Human UGT1A4 and UGT1A3 Conjugate 25-Hydroxyvitamin D₃: Metabolite Structure, Kinetics, Inducibility, and Interindividual Variability

Abstract

25-Hydroxyvitamin D₃ (25OHD₃) is used as a clinical biomarker for assessment of vitamin D status. Blood levels of 25OHD₃ represent a balance between its formation rate and clearance by several oxidative and conjugative processes. In the present study, the identity of human uridine 5′-diphosphoglucuronyltransferases (UGTs) capable of catalyzing the 25OHD₃ glucuronidation reaction was investigated. Two isozymes, UGT1A4 and UGT1A3, were identified as the principal catalysts of 25OHD₃ glucuronidation in human liver. Three 25OHD₃ monoglucuronides (25OHD₃-25-glucuronide, 25OHD₃-3-glucuronide, and 5,6-trans-25OHD₃-25-glucuronide) were generated by recombinant UGT1A4/UGT1A3, human liver microsomes, and human hepatocytes. The kinetics of 25OHD₃ glucuronide formation in all systems tested conformed to the Michaelis-Menten model. An association between the UGT1A4*3 (Leu48Val) gene polymorphism with the rates of glucuronide formation was also investigated using human liver microsomes isolated from 80 genotyped livers. A variant allele dose effect was observed: the homozygous UGT1A4*3 livers (GG) had the highest glucuronidation activity, whereas the wild type (TT) had the lowest activity. Induction of UGT1A4 and UGT1A3 gene expression was also determined in human hepatocytes treated with pregnane X receptor/constitutive androstane receptor agonists, such as rifampin, carbamazepine, and phenobarbital. Although UGT mRNA levels were increased significantly by all of the known pregnane X receptor/constitutive androstane receptor agonists tested, rifampin, the most potent of the inducers, significantly induced total 25OHD₃ glucuronide formation activity in human hepatocytes measured after 2, but not 4 and 24 hours, of incubation. Finally, the presence of 25OHD₃-3-glucuronide in both human plasma and bile was confirmed, suggesting that the glucuronidation pathway might be physiologically relevant and contribute to vitamin D homeostasis in humans.

Vitamin D₃, the major form of vitamin D found in humans, is critical for the regulation of calcium and phosphate homeostasis (1, 2). After its synthesis in the skin or absorption from the diet, vitamin D₃ is delivered through the bloodstream to the liver for conversion to 25-hydroxyvitamin D₃ (25OHD₃). 25OHD₃ is the major circulating form of the hormone and is used as a clinical biomarker for assessment of vitamin D status, primarily because of its relatively high concentration, long systemic half-life, and proximity to the biologically active hormone, 1α,25-dihydroxyvitamin D₃ (1α,25(OH)₂D₃). Conversion of 25OHD₃ to 1α,25(OH)₂D₃ is catalyzed principally by the vitamin D 1α-hydroxylase enzyme, cytochrome P450 27B1 (CYP27B1), found in the kidney and other tissues of the body (3, 4). 25OHD₃ monohydroxylation can also lead to a loss of hormone activity, principally through the action of CYP24A1, which catalyzes sequential side chain oxidation reactions (5). Recently, we have shown that CYP3A4, one of major drug-metabolizing enzymes in the human liver and small intestine, can also catalyze an inactivating oxidation of 25OHD₃ mainly at the 4β- and 4α-positions, providing an alternative and xenobiotic inducible catabolic pathway of 25OHD₃ elimination and potentially affecting vitamin D regulation (6, 7).

Often overlooked in the characterization of vitamin D metabolism are the sulfonation and glucuronidation conjugation reactions that can occur with any of the forms of vitamin D₃. Although the CYP-mediated vitamin D metabolism in human liver and kidney, mediated by CYP2R1, CYP27A1, CYP3A4, or CYP24A1, are important in controlling the effects of vitamin D₃ in the body, conjugation of vitamin D₃ metabolites by phase II enzymes, such as sulfotransferase or uridine 5′-diphosphoglucuronyltransferase (UGT), may be equally important and yet not as thoroughly investigated (8–10). In addition, some of these downstream-conjugated metabolites are excreted in urine and bile, contributing to hormone turnover in the body (11, 12).

The existence of 25OHD₃ glucuronide conjugates in vivo was first described from an analysis of rat bile (13). Subsequent experiments with rat liver microsomes and rat bile led to the identification of 2 glucuronide conjugates formed at the C-3 or C-25 position of 25OHD₃ (14, 15). Moreover, studies using radiolabeled 25OHD₃ dosed in vivo revealed significant biliary excretion of multiple polar vitamin D conjugates (12, 16). However, further characterization of the structure and source of vitamin D conjugates formed in humans has not been reported.

