Disruption of the Oxysterol 7 a -Hydroxylase Gene in Mice*

Mice without oxysterol 7 a -hydroxylase, an enzyme of the alternate bile acid synthesis pathway with a sexually dimorphic expression pattern, were constructed by the introduction of a null mutation at the Cyp7b1 locus. Animals heterozygous ( Cyp7b1 1 / 2 ) and homozygous ( Cyp7b1 2 / 2 ) for this mutation were grossly indistinguishable from wild-type mice. Plasma and tissue levels of 25- and 27-hydroxycholesterol, two oxysterol sub-strates of this enzyme with potent regulatory actions in cultured cells, were markedly elevated in Cyp7b1 2 / 2 knockout animals. Parameters of bile acid metabolism as well as plasma cholesterol and triglyceride levels in male and female Cyp7b1 2 / 2 mice were normal. The cholesterol contents of major tissues were not altered. In vivo sterol biosynthetic rates were unaffected in multi-ple tissues with the exception of the male kidney, which showed a ; 40% decrease in de novo synthesis versus controls. We conclude that the major physiological role of the CYP7B1 oxysterol 7 a -hydroxylase is to metabolize 25- and 27-hydroxycholesterol and that loss of this enzyme in the liver is compensated for by increases in the synthesis of bile acids by other pathways. A failure to catabolize oxysterols in the male kidney may lead to a decrease in de novo sterol synthesis. data indicate that the major physiological role of the CYP7B1 oxysterol 7 a -hydroxylase is to inactivate oxysterols.

7␣-Hydroxylated bile acids are the end products of cholesterol (5-cholesten-3␤-ol) catabolism whose synthesis is required to maintain cholesterol homeostasis. Bile acids are synthesized in the liver by two metabolic pathways that differ in their initial steps (1). The classical pathway is initiated with the conversion of cholesterol to 7␣-hydroxycholesterol (cholest-5ene-3␤,7␣-diol) by the enzyme cholesterol 7␣-hydroxylase (encoded by the Cyp7a1 locus). The alternate pathway involves an initial hydroxylation of the side chain of cholesterol to produce an oxysterol, which is then hydroxylated at the 7-position by an oxysterol 7␣-hydroxylase (encoded by the Cyp7b1 locus). Once these enzymes form 7␣-hydroxylated intermediates, subsequent steps leading to the synthesis of a primary bile acid are shared between the two pathways (2).
The regulation of these pathways and their individual contributions to bile acid metabolism in the mouse are currently being explored using molecular methods. Studies to date show that modulating transcription of the cholesterol 7␣-hydroxy-lase gene regulates output from the classical pathway. Excess bile acids suppress transcription from this gene by acting as ligands for the nuclear hormone receptor FXR (3)(4)(5), whereas oxysterols increase transcription of the gene by activating the nuclear hormone receptor LXR␣ (6 -8). Perturbations in these regulatory inputs have profound consequences for bile acid and cholesterol metabolism in the mouse. Thus, feeding bile acids leads to an approximate doubling of the bile acid pool, suppression of cholesterol 7␣-hydroxylase, and a doubling of the hepatic cholesterol level (9,10). Conversely, deletion of the LXR␣ gene prevents the activation of cholesterol 7␣-hydroxylase and causes a pathological accumulation of hepatic cholesterol (11). The regulation of the alternate pathway of bile acid synthesis is currently less well understood, although several studies suggest that the expression of early enzymes in the pathway may be modulated (12)(13)(14)(15).
