Molecular Regulation of Bile Acid Homeostasis

Bile Acids and Fat Breakdown

The long-known physiologic role of bile acids (as bile salts) is to help in the digestion and absorption of fats and fat-soluble vitamins in the small intestine. In healthy subjects, depending on age, postprandial serum bile acid concentrations increase 2- to 5-fold compared with fasting levels (Linnet, 1983; Salemans et al., 1993). Bile acids act as detergents, emulsifying fats to form micelles facilitating the digestion of fats by pancreatic lipase and subsequent absorption. Lipid digestion is an interfacial process in which the surface area of lipid droplets influences the binding of lipase and thus digestion; smaller droplets provide larger surface area and more efficient digestion by lipase (Armand et al., 1999). In large fat globules, hydrophilic pancreatic lipase cannot reach the core of the globules and can only cleave ester bonds of the triglycerides on the surface of the globules. Therefore, large fat globules need to be broken down to a smaller size to facilitate lipid digestion. When stomach chyme is pushed through the pyloric canal into the duodenum, the emulsion particles are subjected to strong shearing forces that tear the liquid interfaces apart. The emulsion particles that are mixed with bile in the upper small intestine are generally less than 0.5 μm in diameter (Carey et al., 1983). Bile acids are amphipathic and have high surface activity; hence they can mix with larger fat globules, coat fat droplets, and increase the ionization of the interface that facilitates emulsification (Sarkar et al., 2016). Emulsification reduces the particle size and increases the interfacial area of fat droplets. The process eventually leads to a breakup of the bulk oil phase in the presence of low interfacial tensions. Because of the increase in the surface area of the emulsified fat, the lipase digests the ester bonds efficiently. In humans, the total bile acid concentration during lipid digestion remains constant at 2.5–10 mM despite multiple enterohepatic cycles of the bile acid pool (Carey et al., 1983).

Various Bile Acids Have Different Abilities to Enhance Lipid Absorption

Hydrophobic (less hydrophilic) bile acids are more effective in increasing cholesterol absorption compared with more hydrophilic bile acids. Wang et al. (2003) fed mice with chow or chow supplemented with 0.5% (by weight) each of CA, CDCA, deoxycholic acid (DCA), dehydrocholic acid, hyodeoxycholic acid (HDCA), UDCA, ursocholic acid (UCA), ω-muricholic acid (ω-MCA), hyocholic acid (HCA), α-MCA, and β-MCA, and determined cholesterol absorption using a dual-isotope ratio method (Fig. 3). The authors found that compared with chow (mean cholesterol absorption 37%), feeding hydrophilic bile acids, such as β-MCA, resulted in low cholesterol absorption (mean: 11%), whereas feeding a “hydrophobic” (less hydrophilic) bile acid, such as CA, resulted in high cholesterol absorption (mean: 63%). The authors concluded that the balance of hydrophobic and hydrophilic bile salts in bile plays an important role in regulating intestinal cholesterol absorption. The authors hypothesized that bile acids that are more hydrophilic might reduce the solubility of cholesterol by inducing phase separation of the sterol from mixed micelles to a coexisting liquid crystalline vesicle phase, resulting in the reduction of intestinal cholesterol absorption (Fig. 3).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

The effect of hydrophobicity of bile acids on the percentage of cholesterol absorption when fed to mice. The figure is based on estimated mean values from Wang et al. (2003). Moving counterclockwise from β-MCA, the bile acids have increasing hydrophobicity, except for CDCA, which in mice is readily converted to hydrophilic α-MCA and β-MCA. Compared with chow (mean: 37%), the cholesterol absorption was significantly lower by hydrophilic bile acids, such as β-MCA and α-MCA (mean: about 11–12%). Cholesterol absorption increased significantly by hydrophobic (less hydrophilic) bile acids, being highest by CA (mean: 63%). DHCA, a hydrophilic bile acid, increased cholesterol absorption to 45% ± 6% because it was biotransformed mainly to the more hydrophobic taurocholate and taurodeoxycholate.

Classic (Neutral) Pathway

The classic pathway is referred to as the “neutral” pathway because the steroid nucleus modifications occur before the side‐chain oxidation; therefore, most of the intermediates in this pathway do not have a carboxylic acid group until later in the pathway when cytochrome P450 (CYP) 27A1 catalyzes oxidation of the side chain (Fig. 7).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

Bile acid biosynthesis pathways. Both the classic and the alternative pathways are shown.

The classic pathway operates in the liver (Agellon, 2008). In both humans and mice, the classic pathway begins with the hydroxylation of cholesterol at the C7 position by microsomal cholesterol 7α-hydroxylase (CYP7A1), which is the rate-limiting enzyme of the pathway. CYP7A1 was first cloned and characterized from rat liver by Noshiro et al. (1989). The resulting 7α-hydroxycholesterol is converted to 7α-hydroxy-4-cholesten-3-one (named C4) by the microsomal hydroxysteroid dehydrogenase named HSD3B7 (3β-hydroxy-Δ5-C27-steroid dehydrogenase). HSD3B7 then catalyzes the epimerization of the 3β-hydroxyl group of cholesterol to the 3α-hydroxyl of bile acids (Shea et al., 2007). C4 is the common precursor for synthesis of CA and CDCA. The serum C4 concentration is now used as a biomarker for the rate of bile acid synthesis. C4 is metabolized by microsomal sterol 12α-hydroxylase (CYP8B1) to become CA, and without the action of CYP8B1 it becomes CDCA. Therefore, the activity of the CYP8B1 determines the ratio of CA to CDCA. Hydroxylation of C4 at the C12 position by CYP8B1 produces 7α,12α-dihydroxy-4-cholesten-3-one, which goes through NADPH-dependent reductions at the C5 and C3 positions by aldo-keto reductase 1D1 (AKR1D1) and 1C4 (AKR1C4), respectively. Reduction of the C3 double bond by AKR1C4 produces 3α,7α,12α-cholesten-5β-triol. This is followed by CYP27A1-catalyzed side chain oxidation (from ‒OH to ‒COOH) in mitochondria to form 3α,7α,12α-trihydroxycholestanoic acid; and 3α,7α-dihydroxycholestanoic acid without CYP8B1. These bile acid intermediates are activated to form bile acid-CoA thioesters by peroxisomal long chain acyl-CoA synthase (or bile acid-CoA synthase, BACS), which enter peroxisomes via a peroxisomal transporter, an ATP-binding cassette (ABC) D3 for β-oxidation reactions to cleave a propionyl-CoA to form cholyl-CoA and chenodeoxycholyl-CoA, respectively. These bile acid-acyl-CoAs are conjugated to an amino acid taurine or glycine to form conjugated bile acids by bile acid-CoA: amino acid N-acyltransferase (BAAT) (Fig. 7). Conjugated bile acids are secreted into bile.

Human and mouse CYP8B1 were cloned and characterized by Gåfvels et al. (1999). Recently, using a fission yeast-based expression system, Fan et al. (2019) have shown that human CYP8B1 can directly convert CDCA to CA by catalyzing 12α-hydroxylation. However, the relevance of this observation in vivo needs to be established (Fig. 7).

Alternative (Acidic) Pathway

The alternative pathway of bile acid synthesis is also referred to as the “acidic” pathway because the cholesterol side‐chain oxidation occurs early in the pathway and before the modifications to the steroid nucleus. C27 bile acids and oxysterols formed in various cells in the body are transported to the liver to ultimately produce C24 bile acids in hepatocytes. For example, almost all the 24-hydroxycholesterol in the liver originates from the brain (Agellon, 2008). Likewise, 27-hydroxycholesterol is the most abundant oxysterol in the plasma of mice (Li-Hawkins et al., 2000) and humans (Dzeletovic et al., 1995). It is synthesized from cholesterol by sterol 27-hydroxylase in multiple tissues including the liver (Fig. 7).

