Bile Acid Sulfation: A Pathway of Bile Acid Elimination and Detoxification
Abstract
Sulfotransferase-2A1 catalyzes the formation of bile acid-sulfates (BA-sulfates). Sulfation of BAs increases their solubility, decreases their intestinal absorption, and enhances their fecal and urinary excretion. BA-sulfates are also less toxic than their unsulfated counterparts. Therefore, sulfation is an important detoxification pathway of BAs. Major species differences in BA sulfation exist. In humans, only a small proportion of BAs in bile and serum are sulfated, whereas more than 70% of BAs in urine are sulfated, indicating their efficient elimination in urine. The formation of BA-sulfates increases during cholestatic diseases. Therefore, sulfation may play an important role in maintaining BA homeostasis under pathologic conditions. Farnesoid X receptor, pregnane X receptor, constitutive androstane receptor, and vitamin D receptor are potential nuclear receptors that may be involved in the regulation of BA sulfation. This review highlights current knowledge about the enzymes and transporters involved in the formation and elimination of BA-sulfates, the effect of sulfation on the pharmacologic and toxicologic properties of BAs, the role of BA sulfation in cholestatic diseases, and the regulation of BA sulfation.
Key words
- bile acids
- sulfate
- sulfation
- sulfonate
- sulfonation
- sulfotransferase
- SULT
- nuclear receptors
- gender difference
- regulation
- homeostasis
- cholestasis
Bile acids (BAs) are synthesized in the liver, from the oxidation of cholesterol, and secreted into duodenum. BAs are conserved through efficient enterohepatic recycling. BAs have many physiologic functions including the regulation of the expression of genes involved in cholesterol, glucose, and their own homeostasis. However, BAs also have several pathologic effects including carcinogenicity and liver toxicity. Therefore, maintenance of BA homeostasis is essential to achieve their physiologic functions and avoid their toxic effects.
Sulfation is an important metabolic pathway to detoxify and eliminate BAs. BA-sulfates are more water soluble and therefore are more readily excreted in feces and urine. Furthermore, BA-sulfates are less toxic than unsulfated BAs. Therefore, sulfation may serve as an efficient mechanism to maintain BA homeostasis in health and disease.
Despite the importance of sulfation in BA research, there are no recent reviews of this topic (Stiehl, 1974b), whereas hundreds of reviews on other aspects of BA research are available. Therefore, this review focuses on several aspects related to BA sulfation including BA-sulfate formation, elimination, toxicity, and regulation. Because there are major differences in the BA physiologic and pathologic effects among various species, this review focuses on human data, and occasionally presents animal data to compare with or support human data. Because this topic has not been reviewed for a long time, an overview of all data available (regardless of how old they are) is presented, in an attempt to link both old and current data. Finally, as indicated in the text, BA synthesis and enterohepatic recirculation have been the subject for several reviews. However, relevant information from these reviews will still be discussed in some detail to demonstrate the marked influence of sulfation on BA disposition.
BA SYNTHESIS
Cholesterol is the building block of BA synthesis. Detailed description of the enzymes involved in BA synthesis can be found in a recent review (Russell, 2003). The major pathway of BA synthesis is called the neutral or classical pathway and it is initiated with cholesterol hydroxylation at the 7α position by the microsomal CYP7A1 enzyme, which is exclusively expressed in the liver (Chiang, 1998). Next, is the classical oxido-reduction step in steroidogenesis, which includes the oxidation of the 3β-OH and isomerization of the C5-C6 double bond by the microsomal C27-3β-hydroxysteroid dehydrogenase (C27-3β-HSD) (Schwarz et al., 2000). The resulting intermediate is either, hydroxylated at the 12α position by the microsomal CYP8B1 or, passed on directly to the next step (Gafvels et al., 1999). The 12α-hydroxylated intermediates and those that escaped 12α hydroxylation have their C3 oxo and the C4-C5 double bond reduced to yield 3α-OH intermediates by the cytosolic oxosteroid 5β-reductase and 3α-HSD (members of the aldo-keto reducatse family) (Penning et al., 2004). 12α hydroxylation will ultimately produce cholic acid (CA), whereas intermediates that were not hydroxylated will ultimately produce chenodeoxycholic acid (CDCA). CA and CDCA are the primary BAs in humans.
Other primary BAs are synthesized in other species such as, hyocholic acid in pigs (Lundell et al., 2001), α- and β-muricholic acid (MCA) in rodents, and ursodeoxycholic acid (UDCA) in bears (Russell, 2003). Figure 1 shows the chemical structure of the most common BAs in mammals.
The chemical structure of the most common BAs in mammals.
The next step in BA synthesis is the hydroxylation and oxidation to a carboxylic acid at the C27 position by the mitochondrial CYP27A1 enzyme (Pikuleva et al., 1998), followed by ligation to coenzyme A by the bile acid coenzyme-A synthetase (BAS) (Mihalik et al., 2002). The side chains of these C27 intermediates are then shortened to C24 BAs by β-oxidation in the peroxisomes (Norlin and Wikvall, 2007). The final step in BA synthesis is the amidation of the BA-CoA with an amino acid, usually glycine (G) or taurine (T), by the BA-CoA:amino acid N-acyltransferase (BAT) (Solaas et al., 2000). BAs are almost exclusively synthesized and excreted from the liver in the amidated form. G-amidates are predominant in humans (Falany et al., 1994), whereas T-amidates are predominant in rodents (Falany et al., 1997). Amidation increases the acidity of unconjugated BAs, where pKA is reduced from about 5.5 for the unconjugated BAs to 4.5 and 1.5 for those with G- and T-amidation, respectively (Falany et al., 1997; Hofmann and Hagey, 2008; Russell, 2003). This results in complete ionization of BAs at physiological pH, which markedly increases their solubility and therefore, decreases membrane permeability (Gu et al., 1992; Hofmann and Mysels, 1992).
Alternative pathways for BA synthesis, which do not require the initiation by CYP7A1, are also known (Axelson and Sjovall, 1990). These pathways are initiated via the hydroxylation of the cholesterol side chain at the C24, C25, or C27 positions by various enzymes such as, CYP27A1 (Lund et al., 1993) and CYP46A1 (Lund et al., 1999). The resulting oxysterols are then hydroxylated at the 7α position by CYP7B1 (Rose et al., 1997; Schwarz et al., 1998) and CYP39A1 (Li-Hawkins et al., 2000) instead of CYP8B1. In contrast to the classical pathway, these alternative pathways produce predominantly CDCA (Vlahcevic et al., 1999). Alternative BA pathways seem more important in conditions associated with deficiency in CYP7A1 activity (Axelson et al., 1989).
ENTEROHEPATIC RECYCLING OF BAs
BAs synthesized in the liver are secreted into bile, which flows through the bile duct to the intestine. BAs are efficiently absorbed from the intestine, returned to the liver, and resecreted into bile. This cycle is called the enterohepatic recirculation of BAs, and was recently reviewed (Hofmann and Hagey, 2008).
BAs are excreted from the liver into bile via efflux transporters located in the canalicular membrane. Amidated BAs are excreted into bile via the canalicular bile salt export pump (BSEP) (Byrne et al., 2002; Noe et al., 2002). In contrast, unconjugated BAs generally have lower affinity (Gerloff et al., 1998) or no affinity (Noe et al., 2002; Zelcer et al., 2003) for BSEP. In many species including humans, most of the secreted bile is stored in gall-bladder, which, under the influence of cholycystokinin secretion after meal ingestion, contracts to empty its contents into the duodenum. Other species, such as the rat, do not have a gall-bladder and therefore, most of its secreted bile is stored in the small intestine (Hofmann and Hagey, 2008).