Recently, using in vitro recombinant human UGTs, we found that formation of 1α,25(OH)₂D₃-25-glucuronide is catalyzed primarily by UGT1A4 and to a much lesser extent by UGT2B4 and UGT2B7 (17). Based on the results of studies in rats and radiolabel 25OHD₃ dosing studies in humans, one can hypothesize the formation of 25OHD₃ glucuronide conjugate(s) in humans, with transporter-mediated excretion into the bile and possibly blood. Importantly, delivery of 25OHD₃ glucuronides into intestine may contribute to the regulation of intestinal, vitamin D receptor (VDR)-responsive target genes (18). Because circulating blood levels of 25OHD₃ are 1000-fold higher than 1α,25(OH)₂D₃ concentrations, one might expect a priori higher level of 25OHD₃ conjugates into bile. Also, CYP27B1 is expressed in intestinal cells (18, 19) and presumably converts 25OHD₃ to 1α,25(OH)₂D₃ in situ. Free 25OHD₃, released from conjugates after local β-glucuronidase hydrolysis (20), could be reabsorbed into the intestinal mucosa, where it could be either redistributed into blood or bioactivated locally by CYP27B1 (21). Delivery of 1α,25(OH)₂D₃-25-glucuronide to mouse colon in vivo simulates intestinal vitamin D-responsive genes, eg, CYP24A1 or TRPV6 expression (22, 23). Thus, delivery of a 25OHD₃ conjugate to the proximal intestine through a biliary route, followed by deconjugation and 1α-hydroxylation, may serve as an important enteroendocrine or/and paracrine signaling loop and regulator of VDR-dependent gene expression, because VDR is highly expressed in the human intestinal system (21).

In this study, we sought to fully characterize the glucuronidation of 25OHD₃ in humans using recombinant UGT enzymes, liver microsomes, and primary hepatocytes to identify which of the human UGTs is responsible for catalysis. After identification of UGT1A4 and UGT1A3 as principal players in the conjugation reactions, we sought to determine whether a common UGT1A4 gene polymorphism is associated with the formation rates and whether induction of UGT1A4 and UGT1A3 in hepatocytes could enhance production of 25OHD₃ glucuronides. Finally, we have confirmed the presence of 25OHD₃-3-glucuronide in both human plasma and bile.

Materials and Methods

Materials

Alamethicin, carbamazepine, hyperforin, 4-phenyl-1,2,4-triazoline-3,5-dione (PTAD), rifampin, trifluoperazine, uridine-5′-diphosphoglucuronic acid (UDPGA), and β-glucuronidase (Escherichia coli, type IX-A) were purchased from Sigma-Aldrich. Levetiracetam was purchased from LKT Laboratories. 25OHD₃ was obtained from Calbiochem, and its deuterated form, d6–25OHD₃ (containing 6 deuterium atoms at C-26 and C-27), was purchased from Medical Isotope, Inc. 25OHD₃-3-glucuronide was synthesized from the corresponding provitamin D or its derivatives using the Koenigs-Knorr reaction (15). All other buffers and chemicals were of the highest grade commercially available.

Pooled human liver microsomes (HLMs), prepared previously, were generated from the University of Washington School of Pharmacy Human Tissue Bank. Individual preparations of liver microsomes were obtained from the same tissue bank and from livers that are part of a St Jude Children's Research Hospital human tissue repository. Both tissue banks keep anonymous records with no links to the original tissue donor and thus are exempt from Human Subjects review. Basic demographic characteristics of the tissue donors has been described previously (24). The protein concentration in each HLM was measured by BCA assay (Thermo Scientific). Supersomes containing cDNA-expressed human UGT isozymes (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A7, UGT1A8, UGT1A9, UGT1A10, UGT2B4, UGT2B7, UGT2B15, or UGT2B17) were purchased from BD Gentest. Cryopreserved primary human hepatocytes from 3 different human liver donors were obtained from either BD Gentest or Invitrogen. Outdated human plasma was from the Puget Sound Blood Center and anonymous, pooled human bile was kindly provided by Dr Evan D. Kharasch at the Washington University in St Louis. These outdated human samples were analyzed under an institution exempt approval from the University of Washington Human Subjects Review Board.

Formation of 25OHD₃ monoglucuronides by HLM and human UGT isozymes

The following procedures were conducted under low light conditions to avoid potential degradation of vitamin D metabolites. Initial reactions with HLMs were carried out in 50mM Tris-HCl (pH 7.4) solution containing 1-mg/mL pooled HLM, 50-μg/mL alamethicin, 5mM MgCl₂, 40μM 25OHD₃, and 5mM UDPGA, with a 1-mL final volume in glass tubes. Briefly, pooled HLMs were pretreated with alamethicin in Tris-HCl buffer on ice for 15 minutes. MgCl₂ and 25OHD₃ were then added, mixed, and prewarmed at 37°C for 5 minutes. In some experiments, d6–25OHD₃ (10μM) was chosen as the substrate to generate internal standards, the deuterium-labeled 25OHD₃ glucuronides. Reaction was initiated by adding UDPGA and maintained at 37°C for 3 hours. Incubations without UDPGA served as negative controls. Reactions were terminated with 2 volumes of ice-cold acetonitrile, and the mixtures were transferred to Eppendorf tubes and centrifuged for 4 minutes (12 000g). The supernatants were collected, concentrated under a nitrogen stream, and reconstituted in the mobile phase for liquid chromatography-mass spectrometry (LC-MS) analysis.