Deletion of the cholesterol 7␣-hydroxylase gene in mice eliminates the classical pathway of bile acid synthesis and causes an 80% reduction in the bile acid pool size (9,16). This loss also leads to an accumulation of bile acids with different chemical structures from those of animals with two functioning pathways, and it alters several parameters of cholesterol metabolism in the male (9,16). These data indicate that the classical pathway is the major synthetic route to bile acids in mice and that the alternate pathway synthesizes ϳ20% of the normal bile acid pool. Mice deficient in oxysterol 7␣-hydroxylase are not yet available; however, analysis of a human child lacking this enzyme illustrates the overall importance of the alternate pathway. The phenotype of this subject includes a dysfunctional liver (cholestasis), massively elevated serum levels of oxysterols (hyperoxysterolemia), and the absence of 7␣-hydroxylated sterols and bile acids (17). The latter symptom is puzzling as the cholesterol 7␣-hydroxylase gene is intact in this patient. Two explanations are postulated to underlie the absence of bile acids in this individual. First, the outcome may be due to a difference in the ontogeny of the two pathways of bile acid synthesis; the alternate pathway may be activated in humans prior to the classical pathway, leading to a dependence on the alternate pathway in neonates (17). Second, the accumulation of oxysterols may cause liver failure and a consequent loss of the classical pathway.
To resolve this issue and to gain further insight into the role of the alternate pathway in bile acid and cholesterol metabolism, we have produced a line of mice with a disruption in the Cyp7b1 gene encoding oxysterol 7␣-hydroxylase. Bile acid metabolism is normal in these animals, suggesting that either the classical pathway or another synthetic route can fully compensate for the absence of the alternate pathway. Serum and tissue levels of two oxysterols accumulate in micromolar amounts and may adversely affect de novo cholesterol synthesis in the male kidney. The data indicate that the major physiological role of the CYP7B1 oxysterol 7␣-hydroxylase is to inactivate oxysterols.

EXPERIMENTAL PROCEDURES
Construction of Targeting Vector-Two overlapping DNA fragments spanning the 3Ј-end of the Cyp7b1 gene were isolated by standard procedures (18) from a bacteriophage library prepared from 129SvEv genomic DNA. This library was a kind gift from Martin Matzuk (Baylor College of Medicine, Houston, TX). A short arm consisting of a 1.6kilobase pair SacI fragment from the last exon of the gene was inserted on the 3Ј-side of a Neo r gene in the pPolIIshort-neobPA-HSVTK vector (19). A long arm consisting of a 12-kilobase pair SacI fragment spanning introns 4 and 5 of the Cyp7b1 gene was inserted on the 5Ј-side of the Neo r gene to produce the final targeting vector (see Fig. 1A). Homologous recombination should result in deletion of most of exon 6 from the Cyp7b1 gene, which normally encodes amino acids 410 -507 of the protein.
Embryonic Stem Cell Culture-Mouse I1C embryonic stem (ES) 1 cells derived from the 129SvEv mouse strain were cultured on mitotically inactivated, leukemia inhibitory factor-producing STO feeder cells in Dulbecco's minimal essential medium supplemented with 15% (v/v) fetal calf serum, 0.1 mM nonessential amino acids, 2 mM L-glutamine, and 0.1 mM ␤-mercaptoethanol. Growth was at 37°C in an atmosphere of 5% CO 2 . Approximately 2 ϫ 10 7 ES cells were electroporated with 100 g of SalI-linearized targeting vector to initiate homologous recombination. Electroporated cells were cultured in the presence of G418 (190 g/ml) and ganciclovir (2 M) for 12 days prior to subcloning of single ES cell colonies. Those that had undergone homologous recombination were identified by Southern blot analyses of PstI-digested genomic DNA using a Cyp7b1 probe located 3Ј of the short arm fragment in the targeting vector. The frequency of clones with homologous recombination events was 5.7%. Positive clones were expanded and injected into C57BL/6J blastocysts. High percentage male chimeras from two different ES cell clones were crossed with female C57BL/6J mice to generate two independent lines of animals carrying the disrupted Cyp7b1 allele. The experiments reported here were performed in mice of mixed genetic background (C57BL/6J-129SvEv). No phenotypic differences were observed between knockout mice derived from the two independent ES cell lines.
RNA and Protein Blotting-Total RNA was extracted from previously frozen liver tissue using RNA Stat-60 kits (Tel-Test "B", Inc., Friendswood, TX). Polyadenylated mRNA was prepared from pooled total RNA using an mRNA purification kit (Amersham Pharmacia Biotech).