In the small intestine, most amidated BAs are actively absorbed in the ileum (Aldini et al., 1996; Kuipers et al., 1988) by the apical Na+-dependent bile salt transporter (ASBT) located on the apical side of enterocytes (Wong et al., 1995). Organic anion-transporting polypeptides (OATPs), which have a lower affinity for BAs than ASBT (Walters et al., 2000), may also contribute to BA transport across the apical membrane throughout the intestine. Both transporters have higher affinity for amidated rather than unconjugated BAs (Craddock et al., 1998; Walters et al., 2000). Because unconjugated BAs are mostly non-ionized and because they have low affinity for ASBT, passive rather than active absorption is the primary route of absorption throughout the intestinal tract (Aldini et al., 1996; Mekhjian et al., 1979; Schiff et al., 1972; Takikawa et al., 1997b). In contrast, because of their high affinity for ASBT, and because they are mostly ionized at intestinal pH, active transport represents the major route of absorption of amidated BAs (Schiff et al., 1972). Vectorial BA transport from the intestinal lumen to blood is then accomplished by the organic solute transporters (OST-α and β) located on the basolateral side of enterocytes (Ballatori et al., 2005; Dawson et al., 2005; Rao et al., 2008). Partial de-amidation, especially of the G-amidates, takes place by the bacteria in the distal parts of the small intestine, and the liberated unconjugated BAs are passively absorbed (Hepner et al., 1972a,b, 1973).
Unabsorbed BAs are passed along from the small to large intestine. In the large intestine, BAs undergo bacterial transformation, including de-amidation and dehydroxylation. Primary BAs are dehydroxylated at the 7α position by colonic bacteria to produce secondary BAs (Lundeen and Savage, 1990). Deoxycholic acid (DCA) and lithocholic acid (LCA) are secondary BAs produced from the dehydroxylation of CA and CDCA, respectively (Hofmann and Hagey, 2008). Dehydroxylation is performed by strict anaerobic bacteria, whereas de-amidation is performed by both aerobic and anaerobic bacteria (Yesair and Himmelfarb, 1970). Therefore, de-amidation takes place in both the small and large intestines, whereas dehydroxylation only occurs in the large intestine. Using isolated intestinal bacterial strains, it was demonstrated that de-amidation is a pre-requirement for BA dehydroxylation (Gustafsson et al., 1968; Midtvedt, 1974). However, limited dehydroxylation of BAs-amidates still occurs in vivo (Hepner et al., 1972a). ASBT is not abundantly expressed in the colon (Craddock et al., 1998). Therefore, minimal active absorption of the intact BA-amidates takes place in the large intestine (Kuipers et al., 1988; Mekhjian et al., 1979), whereas unconjugated primary and secondary BAs are passively absorbed. BAs that are not absorbed from the small and large intestine undergo further bacterial metabolism, such as reduction and epimerization of their OH groups from the α- to the β- conformation, before excretion in feces (Franklund et al., 1990; Macdonald et al., 1983). Ninety-five percent of BAs excreted in bile are reabsorbed throughout the intestinal tract and less than 5% are excreted in feces, mostly as secondary unconjugated BAs (Hamilton et al., 2007; Hofmann, 1999a; Beher et al., 1984; Mosbach, 1972; Mosbach and Salen, 1974; Owen et al., 1984, 1987a,b; Pacini et al., 1987; Radominska et al., 1993; van Gorkom et al., 2002).
Absorbed BAs are carried in the portal vein and extracted by the liver via the Na+-taurocholate cotrasporting polypeptide (NTCP), members of the OATP family (Hagenbuch and Meier, 1994; Hagenbuch et al., 1991; Jacquemin et al., 1994), fatty acid transport protein (Doege et al., 2006), and/or via passive diffusion (Ko et al., 1994). NTCP has a higher affinity for the amidated compared with the unconjugated BAs (Schroeder et al., 1998). In hepatocytes, most BAs are amidated, but other metabolic pathways take place such as, hydroxylation and sulfate or glucuronide conjugation (Radominska et al., 1993). The reabsorbed and newly synthesized BAs are then excreted in bile to complete the enterohepatic cycle.
Efflux of BAs out of hepatocytes into the systemic blood through the sinusoidal membrane also takes place and is mediated by several transporters including OSTα, and β (Ballatori et al., 2005), and multidrug resistance-associated proteins (MRP1, MRP3, MRP4, and MRP6) (Trauner et al., 2005). Under normal conditions, most BAs are contained within the enterohepatic system, with minimal spill over into blood, and negligible urinary excretion (Baumgartner et al., 1995). Under cholestatic conditions, such as bile duct obstruction, BA efflux through the sinusoidal membrane, facilitated by the upregulation of the sinusoidal efflux transporters, becomes more significant (Baumgartner et al., 1995; Trauner and Boyer, 2003; Zollner et al., 2007). Therefore, urinary excretion becomes the primary route of BA elimination. In the kidney, unbound fractions of BAs are filtered through the glomerulus and reabsorbed efficiently through the apical membrane of proximal tubules in a similar manner to their absorption in the intestine (Barnes et al., 1977), presumably via ASBT (Christie et al., 1996), and OATPs (Trauner and Boyer, 2003). In the opposite direction, active tubular secretion into urine may also take place, presumably via (MRP2) and organic anion transporters (OATs) (St-Pierre et al., 2001).
BA SULFATION IN HUMANS
Sulfation is a phase II conjugation reaction that transfers a sulfonate group (SO3−) from the universal sulfonate donor, 3′-phosphoadenosine 5′-phosphosulfate (PAPS), to a hydroxyl, amine, or carboxylic acid group of a substrate. The resulting sulfate-conjugates carry a permanent negative charge (pKa of sulfate moiety < 1), and therefore, are very water soluble. “Sulfonation” is actually the correct term to describe this metabolic pathway; however, the less proper term “sulfation” was chosen because it has been historically popular, especially in BA research. Sulfation is catalyzed by a group of enzymes called sulfotransferases (SULTs) (Glatt, 2000). BAs have several OH groups that can be targets for sulfation; however, as will be discussed later, mono-sulfation at the 3-OH position is predominant in humans. Figure 2 depicts the enzymatic formation of BA-3 sulfates.
The formation of BA-3-monosulfates catalyzed by SULT2A1.
The formation of BA-sulfated metabolites was recognized early as a mechanism of BA elimination in humans (Palmer, 1967). LCA sulfation was the first to be discovered and it was estimated that 40–75% of LCA in human bile is present in the sulfated form (Palmer and Bolt, 1971). After i.v. or oral administration to healthy humans, LCA and its amidates were shown to be sulfated and primarily excreted in bile in the sulfated-amidated (divalent) form (Allan et al., 1976; Cowen et al., 1975a,c). However, sulfation of CDCA and CA administered to cirrhotic patients was much less than LCA, and the rate of CA sulfation estimated to be half that of CDCA (Stiehl, 1974a; Stiehl et al., 1975). All endogenous BAs may be detected in the sulfated form, but in various proportions. Similar to unsulfated BAs, greater than 90% of BA-sulfates are also amidated (divalent), with G-amidates being predominant in humans (Goto et al., 2007; Stiehl et al., 1985).
Structure of BA-Sulfates in Humans
Conflicting data regarding the structural elucidation of BA-sulfates have been reported. It was reported that in human urine, 77% of BA-sulfates are mono-sulfates, 21% are di-sulfates, and 2% are tri-sulfates (Stiehl, 1974a; Stiehl et al., 1982). In addition, equal proportions of mono- and poly-sulfated BAs in human urine were detected (Raedsch et al., 1981). In human plasma, di- and tri-sulfated BAs were also found to constitute less than 5% of BA-sulfates (Bartholomew et al., 1982). Both BA-3- and -7- sulfates were detected at the ratio of 2:1 (Raedsch et al., 1981).
