Targeted Deletion of the Ileal Bile Acid Transporter Eliminates Enterohepatic Cycling of Bile Acids in Mice*

The ileal apical sodium bile acid cotransporter participates in the enterohepatic circulation of bile acids. In patients with primary bile acid malabsorption, mutations in the ileal bile acid transporter gene (Slc10a2) lead to congenital diarrhea, steatorrhea, and reduced plasma cholesterol levels. To elucidate the quantitative role of Slc10a2 in intestinal bile acid absorption, the Slc10a2 gene was disrupted by homologous recombination in mice. Animals heterozygous (Slc10a2+/–) and homozygous (Slc10a2–/–) for this mutation were physically indistinguishable from wild type mice. In the Slc10a2–/– mice, fecal bile acid excretion was elevated 10- to 20-fold and was not further increased by feeding a bile acid binding resin. Despite increased bile acid synthesis, the bile acid pool size was decreased by 80% and selectively enriched in cholic acid in the Slc10a2–/– mice. On a low fat diet, the Slc10a2–/– mice did not have steatorrhea. Fecal neutral sterol excretion was increased only 3-fold, and intestinal cholesterol absorption was reduced only 20%, indicating that the smaller cholic acid-enriched bile acid pool was sufficient to facilitate intestinal lipid absorption. Liver cholesteryl ester content was reduced by 50% in Slc10a2–/– mice, and unexpectedly plasma high density lipoprotein cholesterol levels were slightly elevated. These data indicate that Slc10a2 is essential for efficient intestinal absorption of bile acids and that alternative absorptive mechanisms are unable to compensate for loss of Slc10a2 function.

Bile acids are synthesized from cholesterol in the liver and secreted into the small intestine where they facilitate the absorption of dietary lipids and fat-soluble vitamins. The majority of bile acids are reabsorbed from the intestine, returned to the liver via the portal venous circulation, and resecreted into bile (1). The enterohepatic circulation of bile acids is an extremely efficient process; less than 10% of intestinal bile acids escape reabsorption and are eliminated in the feces. In the intestine, bile acids are reclaimed through a combination of passive ab-sorption in the jejunum, active transport in the distal ileum, and passive absorption in the colon (2). Bile acids are actively transported in the terminal ileum by the well characterized ileal apical sodium bile acid cotransporter (ASBT 1 ; gene name Slc10a2) (3,4). This sodium-and potential-driven transporter moves bile acids from the lumen of the small intestine across the apical brush border membrane. Bile acids are then shuttled to the basolateral membrane and secreted into the portal circulation.
Several observations support the concept that the terminal ileum is the major site of bile acid reabsorption, including the finding that there is little decrease in intraluminal bile acid concentration prior to the ileum (5) and the appearance of bile acid malabsorption after ileal resection (6). More recent studies using in situ perfused intestinal segments to measure bile acid absorption (7)(8)(9) have also demonstrated that ileal active bile acid transport is a high capacity system sufficient to account for the biliary output of bile acids. Finally, patients with a rare inherited Slc10a2 defect exhibit bile acid malabsorption, refractory infantile diarrhea, steatorrhea, and failure to thrive (primary bile acid malabsorption), underscoring the importance of Slc10a2 in the human infant (10,11). The consensus from these studies was that the ileal active transport system is a major route for bile acid uptake, but passive absorption present down the length of the intestine may also be significant. Thus, the overall quantitative contribution of passive jejunal and/or colonic bile acid absorption is not known. To resolve this question and to further understand the role of Slc10a2 in bile acid and cholesterol homeostasis, we have produced a line of mice with a targeted disruption in the Slc10a2 gene.

EXPERIMENTAL PROCEDURES
Generation of Slc10a2 Ϫ/Ϫ Mice-The targeting vector was constructed as shown (see Fig. 1A). The Slc10a2 region extending from intron 2 through the splice acceptor site of exon 4 was replaced with a neomycin-containing cassette in the pPNT vector (12). The targeting vector was linearized by cleavage with NotI and introduced by electroporation into mouse embryonic stem cells as described (13). Appropriate targeting of the Slc10a2 allele was confirmed by Southern blotting (see Fig. 6A in Supplemental Material). Male chimeric progeny were crossed with female C57BL/6J (Jackson Laboratory) and 129S6/SvEv (Taconic) mice to generate two independent lines of mice carrying the disrupted Slc10a2 allele. Unless indicated, the congenic 129S6/SvEv line was used for the experiments in this study.