Total liver extracts were prepared by homogenizing previously frozen liver tissue in 0.25 M sucrose buffer (0.25 M sucrose, 20 mM Tris acetate (pH 7.4) 1 mM EDTA, 50 g/ml N-acetyl-Leu-Leu-norleucinal, 0.1 mM Pefabloc, 5 g/ml pepstatin A, 10 g/ml leupeptin, and 2 g/ml aprotinin) using a Polytron, followed by centrifugation at 600 ϫ g for 10 min. Homogenates from three to six mice of the same sex and Cyp7b1 genotype were pooled for immunoblot analysis. Mitochondrial membranes were prepared from pooled homogenates by centrifugation at 10,000 ϫ g for 20 min. The resulting pellets, which contained the mitochondria, were washed once with potassium phosphate buffer (10 mM potassium phosphate (pH 7.4), 0.1 mM EDTA, and 5 mM dithiothreitol) and resuspended in the same buffer containing 20% (v/v) glycerol prior to analysis. To prepare microsomal membranes, supernatants from the 10,000 ϫ g centrifugation were subjected to a second centrifugation step at 100,000 ϫ g for 30 min at 4°C. The resulting pellets were resuspended in 50 mM Tris acetate (pH 7.4), 1 mM EDTA, and 20% (v/v) glycerol prior to immunoblot and enzymatic analyses. For immunoblotting, aliquots of protein (75-150 g) were separated by electrophoresis on 10% polyacrylamide gels containing SDS and then transferred to Immobilon-P polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA). A rabbit-generated polyclonal antibody that recognizes amino acids 266 -281 of the mouse CYP7B1 oxysterol 7␣-hydroxylase was used to detect the protein (13). A rabbit-generated polyclonal antibody that recognizes amino acids 476 -490 of the murine cholesterol 7␣-hydroxylase was used to detect the CYP7A1 protein. A polyclonal antibody against the immunoglobulin heavy chain-binding protein was purchased from Stressgen Biotech Corp. (Victoria, Canada), and an antibody that recognizes the mitochondrial carnitine palmitoyltransferase-2 protein was from Vickie Esser and J. Dennis McGarry (University of Texas Southwestern Medical Center).
Enzyme Assays-Total protein homogenates were prepared from liver as described above. CYP7B1 oxysterol 7␣-hydroxylase activity was assayed in a 0.5-ml reaction containing 250 g of protein, 1 M 25-[ 3 H]hydroxycholesterol (delivered in ethanol), 50 mM Tris acetate (pH 7.4), and 1 mM EDTA. The reaction was prewarmed in a 37°C water bath for 5 min prior to the addition of 1.5 mM NADPH. After 15 min at 37°C, 6 ml of methylene chloride was added to terminate the reaction and to extract sterols. The resulting organic phase was evaporated to dryness under nitrogen, and the pellets were dissolved in 40 l of chloroform/methanol (2:1, v/v) and analyzed by thin-layer chromatography on Silica Gel LK5D 150-A plates (Whatman) in a solvent system of toluene/ethyl acetate (2:3, v/v).
Animal Diets-Mice were housed in plastic cages in a temperaturecontrolled room (22°C) with 12-h light cycling. Animals were fed ad libitum a cereal-based rodent diet (diet 7001, Harlan Teklad, Madison, WI) containing 0.2 mg of cholesterol and 40 mg of total lipid per g of diet. Cholesterol absorption, sterol synthesis rates, bile acid pool sizes and compositions, and plasma and tissue cholesterol levels were measured in 3-4-month-old mice.
Sexual dimorphism in CYP7B1 expression was examined in 10-weekold male and female wild-type C57BL/6J mice (stock 000664, Jackson Laboratory, Bar Harbor, ME). Male mice bearing a mutation in the androgen receptor were purchased from the same vendor (stock 001809). Orchiectomies were performed using standard methods of animal surgery. Where indicated, castrated mice were treated with dihydrotestosterone (1.5 mg in a 21-day release pellet; Innovative Research of America, Toledo, OH) or estradiol (0.5 mg in a 21-day release pellet) as described previously (22).
