Abstract
3α-Hydroxy-5β-cholan-24-oic (lithocholic) acid is a relatively minor component of hepatic bile acids in humans but is highly cytotoxic. Hepatic microsomal oxidation offers a potential mechanism for effective detoxification and elimination of bile acids. The aim of the present study was to investigate the biotransformation of lithocholic acid by human hepatic microsomes and to assess the contribution of cytochrome P450 (P450) enzymes in human hepatic microsomes using human recombinant P450 enzymes and chemical inhibitors. Metabolites were identified, and metabolite formation was quantified using a liquid chromatography/mass spectrometry-based assay. Incubation of lithocholic acid with human liver microsomes resulted in the formation of five metabolites, which are listed in order of their rates of formation: 3-oxo-5β-cholan-24-oic (3-ketocholanoic) acid, 3α,6α-dihydroxy-5β-cholan-24-oic (hyodeoxycholic) acid, 3α,7β-dihydroxy-5β-cholan-24-oic (ursodeoxycholic) acid, 3α,6β-dihydroxy-5β-cholan-24-oic (murideoxycholic) acid, and 3α-hydroxy-6-oxo-5β-cholan-24-oic (6-ketolithocholic) acid. 3-Ketocholanoic acid was the major metabolite, exhibiting apparent Km and Vmax values of 22 μM and 336 pmol/min/mg protein, respectively. Incubation of lithocholic acid with a of human recombinant P450 enzymes revealed that all five metabolites were formed by recombinant CYP3A4. Chemical inhibition studies with human liver microsomes and recombinant P450 enzymes confirmed that CYP3A4 was the predominant enzyme involved in hepatic microsomal biotransformation of lithocholic acid. In summary, the results indicate that oxidation of the third carbon of the cholestane ring is the preferred position of oxidation by P450 enzymes for lithocholic acid biotransformation in humans and suggest that formation of lithocholic acid metabolites leads to enhanced hepatic detoxification and elimination.
As major components of bile, bile acids are involved in several important physiological functions, including the excretion of excess hepatic cholesterol and phospholipid, as well as the solubilization and absorption of lipid-soluble nutrients from the diet (Hofmann, 1999, 2002). Bile acids also serve as hepatic signaling molecules through activation of nuclear receptors such as farnesoid X receptor, vitamin D receptor (VDR), pregnane X receptor (PXR), and constitutive androstane receptor, which function in the transcriptional regulation of genes involved in bile acid synthesis, transport, and metabolism (Goodwin and Kliewer, 2002; Saini et al., 2004; Matsubara et al., 2008). Although 3α-hydroxy-5β-cholan-24-oic (lithocholic) acid is a quantitatively minor component of hepatic and biliary bile acids in humans, it is highly cytotoxic (Hofmann, 1999). Treatment of hepatocytes in vitro with lithocholic acid at a concentration of 1 mM increased mitochondrial membrane permeability and caused cellular lysis (Rolo et al., 2000; Palmeira and Rolo, 2004). Administration of lithocholic acid to experimental animals produced liver and biliary tract injury, including bile duct proliferation, multifocal necrosis with vacuolization, inflammation of the portal area, and atrophy of hepatic lobules (Fickert et al., 2006; Beilke et al., 2008). These pathophysiological changes are indicative of intrahepatic cholestasis and cirrhosis (Palmer and Ruban, 1966; Miyai et al., 1971; Fischer et al., 1974). Cholestasis, which is defined as impairment of normal bile flow, is also a common manifestation of many liver diseases, including viral hepatitis, alcoholic liver disease, drug-induced toxicity, and primary biliary cirrhosis (Hofmann, 2002; Vong and Bell, 2004) and is associated with the retention of bile acids within the liver.