The acidic pathway is initiated by mitochondrial/microsomal C-24, C-25, or C-27 sterol hydroxylases that hydroxylate cholesterol to form 24-, 25-, and 27-hydroxycholesterol, respectively, which are then rapidly 7α-hydroxylated by microsomal oxysterol 7 -hydroxylase (CYP7B1) (Pandak and Kakiyama, 2019). In the mammalian central nervous system, sterol 24-hydroxylase (CYP46A1) hydroxylates cholesterol to form 24-hydroxycholesterol in the membranes of the smooth endoplasmic reticulum of neurons (Russell et al., 2009) (Fig. 7). Based on the observation that CYP46A1 and CYP27A1 knockout mice have no apparent alterations in bile acid synthesis, along with the finding that cholesterol 7α-hydroxylase deficient mice have a bile acid content that is 50% of normal because these mice cannot convert cholesterol directly into bile acids, it was concluded that the alternative pathway contributes to about 50% of the bile acids (BA) pool in mice (Schwarz et al., 1998; Schwarz et al., 2001). It is interesting to note that female mice have little or no CYP7B1 (Fu et al., 2012) but do not have decreased BA levels. Analysis of the bile acids in a human subject with cholesterol 7α-hydroxylase deficiency suggests that the alternative pathway contributes to about 5%–10% of the bile acid pool in humans (Pullinger et al., 2002).

Figure 7 shows that the products of the upstream reactions that are part of the alternative pathway feed into the downstream reactions shared by the classic pathway. The alternative pathway for the synthesis of oxysterols exists in numerous tissues. Side-chain oxidation is followed by 7α-hydroxylation of the sterol nucleus by microsomal oxysterol 7α-hydroxylase (CYP7B1) in most tissues. Table 3 compares the classic and the alternative pathway of bile acid synthesis. CYP27A1 was first cloned and characterized from rabbit (Andersson et al., 1989), and CYP7B1 was first cloned and characterized from rat brain (Stapleton et al., 1995). It is important to note that both CYP7A1 and CYP7B1 add a 7-hydroxyl group to their respective substrates; however, CYP7A1 in the classic pathway is a highly specific cholesterol 7α-hydroxylase, whereas CYP7B1 in the alternative pathway is an oxysterol 7α-hydroxylase. CYP7B1 prefers 25-hydroxycholesterol and 27-hydroxycholesterol as substrates, whereas CYP39A1 sterol 7α-hydroxylase is selective for 24-hydroxycholesterol (Agellon, 2008). Table 3 shows the major differences between the classic and the alternative pathway, and Table 4 shows the main differences between human and mouse bile acids.

Setchell et al. (1998) described the clinical presentation of a 10-week-old boy with severe cholestasis, cirrhosis, and liver failure, all due to defects in bile acid synthesis. There was a deficiency in microsomal CYP7B1 in the alternative pathway. Gene sequence analysis revealed a C→ T transition mutation in exon 5 affecting codon 388, thereby creating a premature stop codon and a nonfunctional enzyme. Setchell et al. (1988) had previously reported that early human fetal (16–19 weeks) gallbladder bile is enriched in CDCA and metabolites not found in adults. Based on the predominance of CDCA in the bile, the authors proposed a possible dominance of the alternative pathway in early development. In contrast to humans, CYP7B1 in mice develops slowly during development. Cui et al. (2012) showed that Cyp7b1 mRNA was low in mice before 20 days of age and then increased markedly to adult levels.

Observations from Knockout (Null) Mouse Models in Which Different Genes of the Classic and Alternative Bile Acid Synthesis Pathways Are Inactivated

Five null mouse models are discussed here, CYP7A1-null, CYP7B1-null, CYP8B1-null, CYP27A1-null, and HSD3B7-null.

CYP7A1-Null

CYP7A1-null (Cyp7a1^–/–) mice were genetically engineered by removing exons 3–5 of the gene resulting in a total absence of a functional CYP7A1 enzyme (Ishibashi et al., 1996). Most of the pups arising from crosses of the CYP7A1-null mice died before weaning (3–4 weeks after birth). CYP7A1-null pups defecated fatty stools indicating fat malabsorption (steatorrhea). However, if the pregnant mother’s chow and water were supplemented with both CA and vitamins beginning from gestation day 12 and continued through weaning, CYP7A1-null pups survived (Ishibashi et al., 1996). Once past the weaning stage (3–4 weeks after birth), CYP7B1 expression begins and the alternative pathway of bile acid synthesis is activated, compensating for the absence of CYP7A1 activity (Schwarz et al., 1996; Fu and Klaassen, 2013). Interestingly, Erickson et al. (2003) established their own colony of CYP7A1-null mice. By breeding vendor-supplied CYP7A1-null mice and selecting for pups that did not need vitamin and cholate supplementation, the authors were able to increase the survival to >65% in their CYP7A1-null mice colony. Erickson et al. (2003) speculated that the differential survival of the two groups of CYP7A1-null mice, one needing vitamin and cholate supplementation to survive as pups whereas the others did not, suggested a “leaky” phenotype, that is, the presence of a compensatory mechanism in those pups that survived without vitamin and cholate supplementation. Nevertheless, the complete absence of CYP7A1 activity in CYP7A1-null mice, particularly early in life, causes several phenotypic abnormalities compared with the wild type due to altered lipid homeostasis. The role of CYP7A1 in human bile acid homeostasis has been demonstrated in population studies where an association between high plasma cholesterol concentrations and polymorphisms at the CYP7A1 locus, such as the A→C transversion polymorphism at position –204 of the promoter (–204A>C) was observed (Couture et al., 1999). Ferrell et al. (2016) showed that CYP7A1-null mice in C57BL/6J genetic background survive well and are protected from high-fat-diet induced metabolic disorder.

CYP7B1-Null

CYP7B1-null (Cyp7b1^–/–) mice have no adverse phenotype and are indistinguishable from wild-type mice (Li-Hawkins et al., 2000). Bile acid content and composition, plasma triglycerides and cholesterol concentrations are all normal in male and female CYP7B1-null mice. The cholesterol content of major tissues was not altered. However, plasma and tissue levels of 25- and 27-hydroxycholesterol, two oxysterol substrates of CYP7B1, were markedly elevated in CYP7B1-null mice. In contrast to mice, defects in the CYP7B1 gene in humans give rise to two different disease phenotypes: liver failure in children due to defects in bile acid synthesis and progressive neuropathy (spastic paraplegias) in adults. Lack of CYP7B1 activity results in an elevation of serum oxysterols and other bile acid intermediates, and a corresponding absence of mature bile acids. The spastic paraplegias are associated with lower limb spasticity, and weakness and a corresponding axonal degeneration of motor neurons in the spinal cord (Stiles et al., 2009). It is interesting to note that female mice have very low constitutive expression of CYP7B1 enzyme compared with males and yet they have normal cholesterol and bile acid levels (Li-Hawkins et al., 2000).

Using control (B6/129 genetic background), CYP7B1-null and CYP27A1-null liver cells, Kakiyama et al. (2019) reported a new pathway of bile acid synthesis initiated in liver mitochondria; the pathway begins with the formation of 24-hydroxycholesterol (24HC). This was shown by an increased expression of the steroidogenic acute transport protein (StarD1) that facilitated an increased transport of cholesterol into mitochondrial CYP27A1. In control mice livers, there was a marked downregulation of oxysterol 7α-hydroxylase (CYP7B1), a marked increase in 26-hydroxycholesterol (26HC), and the formation of a vital regulatory oxysterol, 24-hydroxycholesterol (24HC). In CYP7B1-null mice liver, however, a marked increase (13-fold versus control) of 24HC was observed in the StarD1 over expressed mice. No detectable levels of 24HC and 26HC were present in the CYP27A1-null liver.