Cholesterol Balance Studies-The in vivo rate of sterol synthesis in the liver and major extrahepatic organs was measured as described (9). Wild-type and Cyp7b1 Ϫ/Ϫ mice (n ϭ six per genotype and sex) were used in these measurements. Intestinal cholesterol absorption was measured by a fecal dual-isotope ratio method in groups of 10 animals per genotype and sex (9,23). Bile acid pool size and composition were determined using a high pressure liquid chromatography method (9). Plasma total cholesterol and triglyceride concentrations were measured using a reagent mixture (catalog no. 1127771, Roche Molecular Biochemicals). For tissue cholesterol measurements, organs from six mice of each genotype and sex were collected, saponified in ethanol/KOH, and extracted with petroleum ether in the presence of stigmastanol as an internal recovery standard. The amount of extracted cholesterol was measured by gas chromatography and normalized to recovery by comparison with the amount of stigmastanol.

RESULTS
To disrupt the murine oxysterol 7␣-hydroxylase gene, the targeting vector illustrated in Fig. 1A was constructed. Electroporation of linearized vector DNA into 129SvEv-derived I1C embryonic stem cells followed by positive/negative selection and screening produced 10 lines that harbored the desired deletion mutation (Fig. 1A). Four of these cell lines were injected into recipient C57BL/6J blastocysts to produce chimeric males, two of which subsequently transmitted the mutation through the germ line to produce heterozygous mice (Cyp7b1 ϩ/Ϫ ). Crossing heterozygous mice produced a mendelian distribution of wild-type and mutant genotypes. Homozygous Cyp7b1 Ϫ/Ϫ mice were indistinguishable in terms of survival, gross physical appearances, and behaviors from heterozygous and wild-type animals. Backcrossing chimeric males with 129Sv females produced a line of Cyp7b1 Ϫ/Ϫ congenic mice that also did not differ from wild-type mice in these descriptive parameters.
To confirm that deletion of exon 6 from the Cyp7b1 gene eliminated enzyme activity, hepatic lysates were prepared from wild-type and knockout mice of both sexes and assayed for oxysterol 7␣-hydroxylase enzyme activity. As shown in Fig. 1B, wild-type mice expressed an activity that converted 25-hydroxycholesterol into several 7␣-hydroxylated products whose chemical identities were determined previously (13). This activity was dependent on the inclusion of NADPH in the reaction mixture and was more abundant in males than in females. In contrast, extracts prepared from Cyp7b1 Ϫ/Ϫ mice contained little or no oxysterol 7␣-hydroxylase activity, which suggests that the introduced mutation created the expected null allele at the Cyp7b1 locus.
RNA blotting of poly(A) ϩ -enriched mRNA isolated from male and female wild-type mice with an oxysterol 7␣-hydroxylase cDNA probe revealed a single hybridizing mRNA of ϳ2.2 kilobase pairs that was more prevalent in males (Fig. 1C). This mRNA was not detected in RNA isolated from the livers of knockout mice, where instead a series of larger mRNAs were found (Fig. 1C). These aberrant mRNAs may be the products of abnormal splicing from the mutated Cyp7b1 locus since splice acceptor sequences in the last intron of the gene were removed by the introduced deletion (Fig. 1A).
The effects of the mutation on the expression of the CYP7B1 protein and the ability of the abnormal mRNAs to encode the protein were tested by immunoblotting. Homogenates were prepared from the livers of male and female wild-type and mutant mice, separated by gel electrophoresis, and then blotted with a polyclonal antibody that recognizes amino acids 266 -281 of the oxysterol 7␣-hydroxylase enzyme. As shown in Fig.  1D, a protein of ϳ53 kDa was detected in male and female wild-type mouse livers, but not in this tissue from knockout mice. The amount of this protein in wild-type mice was again higher in male versus female liver.