Other reports, however, only detected BA-mono-sulfates. In human urine, only mono-sulfates at the 3-OH position and no poly-sulfates were detected (Alme et al., 1977; Hedenborg and Norman, 1984). Furthermore, it was shown that the small di- and tri-sulfate fractions were actually artifacts due to contamination with the BA-3-mono-sulfate fraction during sample preparation (Alme et al., 1977; Barnes et al., 1979b). Further evidence that only BA-mono-sulfates at the 3-OH position are formed was provided from in vitro studies. Using BA-sulfating protein fractions, purified from human livers, only BA-3-mono-sulfates were formed. Furthermore, incubation with BA-mono-sulfates and -di-sulfates did not result in the formation of BA-poly-sulfates (Loof and Hjerten, 1980)
Structure of BA-Sulfates in Animals
The position at which BAs are sulfated was also studied in other species. Only mono-sulfate metabolites were formed after incubating CDCA with rat and hamster liver homogenates. Sulfation occurred at both the 3-OH and 7-OH positions although the formation rates of BA-3-sulfates were at least five times higher than that of BA-7-sulfates. Di-sulfate formation was negligible, and CDCA-3- and -7-mono-sulfates were not substrates for poly-sulfate formation (Kirkpatrick et al., 1980). It was also shown that, in rat hepatocytes, DCA and CDCA were only metabolized to 3-mono-sulfates (Lambiotte and Thierry, 1980).
Using a rat kidney homogenate, LCA and CDCA were sulfated at the 3-OH and 7-OH positions, respectively (Chen et al., 1978a; Summerfield et al., 1976). In hamsters, incubation of CDCA with various tissue homogenates resulted in the formation of mono-sulfates at the 7-OH position only without any di-sulfates (Barnes et al., 1979b). Mono-sulfates at the 7-OH position were the only BA-sulfates detected in mice feces (Eyssen et al., 1976).
Taking these data collectively, it can be concluded that BA-sulfates are exclusively present in the mono-sulfate form in humans and rodents. Conflicting data have been reported regarding the position of sulfation, but sulfation at the 3-OH and 7-OH is likely to be predominant in humans and rodents, respectively.
EFFECT OF SULFATION ON THE ENTEROHEPATIC RECIRCULATION OF BAs
BA-sulfates undergo limited enterohepatic recirculation because of their limited absorption through the intestine and their limited extraction by the liver. In addition, transporters and bacterial transformation pathways involved in BA-sulfate absorption, metabolism, and excretion are different than those for unsulfated BAs. Figure 3 illustrates the effect of sulfation on the enterohepatic recycling of BAs.
The effect of sulfation on BAs Enterohepatic Recycling: Sulfation of BAs takes place in the liver to produce BA-sulfates [taurine-BA- (T-BA-S), glycine-BA- (G-BA-S), and nonamidated BA-sulfates (BA-S)]. In the small intestine no deconjugation of the sulfate moiety takes place and limited amount of BA-sulfates are absorbed intact. Absorbed BA-sulfates are extracted intact by the liver less efficiently than unsulfated BAs. BA-sulfates in the systemic blood are efficiently excreted in urine by the kidney. The majority of BA-sulfates are not absorbed in the small intestine and are passed intact into the large intestine. BA-sulfates are desulfated, the sulfate moiety is absorbed and largely excreted in urine. BA amidates are mostly deamidated. The released sulfate (S), Glycine (G), and taurine (T) are mostly absorbed. The released unconjugated BAs are biotransformed to secondary BAs and reabsorbed or excreted in the feces as secondary BAs.
It was recognized that BA-sulfates are excreted in bile via a different pathway than unsulfated BAs. The biliary excretion of BA-glucuronide and -sulfate conjugates is reduced in the natural jaundice Groningen yellow (GY) rat and the Eisai hyperbilirubinemic (EHB) rat (naturally Mrp2 deficient rats), whereas the biliary excretion of unconjugated BAs in these rats is not affected (Kuipers et al., 1988, 1989; Takikawa et al., 1991a). In contrast to wild-type rats, membrane vesicles prepared from EHB rat livers were unable to transport BA-sulfates, whereas their ability to transport unsulfated BAs was not affected (Akita et al., 2001). In addition, dibromosulphthalein (DBSP) (Mrp2 substrate) and BA-sulfates inhibit the biliary excretion of each other, whereas no interaction in the biliary excretion of unsulfated BAs and DBSP can be demonstrated (Eng and Javitt, 1983; Uegaki et al., 1999). Using cells double transfected with Mrp2 and Bsep, it was then confirmed that Mrp2 is required to transport divalent BAs (sulfated- or glucuronidated-amidated BAs), whereas Bsep is required to transport monovalent BAs (amidated BAs) (Akita et al., 2001; Stieger et al., 2000). However, it was shown that divalent BAs have equal affinity for human BSEP and MRP2 (Hayashi et al., 2005). Therefore, in contrast to the rat, BA-sulfates in humans might be secreted into bile via both BSEP and MRP2.
Sulfation of BAs markedly decreases their absorption rate, prevents their passive absorption from the small intestine and the colon, and restricts their absorption site to the ileum (De Witt and Lack, 1980; Low-Beer et al., 1969; Walker et al., 1986). In addition, BA-sulfates are poor substrates/inhibitors for ASBT (Craddock et al., 1998). Sulfation at the 7-OH position reduces intestinal absorption more than 3-OH sulfation (De Witt and Lack, 1980; Rodrigues et al., 1995). Therefore, sulfation decreases BA absorption and enhances their fecal excretion in the rat (Palmer, 1971), guinea pig (Low-Beer et al., 1969), and in man (Cowen et al., 1975b). However one report demonstrated that intestinal absorption of BA-sulfates was as complete as unsulfated BAs in rats (Kuipers et al., 1986).
The small portion of BA-sulfates that is absorbed from small intestine is absorbed intact, and is excreted in bile without desulfation during intestinal absorption, hepatic extraction, or biliary excretion (Cowen et al., 1975a,b). However, by double labeling the steroid and the sulfate moieties, and by examining the fecal content, it was demonstrated that BA-sulfates are progressively desulfated before fecal excretion (Cowen et al., 1975b; Palmer, 1971). The released sulfate moiety was mostly recovered in urine. In addition, both rat and human feces have high desulfation activity toward BA-3-sulfates. However, BAs sulfated at the 7 and/or 12 positions are resistant to desulfation, and therefore, if exist, are excreted intact in feces (Pacini et al., 1987; Huijghebaert et al., 1984). In general, BA-sulfates are resistant to any microbial transformations until the sulfate moiety is deconjugated (Pacini et al., 1987; Huijghebaert et al., 1984). Therefore, most BA-sulfates are desulfated first, and the liberated BAs undergo bacterial transformation before being absorbed in intestine or excreted in feces. BA-sulfates are desulfated when incubated with the content of the large, but not the small, intestine indicating that, in contrast to deamidation, desulfation does not take place before the large bowel (Huijghebaert et al., 1984). Therefore, the major contribution of sulfation is to bypass absorption in the small intestine by avoiding bacterial transformation to absorbable BA species (unconjugated and amidated BAs). BA-sulfates are then deconjugated in the large intestine, where BA absorption is much less than that in the small intestine.
Similar to unsulfated BAs, BA-sulfates absorbed in the intestine are carried through the portal vein to the liver. The uptake of BA-sulfates across the sinusoidal side of hepatocytes is not well characterized. Sulfated and unsulfated BAs competitively inhibit the uptake of each other into hepatocytes, and therefore might in part share the same transport system across the sinusoidal membrane (Bartholomew and Billing, 1983). However, hepatocytes have lower uptake capacity toward sulfated BAs compared with unsulfated BAs (Bartholomew and Billing, 1983; Gartner et al., 1990). It was shown that BA-sulfates are substrates for the Oatp1 and Oatp2 localized on the sinusoidal side of hepatocytes in rats (Meng et al., 2002; Reichel et al., 1999). The vectorial transport of sulfated BAs from the sinusoidal to the canalicular side was demonstrated using OATP2/Mrp2 double transfected cells (Sasaki et al., 2002). NTCP may also contribute to BA-sulfates uptake because nor-CA-sulfate was shown to be a substrate for NTCP (Schwab et al., 1997). Along with other BAs, newly synthesized and recirculating BA-sulfates are excreted in bile to complete the enterohepatic cycle.