Animals and Diet-All animal procedures were approved by the Institutional Animal Care and Use Committee. Mice were maintained on a prepared basal diet (14) supplemented with vitamins, cholesterol, or a bile acid binding resin as required. The basal diet contains 16% fat, 20% protein, 64% carbohydrate (% of calories), and 0.017 mg/calorie of cholesterol (0.006%) (w/w). The diets for all pregnant and nursing dams were initially supplemented with vitamins (ADEKs pediatric drops; ScandiPharm, Inc., Birmingham, AL); however, this supplementation was later shown to be unnecessary for survival of the Slc10a2 null progeny. The compositions of the experimental diets used in these studies are shown in Table III of the Supplemental Material. RNA Analyses-Total RNA was extracted from frozen tissue using TRIzol reagent (Invitrogen) as suggested by the manufacturer. For Northern blot analysis, total RNA from individual animals or pools of animals of the same sex and genotype was fractionated on 1.2% (w/v) agarose gels containing 2.2 M formaldehyde, transferred to Nytran (0.45 m; Schleicher & Schuell), and hybridized with the indicated 32 Plabeled random hexamer-primed DNA probes. Expression levels were quantified using a PhosphorImager (Amersham Biosciences).
Preparation of Intestinal BBMV and Measurement of Ileal Bile Acid Transporter Protein and Activity-Small intestines from wild type and Slc10a2 null mice (mixed C57BL/6:129S6/SvEv background) were divided into five sections of equal length and used to isolate brush border membranes by the calcium precipitation method (15). Membrane vesicle transport was assayed in triplicate using freshly isolated ileal BBMV from mice of the indicated genotype (16). Immunoblotting analysis of ileal bile acid transporter protein was performed as described (17) using a rabbit anti-ileal bile acid transporter antibody. The blots were also probed using mouse anti-␤-actin antibody as a control for protein loading.
Fecal Bile Acid and Lipid Excretion-Wild type and Slc10a2 null mice in the congenic 129S6/SvEv background were individually housed in wire bottom cages, and stools were collected for 3 days. The stools were extracted as described by Turley et al. (18) and used to determine the total bile acid content by an enzymatic method (19). For a subset of animals, the fecal bile acid species were also analyzed by mass spectrometry using an Agilent Technologies HP1100 atmospheric pressure ionization-electrospray instrument (20). A second 0.1-g aliquot of dried stool was used to measure neutral sterol and fatty acid content by gas-liquid chromatography (14,21). Intestinal cholesterol absorption was measured using a modification of the fecal dual isotope ratio method (22).
Bile Acid Pool Size and Composition-Pool size was determined as the bile acid content of the small intestine, liver, and gallbladder. These tissues were removed and extracted in ethanol as described (23). The extract was filtered, and bile acid composition was determined using HPLC as described by Rossi et al. (24). Individual bile acid species were detected by measuring refractive index; the total bile acid content of the tissue extracts was measured using an enzymatic assay (19). For a subset of animals, the bile acid species were also analyzed by electrospray mass spectrometry. The lumenal contents of the proximal small intestine were also collected to measure the molar ratio of bile acids to cholesterol in the liquid phase of intestinal contents as described by Repa et al. (25).
Cholesterol 7␣-Hydroxylase-CYP7A1 protein was detected using a chicken polyclonal antibody raised against a maltose-binding protein (New England BioLabs) fusion encompassing amino acids 353 to 436 of the African green monkey CYP7A1 (26). The blots were also probed using a rabbit anti-BIP antibody (27) as a control for protein loading. CYP7A1 protein expression was quantified by scanning the x-ray films using an Alpha Innotec (San Leandro, CA) 5500 imaging system. CYP7A1 enzyme activity was measured in microsomes isolated from individual animals using a reverse phase HPLC assay (28).