The results of Fig. 1 suggested a sexual dimorphism in the expression of hepatic oxysterol 7␣-hydroxylase that could impact on the phenotype of Cyp7b1 Ϫ/Ϫ mice. To confirm these observations, we examined CYP7B1 protein levels in different tissues of male and female mice and the effects of the hormonal environment on these levels. As shown in Fig. 2, immunoblotting of total protein extracts prepared from liver and kidney showed higher levels of CYP7B1 protein in male (lanes 1 and 9) versus female (lanes 2 and 10) tissues. The amount of protein was reduced in the liver and kidney upon castration of wildtype males (lanes 3 and 11) and was reduced to almost undetectable levels in male mice with a mutation in the androgen receptor (Tfm mice) (lanes 6 and 14). The apparent effects of androgens on CYP7B1 expression were confirmed by administration of dihydrotestosterone to castrated males: this androgen increased the levels of CYP7B1 protein in the tissues of previously castrated males (lanes 4 and 5 and lanes 12 and 13). The administration of estradiol to intact male mice elevated the enzyme level in the liver (lane 7), but not the kidney (lane 15); however, castration of female mice did not affect CYP7B1 expression levels in either tissue (lanes 8 and 16).
The experiments summarized in Table I were carried out to determine how loss of oxysterol 7␣-hydroxylase affected cholesterol and bile acid metabolism. Plasma cholesterol and triglyceride levels were not significantly different between adult mice of sundry Cyp7b1 genotypes and sexes (Table I), nor were differences detected in the lipoprotein profiles of pooled plasmas from these animals (data not shown). The cholesterol contents of 12 tissues (adrenal gland, brain, carcass, colon, gonad, heart, small intestine, kidney, liver, lung, spleen, and stomach) were also not significantly different between wildtype and mutant mice. The bile acid pool sizes were similar in control and knockout male mice (Table I). Female mice of both genotypes had identical pool sizes, which were ϳ50% larger than those of their male counterparts ( Table I). Deletion of the Cyp7b1 gene did not change the compositions of the bile acid pools in male and female mice, nor did it affect intestinal cholesterol absorption (Table I).
These data suggested that loss of the CYP7B1 oxysterol 7␣-hydroxylase and hence of the alternate pathway of bile acid synthesis was compensated for by increased output from the cholesterol 7␣-hydroxylase pathway or perhaps another pathway. To test this hypothesis, we determined the expression levels of cholesterol 7␣-hydroxylase as a readout of the classical pathway and of sterol 27-hydroxylase as a measure of the alternate pathway. The levels of cholesterol 7␣-hydroxylase mRNA were not changed between normal and knockout mice of both sexes as judged by blotting (Fig. 3A). However, Cyp7b1 Ϫ/Ϫ knockout mice had elevated levels (30 -40%) of cholesterol 7␣hydroxylase protein (Fig. 3B). There were no changes in sterol 27-hydroxylase mRNA and protein between wild-type and mutant mice (Fig. 3,A and B), although males of both genotypes had ϳ2-fold more sterol 27-hydroxylase protein than females (Fig. 3B).
A chemical phenotype was revealed in Cyp7b1 Ϫ/Ϫ mice when isotope dilution-gas chromatography-mass spectrometry was used to measure levels of side chain oxysterols in the plasma, livers, and kidneys of wild-type and knockout animals of both sexes. As shown in Fig. 4A, the levels of free and esterified 24-hydroxycholesterol in the plasma were not significantly different between animals of various Cyp7b1 genotypes. In contrast, the plasma levels of 25-and 27-hydroxycholesterol were markedly increased in both male and female knockout mice. These increases were larger in Cyp7b1 Ϫ/Ϫ males than in Cyp7b1 Ϫ/Ϫ females, and both free and esterified pools of these two oxysterols were elevated (Fig. 4A). The liver and kidney contents of 25-and 27-hydroxycholesterol also were increased (Fig. 4, B and C, respectively), suggesting that plasma levels dictated tissue oxysterol content. These data also revealed marked differences in the distribution of free and esterified forms of the oxysterols in different tissues. In the plasma, the distribution of free and esterified forms was roughly equal among the three oxysterols (Fig. 4A), whereas in the liver, 24and 25-hydroxycholesterol were largely present in free form and 