Samples were then subjected to LC-UV/MS analysis using an Agilent MSD mass spectrometer coupled with an Agilent 1100 series HPLC system. Chromatographic separation was achieved on a Symmetry C18 (2.1 × 150 mm, 3.5 μm) column (Waters) and a mobile phase consisting of 5mM ammonium acetate (A) (pH 4.6) and acetonitrile (B) at 45°C. A linear gradient from 20% B (0–2 min) to 90% B (17–22 min) in 15 minutes at a 0.25-mL/min flow rate was employed. UV detection was performed at 265 nm. The mass spectrometer was operated in the negative ionization mode. The interface was maintained at 350°C with a nitrogen nebulization pressure of 25 psi, resulting in a flow of 10 L/min. 25OHD₃ glucuronides and d6–25OHD₃ glucuronides were detected by selective ion monitoring at m/z 575 and 581, respectively, after 100-V ion fragmentation.

Isolation and structure identification of 25OHD₃ monoglucuronides

Reaction products were fractionated by HPLC-UV using an Agilent 1100 series system. Chromatographic separations were accomplished using a Waters Symmetry C18 (4.6 × 150 mm, 5 μm) column and a mobile phase consisting of 5mM ammonium acetate (A) (pH 4.6) and acetonitrile (B) at a flow rate of 1.0 mL/min and 45°C. A linear gradient from 45% B (1 min) to 90% B in 6 minutes, maintained at 90% B for additional 4 minutes, and then equilibrated back to 45% B within 3 minutes, was employed. Three peaks designated as metabolites M1 (retention time [RT], 4.2 min), M2 (RT, 5.3 min), and M3 (RT, 5.8 min) were detected by UV absorbance at 265 nm. Elution fractions comprising each peak were pooled and dried under a nitrogen stream and kept at −80°C in methanol.

In order to characterize the structures of 25OHD₃ glucuronides, the 3 isolated peaks, M1 (0.5 μg), M2 (0.2 μg), and M3 (0.05 μg), and chemically synthesized 25OHD₃-3-glucuronide standard (2 μg) were derivatized with PTAD before mass spectral analysis. Acetonitrile (100 μL) containing 100-μg PTAD was added to the dried products and kept at room temperature for 1 hour. The solvent was evaporated after incubation, and the residue was dissolved in the mobile phase (100 μL), 10 μL of which were subjected to the Agilent Electrospray LC-MS system above with a Waters C18 (2.1 × 150 mm, 3.5 μm) column. In order to obtain the characteristic ions that can differentiate the presence of a glucuronic acid moiety on either the A ring (C-1 or C-3) or side chain (C-25) of 25OHD₃, the PTAD-derivatized 25OHD₃ glucuronides were ionized and scanned from m/z 250 to 800 in the positive ion mode.

Ultraviolet spectra of the isolated M3 aglycone were also recorded in ethanol with a Varian Cary 3E UV-Vis spectrophotometer. Briefly, the isolated M3 fractions were pooled and dried under nitrogen stream; β-glucuronidase (2000 U/mL) in 1 mL of 10mM Tris-HCl (pH 7.0) was added into the M3 pellet and incubated at 37°C for 40 minutes. After incubation, the M3 aglycone was extracted by addition of ethyl acetate, dried under nitrogen stream, and reconstituted in ethanol for UV full scan (220–320 nm) analysis.

Kinetic studies of 25OHD₃ monoglucuronide formation

Kinetic studies were conducted using pooled HLM (0.5 mg/mL), UGT1A3 (0.2 mg/mL), and UGT1A4 (0.2 mg/mL) under conditions (worked out in pilot studies) that provided linear product formation with respect to protein concentration and time. The range of substrate 25OHD₃ concentrations was 0.156μM to 40μM. Incubation (200 μL) was conducted at 37°C for 30 minutes and terminated by the addition of 2 volumes of acetonitrile. The isolated internal standard d6–25OHD₃-25-glucuronide was added, and supernatants were collected by centrifugation, dried, and reconstituted in the mobile phase. Each set of data was fit to a simple Michaelis-Menten kinetics model using nonlinear regression data analysis (GraphPad Prism v.5). Each experimental reaction condition was conducted in triplicate and repeated 3 times independently.

Quantification of the 3 25OHD₃ monoglucuronides was performed by LC-MS analysis as described above. Quantification was achieved using a standard solution containing a known amount of 25OHD₃-25-glucuronide (M1) that had been isolated from an enzymatic reaction. To assign the concentration of standard, the isolated M1 was dissolved in 10mM Tris-HCl buffer (pH 7.0) and completely converted to the aglycone, 25OHD₃, using β-glucuronidase (1500 U/mL) treatment for 30 minutes (optimal incubation time). Complete hydrolysis was indicated by full disappearance of the glucuronide peak (m/z 575) that was monitored by LC-MS analysis. The amount of 25OHD₃ released was quantified by LC-MS/MS analysis as described previously (25), and the amount of M1 was assigned assuming that it was equal to that of the released 25OHD₃. Calibration curves were constructed by plotting the peak area ratio of M1 (25OHD₃-25-glucuronide) and the purified, d6–25OHD₃-25-glucuronide internal standard vs the corresponding M1 concentration and fitting a linear regression equation to the data. Formation of 25OHD₃ glucuronides was then quantified by fitting their peak area ratios (metabolites/internal standard) from incubations to the calibration curve.

Interliver variability of 25OHD₃ monoglucuronide formation

In order to evaluate interliver differences in the formation of 25OHD₃ glucuronides, we screened a panel of randomly selected 20 individual HLMs, which were prepared from the University of Washington School of Pharmacy Human Tissue Bank. The assay was carried out as described above, except at a 10μM substrate 25OHD₃ concentration (below the Michaelis constant [K_m] for these reactions). The mixtures were incubated at 37°C for 30 minutes, in a total volume of 0.2 mL. The formation rates of the 3 glucuronide isomers were calculated as described above for the kinetic studies. Duplicate incubations were conducted separately and the mean formation rates for each glucuronide isomer were calculated.