Similar to unsulfated BA, the major route of BA-sulfate elimination is biliary excretion. Less than 5% of an LCA-sulfate dose administered to healthy humans is excreted in urine (Allan et al., 1976; Cowen et al., 1975b). However, in cholestatic conditions, BAs including BA-sulfates are transported out of the liver into systemic blood across the sinusoidal membrane. Sulfated BAs are better substrates than unsulfated BAs for Mrp3 (Hirohashi et al., 2000) and Mrp4 (Zelcer et al., 2003), which expressions are upregulated during cholestasis (Denk et al., 2004; Mennone et al., 2006). Using Mrp4 and Mrp3 null mice, it was demonstrated that Mrp4 rather than Mrp3 may play an important role in the efflux of BA-sulfates into systemic circulation across the sinusoidal side of hepatocytes (Belinsky et al., 2005; Mennone et al., 2006; Sakamoto et al., 2006; Zelcer et al., 2006). Other transporters known to contribute to the BA spill over into systemic blood during cholestasis such as Ostα, and β, Mrp1, and Mrp6 may also contribute to the efflux of BA-sulfates (Ballatori et al., 2005; Trauner et al., 2005)
With complete cholestasis, urinary excretion becomes the primary route of BA and BA-sulfate excretion. Urinary clearance of BA-sulfates in cholestasis, however, is 10- to 100-fold higher than that of the unsulfated BAs (Corbett et al., 1981; Makino et al., 1975; Stiehl, 1974a; Stiehl et al., 1975). Therefore, serum levels of BA-sulfates remain relatively low, whereas urinary levels become very high. The high urinary clearance of BA-sulfates may be due to low extraction by the liver, inhibition of renal reabsorption by the unsulfated BAs (Barnes et al., 1977, 1979a; Summerfield et al., 1977), BA sulfation by the kidney (Summerfield et al., 1977), low affinity to ASBT (St-Pierre et al., 2001), and/or active tubular secretion of BA-sulfates (Corbett et al., 1981).
In summary, sulfation influences the enterohepatic recirculation of BAs through decreasing their intestinal absorption, hepatic extraction, and renal reabsorption. This markedly limits the enterohepatic recirculation of BAs and shifts it toward fecal and renal excretion.
IDENTIFICATION AND CHARACTERIZATION OF BA-SULTs
In humans BA-sulfotransferase activity was mainly detected in liver (Chen et al., 1978b; Kirkpatrick et al., 1988; Loof and Wengle, 1978, 1979), small intestine (Dew et al., 1980; Loof and Wengle, 1979), and adrenal gland (Adams and McDonald, 1981; Loof, 1981). BA-sulfotransferase activity was also characterized in other species such as hamsters (Barnes et al., 1979a), monkeys (Barnes et al., 1986), mice (Takahashi et al., 1990), rats (Chen et al., 1978a; Summerfield et al., 1976; Takahashi et al., 1990), and guinea pigs (Chen et al., 1978b).
Several attempts were made to isolate and purify the BA-SULT fractions from human livers (Chen and Segel, 1985; Falany et al., 1989; Loof and Hjerten, 1980; Radominska et al., 1990) and other species such as rats (Barnes and Spenney, 1982; Barnes et al., 1989; Chen, 1981; Chen et al., 1977; Kane et al., 1988; Ogura et al., 1994), and hamsters (Barnes and Spenney, 1982). Sulfation of hydroxysteroids was well characterized before the discovery of BA sulfation. Early on, it was thought that BA and hydroxysteroid sulfation were performed by distinct SULT isoforms. Evidence was then provided that, in humans, the BA-SULT is the same enzyme as the hydroxysteroid-SULT, which had been known for its high affinity for dehydroepiandrosterone (DHEA) (Radominska et al., 1990). BA/DHEA-SULT was then cloned from human liver (Comer et al., 1993; Kong et al., 1992; Otterness et al., 1992) and is currently known as SULT2A1. The crystal structure of human SULT2A1 has also been determined (Allali-Hassani et al., 2007; Lindsay et al., 2008). In contrast to the presence of a single SULT2A enzyme in humans, at least 2 isoforms of Sult2a have been cloned from mouse; mSult2a1 (Kong et al., 1993) and mSult2a2 (Kong and Fei, 1994), and three isoforms have been cloned from rats; rSult2a1 (Ogura et al., 1989), rSult2a2 (Ogura et al., 1990), and rSult2a3 (Watabe et al., 1994). The various isoforms of Sult2a in rats appear to have similar substrate specificity (Ogura et al., 1994). The SULT2 family (hydroxysteroid SULTS) includes another subfamily of enzymes known as SULT2B, which does not seem to have activity toward BA sulfation (Sakakibara et al., 1998; Shimada et al., 2002).
SULT2A1 has broad substrate specificity for endogenous hydroxysteroid androgens, estrogens, and glucocorticoids (Chatterjee et al., 1994; Falany et al., 1995; Ogura et al., 1994). SULT2A1 substrates also include BAs (Ogura et al., 1994), thyroid hormones (Li and Anderson, 1999), and xenobiotics such as tamoxifen (Shibutani et al., 1998), budesonide (Meloche et al., 2002), and bisphenols (Nishiyama et al., 2002; Pai et al., 2002). Among all substrates, DHEA and epiandrosterone have the highest affinity for Sult2a1 (Ogura et al., 1994). Despite its role in BA and androgen inactivation, SULT2A1 was demonstrated to catalyze the formation of DNA- and protein-reactive metabolites of polycyclic aromatic hydrocarbons and benzylic alcohols (Glatt et al., 1995; Falany et al., 1995).
In humans, SULT2A1 mRNA is abundantly expressed in the liver, intestine, and adrenal gland (Dooley et al., 2000; Falany, 1997; Her et al., 1996; Javitt et al., 2001; Otterness et al., 1995). Some of these studies demonstrated that Sult2a1 was not expressed in kidney, colon, and stomach (Dooley et al., 2000; Otterness and Weinshilboum, 1994). However, others demonstrated high expression of SULT2A1 mRNA in these organs (Dooley et al., 2000; Javitt et al., 2001). Other organs such as, the brain and ovary, also express SULT2A1 mRNA at moderate levels (Dooley et al., 2000; Javitt et al., 2001; Shimada et al., 2001). SULT2A1 Protein and enzyme-activity data also confirmed the presence of SULT2A1 in the human liver, stomach, and small intestine but not in the colon (Chen et al., 2003). In human fetuses, SULT2A1 protein was detected by immunohistochemistry in the adrenal cortex, liver, testis, and intestine, weakly detected in kidney, and was not detected in lung, brain, heart, stomach, skeletal muscle, pancreas, or spleen (Parker et al., 1994). The fetal adrenal gland produces large amounts of DHEA-sulfate to support placental estrogen biosynthesis. This is reflected by the high levels of SULT2A1 expression in the fetal adrenal gland (Coughtrie, 2002).
Based on the plethora of data available on the tissue distribution of SULT2A1 mRNA, protein, and enzyme activity, it can be concluded that the liver is the major site of BA sulfation. However, these same data also suggest that BAs can be sulfated in other organs such as the intestine.