Plasma and Hepatic Lipids-Plasma total cholesterol (Wako), free cholesterol (Wako), and triglyceride (Roche Applied Science) concentrations were determined by enzymatic assay (29). An aliquot of pooled plasma from five animals in each group was fractionated on a 30 ϫ 1-cm Superose 6 column, and the cholesterol content of each lipoprotein fraction was detected using an on-line enzymatic assay. Hepatic triglyceride content was determined by enzymatic assay (29). Hepatic free and total cholesterol was determined by gas-liquid chromatography (14).
Statistical Analyses-Mean values Ϯ S.E. are shown. The data were evaluated for statistically significant differences using the two-tailed Student's t test assuming equal variance. The data from the diet studies were evaluated using analysis of variance with post hoc analysis for individual group differences by the Fisher's protected least significant difference test (Statview; Mountain View, CA). Differences were considered statistically significant at p Ͻ 0.05.

Inactivation of the Slc10a2 Gene in Mice-
The region extending from intron 2 through the splice acceptor site of exon 4 was replaced with a neomycin-containing cassette by homologous recombination (Fig. 1A). Gene disruption was confirmed by Southern blotting (see Fig. 6A in Supplemental Material), Northern blotting (Fig. 1B), and immunoblotting (Fig. 1B). No ASBT mRNA or immunoreactive protein was detected in the distal small intestine of Slc10a2 null mice. ASBT is thought to be the only sodium bile acid cotransporter expressed in the ileal apical brush border membrane (30). To directly test this hypothesis, apical BBMV were isolated and used to measure sodium-dependent taurocholate uptake. In ileal BBMV from wild type mice, [ 3 H]taurocholate (25 M) uptake was increased 42-fold in the presence of an inwardly directed Na ϩ gradient compared (p Ͻ 0.0001). In contrast, there was no Na ϩ -dependent taurocholate uptake in ileal BBMV isolated from Slc10a2 null mice (Fig. 1C). The integrity of the BBMV preparations was confirmed by measuring Na ϩ -alanine cotransport. Na ϩcoupled [ 3 H]alanine uptake (125 M) was similar in BBMV isolated from both mouse genotypes (Slc10a2 ϩ/ϩ , 175 Ϯ 35 pmol/mg protein/15 s versus Slc10a2 Ϫ/Ϫ , 218 Ϯ 60 pmol/mg protein/15 s; p Ͼ 0.5).
Analysis of the Slc10a2 Ϫ/Ϫ Mice-The Slc10a2 Ϫ/Ϫ mice are viable and fertile. Crossing heterozygous mice produced a Mendelian distribution of wild type and mutant genotypes. The Slc10a2 Ϫ/Ϫ mice were indistinguishable in terms of survival, gross appearance, and behaviors from Slc10a2 ϩ/Ϫ and wild type animals. A small growth deficit was observed in male but not female Slc10a2 null mice compared with wild type littermates prior to weaning at day 21 (data not shown). However, the ϳ20% decrease in body weight of male Slc10a2 Ϫ/Ϫ pups was only transitory, and adult body weights and daily fecal excretion (dry weight) were similar between the different genotypes (see Fig. 7, A and B in Supplemental Material). This phenotype is very different from that originally reported for the Cyp7a1 null mice (31,32). The Cyp7a1 Ϫ/Ϫ mice fail to thrive and experience a very high rate of postnatal mortality unless supplemented with exogenous fat-soluble vitamins and cholic acid. In contrast, survival of Slc10a2 Ϫ/Ϫ mice was indistinguishable from wild type littermates regardless of vitamin supplementation for the pregnant and nursing dams.
Possible explanations for the normal postnatal phenotype of the Slc10a2 Ϫ/Ϫ mice include alternative intestinal uptake mechanisms that compensate for the loss of Slc10a2 function, increased bile acid synthesis that is able to compensate for the intestinal bile acid loss, or other metabolic adaptations that are able to maintain fat absorption and lipid homeostasis. To distinguish among these possibilities, bile acid and lipid metabolism was examined in the wild type and Slc10a2 Ϫ/Ϫ mice. As shown in Fig. 2A, fecal bile acid excretion was elevated 24-and 11-fold in male and female Slc10a2 Ϫ/Ϫ mice, respectively. Analysis of the fecal bile acid composition by electrospray mass spectrometry revealed that both genotypes excrete similar bile acid species; however, the total amount of bile acids excreted was greater in the Slc10a2 null mice. These results indicate that alternative intestinal uptake mechanisms are not compensating for the active ileal absorption.
