣ is a required factor for 12␣-hydroxylase promoter activity (27). To determine whether the C-36 peptide inhibits HNF-4␣ binding to the DR-1 site we performed a similar electrophoretic mobility shift assay as shown above using in vitro made HNF-4␣. Fig. 11 shows that C-36 had no effect on HNF-4␣ binding to the 12␣-hydroxylase promoter sequence at concentrations up to 60 M.
To determine whether FTF and C-36 physically interact, we established an ELISA assay. The C-35 (F 3 A) mutant peptide was used as a control. FTF protein was prepared in Escherichia coli as a GST-FTF fusion protein, with amino acids 141-501 from the human FTF fused to GST. Increasing amounts of FTF resulted in increased binding indicating an interaction between FTF and C-36 (Fig. 12A). This interaction is specific for the C-36 peptide since the C-35 (F 3 A) mutant peptide shows significantly lower affinity for FTF. Recombinant GST did not show interaction with the C-36 peptide (data not shown). Furthermore, when both the C-36 and C-35 (F 3 A) peptides were added during the incubation with FTF, C-36 effectively competed for binding to the C-36-coated plate (Fig. 12B). The C-35 (F 3 A) mutant peptide was unable to compete. FIG. 10. Suppression of FTF binding to the 12␣-or 7␣-hydroxylase promoter FTF sites is peptide-specific. Electrophoretic mobility shift assays were performed with in vitro made FTF as described in Fig. 8, adding either the C-36 peptide or a mutated peptide (C-35 (F 3 A)) as indicated. A, the 7␣-hydroxylase promoter probe was used (CYP7A1). B, the 12␣-hydroxylase promoter probe was used (CYP8B1). The average of three experiments Ϯ S.E. is shown, and the data are expressed as a percentage of the binding activity without peptide. A representative experiment is shown at the bottom of each graph.
FIG. 9. C-36 serpin peptide decreases binding of FTF to the 7␣and 12␣-hydroxylase promoters. Electrophoretic mobility shift assays were performed as described under "Experimental Procedures" using 4 l of in vitro made FTF protein, and the indicated amounts of the C-36 peptide were used. Two probes were used, the wild type 12␣-hydroxylase promoter FTF site from nucleotides Ϫ150 to Ϫ121 (CYP8B1) and the wild-type 7␣-hydroxylase promoter FTF site from nucleotides Ϫ69 to Ϫ46 (CYP7A1). The probe sequences are indicated at the bottom. The arrows indicate the retarded bands corresponding to free probes and to either FTF⅐DNA complex.

DISCUSSION
The studies presented here establish a connection between inflammation and cholesterol homeostasis. We demonstrate that a ␣ 1 -AT-derived peptide (C-36) decreases bile acid synthesis in rat primary hepatocytes. We also show that this suppression, at least in part, occurs through a decrease in 7␣-hydroxylase promoter activity. C-36 also decreases 7␣-hydroxylase mRNA expression in injected mice. Furthermore, this suppression of promoter activity seems to occur by preventing binding of FTF to its recognition site within the 7␣-hydroxylase promoter. The regulation of hepatic gene expression by a specific peptide derived from a major serum protein is a novel mechanism of transcriptional regulation.
Considering that 7␣-hydroxylase is the rate-limiting step of bile acid synthesis through the "classic" pathway, which is responsible for about 50% of the total bile acid output (36), the decrease of about 3-fold in bile acid biosynthesis upon incubation of primary rat hepatocytes with the ␣ 1 -AT-derived peptide C-36 ( Fig. 1) correlates well with an almost complete suppression (10-fold) in 7␣-hydroxylase mRNA expression (Figs. 2 and  3). Additionally, the suppression of 12␣-hydroxylase expression by C-36 should result in a more hydrophobic bile, which has higher suppressor potency on 7␣-hydroxylase (19) resulting in a further suppression of 7␣-hydroxylase expression and bile acid synthesis. The effect of the C-36 peptide on 7␣-hydroxylase mRNA levels was highly specific, because a very similar peptide, C-35 (F 3 A), has no effect (Fig. 4). The mutation in the C-35 (F 3 A) peptide is located within the sequence that has been reported to bind to the serpin-enzyme complex receptor (7,37) and to mediate neutrophil chemotaxis (6) and up-regulation of ␣ 1 -antitrypsin mRNA (38). The C-36 peptide, but not the C-35 (F 3 A) control peptide, also suppressed 7␣-hydroxylase mRNA expression in the entire animal, suggesting that this observation made in primary hepatocytes may be physiologically significant.
The C-36 peptide also suppressed expression of other genes involved in cholesterol and lipid metabolism (Fig. 5), although this effect was modest as compared with the effect on 7␣hydroxylase. We speculate that this effect on HMG-CoA reductase, HMG-CoA synthase, LDL receptor, and FAS is a secondary effect due to repression of 7␣-hydroxylase. A decrease in bile acid synthesis should result in higher cellular cholesterol content that in turn should decrease the levels of the mature form for the sterol regulatory element binding proteins, which are transcriptional activators of all these four genes (39).