CYP8B1-Null

In both humans and mice, the CYP8B1 gene is devoid of introns (Gåfvels et al., 1999). Disruption of the Cyp8b1 gene to create CYP8B1-null (Cyp8b1^–/–) mice resulted in a complete absence of CA (Li-Hawkins et al., 2002). Because CDCA in mice is converted to MCA, the bile acid pool in these mice is dominated by MCA. In CYP8B1-null mice, the feedback downregulation of CYP7A1 activity is lost causing expansion of the bile acid content and alterations in cholesterol metabolism. Wild-type female mice typically have a 35%–40% larger bile acid pool compared with males due to the increased production from the CYP7A1 pathway. This sexually dimorphic trait was maintained in the CYP8B1-null mice. Despite a larger bile acid content in female mice, the absorption of dietary cholesterol and lipids is reduced (steatorrhea), suggesting that the composition of the bile acid pool is more important than its size (Li-Hawkins et al., 2002). Subsequent studies further confirmed better lipid profile, glucose homeostasis, and improved liver health in CYP8B1-null mice. CYP8B1-null mice are protected from the following conditions: (1) metabolic impairments, hepatic inflammation, and fibrosis (Patankar et al., 2018); (2) western diet-induced weight gain and hepatic steatosis due to impaired fat absorption (Bertaggia et al., 2017); and (3) weight gain on a high fat diet and liver enlargement/steatosis/serum LDL-cholesterol (Bonde et al., 2016). CYP8B1-null mice also have increased fecal excretion of cholesterol (Bonde et al., 2016) and improved glucose homeostasis because of increased glucagon-like peptide-1 (GLP-1) (Kaur et al., 2015). In mice fed a high cholesterol diet, siRNA-mediated knockdown of CYP8B1 expression was found to significantly decrease steatosis, hepatic lipid content, and hepatic lipid accumulation. Increased clearance of previously accumulated hepatic cholesterol was also observed. The authors concluded that the data demonstrate that inhibition of CYP8B1 could be a viable therapeutic target (Chevre et al., 2018). In human bile, CA and CDCA occur in a molar ratio of approximately 2:1 (Wang et al., 2005).

An absence of functional CYP8B1 results in the absence of CA. For example, in naked mole rats, the CYP8B1 gene has been completely removed by a genomic deletion resulting in a total absence of CA. In both African and Asian elephants, the CYP8B1 gene has accumulated multiple inactivating mutations resulting in a nonfunctional CYP8B1 gene and a total absence of CA (Sharma and Hiller, 2018). Shinde et al. (2019) stated that the loss of CYP8B1 gene in certain species indicates recurrent changes in the selection landscape that is associated with changes in dietary lipid content. However, there are many exceptions, such as phylogenetically closer herbivorous species that show divergence in bile acid production, such as horse and white rhinoceros (both in order Perissodactyla); horse produces both CA and CDCA, but white rhinoceros does not produce either. In contrast, there are phylogenetically more distant herbivorous species that show similarity in bile acid production, such as elephant (order Proboscidea) and manatee (order Sirenia); both do not produce either CA or CDCA. It appears that the loss of CYP8B1 occurred independently in many lineages during mammalian evolution.

CYP27A1-Null

CYP27A1-null (Cyp27a1^–/–) mice have normal plasma levels of cholesterol and normal appearance (Rosen et al., 1998), but the bile acid content is reduced. In CYP27A1-null mice, the bile acid concentration in bile was 0.9 mg/ml compared with 8 mg/ml in controls, with CA being the predominant bile acid. Additionally, the excretion of fecal bile acids was reduced by 80% (78 μg/g in CYP27A1-null mice compared with 430 μg/g in wild type), as determined by the analysis of an equal amount of daily fecal excretion. The reduced synthesis of bile acids in CYP27A1-null mice leads to reduced absorption of cholesterol, which may lead to a compensatory increased synthesis. CYP27A1-null mice have a compensatory upregulation of CYP7A1 expression due to the decrease in negative feedback suppression, resulting in a 4- to 10-fold increase in the concentration of circulating 7α-hydroxycholesterol compared with the wild-type mice (Rosen et al., 1998). CYP27A1-null mice also have hypertriglyceridemia, hepatomegaly, adrenomegaly, and increased hepatic fatty acid synthesis, indicating a more global role of CYP27A1 in lipid homeostasis than previously known (Repa et al., 2000). In humans, mutations in the CYP27A1 gene cause cerebrotendinous xanthomatosis (CTX), which is a rare genetic metabolic disorder of cholesterol and bile acid metabolism causing systemic and neurologic abnormalities (https://www.omim.org/entry/606530; last accessed: June 9, 2021). Patients with CTX have accumulated cholesterol and a bile alcohol, cholestanol. However, CYP27A1-null mice do not have CTX phenotypes.

Rizzolo et al. (2019) produced CYP7A1-CYP27A1 double-null mice models by crossing homozygous CYP7A1-null and homozygous CYP27A1-null mice. The double-null mice, despite lacking two crucial enzymes in bile acid synthesis pathway, still produced bile acids in the liver. Although there was a 40% reduction in bile acid production in the liver, the bile acid pool composition was similar compared with the wild-type mice.

HSD3B7-Null

The development of hydroxysteroid dehydrogenase HSD3B7-null mice (Hsd3b7 ^–/–) (Shea et al., 2007) demonstrated that the stereochemistry of the C-3 OH group of bile acids is important in maintaining the functional and regulatory properties of bile acids in mice and presumably in other species. The HSD3B7 enzyme catalyzes two reactions required for the epimerization of the 3β-hydroxyl group of cholesterol to the 3α-hydroxyl of bile acids. Predictably, the HSD3B7-null mice cannot epimerize the 3β-OH to 3α-OH. In these mice, the bile acid pool is dominated by 3β-OH bile acids, and 90% of these mice die within 3 weeks of birth when maintained on standard feed and water. These mice excrete 2–3 times more bile acids and sterols in the feces compared with wild-type littermates. Consequently, intestinal cholesterol absorption in HSD3B7-null mice is reduced to < 9% of the control mice. Additionally, farnesoid-X-receptor (FXR)-mediated negative feedback regulation is also disrupted. Cholesterol 7α-hydroxylase and sterol 12α-hydroxylase mRNAs are elevated fourfold compared with the wild-type mice. Both these enzymes are normally repressed by bile acids acting through FXR. The expression of FXR target genes, such as small heterodimer partner (SHP) mRNA in the liver, and SHP and fibroblast growth factor 15 (FGF15) mRNAs in the gut are significantly decreased. In the liver, SHP mRNA expression is decreased about 50%, whereas in the gut, the expression of both SHP and FGF15 mRNAs are decreased about 90%. The role of FXR and fibroblast growth factor- (FGF) 15 in bile acid homeostasis is discussed in detail later. Mutations inactivating HSD3B7 gene cause a recessive form of neonatal liver failure in humans (Shea et al., 2007); that is, individuals with homozygous recessive genotype have clinical presentation of the disease phenotype.

Formation of MCA in Mice by CYP2C70 and Observations from CYP2C70-Null Mice

In mice, CDCA (3α,7α) is 6β-hydroxylated by CYP2C70 to produce β-MCA (3α,6β,7α). UDCA (3α,7β), which is the 7β-epimer of CDCA, is a primary bile acid in mice and is produced from CDCA by the epimerase action of CYP2C70 (Fig. 7). β-MCA (3α,6β,7β) is synthesized by two pathways—one by epimerization of the OH group on C-7 (from 7α to 7β) of α-MCA and the other by 6β-hydroxylation of UDCA (Takahashi et al., 2016; de Boer et al., 2017; Honda et al., 2020). de Boer et al. (2017) reported that CYP2C70 catalyzes both the 6β-hydroxylation and C-7 epimerization reactions. Takahashi et al. (2016) used a Cyp2c gene cluster knockout mouse model and found the complete absence of both α-MCA and β-MCA in the liver of these mice. Because the mouse Cyp2c gene cluster contains 16 genes, the authors analyzed several individual recombinant Cyp2c genes, including Cyp2c70, by expressing them in HepG2 cells. The analysis revealed that CYP2C70 produced α-MCA from CDCA and β-MCA from UDCA, respectively. Recently, Honda et al. (2020) generated Cyp2a12 and Cyp2c70 single and double knockout mice and confirmed the finding of Takahashi et al. (2016) that CYP2C70 converts CDCA to α-MCA and UDCA to β-MCA. Mouse Cyp2a12 converts DCA, a secondary BA, to primary BAs as evidenced by the accumulation of DCA in Cyp2a12-null mice (Honda et al., 2020). However, the accumulation of UDCA was found to be much lower than CDCA in Cyp2c70 knockout mice, suggesting that most β-MCA is synthesized from CDCA via α-MCA. Therefore, epimerization of α-MCA to produce β-MCA appears to be the predominant pathway rather than producing β-MCA directly from UDCA (Guo and Chiang, 2020; Honda et al., 2020).