Increased bile acid synthesis is a well recognized mechanism to compensate for disruption of the enterohepatic circulation of bile acids (33)(34)(35). To determine whether synthesis is able to compensate in the Slc10a2 null mice, the bile acid pool size and composition were measured. As shown in Fig. 2B, the total bile acid pool was decreased 83 and 80% in the Slc10a2 Ϫ/Ϫ males and females, respectively. A crude fractional turnover rate (FTR) can be calculated, because the daily rate of fecal bile acid excretion and bile acid pool size were measured in these same animals. The bile acid FTR (daily fecal bile acid excretion/pool size) was ϳ0.13 pools/day in wild type mice (males, 0.13 Ϯ 0.03 pools/day; females, 0.12 Ϯ 0.02 pools/day). The bile acid FTR was remarkably elevated in the Slc10a2 null mice, increasing ϳ150-fold in the males (19.22 Ϯ 2.33 pools/day) and 75-fold in the females (9.26 Ϯ 2.49 pools/day).
Composition of the bile acid pool was significantly altered in the Slc10a2 Ϫ/Ϫ mice. The primary bile acids taurocholate and tauro-␤-muricholate account for ϳ90% of the pool in the male (taurocholate 43%, tauro-␤-muricholate 49%) and female (taurocholate 53%, tauro-␤-muricholate 34%) wild type mice. However, in the Slc10a2 null mice, the fraction of the pool normally occupied by tauro-␤-muricholate is largely replaced by taurocholate, such that taurocholate became the major bile acid species accounting for 81% (males) and 70% (females) of the bile acid pool. The proportion of tauro-␤-muricholate was significantly decreased to ϳ10% of the pool in the Slc10a2 null mice. The remainder of the bile acid pool in the wild type and Slc10a2 null mice consisted primarily of taurodeoxycholate, tauro-␣-muricholate, and tauro--muricholate. No unconjugated bile acids were detected in the bile acid pool by electrospray mass spectrometry (data not shown).
Intestinal Cholesterol and Total Lipid Absorption-Variations in bile acid pool size and composition affect lipid absorption from the small intestine. In general, large decreases in the pool size and/or content of hydrophobic bile acids can decrease lipid absorption (36,37). The diminished bile acid pool in the Slc10a2 null mice raises the possibility that cholesterol and total lipid absorption are impaired. Analysis of the fecal lipid excretion revealed that male and female Slc10a2 Ϫ/Ϫ mice excrete ϳ4-fold more total fat, fatty acids, and neutral sterols than wild type mice (Fig. 2C). However, the total amount of fat excreted is still small. The stool fat content as a percentage of fecal dry weight was ϳ1.5 and ϳ4.2% in wild type and Slc10a2 null mice, respectively. Intestinal lipid absorption was estimated by measuring the daily food consumption and lipid content of the diet. The percent of intestinal lipid absorption decreased from ϳ98 to 93% in males and from ϳ98 to 95% in female Slc10a2 null mice. These studies indicate that despite an 80% decrease in the bile acid pool, total intestinal lipid absorption was decreased less than 10%. Cholesterol absorption has been reported to be more sensitive than total lipid to changes in bile acid pool size or composition (23,36). To further investigate this question in the Slc10a2 null mice, intestinal cholesterol absorption was measured using the fecal dual isotope ratio method. As shown in Fig. 2D, intestinal cholesterol absorption was decreased 26% (males) and 15% (females) in the Slc10a2 Ϫ/Ϫ mice maintained on the low cholesterol-containing basal diet. These results suggest that despite the 80% decrease in bile acid pool size, there are sufficient lumenal bile acids for micellar solubilization of cholesterol. This was confirmed by direct measure of the bile acid to cholesterol ratio in the liquid phase of the proximal small intestine lumenal contents. The molar ratio of bile acids to cholesterol was 296 Ϯ 28 (n ϭ 4) in the wild type mice versus 107 Ϯ 15 (n ϭ 5) in the Slc10a2 null mice (p ϭ 0.0004).