Oxidative biotransformation of lithocholic acid in the liver is a mechanism that can decrease lithocholic acid levels by increasing its elimination. Oxidation catalyzed by hepatic cytochrome P450 (P450) enzymes introduces additional hydroxyl groups into the lithocholic acid molecule. The resulting metabolites are less hydrophobic and contain new functional groups for glucuronide and sulfate conjugation, which facilitates their excretion. A previous study showed that human liver microsomes converted lithocholic acid to three dihydroxy metabolites, namely, 3α,6α-dihydroxy-5β-cholan-24-oic (hyodeoxycholic) acid, 3α,7α-dihydroxy-5β-cholan-24-oic (chenodeoxycholic) acid, and 3α,6β-dihydroxy-5β-cholan-24-oic (murideoxycholic) acid (Xie et al., 2001). Hyodeoxycholic acid, the 6α-hydroxylated product, was identified as the major metabolite, and its formation was shown to be catalyzed by human recombinant CYP3A4 (Araya and Wikvall, 1999; Xie et al., 2001). Subsequently, a different metabolite, 3-oxo-5β-cholan-24-oic (3-ketocholanoic) acid, was identified as the major product of lithocholic acid following incubation with human recombinant CYP3A4 (Bodin et al., 2005). Using rat liver microsomes, we showed that lithocholic acid was extensively metabolized by multiple P450 enzymes to six metabolites (Deo and Bandiera, 2008a). The predominant pathway involved 6β-hydroxylation and produced murideoxycholic acid as the major metabolite. Formation of hyodeoxycholic acid was a minor pathway (Deo and Bandiera, 2008a). Results from our recent investigation of 3α,7α,12α-trihydroxy-5β-cholan-24-oic (cholic) acid and chenodeoxycholic acid biotransformation by human liver microsomes (Deo and Bandiera, 2008b) suggest that neither 6α-hydroxylation nor 6β-hydroxylation would be the preferred P450-catalyzed oxidation pathway for lithocholic acid in human hepatic microsomes. A more thorough investigation of lithocholic acid biotransformation, including the contribution of P450 enzymes other than CYP3A4, is needed to test this hypothesis and to resolve the divergent metabolite profiles reported to date.
Thus, the aim of the present study was to investigate the biotransformation of lithocholic acid by human hepatic microsomes and to assess the contribution of individual P450 enzymes using human recombinant P450 enzymes and chemical inhibitors. Metabolites were identified, and metabolite formation was quantified using a liquid chromatography/mass spectrometry (LC/MS)-based assay. The results obtained indicate that oxidation of the third carbon of the cholestane ring is the preferred pathway for lithocholic acid biotransformation.
Materials and Methods
Chemicals and Reagents. Unconjugated lithocholic acid and unconjugated bile acid standards were purchased from Steraloids Inc. (Newport, RI). The purity of lithocholic acid was >95% as reported by Steraloids Inc. and confirmed by us. 1α,3β,7α,12α-Tetrahydroxy-5β-cholan-4-oic acid and 3α,6β,7β,12α-tetrahydroxy-5β-cholan-4-oic acid were gifts from Dr. Lee R. Hagey (University of California, San Diego, CA). Bile acid standards were dissolved in methanol as 1-mg/ml stock solutions and stored at –4°C. Additional dilutions were made in methanol for the biotransformation assay. Ketoconazole, quercetin, quinidine, sulfaphenazole, and SKF 525A were provided by Dr. T. K. H Chang (University of British Columbia, Vancouver, BC, Canada). Stock solutions of these chemical P450 inhibitors were prepared in methanol. Pooled human liver microsomes were purchased from XenoTech, LLC (Lenexa, KS). Baculovirus-insect cell control microsomes containing expressed human P450 oxidoreductase and baculovirus-insect cell microsomes containing expressed human P450 enzymes (BD Supersomes Enzymes), coexpressed with human P450 oxidoreductase or with human P450 oxidoreductase and human cytochrome b5, were purchased from BD Biosciences (Oakville, ON, Canada). Four individual human liver microsome samples from single donors (HH18, HG95, HH13, and HH837) were obtained from BD Biosciences (Woburn, MA). The four samples were selected based on their testosterone 6β-hydroxylase activity, which is an enzymatic marker for CYP3A, and include two samples with high and two samples with low testosterone 6β-hydroxylase activities. Donor information was provided by BD Biosciences and is summarized here. HH18 (African American female, 78 years of age) and HH837 (Asian female, 52 years of age) had high CYP3A4-dependent testosterone 6β-hydroxylase activity of 12,000 pmol/min/mg protein and 13,700 pmol/min/mg protein, respectively; HG95 (Hispanic female, 47 years of age) and HH13 (Asian male, 55 years of age) had low CYP3A4-catalyzed activities of 650 pmol/min/mg protein and 890 pmol/min/mg protein, respectively, as reported by BD Biosciences. High-performance liquid chromatography-grade chemicals and solvents were purchased from Fisher Scientific (Ottawa, ON, Canada).