Effects of UGT1A4 polymorphism on 25OHD₃ monoglucuronide formation

Allele variants of the UGT1A4 gene are reported to contribute to interindividual difference in enzyme activity, with the UGT1A4*3 (Leu48Val) variant gene product exhibiting increased glucuronidation activity (26, 27). Because UGT1A4 was found to be the predominant enzyme catalyzing 25OHD₃ glucuronidation (Supplemental Table 1), the effect of this UGT1A4 polymorphism on 25OHD₃ glucuronidation was investigated using HLMs, which were isolated from liver samples procured by St Jude Children's Research Hospital (n = 63) and University of Washington (n = 17). Tissue donor characteristics and UGT1A4*3 genotype are summarized in table 2 below. Most the livers were from Caucasian donors, and among the 80 livers genotyped for UGT1A4*3, 32 were heterozygous (TG) and 5 homozygous (GG). Briefly, HLMs (1 mg/mL) were incubated with alamethicin, MgCl₂, 25OHD₃ (25μM), and UDPGA as described above, in a total volume of 0.2 mL for 30 minutes. Duplicate incubations were conducted separately, and the mean formation rates for each glucuronide isomer were calculated.

UGT1A4 and UGT1A3 expression and glucuronidation of 25OHD₃ in human hepatocytes

Cryopreserved human hepatocytes were thawed and cultured, as previously published (6). After plating and culturing of viable cells for 24 hours, the cells were treated with one of the following drugs: rifampin (10μM), hyperforin (0.5μM), phenobarbital (400μM), carbamazepine (50μM), or levetiracetam (200μM) in a 100-μL solution per well. After 48 hours, the treated cells were washed and lysed, and mRNA was stabilized with addition of 140 μL of Ambion lysis/binding buffer provided in the MagMax 96 RNA isolation kit. Total RNA was isolated using the MagMax 96 RNA isolation kit, and RNA quantity was assessed using the NanoDrop spectrophotometer. cDNA was synthesized from RNA template using the High Capacity cDNA Reverse Transcription kit according to manufacturer's protocol. After synthesis, the cDNA reactions were diluted to 80-μL total volume with nuclease-free water. TaqMan reactions were run on a 7900HT Real-Time PCR system, operated with a 384-well optical reaction plate as previously published (6). Probe Identifications for the TaqMan assays were as follows: UGT1A3, Hs04194492_g1; UGT1A4, Hs01655285_s1.

For some of these experiments, after a 48-hour incubation with rifampin (10μM), the cells were washed with saline solution twice and then incubated with 2μM 25OHD₃ for various incubation times (t = 2, 4, or 24 hours). At the end of the treatment period, culture medium was collected and pooled for the quantification of 25OHD₃ monoglucuronides using an optimized LC method to fully separate multiple peaks derived from the culture medium. Separations were achieved after a linear gradient from 30% B (2 min) to 45% B (25 min), hold at 45% B for 1 minute and increased to 60% B (35 min), increased to 90% in 1 minute and hold for 3 minutes (39 min), and then equilibrated back to 30% B in 1 minute.

Detection of endogenous 25OHD₃-3-glucuronide from human plasma and bile

Human plasma (0.5 mL) was first subjected to protein precipitation using 22 volumes of acetonitrile, and then the supernatant was diluted with 2 volumes of 0.1M sodium acetate (pH 4.0) and applied to solid-phase extraction using Waters Oasis WAX (60 mg, 3 cc) anion exchange cartridges as the manufacturer suggested. The eluates were dried and reconstituted in mobile phase solution for LC-MS analysis as described above. In contrast, human bile (5 mL) was first diluted with 2.5 mL water and then subjected to liquid-liquid extraction using an equal volume of hexanes twice. The aqueous solutions were separated, pooled, and buffered with an equal volume of 0.1M sodium acetate (pH 4.0) and subjected to solid-phase extraction using 5 Waters Oasis WAX (60 mg, 3 cc) anion exchange cartridges. The eluates were dried and incubated with 300 μL of PTAD in acetonitrile (1 mg/mL) at room temperature for 1 hour. After incubation, the solution was transferred to a new glass tube, dried, and reconstituted in 50-μL mobile phase solution for LC-MS/MS analysis. Chromatographic separation was achieved on a Symmetry C18 (2.1 × 150 mm, 3.5 μm) column and a mobile phase consisting of water with 0.1% formic acid (A) and acetonitrile (B) at 45°C on an Agilent 1290 series ultra performance LC system coupled with Agilent 6410 triple quadruple mass spectrometer. A linear gradient from 60% B (1 min) to 90% B (15–18 min) in 14 minutes at a 0.3-mL/min flow rate was employed. Injection volume was 20 μL for bile extracts and 1 μL for standard derivatives. The mass spectrometer was operated in the positive ionization mode. The ionization and fragmentation parameters were set as follows: capillary voltage, 4000 V; gas temperature, 300°C; gas flow rate, 12 L/min; nebulizer, 40 psi; fragmentor, 110 V; collision energy, 20 V. 25OHD₃ glucuronides after PTAD derivatization were detected by multiple reaction monitoring channels of m/z 734 → 298 and 734 → 474.