Cyp7a1 Expression-In the mouse, the majority of bile acids are synthesized via the CYP7A1 pathway (23). CYP7A1 activity was significantly increased in both male (2.7-fold) and female (5.2-fold) Slc10a2 Ϫ/Ϫ mice (Fig. 3A). Similar increases were seen for CYP7A1 protein and mRNA levels (Fig. 3B). In addition to the pooled samples shown in Fig. 3B, aliquots of microsomes and mRNA from individual mouse livers were analyzed for CYP7A1 expression. CYP7A1 protein was increased 4.1-fold in male (p ϭ 0.0013) and 6.9-fold in female (p Ͻ 0.0001) Slc10a2 null mice. Similarly, CYP7A1 mRNA was increased 3-fold (p ϭ 0.0013) and 2.3-fold (p ϭ 0.02) in the male and female Slc10a2 null mice, respectively.
Intestinal Bile Acid Absorption-In addition to Slc10a2-mediated transport in the terminal ileum, bile acids may be absorbed by passive non-ionic diffusion in the small intestine, by facilitative transport in the small intestine, or by passive diffusion in the colon following bacterial metabolism. The first mechanism is unlikely to be an important contributor, because mouse bile acids are conjugated to taurine and would be ionized at physiological pH of the small intestine. The quantitative significance of the second mechanism, facilitative transport of conjugated bile acids, is unknown (38 -41). A candidate transporter for this activity, Oatp3, has been identified in proximal small intestine (42). However, Oatp3 is expressed at negligible levels in rodent small intestine (42), and its mRNA expression is not induced in the Slc10a2 null mice (data not shown). It is therefore unlikely that Oatp3 accounts for appreciable intestinal bile acid absorption. A third mechanism involves conversion of taurocholate to deoxycholate following deconjugation and 7-dehydroxylation by the endogenous bacterial flora (43). Deoxycholate produced in the colon undergoes passive nonionic diffusion, and it has been proposed that colonic absorption becomes the major route of intestinal absorption when ileal function is impaired (44). To determine the contribution of these alternative mechanisms for intestinal bile acid absorption, Slc10a2 null mice were fed a semipurified diet containing 2% colestipol, a bile acid binding resin. Bile acid sequestrants should have little effect on fecal bile acid excretion if Slc10a2 is the single major mechanism for intestinal bile acid transport. Conversely if alternative mechanisms such as colonic absorption are important in the mouse, colestipol should increase fecal bile acid excretion. As a control, a separate group of mice were fed a semipurified diet containing 1% cholesterol to determine whether substrate has become rate-limiting for bile acid synthesis in the Slc10a2 null mice. As shown in Fig. 4A, fecal bile acid excretion increased significantly in the wild type mice following cholesterol feeding (3-fold versus basal diet; p ϭ 0.0152) or colestid feeding (9-fold versus basal diet; p Ͻ 0.0001). In contrast, cholesterol feeding (p ϭ 0.2268) and colestipol feeding (p ϭ 0.4860) had little effect on the already elevated fecal bile acid excretion rate in the Slc10a2 null mice. CYP7A1 mRNA expression (Fig. 4B) followed a similar pattern. These data indicate that the non-Slc10a2 mechanisms contribute little to intestinal bile acid absorption in the mouse.
Hepatic and Plasma Lipids-The changes in hepatic lipid and lipoprotein concentrations are described in Table I. Hepatic and plasma triglyceride levels were similar in wild type and Slc10a2 null mice. In contrast, hepatic cholesteryl ester levels were significantly lower in the basal diet-fed Slc10a2 Ϫ/Ϫ mice reflecting the increased bile acid synthesis in these animals (Table I). Surprisingly, the male and female Slc10a2 Ϫ/Ϫ mice had a small but significant (p Ͻ 0.001) increase in their total plasma cholesterol levels. Analysis of the plasma by fast A, fecal bile acid excretion was measured as described under "Experimental Procedures" (n ϭ 10). Bile acid excretion in Slc10a2 Ϫ/Ϫ mice was significantly greater (*, p Ͻ 0.0001 for males and p ϭ 0.0002 for females) than for the wild type mice. B, the mass of bile acid in the enterohepatic circulation was determined by extraction, purification, and quantitation using HPLC (n ϭ 6 -10). The Slc10a2 Ϫ/Ϫ mice had a significantly decreased total bile acid pool size (*p Ͻ 0.0001), as well as decreased amounts of taurocholate (p Ͻ 0.0001) and tauro-␤-muricholate (p Ͻ 0.001 for males and p ϭ 0.0002 for females). The small amount of other bile acids in the pool were not significantly different between the two genotypes (p Ͼ 0.05). C, fecal neutral sterols and fatty acids were quantitated by gas chromatography (n ϭ 10). The Slc10a2 Ϫ/Ϫ mice had significantly increased total fat excretion (*, p Ͻ 0.0001 for males and p ϭ 0.0001 for females), fatty acid excretion (p Ͻ 0.0001 for males and p ϭ 0.0001 for females), and neutral sterol excretion (p Ͻ 0.0001). D, intestinal cholesterol absorption was measured by a dual isotope fecal assay (n ϭ 10). The Slc10a2 Ϫ/Ϫ mice had significantly decreased intestinal cholesterol absorption (*, p ϭ 0.026 for males and p ϭ 0.0009 for females).