Consistent with the role of CYP2C70 in producing α- and β-MCA from CDCA, inactivating liver-specific CYP2C70 function resulted in mice with a human-like bile acid composition. Surprisingly, with a decrease in the FXR antagonist β-muricholic acid and an increase in the FXR agonist CDCA, there was no increase in FXR and decrease in bile acid concentrations in the mice; rather, there was an increase in bile acid concentrations and even hepatotoxicity. There was a blunted FXR-induced cholesterol disposal due to decreased transintestinal cholesterol excretion (de Boer et al., 2020). This finding is consistent with an earlier report by the same authors showing that the pharmacological activation of FXR in mice resulted in a significant increase in β-MCA and a corresponding enhanced cholesterol removal via stimulation of transintestinal cholesterol excretion.

Solubility of Unconjugated and Conjugated Bile Acids

Conjugation of bile acids decreases the pK_a, increases the solubility in polar solvents, and decreases the solubility in nonpolar solvents (Carey, 1984). For example, the pK_a values for unconjugated bile acids in water is about 5 and that for glycine conjugated bile acids is about 3.7 (Fini and Roda, 1987). Therefore, glycine conjugates are moderately strong acids. For comparison, the pK_a of acetic acid is 4.7. In the literature, the reported range of pK_a of unconjugated bile acids is mostly 5–6 and that of glycine-conjugated bile acids is 3–4 . Taurine-conjugated bile acids are stronger acids with pK_a of 1–2 (Kullak-Ublick et al., 2000). Thus, at pH 7.0 unconjugated bile acids have about a 100-fold higher concentration in the water-soluble ionized form (A^‒) than the lipid-soluble unionized or protonated form (HA), about a 10,000-fold higher concentration in the water-soluble than lipid-soluble form for the glycine-conjugated bile acids, and at least 100,000-fold higher concentration in the water-soluble than lipid-soluble form for the taurine-conjugated bile acids.

Hence, most bile acids exist in the water-soluble ionized form at the slightly alkaline pH of the intestinal chyme. As they are ionized, they exist as salts with Na⁺ and other cations. Thus, they should be termed bile salts, but by tradition they are often referred to as bile acids. The pK_a values of bile acids imply that extremely small amounts of bile acids are in the lipid-soluble form to be absorbed passively from the large intestine.

Experimental data show that the aqueous solubility of different bile acids (HA) varies widely, ranging from 5 × 10^‒8 M for lithocholic acid (LCA) to 1.6 × 10^‒3 M for UCA at 37°C. Fully dissociated sodium bile salts are highly soluble in water, attaining molar concentrations as high as 1 to 2 M (Carey, 1984). The critical micellar concentration (CMC) is the concentration of surfactants above which micelles form and all additional surfactants added to the system will form micelles. Therefore, the higher the hydrophobicity of the surfactant, the lower the CMC value. For bile acids, which are surfactants, the CMC values vary from about 0.6 mM to about 10 mM (Carey, 1984). Using microcalorimetric titration, Simonović and Momirović (1997) determined the CMC values of CA, DCA, and CDCA in phosphate buffer (pH: 7.92) to be about 18, 5, and 7 mM, respectively, indicating that DCA is the most hydrophobic (least hydrophilic) of these three bile acids. Micelles formed by the ionized bile salt (A^‒) species can solubilize the unionized (protonated) bile acid (HA) species (Carey, 1984).

Sulfation of Bile Acids

Normally bile acid sulfation is a minor pathway. However, cholestatic conditions significantly increase bile acid sulfation that aids in increased clearance of bile acids through urine and bile. Sulfation is a phase II conjugation reaction catalyzed by sulfotransferases (SULTs). SULTs catalyze the transfer of a sulfonate group (SO₃^‒) from the universal sulfonate donor, 3′-phosphoadenosine 5′-phosphosulfate (PAPS), to a hydroxyl, or amino, or carboxylic acid group of an acceptor molecule (substrate). The resulting conjugates carry a permanent negative charge and are highly water soluble. As a result, the pK_a values of bile acid-sulfates are lower than 1 (https://foodb.ca/compounds/FDB023028; last accessed: June 6, 2021).

Although “sulfonation” is the chemically accurate term to describe this conjugation, “sulfation” is used for historical reasons (Alnouti, 2009). Bile acid sulfation is catalyzed by cytosolic SULTs. The formation of bile acid-sulfated metabolites for excretion by humans was first reported by Palmer (1967). More than 50% of LCA is present in human bile in the sulfated form (Palmer and Bolt, 1971). Bile acid-sulfates are excreted into bile but are also excreted in urine. Biliary excretion of bile acid-sulfates results in their ultimate elimination in the feces because bile acid-sulfates are not well absorbed from the intestine. There are some conflicting reports in the older literature on the proportion of bile acid-sulfate species in humans, but it is now recognized that the primary sulfation site in humans is the 3α-OH whereas that in mice is the 7α-OH, creating bile acid-3-sulfate and bile acid-7-sulfate species, respectively (Zhang and Klaassen, 2010; Chaudhari et al., 2021). Table 5 shows the approximate pK_a values of unconjugated and various conjugated bile acids. In humans, bile acid-3-sulfation is catalyzed by SULT2A1 (Radominska et al., 1990), whereas in mice bile acid-7-sulfation is carried out by SULT2A1 and 2A8 (Feng et al., 2017). Unlike bile acids sulfated at the C-3 position, C-7-sulfates are resistant to hydrolysis and metabolism by the intestinal microbiota, which prevents the C-7-sulfates from being absorbed from the intestine (Robben et al., 1986). Among the secondary bile acids, DCA and LCA predominate in the feces. More than half of LCA in humans is sulfated. Some sulfated LCA is excreted in bile to be ultimately excreted in the feces, whereas some sulfated LCA is effluxed by the liver into the sinusoidal blood by multidrug resistance-associated protein (MRP) 3 and MRP4 for renal excretion (Alnouti, 2009). For example, in cholestasis the bile flow is decreased; this triggers a cellular adaptive response excreting larger amounts of sulfated bile acids in the sinusoidal blood for renal excretion (Rius et al., 2003; Halilbasic et al., 2013).

Cholehepatic Shunt

In addition to the enterohepatic circulation of bile acids between the liver and the intestine, there is a separate shorter path for the recycling of bile acids between hepatocytes and cholangiocytes, which is referred to as the cholehepatic shunt. The cholangiocytes are exposed to high concentrations of bile acids at their apical membrane that faces the flowing bile through bile ductular lumens. In humans, the apical membranes of cholangiocytes have the uptake transporters ASBT and OATP1A2, whereas the basolateral membranes of the cholangiocytes have the efflux transporters Ostα–Ostβ heterodimer and MRP3 (Pauli-Magnus and Meier, 2006). Bile acids transported by the cholangiocytes recycle via the peribiliary plexus back to the hepatocytes for resecretion into bile canaliculus. The cholehepatic shunt prevents the accumulation of bile acids in cholangiocytes, thereby reducing the toxic effects of accumulating bile acids on the cholangiocytes.

Cholehepatic shunt plays an important role in increasing bile flow. Unconjugated bile acids in canalicular bile are protonated by H₂CO₃ generating protonated bile acids and bicarbonate ion (HCO₃⁻). Protonated bile acids are more lipophilic, hence are absorbed by the bile duct epithelial cells and returned to the hepatocytes. These bile acids are subsequently resecreted into bile. Each cycle of absorption and resecretion induces increased bile flow (choleresis) and the generation and secretion of more bicarbonate-rich fluid. An example of hypercholeretic bile acid is norUDCA, which is secreted in part in unconjugated form (Boyer, 2013).