Binding of 25OHD₃-3-glucuronide and 25OHD₃ to vitamin D binding protein (DBP)

The binding assays were performed as previously described (28). Briefly, the rat plasma was diluted 1:5000 (vol/vol) in 0.05M phosphate buffer (pH 7.5) containing 0.01% gelatin and 0.01% merthiolate (PBG buffer). Each assay mixture was placed in a borosilicate glass tube and consisted of 1) 0.5 mL of 1:5000 dilution of vitamin D-deficient rat plasma in PBG buffer, 2) 6000–8000 cpm of [23,24-³H]-25OHD₃ in 20 μL of 100% ethanol, and 3) vitamin D metabolites in 25 μL of ethanol. After 1–2 hours of incubation at 4°C, the bound vitamin D metabolites were separated from free by adding 0.2 mL of a mixture of cold 1.0% Norit A Charcoal and 0.1% Dextran T-70 in PBG buffer to each tube. After 30 minutes at 4°C, the tubes were spun at 1000g for 10 minutes in a refrigerated centrifuge. A portion (0.5 mL) of the supernatant was removed for quantitation of the bound [³H]-25OHD₃.

Statistical analysis

All data are expressed as the mean ± SD unless stated otherwise and were compared using one-way ANOVA with Dunnett's multiple comparison tests. We used GraphPad Prism v.5 for the statistical analyses, and a P < .05 was considered statistically significant.

Results

Formation of 25OHD₃ monoglucuronides by HLMs

25OHD₃ contains 2 hydroxyl groups and presumably is expected to generate 2 different monoglucuronides either at the C-25 or C-3 positions (Figure 1, A and B). However, incubation of 25OHD₃ with pooled HLMs and UDPGA generated 3 distinct metabolites (Figure 1C). As shown in Figure 1C, these 3 different peaks were detected by UV absorbance at 265 nm and selective mass ion monitoring at m/z 575, which corresponds to [M − H]⁻ under the negative ion mode. Formation of these 3 metabolites occurred in a UDPGA-, microsomal protein-, and time-dependent manner. The isolated metabolites were individually treated with β-glucuronidase, and both M1 and M2 released the identical aglycone, 25OHD₃, suggesting that these 2 metabolites are 25OHD₃ glucuronides (data not shown). However, hydrolysis of M3 did not generate detectable amount of 25OHD₃ that matches with 25OHD₃ standard by LC-MS/MS, suggesting that it might be a conjugate of a different monohydroxy isomer of vitamin D₃.

Figure 1.

Formation of 25OHD₃ monoglucuronides with HLMs. The proposed chemical structure of 25OHD₃ monoglucuronide, conjugated at either C-25 (A) or C-3 (B) position of 25OHD₃ (M1 and M2) is shown. C, Representative chromatogram of 3 glucuronide isomers. Glucuronides were analyzed by LC-MS at m/z 575 under the negative mode. D, Chromatogram of the isolated d6–25OHD₃-25-glucuronide as an internal standard. The corresponding peak was isolated by LC-UV and then analyzed by LC-MS at m/z 581 under the negative mode.

Incubations of a deuterated form, d6–25OHD₃, also resulted in 3 glucuronides with a selective mass ion at m/z 581. The deuterated monoglucuronides were separated and purified by HPLC-UV, and the major peak (d6–25OHD₃-25-glucuronide, assigned as below) served as an internal standard for quantification purposes (Figure 1D).

Structure identification of 25OHD₃ monoglucuronides

In order to identify the structures of the 3 25OHD₃ monoglucuronides, the corresponding peaks were isolated and purified using HPLC-UV detection. LC-MS was performed on the isolated fractions to characterize the mass ions of these metabolites after incubation with PTAD, as previously published (17). A chemically synthesized 25OHD₃-3-glucuronide served as a standard for structure identification; the 2D-NMR ROESY spectrum of that metabolite is shown in Supplemental Figure 1. The PTAD adducts of 25OHD₃-3-glucuronide (2 μg) and M2 (0.2 μg) under the positive ion mode (full scan) showed the presence of ions at m/z 790 and 734, corresponding to the [M + K]⁺ and [MH − H₂O]⁺ adducts (Figure 2, A and B). The mass spectra of the PTAD-derivatized M2 and 25OHD₃-3-glucuronide were essentially the same, and both contained a characteristic fragment ion at m/z 474. This fragment originates from cleavage of the bond between C-6 and C-7 of the A ring skeleton that contains a glucuronic acid at C-3 position (Figure 2C). Moreover, M2 always coeluted with 25OHD₃-3-glucuronide standard under multiple LC conditions. These results suggest that M2 is the 25OHD₃-3-glucuronide. In contrast, the PTAD adduct of M1 (0.5 μg) showed the presence of ions at m/z 790 and 774, corresponding to the [M + K]⁺ and [M + Na]⁺ adducts, respectively (Figure 2D). The mass spectrum of the PTAD-M1 adduct had 2 characteristic ions at m/z 558 and 298 (25) but lacked the characteristic ion at m/z 474 that was observed for 25OHD₃-3-glucuronide. This suggests that M1 is 25OHD₃-25-glucuronide, which is in agreement with previous reports (7, 29).