SULFATION AS A BA DETOXIFICATION PATHWAY IN VARIOUS SPECIES
BAs are well known for their role in fat absorption. BAs form mixed micelles with phospholipids creating an emulsion that aids in the intestinal absorption of fat and fat-soluble vitamins (Hofmann and Hagey, 2008). Recently, BAs have been recognized as signaling molecules and hormones. BAs have been shown to regulate the expression of several target genes, which play important roles in apoptosis, proliferation (Jones et al., 1997; Sodeman et al., 2000), liver regeneration (Monte et al., 2002), as well as cholesterol, triglyceride (Kalaany and Mangelsdorf, 2006), energy (Watanabe et al., 2006), glucose (Claudel et al., 2005), and their own homeostasis (Staudinger and Lichti, 2008). In addition, BAs can be used as biomarkers for hepatobiliary diseases in infants (Mills et al., 1998; Mushtaq et al., 1999), and to monitor the functions of organs involved in their synthesis and elimination, such as liver and kidney (Alme et al., 1977). BAs have also been shown to play a role in the prevention of obesity and resistance to insulin, which can be a novel target to improve and diagnose metabolic disorders (Libert et al., 1991; Tsukada et al., 1994). Furthermore, UDCA is used therapeutically in patients with gallstone and primary biliary cirrhosis (Poupon et al., 1997). The physiologic roles of BAs have been recently reviewed (Hofmann and Hagey, 2008).
In contrast to their physiologic functions, BAs are also cytotoxic. The pathologic effects of BAs are believed to be due to their detergent properties, which enable them to bind to and solubilize membrane lipids (Plaa and Priestly, 1976). BAs cause plasma and mitochondrial membrane disruption (Garner et al., 1991; Lee and Whitehouse, 1965; Weissmann and Keiser, 1965; Sanchez Pozzi et al., 2003). BAs exert a cathartic effect by altering water and salt transport in the colon (Mekhjian and Phillips, 1970). LCA and conjugates are potent pyrogens in humans (Dillard and Bodel, 1970). BAs are also genotoxic, tumor promoters (Fukase et al., 2008; Rosignoli et al., 2008), and are believed to be involved in colon cancer (Narisawa et al., 1974; Reddy et al., 1976, 1977). The common factor between the various toxicities caused by BAs is that secondary BAs, LCA, and to a lesser extent DCA are the most toxic, whereas the primary BAs, CA, and CDCA, are not as toxic.
The pathologic effects on the liver were early recognized as a major toxicity of BAs (Holsti, 1956). It was then demonstrated that LCA is the most hepatotoxic component of bile (Palmer, 1972). A variety of pathologic-hepatobiliary changes induced by LCA, including bile duct infarction, liver fibrosis, and cirrhosis, were demonstrated in several species (Fickert et al., 2006; Hofmann, 1999b; Palmer, 1972). A major manifestation of the hepatobiliary toxicity of BAs is cholestasis. Cholestasis is the decrease in bile flow and biliary BA excretion, which ultimately can lead to a complete cessation of bile flow (Plaa and Priestly, 1976). All BAs have cholestatic effects at high doses. However, the cholestatic activity of BAs is proportional to their detergent properties, and is inversely proportional to the number of hydroxyl groups on the steroidal backbone, with LCA (mono-hydroxy BA) being the most potent cholestatic agent (Baumgartner et al., 1992; Drew and Priestly, 1979; Yousef et al., 1987b). LCA was demonstrated to induce cholestasis in several species including rats (Fisher et al., 1971; Javitt, 1966; Kakis and Yousef, 1978), hamsters (Schaffner and Javitt, 1966), and mice (Fickert et al., 2006).
The high cholestatic activity of LCA is in part attributed to its water insolubility. BAs with low solubility may form insoluble BA salts, such as Ca+2 salts, which precipitate in the canaliculi (Fickert et al., 2006; Yousef et al., 1997). Another mechanism of liver toxicity is that BAs increase the biliary secretion of phospholipids until the hepatic phospholipid pool is exhausted. Then, due to their detergent activity, BAs start to solubilize membrane phospholipids, which disturbs the integrity of the canalicular membranes (Kakis and Yousef, 1978; Miyai et al., 1977; Small and Admirand, 1969; Yousef et al., 1987a). Several molecular mechanisms underlying the cholestatic effect of LCA have also been suggested (Beuers et al., 2003). Sulfate-conjugation adds a permanent negative charge to BAs increasing their water solubility and therefore, enhance their elimination (Oelberg et al., 1984). In addition, sulfation increases the BA-micellar critical concentration, prevents cholesterol and phospholipid solubilization, and therefore, abolishes the BA-detergent activity against membranes (Donovan et al., 1993). Therefore, only based on the physicochemical properties of BA-sulfates, sulfation is expected to decrease BA toxicity.
In deed, Sulfation abolishes the cholestatic activity of CDCA, CA, and DCA. In fact, the sulfate metabolites produce a choleretic effect (increase bile flow) in rats (Eng and Javitt, 1983; Yousef et al., 1987b). At the cellular level, sulfation decreases or abolishes LCA cytotoxicity against hepatocytes (Takikawa et al., 1991b) and vascular endothelial cells (Garner et al., 1991). Also sulfation reduces phospholipid biliary excretion, and prevents solubilization of the cellular membrane phospholipids (Yousef et al., 1987b, 1992). Furthermore, sulfation decreases the carcinogenicity of LCA (Takahashi et al., 1993) and prevents the cathartic effect of CDCA and DCA on the colon (Breuer et al., 1983). In contrast to sulfation, G- or T-amidation does not decrease the cholestatic and cytotoxic effects of BAs (Dawson and Isselbacher, 1960; Javitt and Emerman, 1968).
Sulfation of LCA and T-LCA markedly decreases, but does not prevent their cholestatic activity (Fisher et al., 1971; Yousef et al., 1981, 1987b). In contrast, G-LCA-sulfate is as cholestatic as G-LCA in rats (Fisher et al., 1971; Yousef et al., 1981). It was suggested that the detoxification effect of sulfation on BAs may vary in different species depending on the prevalence of T- or G-amidation (Yousef et al., 1981). Therefore, LCA-sulfate induces cholestasis in guinea pigs, which primarily conjugate BA with G (Yousef et al., 1981). Feeding guinea pigs with a T-enriched diet prevents the cholestatic effect of G-LCA-sulfate by inducing its conversion to T-LCA-sulfate (Dorvil et al., 1983). T-amidation alone does not prevent the cholestatic effect of LCA, but T-amidation and sulfation combined does (Javitt and Emerman, 1968). Sulfation of LCA and T-LCA, but not G-LCA, was also shown to prevent the pathologic and morphologic effects that LCA inflicts on the rat liver (Leuschner et al., 1977; Yousef et al., 1981). The difference in the toxicity of G-LCA- and T-LCA-sulfates may be due to the fact that although sulfation increases LCA and T-LCA solubility, G-LCA-sulfate is not any more soluble than G-LCA (Carey et al., 1979). The fact that G-LCA-sulfate is the major end-product metabolite of LCA in humans does not mean that sulfation cannot protect against G-LCA toxicity in humans. Even if G-LCA-sulfate is as cholestatic/hepatotoxic as G-LCA, sulfation is still an effective pathway for BA detoxification via enhancing their fecal and renal excretion.
The most toxic BA, LCA, is a minor component (< 2%) of human bile and under normal conditions, does not exist in the enterohepatic system at levels high enough to produce any toxicity. However, LCA toxicity becomes relevant under conditions of high exposure, such as in cholestasis and during CDCA or UDCA therapy. CDCA and UDCA decrease the cholesterol saturation of bile, and therefore, are used clinically for gallstone dissolution (Danzinger et al., 1973). LCA, formed by intestinal dehydroxlation, is the major metabolite of these BAs. Therefore, CDCA produces severe hepatic toxicity, as a result of the accumulation of the LCA metabolite in rhesus monkeys (Dyrszka et al., 1976; Gadacz et al., 1976), baboons (Morrissey et al., 1975), and rabbits (Fischer et al., 1974). In humans, however, CDCA therapy is relatively safe, and is not associated with hepatic injury (Schoenfield and Lachin, 1981). The difference in CDCA toxicity among various species is due to the ability of humans to sulfate the resulting LCA metabolite, which enhances LCA fecal excretion, and prevents its accumulation in the enterohepatic system. Whereas, the other species lack such efficient-sulfating capabilities (Dew et al., 1982; Gadacz et al., 1976; Hofmann, 2004; Stellaard et al., 1985).