Lithocholic Acid Biotransformation Assay. Reaction mixtures contained 50 mM potassium phosphate buffer, pH 7.4, 3 mM magnesium chloride, 0.5 mg of human hepatic microsomal protein, 1 mM NADPH, and varying concentrations (1–250 μM) of lithocholic acid in a final volume of 1 ml. After preincubation for 10 min at room temperature, reactions were initiated with NADPH and allowed to proceed for 30 min at 37°C. Reactions were terminated with 8 ml of dichloromethane/isopropanol (80:20 v/v). Internal standard (cholic-2,2,4,4-d4 acid, 0.4 μg) was then added to each sample. Sample extraction, evaporation, and reconstitution in preparation for analysis by LC/MS were carried out as described previously (Deo and Bandiera, 2008a). Reaction mixtures that were devoid of substrate, NADPH, or microsomes, as well as reaction mixtures that contained defined concentrations of authentic bile acid standards, were routinely included in each assay.
Assay conditions were tested using pooled human liver microsomes to ensure that substrate and cofactor concentrations were saturating and that product formation was linear with respect to incubation time (1–60 min) and protein concentration (0.25–2 mg/ml reaction mixture). To determine whether metabolite formation was P450-mediated, preliminary experiments were conducted with carbon monoxide-treated hepatic microsomes or heat-denatured microsomes or by replacing NADPH with NADH or by adding SKF 525A, a P450 inhibitor.
Incubations with human recombinant P450 enzymes were also carried out. Reaction mixtures contained 30 pmol of recombinant P450 enzyme (CYP1A1, CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, CYP3A5, or CYP4A11) or, in the case of insect cell control microsomes and reductase control, an equivalent amount of protein (0.15 mg).
Analytical Methods. LC/MS conditions to resolve and quantify the metabolites were used as outlined previously (Deo and Bandiera, 2008a,b). A mixture consisting of 19 bile acid standards was prepared for identification of bile acids. Under these conditions, 1α,3β,7α,12α-tetrahydroxy-5β-cholan-4-oic acid (molecular mass, 424.6) and 3α,6β,7β,12α-tetrahydroxy-5β-cholan-24-oic acid (molecular mass, 424.6) typically eluted at 4 and 5 min, respectively, and were monitored at m/z 423. 3α,6β,7α-Trihydroxy-5β-cholan-24-oic (α-muricholic) acid (molecular mass, 408.6), 3α,6β,7β-trihydroxy-5β-cholan-24-oic (β-muricholic) acid (molecular mass, 408.6), 3α,6α,7α-trihydroxy-5β-cholan-24-oic (γ-muricholic) acid (molecular mass, 408.6), and cholic acid (molecular mass, 408.6) eluted at 12, 13, 15, and 16 min, respectively, and were monitored at m/z 407. 7α,12α-Dihydroxy-3-oxo-5β-cholan-24-oic (3-dehydrocholic) acid (molecular mass, 406.6) eluted at 13 min and was monitored at m/z 405. Murideoxycholic acid (molecular mass, 392.6), 3α,7β-dihydroxy-5β-cholan-24-oic (ursodeoxycholic) acid (molecular mass, 392.6), hyodeoxycholic acid (molecular mass, 392.6), chenodeoxycholic acid (molecular mass, 392.6), and 3α,12α-dihydroxy-5β-cholan-24-oic (deoxycholic) acid (molecular mass, 392.6) eluted at 12, 13.4, 15, 19.5, and 20 min, respectively, and were monitored at m/z 391. 3α-Hydroxy-6-oxo-5β-cholan-24-oic (6-ketolithocholic) acid (molecular mass, 390.6), 3α-hydroxy-7-oxo-5β-cholan-24-oic (7-ketolithocholic) acid (molecular mass, 390.6), and 7α-hydroxy-3-oxo-5β-cholan-24-oic acid (molecular mass, 390.6) eluted at 13.4, 14.5, and 16 min, respectively, and were monitored at m/z 389. 3β-Hydroxy-5β-cholan-24-oic (isolithocholic) acid (molecular mass, 376.6) and lithocholic acid (molecular mass, 376.6) eluted at 20.5 and 22.5 min, respectively, and were monitored at m/z 375. 3-Ketocholanoic acid (molecular mass, 374.6) eluted at 21.5 min and was monitored at m/z 373. The internal standard, cholic-2,2,4,4-d4 acid, eluted at 15.4 min and was monitored at m/z 411. Metabolites were quantified from calibration plots of the peak area ratio of authentic standard and internal standard plotted against the concentration of the authentic standard.