Figure 2.

Mass spectra of the PTAD derivatives of 25OHD₃ glucuronides. A, Mass spectrum of the PTAD-25OHD₃-3-glucuronide standard. B, Mass spectrum of the PTAD-M2. C, Proposed fragmentation of M2 to generate ion m/z 474. D, Mass spectrum of the PTAD-M1. The PTAD-derivatized 25OHD₃ glucuronides were ionized and scanned from m/z 250 to 800 in the positive ion mode. Experimental details have been described in Materials and Methods.

Because 25OHD₃ contains only 2 hydroxyl groups, we considered the possibility that M3 is an unknown vitamin D₃ glucuronide isomer. Multiple precursor ions at m/z 790, 774, 752, or 734 for a PTAD derivative of M3 (0.05 μg) were undetectable at that time point, but later detected after a larger amount of M3 (∼0.3 μg) was applied (Supplemental Figure 2). However, the structure of M3 could not be assigned based on its mass spectra alone. In order to explore the possible structures of M3, we first tested whether or not M3 is a 3-epi-25OHD₃ glucuronide. HLM incubations with 3-epi-25OHD₃ were performed, resulting in 3 glucuronide peaks, as seen for 25OHD₃ (Supplemental Figure 3). By comparing the respective 25OHD₃ products, and assuming corresponding identity for the regional isomers of 25OHD₃ and 3-epi-25OHD₃, only the “second peaks” from 25OHD₃ and 3-epi-25OHD₃ were resolved. From first principles, we propose that both metabolite peaks are conjugated at the 3 position, enhancing the stereochemical difference of 25OHD₃ and 3-epi-25OHD₃ isomers. On the other hand, one can assume by inference that the “first peaks” (M1) and “third peaks” (M3) are 25-glucuronides. With this in mind, we considered that M3 might be a 5,6-trans-25OHD₃ glucuronide. To test this hypothesis, the isolated M3 fractions were pooled and incubated with β-glucuronidase, extracted and subjected to full UV scan analysis. As shown in Supplemental Figure 4, the UV spectrum of the peak from M3 aglycone showed a wavelength of maximum absorption (λ_max) shift to 272–273 nm from the reference 265 nm for 25OHD₃; this shift is a characteristic of 5,6-trans-25OHD₃ (30). On the basis of the above observations, we suggest that M3 is 5,6-trans-25OHD₃-25-glucuronide (t-25OHD₃-25-glucuronide).

Identification of human UGT isozymes catalyzing 25OHD₃ glucuronidation

In order to determine which human UGT isozymes were responsible for the 25OHD₃ glucuronidation reactions, we screened the 25OHD₃ glucuronosyltransferase activity of 12 human UGT isozymes (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A7, UGT1A8, UGT1A9, UGT1A10, UGT2B4, UGT2B7, UGT2B15, and UGT2B17) that were incubated with 20μM 25OHD₃ individually, as previously published (17). Two isozymes, UGT1A4 and UGT1A3 (the relative activity ratio of UGT1A4/1A3 was 2.7), exhibited significant 25OHD₃ glucuronidation activity, whereas the other UGT isozymes had no detectable catalytic activity under the incubation conditions that were employed (Supplemental Table 1).

Determination of kinetic parameters for 25OHD₃ monoglucuronide formation

Kinetic studies of 25OHD₃ glucuronide formation were performed in vitro using pooled HLMs, UGT1A3 and UGT1A4 supersomes (Figure 3). The K_m and the maximum velocity (V_max) values of 25OHD₃ glucuronide formation by pooled HLMs were estimated using a simple Michaelis-Menten kinetics model, as following: 19.1 ± 1.19μM (M1, 25OHD₃-25-glucuronide), 22.6 ± 3.36μM (M2, 25OHD₃-3-glucuronide), and 20.3 ± 3.31μM (M3, t-25OHD₃-25-glucuronide) and 4.09 ± 0.13, 1.33 ± 0.19, and 0.25 ± 0.03 pmol/min·mg protein, respectively (Table 1). The kinetic parameters (K_m and V_max) for 25OHD₃-25-glucuronide, 25OHD₃-3-glucuronide, and t-25OHD₃-25-glucuronide formation by UGT1A3 were, respectively, 3.95 ± 0.35μM, 5.90 ± 0.38μM, and 9.35 ± 2.38μM and 1.74 ± 0.10, 0.56 ± 0.03, and 0.22 ± 0.01 pmol/min·mg protein. For UGT1A4, they were, respectively, 6.39 ± 0.38μM, 2.16 ± 0.16μM, and 4.19 ± 0.30μM and 7.35 ± 0.20, 0.93 ± 0.02, and 0.26 ± 0.01 pmol/min·mg protein.

Figure 3.

Substrate concentration-rate profiles of 25OHD₃ monoglucuronides by pooled HLMs (A), UGT1A3 (B), and UGT1A4 (C). Each set of data was fit to a simple hyperbolic model using nonlinear regression data analysis (Prism v.5). Experimental details have been described in Materials and Methods. Closed circle, open circle, and closed square indicate the catalytic activities of 25OHD₃-25-glucuronide (M1), 25OHD₃-3-glucuronide (M2), and 5,6-trans-25OHD₃-glucuronide (M3) formation, respectively.

Table 1.