Sulfation is a minor pathway of BA metabolism in rats and mice. In rat primary hepatocytes, very low (Kirkpatrick and Belsaas, 1985) or no (Lambiotte and Thierry, 1980) BA-sulfating activity was detected. Furthermore, sulfation of LCA by rat liver was more than a 100-fold less than that of humans (Kirkpatrick et al., 1988). In addition, only small amounts to none of LCA (Palmer, 1971), CDCA, or CA (Cleland et al., 1984; Takita et al., 1988) administered to rats were sulfated. Also, levels of endogenous BA-sulfates in rat urine were undetectable or constituted less than 0.001% of total BAs. In bile duct-ligated (BDL) rats and mice, the undetectable or low levels of urinary BA-sulfates remained unchanged (Lee et al., 2001) or increased to 0.2–2.7% of the total BAs in urine (Kinugasa et al., 1981; Marschall et al., 2006; Takita et al., 1988). Instead, hydroxylation at the 6β position represents the major pathway of BA detoxification in rats, which is induced in cholestatic conditions to enhance the urinary excretion of accumulated BAs (Danielsson, 1973; Greim et al., 1972b; Kinugasa et al., 1981; Takita et al., 1988). Similar to sulfation in humans, hydroxylation also protects against BA toxicity in rodents (Bagheri et al., 1978; Hunt et al., 1964; Takikawa et al., 1997a). Other reports however, demonstrated the capability of rodents to sulfate BAs to a significant extent under cholestatic conditions (Leuschner et al., 1977). Furthermore, later reports demonstrated that 5% of serum and 40% of urinary BAs are present in the sulfate form in rats (Purucker et al., 2001).
Rats do not only have lower BA sulfation capabilities compared with humans, but they are also less capable of eliminating BA-sulfates. Despite the obstructed biliary pathway in BDL rats, urinary excretion of CDCA-, T-CDCA-, LCA-, and T-LCA- sulfates was less than 5% of the dose administered (Cleland et al., 1984; Little et al., 1991; Takada et al., 2003). Instead, most of the dose was retained in plasma and peripheral tissues such as skin and muscle (Little et al., 1991). The low urinary excretion of BA-sulfates in rats might be due to the fact that sulfation prevents hydroxylation of BAs. Unsulfated BAs are more efficiently hydroxylated and excreted in urine than sulfated BAs.
BA-SULFATES IN HUMANS IN HEALTH AND DISEASE
Data on the level of BA-sulfates vary markedly, in large part, because of the variety of analytical techniques used to quantify them. Table 1 summarizes the reported proportions of total and individual BAs present in the sulfate form in human serum, bile, and urine. The amount of total BAs excreted in urine is very low (< 1μM/day) under normal conditions. In various occasions, it was reported that 40–70% of BAs in urine are sulfated, Table 1. The concentration of total BAs in serum of a healthy man is also very low (< 3μM), with less than 15% present in the sulfated form, Table 1. Wide inter-individual variations also exist in the levels of BA-sulfates in serum, where proportions as high as 64% were reported in some healthy individuals. Very low proportions of BAs in bile are sulfated, with available data ranging from undetectable up to 4%. From Table 1, it can also be seen that most BAs in urine, small proportions of serum BAs, and negligible proportions of biliary BAs are sulfated. Furthermore, a common finding of these studies is that the proportion of individual BAs that are sulfated decreases with the increase in the number of hydroxyl groups on the BA. Therefore, the LCA (mono-hydroxy BA) is almost exclusively present in the sulfated form, whereas CA (tri-hydroxy BA) is mostly present in the unsulfated form. Structural factors, in addition to the number of hydroxyl groups, also play a role in determining the extent of sulfation. The percentage of DCA present in the sulfated form is generally higher than that of CDCA, despite the fact that they are both dihydroxy BAs. Despite the high proportion of LCA present in the sulfate form, the proportion of total BA-sulfates remains very low in serum and bile, because LCA, in all of its forms, does not constitute more than 2% of total BAs (Perwaiz et al., 2001).
The Percentage of Individual and Total BAs Present in the Sulfate Form
In hepatobiliary/cholestatic diseases, due to the impairment of biliary excretion, urinary excretion of BAs increases more than 100 fold. Similar to the situation under normal conditions, a large proportion of BAs (25–80%) in urine is excreted in the sulfated form under cholestatic conditions, Table 1. The amount of BA-sulfates excreted in urine was suggested to be used as a specific biomarker for the diagnosis of intrahepatic cholestasis in pregnant women (Huang et al., 2007). Under the same pathologic conditions, serum BA levels also increase however, the concentration of sulfated BAs does not increase in parallel with the increase in unsulfated BAs in the serum. This indicates the efficient renal excretion of BA-sulfates, where BA-sulfates are more efficiently cleared from serum into to urine than unsulfated BAs. Therefore, the percentage of sulfated BAs in serum either remains unchanged, or even decrease, with hepatobiliary diseases in humans. Other studies however, have indicated that the proportion of sulfated BAs in serum in hepatobiliary diseases increases up to 40% of total BAs. In cholestatatic patients, sulfated BAs in bile remains low, and ranges from undetectable up to 4% of total BAs. The Percentage of BA-sulfates in the human liver was estimated to be less than 10%, which further decreases under cholestatic conditions (Fischer et al., 1996; Greim et al., 1972a).
Due to the impairment of the biliary excretion route under cholestatic conditions, the amount of BAs including BA-sulfates increases in serum, tissues, and urine. However, the ratio of BA-sulfates:total BAs remains constant or may decrease. This may be ascribed to the impairment of the liver capability to sulfate BAs as a result of the deteriorated liver function under these pathologic conditions (Fischer et al., 1996; Kato et al., 1996; Murata et al., 1983; Takikawa et al., 1983). It was thought that the increase in the amount of BA-sulfates excreted in urine might be due to a compensatory increase in BA sulfation as a result of the upregulation of the hepatic SULT2A activity/expression during cholestatic conditions. However, it was then demonstrated that hepatic SULT activity toward BAs and other substrates (Iqbal et al., 1990; Loof and Nyberg, 1983; Loof and Wengle, 1982), and SULT2A1 protein expression (Elekima et al., 2000) are actually reduced during cholestatic diseases in humans. Others, however, reported that SULT2A1 mRNA expression was not altered in humans in cholestatic diseases (Zollner et al., 2007). In animal models, hepatic BA-SULT activity and the percentage of BA-sulfates were shown to be reduced in BDL-hamsters (Barnes et al., 1979b) and in BDL-pregnant rats (Chen et al., 1982). Furthermore, cholestasis induced by LCA administration did not induce BA-Sult activity in rats (Balistreri et al., 1984). Also both LCA and BDL suppressed Sult2a1 expression in female mice (Uppal et al., 2007).
Therefore, it can be concluded that the increase in the formation and urinary excretion of BA-sulfates during cholestatic diseases is unlikely to be due to enhanced SULT2A1 activity or expression, but rather, due to the increase in the substrate (BAs) availability for sulfation (Barnes et al., 1979b; Galle et al., 1989; Kirkpatrick and Belsaas, 1985). It is also possible that sulfation of BAs is enhanced as an adaptive/compensatory mechanism at early stages of cholestasis but, with additional damage at advanced stages, the liver capability to sulfate BAs is decreased (Galeazzi and Javitt, 1977; Stiehl et al., 1978a).
GENDER DIFFERENCES IN BA SULFATION
Gender differences were observed in relation to BA sulfation in several animal species. In germfree rats, the percentage of BA-sulfates was 10-fold higher in female cecal content and feces compared with males (Eyssen et al., 1977; Parmentier et al., 1981). In female rats, the percentage of BA-sulfates in bile is about 1%, whereas less than 0.2% of biliary BAs are sulfated in males (Eriksson et al., 1978).