Data Analysis and Calculation of Enzyme Kinetic Parameters. Data were analyzed using the SigmaPlot Enzyme Kinetics Module (version1.1; Systat Software, Inc., San Jose, CA). Metabolite formation as a function of substrate concentration was analyzed by nonlinear regression analysis, and apparent Km, K′, and Vmax values were generated using the Michaelis-Menten equation or the Hill equation, as described previously (Deo and Bandiera, 2008a). Several criteria such as sum of squares (R2), S.D. of residuals (Sy.X) for each equation, Akaike information criterion, and visual inspection of the fit were used to choose the appropriate fitting model (Tracy and Hummel, 2004).
Chemical Inhibition Studies. P450 chemical inhibition studies were carried out using human liver microsomes. Chemical inhibitors used were quercetin (a CYP2C8 inhibitor, at 2, 20, and 50 μM), sulfaphenazole (a CYP2C9 inhibitor, at 1, 10, and 20 μM), quinidine (a CYP2D6 inhibitor, at 0.1, 1, and 10 μM), and ketoconazole (a CYP3A4 inhibitor, at 0.01, 0.1, 1, and 10 μM) at the final concentrations indicated in parentheses. Reactions were initiated with NADPH after initial first preincubation of microsomes with lithocholic acid and the P450 inhibitor for 5 min at 37°C.
Statistical Analysis. Comparison of chemical inhibition data with the control was carried out using analysis of variance. A p value of ≤0.05 was considered statistically significant. Spearman rank correlation analysis (nonparametric) was used to determine the correlation between individual rates of lithocholic acid metabolite formation and testosterone 6β-hydroxylase activities of the four human liver microsome samples.
Results
Lithocholic Acid Biotransformation by Human Liver Microsomes. Incubation of lithocholic acid with human liver microsomes yielded five metabolites identified as 3-ketocholanoic acid, hyodeoxycholic acid, ursodeoxycholic acid, murideoxycholic acid, and 6-ketolithocholic acid. The major metabolite was 3-ketocholanoic acid (Fig. 1). We also tested for formation of tetrahydroxy metabolites of lithocholic acid such as 1α,3β,7α,12α-tetrahydroxy-5β-cholan-4-oic acid and 3α,6β,7β,12α-tetrahydroxy-5β-cholan-24-oic acid, but metabolite peaks corresponding to the m/z value of tetrahydroxy metabolites (m/z 423) were not observed.
No metabolite formation was observed in reaction mixtures devoid of substrate, NADPH, or microsomes. Peaks with retention times of 19.5 min (m/z of 391) and 14.5 min (m/z 389), which corresponded to chenodeoxycholic acid and 7-ketolithocholic acid, respectively, were determined to be minor contaminants of lithocholic acid, as previously reported (Deo and Bandiera, 2008a). We did not detect metabolites of chenodeoxycholic acid (Deo and Bandiera, 2008b) in the present study.
Metabolite formation was evaluated over a substrate concentration range of 1 to 250 μM. An incubation time of 30 min and a microsomal protein concentration of 0.5 mg/ml were found to be optimal and were used for all the subsequent experiments. A lithocholic acid concentration of 100 μM was found to be saturating for all the metabolites under consideration, and this was used for any further experiments. Plots of metabolite formation versus substrate concentration showed that hepatic microsomal formation of 3-ketocholanoic acid, hyodeoxycholic acid, ursodeoxycholic acid, 6-ketolithocholic acid, and murideoxycholic acid followed Michaelis-Menten kinetics (Fig. 2).