Kinetic Parameters Associated with 25OHD₃ Glucuronide Formation in HLMs, UGT1A4, and UGT1A3

	Km (μM)	Vmax (pmol/min·mg protein)	Vmax/Km (μL/min·mg protein)
25OHD3-25-glucuronide
HLMs	19.1 ± 1.19	4.09 ± 0.13	0.21
UGT1A4	6.39 ± 0.38	7.35 ± 0.20	1.15
UGT1A3	3.95 ± 0.35	1.74 ± 0.10	0.44
25OHD3-3-glucuronide
HLMs	22.6 ± 3.36	1.33 ± 0.19	0.06
UGT1A4	2.16 ± 0.16	0.93 ± 0.02	0.43
UGT1A3	5.90 ± 0.38	0.56 ± 0.03	0.09
5,6-trans-25OHD3-25-glucuronide
HLMs	20.3 ± 3.31	0.25 ± 0.03	0.01
UGT1A4	4.19 ± 0.30	0.26 ± 0.01	0.06
UGT1A3	9.35 ± 2.38	0.22 ± 0.01	0.02

Kinetic parameters were derived by fitting data to a simple hyperbolic model (Michaelis-Menten equation) using nonlinear regression data analysis (Prism v.5). Data represent as mean ± SD from 6 replicates in 3 independent experiments.

Interliver variability in 25OHD₃ glucuronide formation

To evaluate interliver variability in 25OHD₃ glucuronide formation, a panel of 20 HLMs was prepared and incubated with 25OHD₃. As shown in Figure 4A, the variability in 25OHD₃-25-glucuronide, 25OHD₃-3-glucuronide, and t-25OHD₃-25-glucuronide formation was about 3.2-fold (0.47–1.49 pmol/min·mg protein; mean = 0.92 pmol/min·mg protein), 5.7-fold (0.11–0.63 pmol/min·mg protein; mean = 0.28 pmol/min·mg protein), and 3.0-fold (0.03–0.09 pmol/min·mg protein; mean = 0.06 pmol/min·mg protein), respectively. The mean product ratio of 25OHD₃-25-glucuronide to 25OHD₃-3-glucuronide to t-25OHD₃-25-glucuronide was 17:5:1, and a strong correlation of formation rates among these 3 glucuronides was observed (r > 0.8, P < .01). In addition, formation of each glucuronide was also strongly correlated with N-glucuronide formation from trifluoperazine, which is a UGT1A4 probe substrate (25OHD₃-25-glucuronide, r = 0.93; 25OHD₃-3-glucuronide, r = 0.82; t-25OHD₃-25-glucuronide, r = 0.69) (data not shown).

Figure 4.

Interindividual difference in 25OHD₃ monoglucuronide formation and inductive expression of UGT1A3 and UGT1A4 genes in human hepatocytes. A, Twenty HLMs isolated from different donors were incubated with 25OHD₃ (10μM) as described above. Three monoglucuronides 25OHD₃-25-glucuronide (25OHD₃-25-G), 25OHD₃-3-glucuornide (25OHD₃-3-G), and 5,6-trans-25OHD₃-25-glucuronide (t-25OHD₃-25-G) were detected, and the formation rates represent the total amounts of product formed per minute per milligram of protein. B, Effects of the UGT1A4*3 genotype on the rates of 25OHD₃ glucuronidation. UGT1A4*3 (Leu48Val) variant gene product exhibits an increased glucuronidation activity. Subject (n = 80) characteristics and UGT1A4*3 genotype are summarized in Table 2, with 43 wild type (TT), 32 heterozygous (TG), and 5 homozygous (GG). The major product, 25OHD₃-25-glucruonide, was quantified using LC-MS, and the formation rate represents the total amounts of product generated per minute per milligram of protein. Statistical analysis was performed using one-way ANOVA with Dunnett's multiple comparison test; *, P < .05 was considered statistically significant. C, Inductive expression of UGT1A3 and UGT1A4 genes in human hepatocytes. Human hepatocytes from 3 different donors were treated with 10μM rifampin (RIF), 400μM phenobarbital (PB), 0.5μM hyperforin (HF), 50μM carbamazepine (CBZ), 200μM levitiracetam (LEV), or vehicle control (CTR, 0.1% vol/vol) for 48 hours, as indicated. Total RNA from each sample was isolated, and the expression of UGT1A3 and UGT1A4 was determined by quantitative RT-PCR assay. Data represent mean ± SE (from 4 replicate determinations of 3 different liver donors) of the fold induction in treated cells, compared with those of vehicle treated cells, after normalization to the 18s ribosomal RNA level. Statistical analysis was performed using an unpaired t test. *, P < .05 for the inductive effect of drugs, compared with corresponding control group.