27-hydroxycholesterol in an esterified form (Fig. 4B). A FIG. 2. Sexual dimorphism in hepatic and renal CYP7B1 expression. Microsomal membranes were prepared from the livers or kidneys (n ϭ 3) of male or female mice treated in the indicated ways. Aliquots (75 g) of protein were subjected to immunoblotting using a polyclonal anti-peptide antibody directed against amino acids 266 -281 of the CYP7B1 oxysterol 7␣-hydroxylase. Lanes 1 and 9, extracts from untreated male mice; lanes 2 and 10, extracts from untreated female mice; lanes 3 and 11, extracts from male mice castrated 7 days prior to harvest; lanes 4 and 12, extracts from male mice castrated 7 days prior to harvest and then treated with dihydrotestosterone (DHT) for 1 day; lanes 5 and 13, extracts from male mice castrated 7 days prior to harvest and then treated with dihydrotestosterone for 3 days; lanes 6 and 14, extracts from male mice harboring a mutation in the X-linked androgen receptor gene (Tfm); lanes 7 and 15, extracts from male mice treated with estradiol (E) for 7 days; lanes 8 and 16, extracts from female mice castrated 7 days prior to harvest. Bip, immunoglobulin heavy chain-binding protein. majority of each oxysterol was present in free form in the kidney (Fig. 4C). The concentrations of serum 24-and 27-hydroxycholesterol measured in male wild-type mice (Fig. 4A) were comparable to those reported in a previous study (24). However, the values determined for liver oxysterol contents (Fig. 4B) were substantially lower (16 -30-fold) than those measured in two earlier studies (25,26). These differences may reflect dietary influences (the mice used here were fed normal chow, whereas those in the earlier study (25) were fed normal chow supplemented with 20% peanut oil) or differences in the detection methods (26). Although precautionary measures to exclude autoxidation of cholesterol were not taken during sample workup in the experiments of Fig. 4, the three oxysterols measured here do not form spontaneously or are minor autoxidation products of cholesterol (21). Thus, the nonenzymatic formation of these sterols contributed very little to the final values measured.
To determine the consequences of hyperoxysterolemia in the Cyp7b1 Ϫ/Ϫ mice for cholesterol metabolism, we measured in vivo sterol synthetic rates by following the incorporation of [ 3 H]water into digitonin-precipitable sterols. The data of Fig.  5A show that there were no significant differences in sterol synthetic rates between male wild-type and knockout mice in the adrenal gland, brain, heart, lung, and spleen. However, in the experiment shown, a 46% reduction in de novo sterol synthesis was detected in the kidneys of male Cyp7b1 Ϫ/Ϫ mice. In a second experiment involving six animals per group, sterol synthesis was reduced 34% in the knockout male kidney (data FIG. 3. Cholesterol 7␣-hydroxylase and sterol 27-hydroxylase levels in mice of different Cyp7b1 genotypes. A, equal amounts of hepatic poly(A) ϩ -enriched mRNA from six animals of the indicated sexes and genotypes were pooled, and aliquots (5 g) were separated by electrophoresis through glyoxal-agarose gels. After transfer to nylon filters, cholesterol 7␣-hydroxylase mRNA (CYP7A1; left panel) and sterol 27-hydroxylase mRNA (CYP27; right panel) were detected by hybridization using radiolabeled cDNA probes. The filters were stripped and reprobed with a second radiolabeled cDNA to detect the cyclophilin mRNA (lower panels). The relative amounts of CYP7A1 and CYP27 mRNAs (Level) in each lane were determined by phosphoimaging after normalization to the cyclophilin signal. B, equal amounts of liver were pooled from six animals, and microsomal (CYP7A1) and mitochondrial (CYP27) proteins were prepared as described under "Experimental Procedures." Aliquots (150 g of protein) were separated on SDS-polyacrylamide gels, transferred to membranes, and blotted with anti-peptide antibodies directed against cholesterol 7␣-hydroxylase (CYP7A1; left panel) or sterol 27-hydroxylase (CYP27; right panel). Control (Ctl) lanes contained aliquots of hepatic microsomal or mitochondrial proteins isolated from Cyp7a1 Ϫ1Ϫ /Cyp27 Ϫ1Ϫ double knockout mice. Membranes were stripped and reprobed with an antibody against the immunoglobulin heavy chain-binding protein (BIP; left panel) or against carnitine palmitoyltransferase-2 (CPT2; right panel) for normalization purposes (Level). kb, kilobase pairs.