Observations from BSEP-Null Mice

Wang et al. (2001a) developed BSEP-null (Spgp^–/–) mice. Surprisingly, BSEP-null mice did not show any signs of cholestasis except for a slight reduction in bile flow. Cholestasis involves a marked reduction in bile flow, and bile acid secretion is a major driving force of bile flow. Thus, functional inactivation of BSEP in mice, in contrast to humans, does not exhibit any severe phenotype, such as severe cholestasis or progressive liver injury. This is because mice can detoxify hydrophobic (less hydrophilic) bile acids by hydroxylation producing polyhydroxylated bile acids and use alternative mechanisms of bile acid transport (Lam et al., 2005). Interestingly, BSEP-null mice have a sharply reduced secretion of hydrophobic bile acids (5%–6% of the wild type) but maintain nearly normal amounts of the more hydrophilic bile acids, such as α-MCA and β-MCA, and greatly increased amounts of tetrahydroxy bile acids (18% of total biliary bile acids) that are not normally detected in wild-type control mice. Tetrahydroxy bile acids are produced by hydroxylation of bile acids, such as 12α-hydroxylation of MCAs, and they help avoid cholestasis-induced severe liver damage. The presence of tetrahydroxylated bile acids had previously been reported in humans with cholestasis. For example, CDCA, CA, and DCA can be hydroxylated and transformed into tri- or tetrahydroxylated bile acids in patients with cholestasis. CA, which is already a trihydroxy bile acid can be further hydroxylated at the C-1 or C-6 position, producing a tetrahydroxylated bile acid (Bremmelgaard and Sjövall, 1980). Despite the absence of a typical cholestasis phenotype, increased hydroxylation of bile acids in BSEP-null mice is consistent with a cholestatic phenotype, which is attenuated by a greatly increased secretion of polyhydroxylated bile acids into the bile.

Polyhydroxylated bile acids (e.g., tetrahydroxylated bile acids) are less injurious and are exported into the canalicular space by the multidrug resistance [MDR/P-glycoprotein (P-gp)] 1, and MRP2 (Wang et al., 2009; Megaraj et al., 2010). Table 6 summarizes some information on rodent and human BSEPs.

Observations from MDR1-Null and MRP2-Null Mice

Because BSEP-null mice maintain substantial canalicular bile acid secretion and do not develop severe cholestasis, this suggests the presence of an alternative transport mechanism that maintains bile acid secretion into bile canaliculi and allows these mice to remain physiologically relatively normal. These alternative canalicular transport systems involve P-glycoprotein/ (P-gp) MDRP1 and MRP2.

MDR1/P-gp-Null Mice

Because the expression of both P-gp isoforms in mice, MDR1A (gene symbol: Abcb1a) and MDR1B (gene symbol: Abcb1b), was markedly increased in BSEP-null mice, Wang et al. (2009) postulated that P-gp might play a compensatory role in BSEP-null mice in maintaining the alternative bile acid transport process. The authors crossed MDR1A/1B-double knockout (Abcb1a/1b^–/–) mice with BSEP-null mice to create a triple knockout mouse model lacking functional BSEP as well as MDR1A and MDR1B. These mice exhibited signs of cholestasis, including histologic changes typical of cholestatic stress, blockage of bile flow, signs of hepatic inflammation, elevated plasma bilirubin, and mortality. Earlier, Lam et al. (2005) demonstrated that plasma membrane vesicles isolated from a cell line expressing high levels of Chinese hamster P-gp1 (an MDR1 ortholog) was capable of transporting bile acids, albeit with one-fifth the affinity compared with BSEP. This work demonstrated for the first time that MDR1/P-gp in mice could transport bile acids, thus playing a compensatory role in BSEP-null mice in maintaining an alternative bile acid transport process.

MRP2-Null Mice

MRP2 is an efflux transporter that transports glucuronides, glutathione, and sulfate conjugates into bile (Jedlitschky et al., 2006). Mice with a functional BSEP but lacking a functional MRP2 (gene symbol: Abcc2) are not likely to have any major impairment of bile acid transport, as demonstrated by serum bile acid concentrations in MRP2-null (Abcc2^‒/‒) mice. Although bile acid concentrations were higher in the serum of some MRP2-null mice, interanimal variation was high and not statistically different from control mice (Chu et al., 2006). However, functional disruption of MRP2 in both humans and rats causes conjugated hyperbilirubinemia. The expression of MRP2 (as well as MDR1A, MDR2, and MRP3) is markedly induced in BSEP-null mice (Lam et al., 2005). To determine the role of MRP2 in the canalicular transport of bile acids, Megaraj et al. (2010) characterized the transport of two isomers of a tetrahydroxylated bile acid 6-OH-taurocholic acid (6-OH-TC) (3α,6α,7α,12α and 3α,6β,7β,12α) by membrane vesicles overexpressing mouse MRP2, MDR1A, or BSEP. They found that all these transporters could transport 6-OH-TC, but MRP2 showed the highest affinity and capacity for 6-OH-TC transport. They also studied the biliary excretion of 6-OH-TC in wild-type and MRP2-null mice in the presence or absence of a P-gp inhibitor. They also used a P-gp and breast cancer resistance protein transporter inhibitor (GF120918, aka Elacridar) to rule out any potential role of breast cancer resistance protein in the transport. These experiments by Megaraj et al. (2010) demonstrated that MRP2 plays an important role for the lost function of BSEP in mice. The relative affinity for 6-OH-TC transport was in the order of MRP2 > P-gp > BSEP, suggesting that MRP2 likely plays a beneficial role in cholestasis in mice.

In summary, BSEP transports most bile acids, whereas unusual bile acids, such as sulfated bile acids or tetrahydroxylated bile acids, are transported by other ABC transporters, such as MDR1 and MRP2 (Dawson et al., 2009). MRP3 and MRP4 also play an important role in the basolateral efflux of sulfated bile acids into the systemic circulation for renal excretion (Zelcer et al., 2003; Kullak–Ublick et al., 2004). Additionally, when BSEP function is compromised, MDR1 and MRP2 play a compensatory role in canalicular bile acid transport.

Observations from Mdr2-Null and G5/G8-Null Mice

In mice, MDR2 (gene symbol Abcb4) is the ortholog of human MDR3. MDR2-null (Abcb4^‒/‒) mice develop cholestasis, hepatic lesions, injury to the bile canaliculi, and portal inflammation. In homozygous null (Abcb4^‒/‒) mice, the bile is almost completely devoid of PC whereas in heterozygous null (Abcb4^+/‒) mice, the PC content of the bile is 60% of that in the wild-type mice (Smit et al., 1993). MDR2 in mice does not transport cholesterol, which is transported by G5/G8. Nevertheless, cholesterol excretion in MDR2-null mice is severely impaired because the absence of PC makes the bile acid micelles poor cholesterol solubilizers (Oude Elferink and Paulusma, 2007). The role of PC in solubilizing cholesterol in bile is further evidenced by G5/G8-null mice (with both Abcg5 and Abcg8 genes disrupted). Wittenburg and Carey (2002) suggested that the G5/G8 heterodimer translocates cholesterol from the inner to the outer leaflet of the canalicular membrane. Thus, G5/G8-null mice have extremely low biliary cholesterol concentrations, but biliary phospholipids and bile acid concentrations are not altered (Yu et al., 2002).

ATP8B1 is a P4-type ATPase and another membrane lipid translocator. It acts as a flippase that translocates PS and PE from the outer to the inner leaflet of the plasma membrane to maintain the asymmetric lipid distribution. This lipid asymmetry plays an important protective role and makes the membrane more resistant to bile acids. Because PS is not translocated from the outer to the inner leaflet in ATP8B1-null mice, it is readily extracted by bile salts. Hence, ATP8B1-null mice show enhanced biliary excretion of PS and cholesterol (Paulusma et al., 2006).

The current model of canalicular lipid extraction depicts that bile salts transported by BSEP in the bile canaliculi reach a concentration in the bile above the critical micellar concentration and form micelles. MDR3 in humans (MDR2 in mice) translocates PC from the inner to the outer leaflet of the canaliculi so that it can be extracted by bile salt micelles to form mixed micelles. Likewise, G5/G8 translocates cholesterol from the inner to the outer leaflet and allows it to be extracted in the same manner as PC by bile salt micelles. Mixed micelles of bile salts and phospholipids extract and solubilize cholesterol more efficiently than simple bile salt micelles. In this model, ATP8B1 translocates PS and PE from the outer to the inner leaflet of the canalicular membrane of the hepatocyte, thereby increasing the relative content of cholesterol in the outer leaflet, which makes the membrane resistant to bile salts (Oude Elferink and Paulusma, 2007).