The apparent Vmax value for 3-ketocholanoic acid formation was seven times larger than for hyodeoxycholic acid and 72 times larger than for 6-ketolithocholic acid formation. The apparent Km values associated with formation of 3-ketocholanoic acid and murideoxycholic acid were similar (21.9 ± 2.6 and 18.2 ± 3.2 μM, respectively), but the Vmax/Km ratio for 3-ketocholanoic acid formation was 15 times larger than for hyodeoxycholic acid and 72 times larger than for murideoxycholic acid formation (Table 1), indicating that formation of 3-ketocholanoic acid was catalyzed more efficiently by human hepatic microsomes than formation of the other four metabolites.
Lithocholic Acid Biotransformation by Human Recombinant P450 Enzymes. The contribution of individual P450 enzymes to lithocholic acid biotransformation was evaluated using a panel of 12 human recombinant P450 enzymes. Initial experiments were conducted to determine P450 concentrations that would ensure linearity of product formation. An incubation time of 20 min and a recombinant P450 enzyme concentration of 30 pmol of P450/ml were found to be optimal at a substrate concentration of 100 μM. Under these experimental conditions, conversion of lithocholic acid to 3-ketocholanoic acid, hyodeoxycholic acid, ursodeoxycholic acid, murideoxycholic acid, and 6-ketolithocholic acid was catalyzed by CYP3A4 but not by any of the other P450 enzymes tested (Fig. 3).
Formation of all five metabolites by recombinant CYP3A4 was evaluated over a substrate concentration range of 1 to 250 μM. Metabolite formation exhibited typical Michaelis-Menten-type kinetics with the exception of murideoxycholic acid, which followed sigmoidal kinetics (Fig. 4). Kinetic parameters for the formation of these metabolites by recombinant CYP3A4 are listed in Table 2. 3-Ketocholanoic acid was the major metabolite produced by recombinant CYP3A4 (27.2 pmol/min/pmol P450; see Fig. 3), followed by hyodeoxycholic acid, ursodeoxycholic acid, murideoxycholic acid, and 6-ketolithocholic acid.
Chemical Inhibition Studies. The effect of chemical inhibitors on lithocholic acid biotransformation was evaluated by incubating lithocholic acid with human hepatic microsomes in the presence of different concentrations of ketoconazole (an inhibitor of CYP3A4), quercetin (an inhibitor of CYP2C8), sulfaphenazole (an inhibitor of CYP2C9), and quinidine (an inhibitor of CYP2D6). The rate of formation of 3-ketocholanoic acid was inhibited by 70% in the presence of 10 μM ketoconazole and by 48% in the presence of 50 μM quercetin (Fig. 5A). The rate of formation of hyodeoxycholic acid was inhibited by 95% in the presence of 10 μM ketoconazole and by 60% in the presence of 50 μM quercetin (Fig. 5B). 6-Ketolithocholic acid was not detected at higher concentrations of ketoconazole (e.g., 10 μM). There was no effect of sulfaphenazole or quinidine at all the concentrations tested on rates of formation of any of the metabolites as compared with the control (no inhibitor).
To determine whether quercetin was able to inhibit CYP3A4, human recombinant CYP3A4 was incubated with lithocholic acid in the presence of various concentrations of quercetin, and metabolite formation was monitored. Decreased formation of all five metabolites was observed in the presence of quercetin at final concentrations of 20 and 50 μM. For example, the rate of formation of 3-ketocholanoic acid was decreased by 60%, and the rate of formation of hyodeoxycholic acid was decreased by 70% in the presence of 50 μM quercetin (data not shown).
Lithocholic Acid Biotransformation in Human Hepatic Microsomes with Varying CYP3A4 Levels. Hepatic microsome samples obtained from individual human hepatic donors were used to assess the importance of microsomal CYP3A4 levels, which are reflected by testosterone 6β-hydroxylase activities, on metabolite formation. Formation of all five lithocholic acid metabolites was greater in microsomal samples HH18 and HH837, which had relatively high CYP3A4-linked testosterone 6β-hydroxylase activities (12,000 and 13,700 pmol/min/mg protein, respectively), than in microsomal samples HG95 and HH13, which had relatively low (650 and 890 pmol/min/mg protein, respectively) CYP3A4-linked testosterone 6β-hydroxylase activities (Table 3). The correlation between testosterone 6β-hydroxylase activities and rates of metabolite formation was greater than 0.95 for all five lithocholic acid metabolites, suggesting that the same P450 enzyme catalyzed testosterone 6β-hydroxylation and lithocholic acid biotransformation.