protein liquid chromatography revealed no lipoprotein particle size differences between the null and wild type mice (data not shown), and almost all the cholesterol was found in the high density lipoprotein (HDL) fraction (Table I).
Alterations in Gene Expression-To identify possible compensatory mechanisms engaged in response to bile acid malabsorption, we quantified mRNA levels for candidate genes involved in cholesterol and bile acid metabolism (Fig. 5). An increased level of sterol 12␣-hydroxylase (Cyp8b1) was detected in the Slc10a2 Ϫ/Ϫ mice, consistent with the increased proportion of cholate in the bile acid pool (Fig. 2B). An increased level of sterol 27-hydroxylase mRNA was also detected. Expression of the hepatic bile salt export pump (Abcb11) was reduced in the Slc10a2 Ϫ/Ϫ mice, presumably in response to the decreased hepatic flux of bile acids in the enterohepatic circulation. The greatest mRNA increases were observed for enzymes in the cholesterol biosynthetic pathway. The mRNA for HMG CoA synthase was elevated 5-to 7-fold, where HMG CoA reductase was increased 3-to 6-fold. Relatively small changes were observed for several of the nuclear receptors important for cholesterol and bile acid homeostasis, including LXR␣, FXR, and LRH. In contrast, SHP expression was unchanged in the males and dramatically reduced in the female Slc10a2 null mice.

DISCUSSION
The major finding of this study is that disruption of Slc10a2 resulted in profound bile acid malabsorption, and jejunal/colonic absorption contributed little to intestinal conservation of bile acids in the mouse. Intestinal uptake of bile acids was thought to be the result of a combination of passive absorption in the jejunum, active transport in the distal ileum, and passive absorption in the colon (2). Uptake is transcellular, and there is little evidence of paracellular absorption of bile acids in any of these compartments (45,46). Slc10a2 is responsible for active ileal bile acid uptake (30), and most studies support the concept that the terminal ileum is an important site of bile acid absorption, including the finding that inherited mutations in human Slc10a2 causes bile acid malabsorption (10). However, the non-Slc10a2 contribution to intestinal bile acid absorption was not quantified in primary bile acid malabsorption subjects, and there is a persistent body of evidence for significant jejunal absorption (38 -41). In addition, the endogenous bacterial flora  4 -5). A, CYP7A1 enzyme activity was measured using an HPLC-based assay and hepatic microsomes isolated from mice of the indicated sex and Slc10a2 genotype. CYP7A1 activity was significantly greater in the Slc10a2 Ϫ/Ϫ mice (*, p ϭ 0.0136 for males and p Ͻ 0.0001 for females) than for the wild type mice. B, CYP7A1 protein and mRNA levels in liver of wild type and Slc10a2 Ϫ/Ϫ mice. Microsomes were pooled from four to five animals, and ϳ100 g of microsomal protein was subjected to immunoblotting as described under "Experimental Procedures." The blots were probed using anti-BIP antibody as a control. CYP7A1 protein was elevated ϳ5-fold in the Slc10a2 Ϫ/Ϫ mice. Pooled aliquots of total RNA were subjected to Northern blot hybridization using radiolabeled probes for CYP7A1 and cyclophilin. CYP7A1 mRNA was elevated ϳ4-fold in the Slc10a2 null mice. Fecal bile acid excretion was measured as described under "Experimental Procedures." Fecal bile acid excretion was significantly increased in the wild type mice by cholesterol (*, p ϭ 0.0152) and cholestyramine (*, p Ͻ 0.0001) feeding. In contrast, cholesterol or colestid feeding did not significantly change bile acid excretion in the Slc10a2 Ϫ/Ϫ mice (p Ͼ 0.2). B, CYP7A1 mRNA levels in liver of wild type and Slc10a2 Ϫ/Ϫ mice fed the indicated diets. CYP7A1 and cyclophilin mRNA levels were measured in pooled aliquots (n ϭ 4 -5) of liver total RNA by Northern blot hybridization. The level of CYP7A1 mRNA expression relative to the wild type male mice is indicated.