Some Gut Microbiota Are Harmed by Bile Acids

Some bile acids can inhibit bacterial growth. The inhibitory effect of bile acids on bacterial growth was observed with Bacteroides, Clostridia, Lactobacillus, Bifidobacteria, Escherichia coli, Enterococcus fecalis, and 7α-dehydroxylating bacteria, such as Clostridium scindens, Clostridium hylemonae, and Clostridium hiranonis (Tian et al., 2020). The cytotoxic effect of bile acids is associated with their amphipathic nature. Bile acids can disrupt bacterial cell membranes, damage DNA, alter the conformation of proteins, and chelate iron and calcium. Therefore, bacteria colonizing the intestinal tract, such as Lactobacillus, Bifidobacterium, and Bacillus can repair bile acid-induced damage. Bacteria chronically exposed to bile acids overexpress proteins that help counteract the negative effects of bile acids, such as transporters that efflux bile salts, and enzymes that modify the global metabolism (Tian et al., 2020). Using a combination of flow cytometry, growth rate measurements (OD₆₀₀), nuclear magnetic resonance and mass spectrometry based metabolomics to investigate the effects of bile acids on individual bacteria and cecal bacterial community, Tian et al. (2020) showed that (1) unconjugated bile acids possess more potent antibacterial activity than conjugated bile acids and (2) gram-positive bacteria are more sensitive to bile acids than gram-negative bacteria.

Observations from ASBT-Null Mice

ASBT-null mice (Slc10a2^‒/‒) have impaired intestinal bile acid absorption and, consequently, increased fecal bile acid loss (Dawson et al., 2003). No ileal histologic or ultrastructural changes were noted in ASBT-null mice, and fecal bile acid excretion was increased 10- to 20-fold. Because of the poor bile acid absorption and the increased bile acid loss in feces, bile acid synthesis is increased, but the bile acid content is still decreased by 80% and selectively enriched in CA, reflecting an increased contribution of the CYP7A1/CYP8B1 pathway (Dawson et al., 2003). Because the bile acid pool in mice is rich in hydrophilic bile acids and taurine conjugates, mice appear to be more dependent on ASBT function than humans (Li and Dawson, 2019).

ASBT is also found in the apical membrane of cholangiocytes that form the epithelial lining of bile ducts. Here ASBT plays an important role in cholehepatic shunting in which bile acids, after being secreted from hepatocytes, are reabsorbed by cholangiocytes and cycled back to hepatocytes. Cholehepatic shunting is triggered by bile duct obstruction and cholestasis. During cholehepatic shunting, taurine-conjugated bile acids are transported by ASBT, whereas unconjugated bile acids are thought to directly permeate through the membrane (Glaser and Alpini, 2009).

Observations from IBABP-Null Mice

Praslickova et al. (2012) developed an IBABP-null (Fabp6^‒/) mouse model. The IBABP-null mice had a decreased bile acid content, but there were sex dimorphic effects on the retention and excretion of bile acids. Using exogenously administered radiolabeled bile acid (³[H]TCA) to trace the transit and fate of bile acids in the small intestine, it was found that female IBABP-null mice had an increased excretion of bile acids in the feces, whereas IBABP-null males had an increased retention of bile acids in the small intestine. The authors concluded that the loss of IBABP impairs the intestinal transit of bile acids in male mice, but female mice have greater dependence on IBABP for the proper transfer of bile acids across the enterocytes. The authors also concluded that IBABP may not be an absolute prerequisite for the transfer of bile acids within enterocytes; however, it is required for efficient transport of bile acids from the intestinal lumen (apical part of the enterocyte) to the portal circulation (basolateral part of the enterocyte) (Praslickova et al., 2012).

Observations from OSTα‒OSTβ-Null Mice

OSTα-null (Slc51a ^‒/‒) mice were developed to study the in vivo functions of this unusual transporter (Rao et al., 2008; Ballatori et al., 2008). In OSTα-null mice, OSTβ protein was also markedly reduced in addition to the loss of OSTα protein, demonstrating that the expression of α and β subunits are coordinately regulated so that the loss of OSTα resulted in a dramatically reduced expression of OSTβ. Because of the loss of OSTα–OSTβ heterodimer expression, bile acids accumulate inside the enterocytes, causing damage to the enterocytes. As a result, OSTα-null mice have morphologic changes in the ileum secondary to enterocyte injury (Rao et al., 2008; Ballatori et al., 2008). Total bile acid pool from liver, gallbladder, and small intestine is 90% lower, and serum bile acid pool is 60% lower in OSTα-null mice compared with controls (Ballatori et al., 2008). Unlike ASBT-null mice that have a selective enrichment of CA in the bile acid pool, reflecting an increased contribution of the CYP8B1 pathway, the composition of the bile acid pool is not altered in OSTα-null mice (Dawson et al., 2003; Rao et al., 2008). Fecal cholesterol excretion is increased about fourfold in OSTα-null mice, and the same trend is observed in ASBT-null mice (Dawson et al., 2003; Rao et al., 2008). When OSTα-null mice were fed CA, the fecal excretion of CA was increased about fivefold by day 2. This indicated not only a reduction of OSTα–OSTβ mediated CA transport to the portal circulation, but also a reduction in the uptake of CA into the enterocytes by ASBT, strongly suggesting a corresponding reduced ASBT expression (Rao et al., 2008). A reduction in ASBT expression secondary to the loss of OSTα–OSTβ expression suggests the existence of a sensing mechanism that prevents the accumulation of bile acids in enterocytes, thereby protecting them from structural damage.

By examining intestinal fibroblast growth factor 15 (FGF15) expression in wild-type, OSTα-null, FXR-null, and OSTα‒FXR-double null mice, Lan et al. (2012) found that the changes in bile acid homeostasis in OSTα-null mice were due to the altered gut-liver bile acid signaling through the FXR‒FGF15‒FGFR4 pathway (discussed in more detail later under Observations in FXR-null mice).

The importance of OSTα‒OSTβ in human intestinal bile acid transport has been demonstrated by two brothers bearing an inherited OSTβ-deficiency. The clinical presentation of these pediatric patients included congenital chronic diarrhea, fat-soluble vitamin deficiency, and features of cholestatic liver disease (Sultan et al., 2018). The authors performed whole exome sequencing and identified a homozygous single nucleotide deletion, which is in codon 27 in the first coding exon of the OSTβ gene. This deletion resulted in a frameshift and corresponding formation of a premature stop in codon 50. Cells containing the truncated OSTβ protein have a reduced expression of the OSTα partner protein and markedly reduced taurocholate uptake.

Observations from NTCP-Null Mice

In NTCP-null mice (Slc10a1^–/–), serum bile acid concentrations remain normal in the majority (60%–75%) of the animals. Only a subset of NTCP-null mice has markedly elevated serum bile acid levels, particularly conjugated bile acid concentrations (millimolar concentration compared with controls, which were <20 µM) (Slijepcevic et al., 2015). The results in NTCP-null mice indicate the existence of a parallel bile acid uptake system in mice. Feeding UDCA to NTCP-null mice resulted in serum bile acid accumulation in the null mice that previously had normal serum bile acid concentrations. This demonstrates the existence of a Na⁺-independent bile acid uptake system that operates close to its maximum capacity in NTCP-null mice and can be easily saturated (Slijepcevic et al., 2015).

In contrast to NTCP-null mice, humans that inherited NTCP deficiency have nearly a 100-fold increase in serum bile acid concentrations, primarily conjugated bile acids (Vaz et al., 2015). Likewise, NTCP inhibition by myrcludex B, a drug used to block the uptake of viruses into the liver, resulted in a 19-fold increase in total serum bile acids, whereas conjugated bile acids increased more than 100-fold (Blank et al., 2018). These reports underscore a crucial role of NTCP in the hepatic uptake of bile acids, especially conjugated bile acids, in humans.

Observations from OATP-Null Mice

van de Steeg et al. (2010) developed OATP1A/1B-null (Slco1a/1b2^–/–) mice. Because of the deletion of the entire Slco1a/1b locus, OATP1A/1B-null mice had all five OATPs deleted— OATP1A1, -1A4, -1A5, -1A6, and OATP1B2. The serum levels of unconjugated bile acids were elevated by about 13-fold in OATP1A/1B-null mice compared with controls, suggesting an important role of mouse OATP transporters in the uptake of unconjugated bile acids. Using OATP1B2-null (Slco1b2^–/–) mice, Csanaky et al. (2011) showed that these mice have a 15-fold higher unconjugated serum bile acid levels, thereby clearly identifying OATP1B2 as a key transporter that mediates the basolateral uptake of unconjugated bile acids into the liver.