Discussion
Biotransformation of lithocholic acid by human hepatic microsomes generated five metabolites. The major metabolite was 3-ketocholanoic acid, indicating that the predominant biotransformation pathway for lithocholic acid was oxidation at the C-3 position. Hydroxylation at the 6α position, which produced hyodeoxycholic acid, represented a quantitatively minor biotransformation pathway. Formation of all five metabolites was found to be catalyzed by CYP3A4.
There are several significant differences between our results and previous studies that investigated lithocholic acid biotransformation in human hepatic microsomes. For example, Xie et al. (2001) reported that hyodeoxycholic acid was the major metabolite of lithocholic acid following incubation with human liver microsomes or recombinant CYP3A4, and hyodeoxycholic acid was the only metabolite of lithocholic acid identified in an earlier study (Araya and Wikvall, 1999). In contrast, we showed that hyodeoxycholic acid is not the major metabolite of lithocholic acid in human liver microsomes. Our results are more consistent with those of Bodin et al. (2005), who reported the formation of the major metabolite was 3-ketocholanoic acid using recombinant CYP3A4. However, they identified 1β-hydroxylithocholic acid and hyodeoxycholic acid as lithocholic acid metabolites and did not report formation of murideoxycholic acid, ursodeoxycholic acid, or 6-ketolithocholic acid (Bodin et al., 2005), which differs from our study. A partial explanation for the differences in the metabolite profiles observed could be the use of LC/MS in our study versus thin-layer chromatography or gas chromatography/MS procedures in the other studies (Araya and Wikvall, 1999; Xie et al., 2001).
The present study also focused on the contribution of various P450 enzymes to the formation of the five metabolites. Using a of human recombinant P450 enzymes, we showed that CYP3A4 was the only P450 enzyme among those tested that catalyzed oxidation of lithocholic acid to 3-ketocholanoic, hyodeoxycholic, murideoxycholic, ursodeoxycholic, and 6-ketolithocholic acids. The role of CYP3A4 in metabolite formation was further assessed using chemical inhibitors. Inhibitor concentrations were chosen based on previous reports (Liu et al., 2007; Wang et al., 2008). Formation of all five metabolites was inhibited by the CYP3A4 inhibitor ketoconazole. Quinidine and sulfaphenazole had no effect, but quercetin, at concentrations of 20 and 50 μM, inhibited lithocholic acid metabolism. Quercetin also inhibited metabolite formation when lithocholic acid was incubated with recombinant CYP3A4, which shows that quercetin is nonselective at these concentrations and that CYP3A4 is responsible for formation of all five metabolites. Quercetin is known to inhibit CYP2C8, CYP3A4, and CYP1A2 at concentrations similar to those used in our study (Walsky et al., 2005).
The dominant role of CYP3A4 in hepatic microsomal lithocholic acid metabolism was apparent when human liver samples expressing high or low testosterone 6β-hydroxylase activity were evaluated. The results clearly show that rates of formation of lithocholic acid metabolites parallel CYP3A4-mediated testosterone 6β-hydroxylase activities in human liver samples. This finding confirms that other hepatic microsomal P450 enzymes do not contribute to lithocholic acid biotransformation when expression of CYP3A4 is reduced. A physiological implication is that interindividual variability in CYP3A4 levels, which is well documented (Shimada et al., 1994; Lampen et al., 1995), will affect intrinsic clearance and elimination of lithocholic acid. Moreover, CYP3A4 inhibition by ketoconazole, a widely used drug for treating skin and antifungal infections, may be of clinical importance for lithocholic acid homeostasis. The currently recommended daily oral dose for ketoconazole is 200 to 400 mg (Daneshmend and Warnock, 1988; Schäfer-Korting, 1993; Gupta et al., 1994), and peak plasma concentrations of ketoconazole of 2.5 to 4 μg/ml or 4.7 to 7.5 μM were measured in healthy volunteers after administration of a single oral 200-mg dose (Heel et al., 1982; Gupta et al., 1994). This serum concentration is close to the highest concentration of ketoconazole used in the present study that produced maximal inhibition of lithocholic acid biotransformation. Ketoconazole is implicated in a number of hepatic dysfunctions, and ketoconazole-induced hepatic cholestatic injury has been reported in humans (Stricker et al., 1986; Bensaude et al., 1988; Findor et al., 1998). In vivo studies involving rats treated with ketoconazole (25–50 mg/kg) showed elevated levels of bile acids such as cholic acid, taurocholic acid, chenodeoxycholic acid, and taurochenodeoxycholic acid in serum (Azer et al., 1995). Considering the potential of ketoconazole and related antifungal agents such as clotrimazole, itraconazole, and fluconazole to cause similar effects in humans, an implication of our results is that prescribing ketoconazole during cholestasis may increase the risk of liver damage.