in the distal small intestine, cecum, and colon converts taurocholate to deoxycholate, a more hydrophobic dihydroxy bile acid that can undergo significant passive absorption (43). Although only a minor contributor under normal physiologic conditions, it had been proposed that colonic absorption of unconjugated bile acids becomes a major route of intestinal absorption when ileal function is impaired and bile acid flux into the colon is high (44,46). Generation of the Slc10a2 null mouse made it possible to directly quantify the in vivo contribution of the jejunal and colonic absorptive pathways. Surprisingly, physically sequestering bile acids using binding resins, a treatment that interferes with intestinal bile acid absorption, had no effect on fecal bile acid excretion in the Slc10a2 null mice.
There are several issues that must be considered before extrapolating these findings to wild type mice or to humans. First, mouse bile acids are conjugated with taurine and do not undergo appreciable non-ionic passive diffusion across the small intestinal epithelia. The bile acid glycine conjugates present in man and many other species may undergo more passive diffusion. Second, jejunal bile acid absorption in vivo may be a low affinity process and could have been masked by the low intralumenal bile acid concentrations in the Slc10a2 null mouse. Third, colonic absorption of bile acids is dependent on composition of the intestinal flora, pH of the lumenal contents, and the intestinal transit time (47). Because only a very small fraction of the intestinal flora harbors the bacterial genes nec-essary to 7␣-dehydroxylate bile acids (43), differences in the content of those specific anaerobic bacteria between species or even between different mouse strains could alter deoxycholate production. Finally, bile acids are bacteriostatic (48,49), and their increased flux into the cecum and colon in the Slc10a2 null mice may have reduced the bacterial content and deoxycholate production in the distal small intestine and colon.
Variations in bile acid pool size and composition are known to affect the amount of cholesterol absorbed from the small intestine. A comparison of intestinal cholesterol absorption in the Cyp7a1 Ϫ/Ϫ , Cyp8b1 Ϫ/Ϫ , and Slc10a2 Ϫ/Ϫ mice nicely illustrates these relationships. Slc10a2 Ϫ/Ϫ and Cyp7a1 Ϫ/Ϫ mice have a similar ϳ80% decrease in their bile acid pool (22,35). However, intestinal cholesterol absorption is dramatically reduced in the male Cyp7a1 Ϫ/Ϫ mice (decreased greater than 95%) but only mildly affected in the male Slc10a2 Ϫ/Ϫ mice (decreased ϳ26%). Conversely, Cyp8b1 Ϫ/Ϫ and Slc10a2 Ϫ/Ϫ mice show similar decreases in intestinal cholesterol absorption (ϳ40% in Cyp8b1 Ϫ/Ϫ and ϳ25% in Slc10a2 Ϫ/Ϫ ) (36), but the bile acid pool is increased in Cyp8b1 Ϫ/Ϫ mice (ϳ37%) and dramatically reduced in the Slc10a2 Ϫ/Ϫ mice (ϳ80%). The differences between these genotypes can be attributed to the increased relative content of cholate in the Slc10a2 Ϫ/Ϫ mice. Expression of the hepatic enzyme responsible for cholate synthesis, Cyp8b1, is induced in response to decreased return of bile acids in the enterohepatic circulation. This important ad- FIG. 5. Gene expression in wild type and Slc10a2 null mice. Pooled aliquots of total RNA (5 g or 25 g) isolated from the liver (n ϭ 4 -5) were analyzed by Northern blot hybridization using cDNA probes for the indicated genes. The asterisk (*) indicates that 25 g of total RNA was analyzed. Each blot was also hybridized with a cyclophilin cDNA probe to normalize relative expression levels. For each gene, the mRNA expression level is indicated relative to the level found in wild type male mice. aptation enriches the smaller pool in Slc10a2 null mice with a more hydrophobic bile acid, thereby supporting near normal levels of intestinal cholesterol absorption. In contrast the bile acid pool in the Cyp7a1 Ϫ/Ϫ and Cyp8b1 Ϫ/Ϫ mice are enriched in more hydrophilic bile acids such as ␤-muricholate that may antagonize the ability of hydrophobic bile acids to solubilize and deliver cholesterol to the intestinal epithelial cell (37). Taken together, these data support the concept that bile acid composition is more important than pool size for affecting solubilization and absorption of cholesterol from the intestinal lumen. It is important to note that the mice in this study are maintained on a diet low in fat (16% of calories) and cholesterol (0.006%). The intestinal lipid absorption-sparing effect of cholate enrichment in the Slc10a2 null mice may be lost when challenged with diets containing more fat, such as a typical Western diet.