FIG. 4. Plasma and tissue levels of oxysterols.
A, a pool of plasma was prepared from mice (n ϭ 6) of the indicated sex and Cyp7b1 genotype, and levels of oxysterols were determined by isotope dilutiongas chromatography-mass spectrometry as described under "Experimental Procedures." White bars on the histogram indicate total oxysterol levels (free plus esterified), and shaded bars indicate free oxysterol levels. The levels of free and total 25-hydroxycholesterol in the livers of male Cyp7b1 ϩ/ϩ mice were the same. B, same as in A, except oxysterols were extracted from the livers of mice (n ϭ 3) of the indicated Cyp7b1 genotype and sex. C, same as in A, except oxysterols were extracted from the kidneys of mice (n ϭ 3) of the indicated Cyp7b1 genotype and sex. In several cases, the levels of free oxysterols slightly exceeded or were identical to those of total oxysterols. not shown). In both experiments, this reduction was sex-specific, as renal sterol synthetic rates were similar in wild-type and knockout females, as were rates in the heart, liver, lung, and spleen (Fig. 5B).

DISCUSSION
In this work, we describe the phenotype of mice lacking the oxysterol 7␣-hydroxylase encoded by the Cyp7b1 gene. The most striking finding in these animals is a large increase in circulating and tissue levels of two oxysterols, 25-and 27hydroxycholesterol. When added to the medium of cultured cells at low micromolar concentrations, these oxysterols repress the synthesis of cholesterol and the expression of the low density lipoprotein receptor (27,28). However, in the Cyp7b1 Ϫ/Ϫ mice, the accumulation of these oxysterols to similar levels does not appear to adversely affect cholesterol synthesis or accumulation in most tissues. The exception to this finding is the male kidney, in which de novo cholesterol synthesis is decreased by ϳ40% in knockout mice.
There are several possible reasons for why the hyperoxysterolemia induced in Cyp7b1 Ϫ/Ϫ mice did not have adverse consequences for overall cholesterol metabolism. First, the levels of oxysterols reached in these animals may not be high enough to trigger repression of sterol-regulated genes. Although concentrations of 25-hydroxycholesterol in the 0.1-1.0 M range are sufficient to down-regulate enzymes involved in cholesterol biosynthesis in vitro (28), and the circulating levels of both 25and 27-hydroxycholesterol are in this range (Fig. 4), a still higher concentration may be required for an in vivo effect. We are currently testing this hypothesis by crossing transgenic mice that overexpress the cholesterol 25-hydroxylase gene (29) with Cyp7b1 Ϫ/Ϫ mice in an effort to further increase circulating levels of this oxysterol. Second, the oxysterols may not be present in an active form within cells. Oxysterol levels in the blood of knockout animals are similar to those measured in tissues (Fig. 4), indicating that the sterols are effectively delivered to cells. Once in the cell, however, binding proteins that render them inactive as regulatory molecules may sequester oxysterols. One candidate is the oxysterol-binding protein, which has selectivity for 25-and 27-hydroxycholesterol (30,31). There may also be other binding proteins that accomplish this task and that are expressed in a tissue-specific fashion, thus explaining the selective decrease in sterol synthesis observed in the kidneys of male knockout animals (Fig. 5A).
A third explanation for these observations is that the form in which the oxysterols exist in the tissue may determine their ability to regulate cholesterol synthesis and other parameters of lipid metabolism. An esterified oxysterol may be a less potent regulator than one in free form, and esterification may represent a general means to inactivate oxysterols (26,32). This idea is supported by the current findings that excess oxysterols do not appear to affect cholesterol synthesis in the livers of Cyp7b1 Ϫ/Ϫ mice, where they exist mostly in an esterified form, but do reduce de novo cholesterol synthesis by ϳ40% in the male kidney, where they are present in largely free form (Figs. 4 and 5). Why this decrease in sterol synthesis is not detected in the female kidney despite an elevation of free oxysterols is not clear.