To determine whether lithocholic acid biotransformation constitutes a detoxification pathway, the biological activities of the metabolites need to be considered. Makishima et al. (2002) showed that PXR and VDR can be activated by lithocholic acid and two of its metabolites, 3-ketocholanoic acid and 6-ketolithocholic acid (Makishima et al., 2002). Activation of PXR and VDR leads to up-regulation of hepatic and intestinal CYP3A4, thereby increasing biotransformation of lithocholic acid. Ursodeoxycholic acid, another metabolite identified in the present study, was previously shown to activate PXR in reporter gene assays and to induce CYP3A4 expression in primary cultures of human hepatocytes (Schuetz et al., 2001). Ursodeoxycholic acid is hepatoprotective and is effective in the treatment of cholestatic patients (Poupon and Poupon, 1995; Jacquemin et al., 1997). The remaining metabolites, hyodeoxycholic and murideoxycholic acids, are considerably more hydrophilic than lithocholic acid (Fini et al., 1985) and have been reported to be rapidly eliminated from liver, bile, and urine (Summerfield et al., 1976; Khallou et al., 1993). Thus, we suggest that formation of lithocholic acid metabolites leads to enhanced hepatic detoxification and elimination by two mechanisms, namely, formation of more hydrophilic, noncholestatic metabolites such as ursodeoxycholic acid, hyodeoxycholic acid, and murideoxycholic acid, and by formation of metabolites that bind and activate PXR and VDR, which leads to increased transcription of CYP3A4 and enhanced oxidation of lithocholic acid.
The physiological relevance of our study and, in particular, the in vitro to in vivo extrapolation of the results obtained are unknown. The concentration of lithocholic acid in normal human liver is approximately 5 μM or less (Setchell et al., 1997; Hofmann, 2002), which is much lower than the concentration (100 μM) found to be saturating for microsomal lithocholic acid biotransformation in the present study. At a concentration of 5 μM, the rate of formation of 3-ketocholanoic acid by human liver microsomes was approximately 25 to 50 pmol/min/mg protein (Fig. 2). This rate is greater than the rates of formation of the other four metabolites. Consequently, the results suggest that 3-ketocholanoic acid is the predominant lithocholic acid metabolite formed by P450 enzymes in human liver even at physiological concentrations. However, our study did not address the relative importance of hepatic P450-catalyzed lithocholic acid oxidation compared with conjugation reactions catalyzed by phase II enzymes such as sulfotransferases, glucuronosyl transferases, and amino acid transferases. Sulfate conjugation occurs on the 3α-hydroxyl group of bile acids and is considered to be an effective pathway for bile acid excretion by the urinary route, especially in humans (Hofmann, 2002). Likewise, glucuronide conjugation of bile acids at the C-24 carboxyl, 3α-hydroxyl, and 6α-hydroxyl groups can also occur. Thus, sulfate or glucuronide conjugation may compete with P450-catalyzed oxidation of lithocholic acid, especially at the 3α-hydroxyl position, and formation of 3-ketocholanoic acid may be only a minor reaction in vivo. On the other hand, sulfate or glucuronide conjugation of the other oxidized lithocholic acid metabolites such as hyodeoxycholic, murideoxycholic, and ursodeoxycholic acids is highly probable, leading to enhanced elimination of the metabolites and helping to regulate hepatic lithocholic acid levels. The relative contribution of oxidation and conjugation reactions to bile acid biotransformation is an area that requires further investigation.