Further studies (Slijepcevic et al., 2017) revealed an important role of OATP, along with NTCP, in the uptake of conjugated bile acids. When wild-type mice were administered myrcludex B (NTCP inhibitor), there was a small effect on serum taurocholate clearance compared with controls. In contrast, treating OATP1A/1B-null mice with myrcludex B caused a complete abrogation of taurocholate transport, suggesting that NTCP and OATP together mediate hepatic uptake of conjugated bile acids in mice. Next, Slijepcevic et al. (2017) used OATP1A/1B-null mice for expressing the human OATP1B1. The authors found that hepatic expression of human OATP1B1 reversed the conjugated hyperbilirubinemia in OATP1A/1B-null mice but could not block the rise in plasma conjugated bile acids after inhibition of NTCP with myrcludex B. These observations strongly suggest that OATP1B1 can mediate the hepatic uptake of conjugated bile acids, but NTCP is the primary uptake transporter of conjugated bile acids in humans. Using bile acid uptake in U2OS cells transiently overexpressing mouse NTCP, OATP1A1, OATP1B2 and rat OATP1A4, Slijepcevic et al. (2017) demonstrated that mouse OATP1A1 and OATP1B2 were capable of significant taurocholate uptake (10–12 times greater than in untransfected U2OS cells).

Observations in LXR-Null Mice

The role of LXR was demonstrated using LXR-null (Lxrα^–/–) mice. When LXR-null mice were fed a high cholesterol diet, the mice accumulated massive amounts of cholesterol and had liver abnormalities because these mice could not mount the compensatory increase in bile acid synthesis and elimination due to the lack of activation of the Cyp7a1 gene (Peet et al., 1998). However, the accumulation of more cholesterol in the liver of LXR-null mice may not be solely due to the lack of induction of CYP7A1 but may also be, in part, due to impaired cholesterol elimination pathways that are also regulated by LXR, such as ABCA1, ABCG1, and ABCG5/G8. The effect of LXRα on human CYP7A1 gene transcription is less pronounced. This was determined by the transcriptional regulation of luciferase reporter gene expression under the control of human CYP7A1 promoter in HepG2 cells, a model established for the study of regulation of cholesterol 7α-hydroxylase (Chiang et al., 2001).

Observations in FXR-Null Mice

The inhibitory role of bile acid activated FXR on the transcription of Cyp7a1 gene and resulting bile acid synthesis was confirmed using FXR-null (Fxr^–/–) mice (Sinal et al., 2000). When challenged with a diet containing CA, FXR-null mice had severe hepatotoxicity and wasting compared with wild-type mice (Sinal et al., 2000). Bile acid activated FXR is now known to decrease the expression of CYP7A1, CYP8B1, CYP27A1, and CYP7B1 (Chiang, 2015). de Aguiar Vallim et al. (2015) reported that FXR activation induces the expression of many transcriptional repressors including the musculoaponeurotic fibrosarcoma oncogene homolog G (MAFG), which directly represses Cyp8b1 gene expression by binding to the MAFG response element (MAREs) in the promoter of the Cyp8b1 gene.

By examining the FGF15 expression in the intestine of wild-type, OSTα-null, FXR-null, and OSTα‒FXR-double null mice, Lan et al. (2012) concluded that the changes in bile acid homeostasis (decreased bile acid content) in OSTα-null mice were due to altered gut-liver bile acid signaling through the FXR‒FGF15‒FGFR4 pathway. The authors found that the expression of CYP7A1 in the liver was decreased in OSTα-null mice, and ileal FGF15 protein was elevated almost 20-fold. These effects were expected because the accumulated bile acids should activate FXR expression in enterocytes, which in turn should upregulate FGF15 expression. FGF15, after being transported to the liver, should ultimately downregulate the expression of CYP7A1 (Figure 9). Conversely, OSTα‒FXR-double null mice have increases in hepatic CYP7A1 expression in the liver due to the loss of FXR and the loss of downstream FXR signaling. Surprisingly, OSTα‒FXR-double null mice had restoration of intestinal cholesterol absorption. The authors concluded that signaling via FXR is required for the paradoxical repression of hepatic bile acid synthesis but not the complex intestinal adaptive changes in OSTα-null mice.

Interestingly, knowledge gained from the whole body FXR-null mice (Sinal et al., 2000), discussed above, was further enriched using tissue-specific FXR-null mice, namely as liver- and intestine-specific FXR-null mice. Using these mice, Kong et al. (2012) showed that the intestinal FXR/FGF15 pathway was critical for suppressing both Cyp7a1 and Cyp8b1 gene expression, but the liver FXR/SHP pathway was important for suppressing Cyp8b1 gene expression and had a minor role in suppressing Cyp7a1 gene expression. Figure 9 illustrates how bile acids negatively regulate their own synthesis. Bile acid activated intestinal FXR induces intestinal FGF15 expression. FGF15 is carried through portal blood to the liver, where it binds to FGFR4 and activates it. Activated FGFR4 in the liver strongly triggers the extracellular signal-regulated kinase 1/2 (ERK1/2), and to a lesser extent c-Jun N-terminal kinase 1/2 (JNK1/2) to downregulate both Cyp7a1 and Cyp8b1 gene expression in the liver (Kong et al., 2012). Bile acid also activates hepatic FXR that induces hepatic SHP expression. Increased SHP expression contributes to the downregulation of both Cyp7a1 and Cyp8b1 gene expression (Kong et al., 2012).

FXR regulates the expression of various bile acid transporters, such as BSEP, NTCP, ASBT, OSTα–OSTβ, and the bile acid conjugation enzymes (e.g., BACS, BAAT) (for an extensive list of the FXR target genes, see Lee et al., 2006). Of these FXR targets, BSEP, OSTα–OSTβ heterodimer, BACS, and BAAT expressions are induced, whereas NTCP and ASBT expressions are repressed. FXR is activated by both free and conjugated bile acids. A highly hydrophilic bile acid, taurine conjugated β-muricholic acid, acts as an FXR antagonist (Sayin et al., 2013). In humans, serum FGF19 peaks about 90–120 minutes after the postprandial release of bile acids, and this peak precedes the repression of bile acid synthesis. FGF19 decreases in subjects administered the bile acid sequestrant, cholestyramine (Lundasen et al., 2006).

Observations from TGR5-Null and TGR5-Transgenic Mice

Maruyama et al., (2006) developed TGR5-null (Tgr5^‒/‒) mice. In these mice, the expression of proteins involved in the hepatic synthesis and transport of bile acids was higher than wild-type controls, but the bile acid pool was not increased; in fact, it was decreased by about 25%, suggesting that TGR5 contributes to bile acid homeostasis. Interestingly, both homozygous and heterozygous TGR5-null females fed a high-fat diet for 2 months showed significant fat accumulation with body weight gain compared with the wild-type mice, suggesting that TGR5 plays an important role in energy homeostasis. TGR5-null mice fed a lithogenic diet are protected against cholesterol gallstone formation (Maruyama et al., 2006; Duboc et al., 2014).

Thomas et al. (2009) showed that glucose tolerance was improved in transgenic mice overexpressing TGR5, and it was accompanied by increased GLP-1 and insulin secretion. This observation strongly suggests a role of bile acids and the bile acid receptor TGR5 in glucose metabolism through the Bile Acid–TGR5–GLP-1 pathway. When the authors added a TGR5 agonist (6-alpha-ethyl-23(S)-methyl-cholic acid) to a high-fat diet, glucose tolerance, insulin and GLP-1 secretion, and insulin sensitivity in both liver and muscle of TGR5-overexpressing mice improved, but not in TGR5-null (Tgr5^–/–) mice. TGR5-overexpressing mice have enhanced GLP-1 secretion and insulin release in response to a glucose load, and TGR5-null mice have impaired glucose tolerance (Thomas et al., 2009). Therefore, the data from TGR5-overexpressing and TGR5-null mice collectively suggest that TGR5’s role in glucose metabolism is bile acid dependent (Pols et al., 2011).