A fourth hypothesis that must be considered is that the ability of oxysterols to down-regulate genes may be only an in vitro phenomenon. We believe this notion is unlikely based on the existence of oxysterol biosynthetic enzymes in many species and of enzymes like the CYP7B1 oxysterol 7␣-hydroxylase that specifically catabolizes these compounds. The striking phenotype of the human subject lacking CYP7B1 also argues for an important role for oxysterols in lipid metabolism (17).
The CYP7B1 enzyme is expressed at higher levels in the liver and kidney of the male compared with the female (Fig. 2). Androgens appear to be at least partially responsible for these differences, as castration and an inherited mutation in the androgen receptor reduce CYP7B1 expression in males. This sexually dimorphic expression pattern may explain the finding that de novo cholesterol synthesis is reduced by ϳ40% in the kidneys of Cyp7b1 Ϫ/Ϫ males, but not females (Fig. 5). In wildtype animals, the rates of sterol synthesis in the male kidney are approximately twice those in the female kidney, whereas in Cyp7b1 Ϫ/Ϫ mice, synthesis rates are similar in males and females and resemble those of the wild-type female (Fig. 5). The increased sterol synthesis in the normal male may thus be related to the need to replace cholesterol catabolized by the oxysterol 7␣-hydroxylase enzyme. In agreement with this idea, deletion in mice of the sterol 27-hydroxylase enzyme, which synthesizes a substrate for the CYP7B1 oxysterol 7␣-hydroxylase enzyme, also causes a 50% reduction in sterol synthesis in the male (but not female) kidney. 2 These data imply that the male kidney in the mouse plays a unique and as yet to be determined role in cholesterol metabolism, which we speculate may be related to the synthesis and excretion of pheromones in the urine.
The absence of a bile acid phenotype in Cyp7b1 Ϫ/Ϫ mice is easier to explain. The output from the classical pathway is regulated by the level of bile acids (2). When this level is decreased, the activity of cholesterol 7␣-hydroxylase and the synthesis of bile acids increase to restore homeostasis. Thus, a decrease in the output from the alternate pathway by inactivation of the CYP7B1 oxysterol 7␣-hydroxylase should be compensated for by an increase in synthesis via the classical pathway. The alternate pathway synthesizes ϳ20% of the bile acid FIG. 5. Rates of sterol synthesis in the tissues of Cyp7b1 wildtype and knockout mice. Mice of both genotypes and sexes (3 months of age) were maintained on a basal chow diet. A bolus of [ 3 H]water (40 mCi) was administered by intraperitoneal injection, and the animals were killed 1 h later. The indicated tissues were dissected, and the amount of [ 3 H]water incorporated into digitonin-precipitable sterols was determined as described under "Experimental Procedures." Values represent the mean Ϯ S.E. of the data. A, male mice. The asterisk denotes that the value for the Cyp7b1 Ϫ/Ϫ animals was significantly different (p Ͻ 0.0005) from that for the wild-type animals. B, female mice. pool in the male mouse (9,16), meaning that the level of cholesterol 7␣-hydroxylase activity should be increased by a corresponding amount in Cyp7b1 Ϫ/Ϫ animals. In agreement with this general hypothesis, the bile acid pool size in knockout mice is the same as that in wild-type controls (Table I), and male and female Cyp7b1 Ϫ/Ϫ mice have slightly elevated levels of cholesterol 7␣-hydroxylase enzyme (Fig. 3B). This elevation is more evident in males than in females (Fig. 3B), suggesting that the CYP7B1 pathway contributes more to the bile acid pool in males. Furthermore, the hepatic level of CYP7B1 in wild-type mice is higher in males (Figs. 1 and 2), as is the level of sterol 27-hydroxylase (Fig. 3B). We have not been able to consistently demonstrate an increase in cholesterol 7␣-hydroxylase mRNA in knockout mice. This failure is probably due to the small size of the change and to large intra-animal variation in cholesterol 7␣-hydroxylase mRNA levels.
The accompanying paper (33) reports a second oxysterol 7␣hydroxylase whose expression is unchanged between wild-type and Cyp7b1 Ϫ/Ϫ mice. The results of these two studies thus suggest that the cholesterol 7␣-hydroxylase pathway is the major regulated pathway in mice, whereas the oxysterol 7␣hydroxylase pathways are constitutively active.