The ␣ 1 -AT C-36 peptide regulates 7␣-hydroxylase mRNA expression at the transcriptional level, at least in part, based on our promoter studies (Fig. 7). It is still possible that other mechanisms are also involved, because the C-36 peptide regulates 7␣-hydroxylase promoter activity only 2.6-fold, compared with a 10-fold regulation of 7␣-hydroxylase mRNA expression. Nevertheless, the C-36 peptide not only regulates the 7␣-hy-droxylase promoter, but also other promoters that are dependent on FTF for their expression, such as the 12␣-hydroxylase (Fig. 8) and the AFP promoters.
Several lines of evidence support the notion that the C-36 peptide targets FTF to suppress bile acid biosynthesis. First, FTF is the major factor required for 7␣-hydroxylase expression, because mutations within its recognition element render an inactive promoter (22), whereas mutations anywhere else in the promoter have a relatively minor effect (23). A strong suppression of 7␣-hydroxylase, such as the 10-fold suppression caused by C-36, should work through a factor that plays a major role in 7␣-hydroxylase promoter activity, such as FTF. Second, two other promoters also dependent on FTF for its activity, the 12␣-hydroxylase promoter (26) (Fig. 8) and the AFP promoter (21) (data not shown), are also regulated by C-36. Third, mutations of different parts of the 7␣-and 12␣hydroxylase promoters, excluding the FTF sites, have no effect on the C-36-mediated regulation of their activity. Unfortunately we could not directly test whether mutation of the FTF sites would have an effect on the C-36 regulation, because those mutations inactivate both promoters (22,26) (Figs. 7 and 8).
Another line of evidence that adds further support to FTF being the factor involved in the C-36-mediated suppression of bile acid synthesis came from our in vitro binding studies of FTF to its recognition sites on both the 7␣-and the 12␣hydroxylase promoters (Fig. 9). FTF binding to those sites correlates with promoter activities (22,26). When we included the C-36 peptide in the binding reaction, FTF binding to both promoters dramatically decreased, and this suppression was highly specific for both the peptide and the transcription factor, because a mutant peptide, C-35 (F 3 A), had no effect on FTF binding to either promoter (Fig. 10). Another nuclear receptor, FIG. 11. Inability of the C-36 peptide to alter HNF-4␣ binding to the 12␣-hydroxylase promoter. Electrophoretic mobility shift assays were performed as described in Fig. 9, but using the 12␣-hydroxylase promoter FTF probe and in vitro made HNF-4␣. HNF-4␣, also binds to overlapping sequences with FTF within the 12␣-hydroxylase promoter, and it plays a role in basal promoter activity (27). However, the C-36 peptide had no effect on binding of HNF-4␣ to the 12␣-hydroxylase promoter probe (Fig. 11).
The C-36-mediated inhibition of FTF binding to its DNA recognition site suggests that the C-36 peptide suppresses promoter activity by preventing FTF binding to its recognition site. This data could be explained either by direct binding of the C-36 peptide to the FTF recognition site or by direct interaction between FTF and C-36. We were unable to observe binding of the C-36 to the FTF recognition site (data not shown), but we did observe physical interaction between FTF and C-36 using an ELISA assay (Fig. 12). This interaction does not require the FTF DNA binding domain, because we used a GST-FTF-(141-501) fusion protein that has its DNA binding domain deleted, which suggests that the C-36 peptide inhibits FTF activity by inducing a conformational change in its structure such that it cannot bind to its DNA recognition site. This is reminiscent of the ligand-induced conformational changes that are believed to lead to the dissociation of multiprotein complex formed between the ligand binding domain of some nuclear receptors and hsp90, hsp70, p59, and other factors (40,41). It should be pointed out that, although there is no known ligand for FTF, it has a ligand domain similar to other nuclear receptors located at the same relative position with respect to its DNA binding domain. This proposed mechanism of action is similar to the one proposed for the SHP-mediated suppression of 12␣-hydroxylase promoter activity that is involved in the bile acid-mediated suppression of its activity (27). Whether the C-36 peptide and SHP interact with the same FTF domain has not been determined. It should be mentioned that attempts to show differential FTF binding activity in nuclear extracts prepared from cells treated with the C-36 peptide failed (data not shown), presumably due to loss of the C-36 peptide during nuclei isolation. Interestingly, differential FTF binding has not been demonstrated in liver nuclear extracts prepared from cells treated with bile acids.
The physiological role and potential significance of the C-36mediated suppression of bile acid synthesis is unknown at this point, but it is tempting to speculate that the increase in ␣ 1 -AT synthesis that occurs during the acute phase response could play a role in the development of atherosclerosis. A decrease in 7␣-hydroxylase expression and bile acid synthesis should result in higher plasma cholesterol levels due to less cholesterol being removed from the body. Consistent with this, plasma levels of ␣ 1 -AT are correlated with the development of both early and advanced atherosclerosis in the carotid arteries (42). Other studies have previously shown a down-regulation of either the classic (43) or the alternative (44) bile acid biosynthesis pathways by acute phase response factors that together with the effect of ␣ 1 -AT-derived C-36 peptide could contribute to the hypertriglyceridemia observed in response to inflammation or injury. Further studies on the role and mechanism of action of ␣ 1 -AT-derived peptides on bile acid synthesis should provide answers to these questions and perhaps support the notion that, under the inflammatory conditions associated with atherosclerosis, ␣ 1 -AT may play a role not only as a protease inhibitor, but also as a reservoir of physiologically active peptide degradation products.