At the enzyme-activity level, livers from female rats have threefold higher Sult-activity toward T-LCA than that in males (Hammerman et al., 1978). In hamsters, female livers have a four-fold higher Sult activity toward G-CDCA than that in males, this higher enzymatic activity however is not reflected on the percentage of BA-sulfates in bile, urine, or blood (Barnes et al., 1979a, b). In mice, a 70-fold higher Sult activity toward DHEA was demonstrated in livers from females compared with that in males (Borthwick et al., 1995a). The gender-specific Sult2a1 activity and expression were shown to be due to stimulatory effects by female estrogens, and suppressive effects by male androgens and male-pattern growth hormone (Borthwick et al., 1995b; Carlstedt-Duke and Gustafsson, 1973; Collins et al., 1987; Hammerman et al., 1978; Kane et al., 1984; Kirkpatrick et al., 1985; Labrie et al., 1994; Liu and Klaassen, 1996; Torday et al., 1971; Yamazoe et al., 1987). The sex differences start at puberty, where Sult2a1 activity begins to decline while female's stays constant or does not decrease as fast (Balistreri et al., 1984; Chen et al., 1982; Kane et al., 1988). However, the gender differences in Sult-activity toward T-LCA was liver specific, and there were no gender differences in kidney activity (Hammerman et al., 1978).
This gender difference also exists at the Sult mRNA and protein levels. In rats, Sult2a1 mRNA levels are at least 6-fold higher in female than male livers (Chatterjee et al., 1987; Dunn and Klaassen, 1998; Runge-Morris and Wilusz, 1991). Also, a 15-fold higher Sult2a1 protein expression was reported in livers from female rats than that in males (Chen et al., 1995; Homma et al., 1992). In mice, the mRNA expression in female livers is much higher than that in males (Alnouti and Klaassen, 2006, 2008; Saini et al., 2004; Wu et al., 2001).
Sult2a1 expression is also age-dependent. Sult2a1 mRNA levels are very low in livers obtained from infant rats before and right after birth. The expression starts increasing until puberty, where male expression declines in response to rising androgen levels, whereas female expression remains high in rats, mice and hamsters (Alnouti and Klaassen, 2006; Balistreri et al., 1984; Chen et al., 1982; Demyan et al., 1992; Kane et al., 1988; Song et al., 1991). This establishes the gender difference in Sult2a1 expression throughout the animal life (Chatterjee and Roy, 1990; Echchgadda et al., 2004a; Homma et al., 1992; Runge-Morris and Wilusz, 1991; Song et al., 1998). Therefore, Sult2a1 is considered a senescence enzyme, the expression of which is regulated by androgens via the androgen receptor (Demyan et al., 1992; Song et al., 1990, 1998).
The gender differences in BA-sult activity, sult2a expression, and levels of BA-s reported in animals, may not apply to humans. It was demonstrated that no age or gender differences in hepatic SULT activity toward DHEA, SULT2A1 protein expression in liver (Aksoy et al., 1993; Her et al., 1996), or in small intestine exist in humans (Chen et al., 2003). More importantly, there were no gender differences in BA-sulfate levels in human plasma (Setchell and Matsui, 1983) or bile (Rossi et al., 1987). In addition, there were no gender differences in the levels of LCA-sulfate after an exogenous LCA dose administered to humans (Allan et al., 1976).
REGULATION OF SULT2A EXPRESSION
The regulation of SULT2A expression has been extensively studied as a part of the efforts to understand BA homeostasis. Molecular mechanisms of the regulation of BA homeostasis have been reviewed on several occasions (Chiang, 1998; Makishima, 2005; Trauner et al., 2005; Trottier et al., 2006; Zollner et al., 2006). Maintenance of BA homeostasis is essential because these molecules are toxic when they accumulate in the liver during cholestatic diseases. However, the development of cholestatic liver damage is counteracted by a variety of intrinsic hepato-protective mechanisms to maintain BA homeostasis and reduce intrahepatic accumulation of BAs. BA homeostasis is thought to be tightly controlled by nuclear receptor-mediated mechanisms, including farnesoid X receptor (FXR), constitutive androstane receptor (CAR), pregnane X receptor (PXR), vitamin D receptor (VDR), peroxisome proliferator-activated receptor (PPAR-α), hepatocyte nuclear factor 4 (HNF4α), and liver X receptor (LXR). These nuclear receptors are triggered into action by BAs themselves to regulate the expression of target genes. This nuclear receptor-mediated maintenance of homeostasis is the result of coordinated feedback and feed-forward regulation of genes involved in BA synthesis, metabolism, and transport. Therefore, under conditions of BA loading (cholestasis, bile duct ligation, BA feeding), elevated BA concentrations activate nuclear receptors, which up- and/or downregulate these target genes to eventually bring BAs back to normal levels.
BA sulfation is suggested to be a primary target for regulation under cholestatic conditions because of its capability to detoxify and enhance BA elimination. The marked increase in BA-sulfates excreted in urine during cholestatic diseases in humans suggested that BA sulfation is induced as a part of the adaptive changes to alleviate BA accumulation.
The nuclear receptor, FXR, was suggested to play an important role in BA homeostasis through the regulation of Sult2a1 expression. In vitro, CDCA activates the mouse Sult2a1 promoter via FXR. In contrast to other FXR genes, which contain an IR-1 (inverted repeats with a single nucleotide spacer), Sult2a1 response element to FXR activation was identified as an IR-0 (inverted repeats without a spacing) (Song et al., 2001). Female FXR-null animals are more resistant to LCA-induced liver toxicity than wild-type mice. The null mice were shown to have higher Sult2a1 expression, higher LCA hepatic Sult-activity, and higher biliary and fecal excretion of LCA-sulfate (Kitada et al., 2003; Miyata et al., 2006). Therefore, it was concluded that enhanced BA sulfation in these animals is responsible for their protection against LCA toxicity.
However, wild-type mice females also have a higher Sult2a1 expression than males, yet the wild-type females are not more protected against LCA toxicity. This indicates that the protection of FXR-null females cannot be explained solely based on enhanced BA sulfation. Actually, wild-type female mice are more susceptible to LCA-induced liver toxicity than wild-type males (Uppal et al., 2007). Also CDCA was shown to be more cholestatic in female than in male rats due to the less efficient hydroxylation of CDCA to β-MCA (Fisher et al., 1972). Furthermore, the fact that Sult2a1 expression is upregulated in FXR-null mice is contradictory with the reports suggesting the positive regulation of Sult2a1 by FXR.
The role of PXR in the regulation of Sult2a1 and BA sulfation was also studied, and it was found that Sult2a1 expression is downregulated in PXR-null mice (Kitada et al., 2003). The PXR ligand, pregnenolone-16α-carbonitrile (PCN), was shown to induce the expression of Sult2a1 in wild-type but not in PXR-null mice. PCN was also shown to induce the expression of hepatic Sult2a1 in mice via the activation of PXR (Echchgadda et al., 2004a). An activated form of human and mouse PXR also binds to and activates the mouse Sult2a1 promoter in vitro and in vivo, and the same IR-0 response element for FXR was identified as a response element for PXR (Echchgadda et al., 2004a; Sonoda et al., 2002). Transgenic mice expressing the hyperactive PXR were more protected against LCA-induced liver toxicity, indicating that the sustained activation of PXR is sufficient to protect against BA-induced hepatotoxicity. The protection was suggested to be due to Cyp3a11 induction, which in turns induces BA hydroxylation (Staudinger et al., 2001; Xie et al., 2001). Both PXR and glucocorticoid-receptor (GR) were also demonstrated to mediate glucocorticoid induction of the rat Sult2a1 promoter in vitro (Runge-Morris et al., 1999). PCN and dexamethasone, both PXR ligands, were shown to induce human SULT2A1 mRNA and protein expression in primary hepatocytes (Duanmu et al., 2002). PCN was also shown to protect against LCA-induced hepatotoxicity by inducing Sult2a expression and therefore increasing biliary and fecal excretion of LCA-sulfates (Miyata et al., 2006). However, the PXR ligand, rifampicin, induced the expression of SULT2A1 in 12 out of 23 human primary hepatocytes cultures, whereas it either suppressed or had no effect on SULT2A1 expression in the remaining samples. It was concluded that PXR actually inhibits the expression of SULT2A1, and that the stimulatory effect of rifampicin observed in some samples is PXR-independent, which contradicts with the suggested positive regulation of Sult2a1 by PXR (Fang et al., 2007).