Comparison of metabolite profiles of lithocholic acid with those of cholic acid and chenodeoxycholic acid (Deo and Bandiera, 2008b) reveals that formation of 3-oxo metabolites is the dominant biotransformation pathway for bile acids in human liver microsomes. As reported, 3-dehydrocholic acid was the only metabolite of cholic acid, and 7α-hydroxy-3-oxo-5β-cholanoic acid was the major metabolite of chenodeoxycholic acid (Deo and Bandiera, 2008b), whereas 3-ketocholanoic acid was the major metabolite of lithocholic acid in this study. Formation of all three metabolites was mediated by CYP3A4. Hydroxylation at the 6α position, which was proposed to be a major bile acid hydroxylation pathway in humans (Setchell et al., 1997; Araya and Wikvall, 1999), is also catalyzed by CYP3A4 but is relatively less important. Comparison of Vmax and Km values, obtained using human liver microsomes, for chenodeoxycholic and cholic acid metabolite formation (Deo and Bandiera, 2008b) revealed that lithocholic acid was biotransformed as readily as chenodeoxycholic acid and slower than cholic acid but at lower substrate concentrations. In contrast, the rate of lithocholic acid biotransformation by rat liver microsomes was much greater than by human liver microsomes. For example, the rate of murideoxycholic acid formation by rat liver microsomes, expressed in terms of the Vmax value (Deo and Bandiera, 2008a), was approximately 18 times faster than the rate of 3-ketocholanoic formation by human liver microsomes.
Based on the results obtained, a scheme for the biotransformation of lithocholic acid in human hepatic microsomes is proposed in Fig. 6. The results provide compelling evidence that lithocholic acid can serve as a physiological substrate for CYP3A4 with 3-ketocholanoic acid as its major metabolite. Because of the dominant role of CYP3A4 in bile acid metabolism in human liver, factors that affect CYP3A4 levels or activity will probably affect the clearance and elimination of bile acids such as lithocholic acid. CYP3A4 expression has been shown to be altered as a result of changes in nutrition, diseases, environmental factors, and exposure to chemicals such as phenobarbital, rifampicin, and phenytoin or, alternatively, as a result of inhibition of CYP3A4 activity by triazole antifungal agents or macrolide antibiotics. These factors may contribute to a change in lithocholic acid biotransformation resulting in clinically significant drug interactions.
Acknowledgments
We thank Andras Szeitz (University of British Columbia, Faculty of Pharmaceutical Sciences) for technical help with the LC/MS analysis and Dr. Eugene Hrycay for helpful suggestions during the preparation of the manuscript.
Footnotes
-
This work was supported by the Canadian Institutes of Health Research [MOP-81174]. Partial support was provided by a training grant from Merck Research Laboratories (Merck & Co., Whitehouse Station, NJ) (to A.K.D.).
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.109.027763.
-
ABBREVIATIONS: VDR, vitamin D receptor; PXR, pregnane X receptor; lithocholic, 3α-hydroxy-5β-cholan-24-oic; P450, cytochrome P450; hyodeoxycholic, 3α,6α-dihydroxy-5β-cholan-24-oic; chenodeoxycholic, 3α,7α-dihydroxy-5β-cholan-24-oic; murideoxycholic, 3α,6β-dihydroxy-5β-cholan-24-oic; 3-ketocholanoic, 3-oxo-5β-cholan-24-oic; cholic, 3α,7α,12α-trihydroxy-5β-cholan-24-oic; LC/MS, liquid chromatography/mass spectrometry; SKF 525A, 2′-diethylaminoethyl 2,2-diphenylpentanoate hydrochloride; α-muricholic, 3α,6β,7α-trihydroxy-5β-cholan-24-oic; β-muricholic, 3α,6β,7β-trihydroxy-5β-cholan-24-oic; γ-muricholic, 3α,6α,7α-trihydroxy-5β-cholan-24-oic; 3-dehydrocholic, 7α,12α-dihydroxy-3-oxo-5β-cholan-24-oic; ursodeoxycholic, 3α,7β-dihydroxy-5β-cholan-24-oic; deoxycholic, 3α,12α-dihydroxy-5β-cholan-24-oic; 6-ketolithocholic, 3α-hydroxy-6-oxo-5β-cholan-24-oic; 7-ketolithocholic, 3α-hydroxy-7-oxo-5β-cholan-24-oic; isolithocholic, 3β-hydroxy-5β-cholan-24-oic.
- Accepted May 27, 2009.
- Received March 31, 2009.
- The American Society for Pharmacology and Experimental Therapeutics