In the Slc10a2 null mice, CYP7A1 expression and hepatic bile acid synthesis was increased in response to the interruption of the enterohepatic circulation of bile acids. This increased demand on hepatic cholesterol stores was only partially offset by increased hepatic sterol synthesis. Hepatic total cholesterol and cholesteryl ester levels were significantly reduced in the male (40% decrease in cholesteryl ester) and female (75% decrease in cholesteryl ester) Slc10a2 null mice. Despite this decrease in hepatic total cholesterol, the Slc10a2 null mice had a small but significant increase in total plasma cholesterol levels. This increase was selective for the HDL fraction, and no change was observed for plasma VLDL or LDL concentrations. An increase in HDL cholesterol has been reported in the Fxr Ϫ/Ϫ mice and was attributed to impaired hepatic selective removal of HDL cholesterol ester by SRBI (50). It is not clear if a similar mechanism is responsible in the Slc10a2 null mice. In contrast to the FXR null mice, SRBI mRNA expression was unchanged in the Slc10a2 Ϫ/Ϫ mice. Another possible explanation for the increase in HDL is that apoA-I expression is derepressed in the Slc10a2 Ϫ/Ϫ mice. Several studies have found that bile acids decrease apoA-I expression (51,52), though this bile acid repression has not been a universal finding (50).
Several lines of evidence indicate that changes in bile acid metabolism affect plasma triglyceride levels. In humans, increased plasma triglyceride levels are associated with interruption of the enterohepatic circulation using bile acid binding resins or after ileal exclusion (53)(54)(55). Conversely, patients administered chenodeoxycholic acid exhibited decreased plasma VLDL triglyceride levels (56,57). Familial hypertriglyceridemia, a disease characterized by elevated plasma VLDL triglyceride levels, is associated with hepatic bile acid overproduction, impaired absorption of bile acids (58,59), and decreased ASBT expression (60). Finally, hypertriglyceridemia is observed in subjects with rare inherited defects in bile acid synthesis (61,62), suggesting that a decreased hepatic concentration or flux of bile acids is the critical factor rather than increased bile acid synthesis rates. In contrast to these human studies, hepatic and plasma triglyceride levels were unchanged in the Slc10a2 null mice. This finding is in agreement with previous mouse studies that reported cholestyramine feeding had no effect (63) or only weakly raised plasma triglyceride levels in wild type mice (64). Although the mechanism responsible for this species difference is not clear, hypertriglyceridemia is induced in gene-targeted mice with a defect in bile acid biosynthesis (Cyp27 Ϫ/Ϫ mice, 63) or in the ability of the hepatocyte to respond to bile acids (Fxr Ϫ/Ϫ mice, 50). Taken together, these results indicate that simply disrupting the return of bile acids to the liver in the enterohepatic circulation (Slc10a2 Ϫ/Ϫ mice) is not sufficient to induce hypertriglyceridemia, and a second metabolic perturbation is required.
In conclusion, the results of these studies indicate that Slc10a2 is the major mechanism for intestinal reclamation of bile acids and crucial for maintenance of the bile acid pool. The Slc10a2 null mice generated here provide a useful animal model to identify the physiological mechanisms engaged to maintain lipid homeostasis under conditions of chronic bile acid malabsorption.