The role of CAR in BA sulfation as a mechanism for the maintenance of BA homeostasis was demonstrated at several levels. The CAR ligands, 1,4-bis[2-(3,5-dichloropyridyloxy)]benzene (TCPOBOB) and phenobarbital (PB), induced Sult2a1 expression in wild-type but not in CAR-null mice (Assem et al., 2004; Maglich et al., 2004; Zhang et al., 2004). Sult2a1 expression and activity were induced in mice transfected with an activated form of mouse CAR (VP-CAR). This induction disappeared with CAR inhibition. These transgenic mice were more protected against LCA-induced liver toxicity, indicating that the sustained activation of CAR is sufficient to protect against BA-induced hepatotoxicity via Sult2a1 induction (Saini et al., 2004). CAR was also shown to bind to and activate the mouse and rat Sult2a1 promoter, and the same IR-0 was identified as a shared response element for FXR, PXR, and CAR (Assem et al., 2004; Saini et al., 2004). TCPOBOB potentiated the activation of mouse and rat Sult2a1 promoter by CAR, whereas androstenol (CAR antagonist) inhibited the CAR effect (Saini et al., 2004). It was further suggested that the expression of both Sult2a1 and Mrp4 is simultaneously induced by CAR under cholestatic conditions to facilitate the formation and excretion of BA-sulfates (Assem et al., 2004). Double CAR-PXR-null female mice were shown to be more sensitive to LCA-induced liver toxicity than the male double null-mice, which was suggested to result from the higher expression of Sult2a1 in the transgenic females compared with males (Uppal et al., 2005). Therefore, it was concluded that both PXR and CAR combined play an important role in the protection against BA-induced hepatotoxicity via the activation of several genes involved in BA homeostasis including Sult2a1. In addition to its role in BA homeostasis, CAR regulation of Sult2a1 was shown to be involved in energy homeostasis. HNF4α was shown to induce expression of CAR during fasting and caloric restriction, which in turn induced the expression of CAR target genes such as, Sult2a1 (Ding et al., 2006). Sult2a1 catalyzes the sulfation of tetraiodothyronine (T4), which prevents its conversion to the more active triiodothyronine (T3) (Li and Anderson, 1999; Mol and Visser, 1985). The reduction in T3 levels reduces the basal metabolic rate and therefore, prevents weight loss during fasting (Maglich et al., 2004). It was shown later that HNF4α synergizes CAR and PXR regulation of Sult2a1, but it was not essential for CAR or PXR responses (Echchgadda et al., 2007).
However, it was reported that the CAR ligands, PB, TCPOBOB, and/or diallyl sulfide do not induce any of the expression of Sult2a1 in mice (Alnouti and Klaassen, 2008; Ueda et al., 2002). In addition, PB did not affect hepatic expression of Sult2a1 in rats (Hellriegel et al., 1996; Rushmore and Kong, 2002).
Other nuclear receptors suggested to be involved in Sult2a1 regulation are VDR, LXR, and PPARα. The VDR ligand, vitamin D, binds to and activates the human, rat, and mouse SULT2A1/Sult2a1 promoter and induces the mRNA and protein expression in vitro (Chatterjee et al., 2005; Echchgadda et al., 2004b; Song et al., 2006). In addition the same IR-0 was identified in the Sult2a1 promoter as a shared response element for FXR, PXR, CAR, and VDR (Echchgadda et al., 2004b). Female mice expressing a hyperactive form of LXR were shown to be more resistant to LCA- and BDL-liver induced toxicity than both the transgenic males and wild-type females. This protection was suggested to be due to the enhanced Sult2a1 expression. However, basal expression of Sult2a1 is also higher in wild-type female compared with male mice yet, wild-type female mice are not more protected against LCA toxicity than wild-type males. In addition the same IR-0 was identified in the Sult2a1 promoter as a shared response element for FXR, PXR, CAR, VDR, and LXR (Uppal et al., 2007). The PPAR-α ligand, ciprofibrate, was shown to induce Sult2a1 mRNA expression in primary human hepatocyes. PPARα binds to and activates the human SULT2A1 promoter but not the rat promoter. A DR-1 (direct repeat with 1 nucleotide spacer) was identified as the peroxisome proliferator response element in human SULT2A1 (Fang et al., 2005).
In conclusion, several lines of evidence have been provided to support the role of various nuclear receptors in the maintenance of BA homeostasis via the regulation of Sult2a1 expression. However, little functional data are provided to support the idea that regulation of Sult2a1 expression is translated into significant changes in the concentration of BA-sulfates, in the hepatic or extrahepatic tissues and fluids, as an adaptive mechanism to decrease the toxicity of the accumulated BAs during cholestasis. Most of these studies were performed using the mouse or rat model. As mentioned earlier, most studies in which BA-sulfates were measured in rats and mice, reported undetectable or very low levels of BA-sulfates in urine, which remained unchanged or slightly increased under cholestatic conditions (Kinugasa et al., 1981; Lee et al., 2001; Marschall et al., 2006; Takita et al., 1988; Purucker et al., 2001).
SUMMARY
BA sulfation is catalyzed by SULT2A1. Sulfation alters the physicochemical properties of BAs, which markedly alters their pharmacologic and toxicologic profiles. Sulfation enhances the fecal and urinary excretion of BAs, and decreases many of their toxic effects. Therefore, sulfation may be an effective pathway for BA detoxification, especially in hepatobiliary diseases. In humans, most BAs excreted in urine are present in the sulfate form, whereas small proportions of BAs in serum and bile are sulfated. This indicates the efficient urinary excretion of BA-sulfates. LCA is almost completely present in the sulfate form, whereas CA is mostly unsulfated under both healthy and pathologic conditions. In cholestatic diseases, the formation of BA-sulfates increases due to the accumulation of unsulfated BAs available for sulfation in the liver, and not due to the induction of SULT2A1 expression or function. There are major species differences in BA sulfation. Sulfation is a major pathway of BA metabolism in humans, whereas hydroxylation is the major pathway of BA elimination in other species, such as the rat. Sult2a1 expression and BA sulfation may be gender-specific in other species such as rats and hamsters, but not in humans. Several lines of evidence exist regarding the role of various nuclear receptors in the regulation of SULT2A1 expression under cholestatic conditions. However, there is a lack of functional data to support the role of nuclear receptors in the regulation of sulfation and BA homeostasis as a defense mechanism against BA accumulation in cholestatic diseases. Future studies may uncover new treatments of cholestatic diseases using novel targets to enhance BA sulfation.
- Department of Pharmaceutical Sciences, College of Pharmacy, University of Nebraska Medical Center, Omaha, Nebraska 68198
- 1For correspondence via fax: Fax: (402) 559-9543. E-mail: yalnouti{at}unmc.edu.
- Received October 14, 2008.
- Accepted December 24, 2008.
Acknowledgments
We would like to thank Drs Alan F. Hofmann (University of California-San Diego), Edward B. Roche, and Dennis H. Robinson (University of Nebraska Medical Center) for their invaluable suggestions and critical reading of this article.
- © The Author 2009. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please email: journals.permissions@oxfordjournals.org
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