The Value of Serum Bile Acid Concentrations

The serum bile acid concentrations have often been thought to reflect the concentration of bile acids in the liver. However, as shown in Table 13 this is not always true. While there was a similar percent change in bile acid concentrations in the serum as in the liver for a few of the experimental protocols, this was not true for most of the experiments. For example, the bile acid concentrations in the liver of the male and female CAR-null mice and the CAR agonist treated mice decreased 35%–50%, but the concentration in serum did not change. An even more extreme example is the female mice administered the AhR agonist that resulted in a decrease in the concentrations of bile acids in the liver but an increase in the serum. It is not surprising that the bile acid concentration in the serum does not reflect the concentration in the liver as the concentration in the serum is dependent on their absorption from the intestine and the first-pass extraction by the bile acid transporters in the liver. In addition, less than 0.05% of the bile acids in the body is in the serum.

Maintenance of Bile Acid Concentrations in the Liver

The maintenance of bile acid concentrations in the liver is thought to be important to provide sufficient bile acids for the intestinal absorption of lipids and lipid-soluble vitamins, and yet not so high that they will produce liver injury. Bile acid concentrations in the liver result from the balance of uptake of bile acids from the portal circulation and synthesis of new bile acids in the liver, and that excreted into the bile. Under the protocols of these 18 experiments, surprisingly 16 protocols resulted in a decrease in bile acid concentrations in the liver (Table 13). Thus, bile acid concentrations in the liver are not tightly regulated as usually thought.

Biliary Excretion of Bile Acids

Biliary excretion of bile acids is essential for their transfer to the intestine. However, most experimentalists do not cannulate the bile duct and thus there is minimal data available on biliary excretion. As noted in Table 13, the biliary excretion of bile acids was not altered by 10 of the 18 experimental protocols. Biliary excretion of the bile acids remained constant or even increased when the bile acid concentrations in the liver decreased about 50%. This might suggest that the primary factor the body is attempting to regulate is to maintain the correct concentration of bile acids in the gut. This can be achieved by regulating bile acid secretion.

The Bile Acid/FXR/FGF15/CYP7a1 Paradigm after Activation or Knockout of Various Transcription Factors

As noted previously, the homeostasis of bile acids in mice is thought to involve the following steps in sequence: (I) bile acid-mediated activation of the intestinal FXR, (II) increases in FGF15 expression, (III) secretion of bile acids and FGF15 into the portal blood, (IV) interaction of FGF15 with the FGFR4/β-KL in the liver, and (V) decreased expression of CYP7A1, resulting in decreased synthesis of bile acids.

Bile acid-mediated expression of FXR not only decreases CYP7A1 expression in the liver, but also (I) decreases NTCP expression in the liver (hence reduces uptake of bile acids into the liver) and (II) increases BSEP expression in the liver (hence increases efflux of bile acids out of the liver). The combined effects of such differential expression of these transporters in the liver should decrease the concentration of bile acids in the liver, resulting in prevention/reduction of bile acid-induced hepatotoxicity.

In the intestine, in addition to the bile acid-FXR induction of FGF15, bile acids (I) decrease the expression of ASBT (hence reduced uptake of bile acids into the enterocytes) and (II) increase the expression OSTα/OSTβ (hence increased efflux of bile acids out of the enterocytes into the blood). The combined effects of such differential expression of these transporters in the intestine is a decrease in the amount of bile acids in the enterocytes and prevention/reduction of bile acid-induced enterocyte damage.

Of the 18 groups of mice shown in Table 13, 10 groups had altered expression of FXR in the intestine. Five of the groups had a decrease in FGF15 expression (AhR agonist in female mice, PXR agonist in female mice, PPARα agonist in male mice, PPARα null in female mice, and NRF2 null in male and female mice). According to the intestinal FGF15-CYP7A1 paradigm, one would expect corresponding increases in CYP7A1 in these same groups of mice. However, only in two of these five groups was there an increase in CYP7A1. In six groups of mice, an increase in FGF15 expression was observed in the intestine (AhR agonist in male mice, PXR null in male and female mice, PXR agonist in male mice, PPARα null in female mice, and NRF2 agonist in female mice). Again, according to the FGF-CYP7A1 paradigm, one would anticipate that the CYP7A1 would decrease in these mice; however, only in three of the six groups was there a decrease in CYPA1. Thus, in many of these groups of mice, there is a poor inverse expression pattern between FGF15 and CYP7A1.

The intestinal FXR-FGF15 pathway not only is thought to have a downregulating effect on CYP7A1 expression and a corresponding decrease in bile acid synthesis, but also a downregulating effect on NTCP expression to decrease the uptake of bile acids into the liver. However, of the 18 groups of mice, only three groups showed an inverse expression pattern between FGF15 and NTCP. It is also thought that the FXR-FGF15 pathway increases BSEP to enhance the transport of bile acids into bile. However, only four groups showed a direct association between the expression of FGF15 and BSEP whereas two groups showed an inverse association. Thus, overall, there is a poor association between FGF15 and the expression of bile acid transporters in the liver.

Bile acids in the enterocytes not only activate FXR to increase the expression of FGF15, but also decrease ASBT expression to reduce the uptake of bile acids into enterocytes and increase the expression of OSTα/OSTβ to increase the efflux of bile acids out of the enterocytes into the portal circulation. The opposite expression of these two transport systems is thought to help decrease the concentration of bile acids in the enterocytes and protect them from bile acid toxicity. However, the data in Table 13 indicate that the observed expression of both ASBT and OSTα/OSTβ do not correlate, inversely and directly respectively, with the expression of FGF15 in most groups of mice to regulate bile acid concentrations in the enterocytes and hepatocytes.

Lack of Association between Intestinal FXR Expression and Bile Acid Concentrations in the Liver

Of the 18 protocols, only six exhibited an inverse association between the FXR mRNA in the intestine and BA concentrations in the liver.

Lack of Association between the Expression of CYP7A1, CYP7B1, and CYP8B1

As noted previously, CYP7A1 is the enzyme that hydroxylates the C-7 position in the classic route of bile acid synthesis, whereas CYP7B1 does the same reaction for the alternative route of bile acid synthesis. CYP8B1 is the 12-hydroxylase that is responsible for the formation of cholic acid. There was no association between CYP7A1 and CYP7B1 expression as seven of the protocols increased and three decreased CYP7A1, whereas only two resulted in an increase of CYP8B1 and 12 decreased it. CYP8B1 was the most stably expressed of these three enzymes as only four protocols increased it and three decreased CYP8B1 (Table 13).

Changes in Serum and Liver Bile Acid Concentrations

A poor association was observed between changes in the concentrations of bile acids in the serum and liver after various protocols (Table 14). At 12 hours of bile duct ligation (BDL), the bile acid concentration in the liver increased 18-fold and in serum 576-fold. The bile acid resin decreased liver bile acid concentrations 85%, but serum concentrations remained constant. Germ-free mice have twice the bile acid concentration in the liver and a fourfold higher concentration in the serum.

The Bile Acid/FXR/FGF15/CYP7A1 Paradigm

The bile acid/FXR/FGF15/CYP7A1 paradigm exhibited a poor inverse association between FGF15 and CYP7A1, seen after the alteration of various receptors described in the previous section (Fig. 13); however, a clear inverse association was observed in some of these experimental models (Fig. 14). For example, mice fed the 2% resin diet and the CYP450-reductase-null mice both had more than an 85% reduction in FGF15 mRNA in the ileum and a 3- to 4-fold increase in CYP7A1 mRNA in the liver. Most markedly, the HMG-CoA reductase inhibitor decreased FGF15 by 86% in the intestine and increased CYP7A1 11-fold in the liver. A 2- to 4-fold increase in FGF15 in the intestine in CYP7A1-TG mice and in mice fed 0.10% CA decreased CYP7A1 by more than 95% in the liver. However, there was no inverse association between intestinal FGF15 and CYP7A1 in the liver in the caloric-restriction experiments, the circadian rhythm experiments, and female/male mice.

An increase in FGF15 is thought not only to decrease CYP7A1 but also increase BSEP and decrease NTCP expressions in the liver, collectively resulting in less bile acid accumulation in hepatocytes. But this association of expressions was seldom observed in these various experimental models. Likewise, a decrease in FGF15 in the intestine is thought to decrease ASBT and increase OSTα and OSTβ expressions in the intestine, collectively resulting in less bile acid accumulation in the enterocytes, but this too was seldom observed in the various experimental models.