Effects of hepatic UGT1A4 polymorphism on 25OHD₃ monoglucuronide formation

Allele variants of the UGT1A4 gene, such as UGT1A4*3 (GG), are closely associated with interindividual UGT1A4 enzyme activity (26, 27). As reported above, UGT1A4 is the predominant catalyst of 25OHD₃ glucuronidation in human liver; thus, it was of interest to us to investigate whether the UGT1A4 polymorphism is associated with 25OHD₃ glucuronidation in vitro. Using 80 HLMs isolated from human livers genotyped for UGT1A4*3 (summarized in Table 2), the formation rates of 25OHD₃-25-glucuronide were compared. As shown in Figure 4B, the mean rate of 25-glucuronidation of 25OHD₃ was significantly higher, by 85% (P = .03), in microsomes from homozygous UGT1A4*3 (GG) livers, compared with the mean rate from wild-type (TT) livers. Although glucuronidation activity among the heterozygous livers (0.152 pmol/min·mg protein) was not significantly different from wild type, an increasing trend for a allelic variant-dose effect was observed, with wild type having the lowest activity (0.123 pmol/min·mg protein) and homozygous UGT1A4*3 livers having the highest activity (0.227 pmol/min·mg protein). In contrast, the sex and age of the liver donor were not associated with glucuronidation activity (data not shown).

Table 2.

Subject Characteristics According to UGT1A4*3 Genotype

UGT1A4*3 genotype	TT	TG	GG	All genotypes
Number of subjects	43	32	5	80
Age (y), median	48	43	44	44
Sex, male/female	26/16	17/15	3/2	46/33
Race/ethnicity
White	40	30	4	74
Black	1	2	1	4
Hispanic	1	0	0	1

Eighty subjects were randomly drawn out from the University of Washington Tissue Bank and from livers that are part of a St Jude Children's Research Hospital human tissue repository. UGT1A4*3 genotype was determined by DNA sequencing.

Induction of UGT1A3 and UGT1A4 gene expression and catalytic activity in human hepatocytes

Many drugs are potent inducers of drug-metabolizing enzymes in the liver. We have shown that induction of CYP3A4 by certain drugs that are pregnane X receptor (PXR) or constitutive androstane receptor (CAR) activators can accelerate the 4-hydroxylation of vitamin D (6). However, it is unknown whether UGT1A3 or UGT1A4 activity could be also induced by drugs such as rifampin, thereby accelerating phase II conjugation of 25OHD₃. Accordingly, cryopreserved human hepatocytes from 3 different donors were treated with rifampin, phenobarbital, hyperforin, and carbamazepine. Levetiracetam, a new generation antiepileptic drug, which is not a PXR/CAR ligand, was used as a negative control (Figure 4C). After 48 hours of exposure, both UGT1A3 and UGT1A4 mRNA contents were significantly elevated by rifampin (2.3- and 7.0-fold, respectively), phenobarbital (1.9- and 5.1-fold, respectively), and carbamazepine (1.6- and 2.3-fold, respectively). Hyperforin significantly increased the levels of UGT1A4 mRNA but not the UGT1A3 mRNA levels. As expected, neither UGT1A3 nor UGT1A4 was induced by levetiracetam. These results suggest that both UGT1A3 and UGT1A4 are inducible by PXR agonists.

In order to determine whether 25OHD₃ glucuronides are produced in human hepatocytes and whether their formation is inducible after treatment with PXR agonists, a 2-stage sequential human hepatocyte treatment experiment was performed. Rifampin was selected as a prototypical inducer because of its highest inductive capacity, as shown in Figure 5. After incubation of hepatocytes with 10μM rifampin or vehicle for 48 hours, the cells were washed and incubated with 2μM 25OHD₃ for 2, 4, and 24 hours. As shown in Figure 5A, the same 3 isomers of 25OHD₃ glucuronide observed in HLM incubations were detected in the human hepatocyte incubations. Interestingly, the rank order for product formation of glucuronides was time dependent, changing from a preference for 25OHD₃-25-glucuronide > 25OHD₃-3-glucuronide > t-25OHD₃-25-glucuronide to t-25OHD₃-25-glucuronide > 25OHD₃-3-glucuronide > 25OHD₃-25-glucuronide with increased exposure times between 2 and 24 hours. For a simple comparison, we summed the concentrations of these 3 identical peaks generated from each cell treatment. As shown in Figure 5B, formation of total glucuronides increased in a time-dependent manner. However, a statistically significant increase of glucuronide formation by rifampin pretreatment was only observed in hepatocytes treated with 25OHD₃ for 2 hours.

Figure 5.

Measurement of 25OHD₃ monoglucuronides in human hepatocytes and plasma. A, Representative chromatograms of 3 monoglucuronides in human hepatocytes treated with 25OHD₃. Solid line, cell culture medium extracts after a 24-hour incubation; dash line, cell culture medium extracts after a 2-hour incubation. Arrows indicate the presence of 3 25OHD₃ monoglucuronides. B, Time-dependent formation of 25OHD₃ glucuronides in human hepatocytes with or without rifampin (10μM) pretreatment. Statistical analysis was performed using an unpaired t test (rifampin vs vehicle treated at different time points). C, Representative chromatogram of 25OHD₃-3-glucuronide (m/z 575) in human plasma. Human plasma (0.5 mL) was subjected to protein precipitation, solid-phase extraction using Oasis WAX anion exchange columns, and then the eluates were dried and reconstituted for LC-MS analysis. Solid line, plasma extracts indicating the presence of 25OHD₃-3-glucuronide; dash line, reaction products from HLM incubations with 25OHD₃ indicating 3 major monoglucuronides.

The presence of endogenous 25OHD₃-3-glucuronide in human plasma and bile

The presence or absence of 25OHD₃ monoglucuronides in human plasma and bile was also determined. Potential 25OHD₃ glucuronides were extracted by an anion exchange cartridge from 3 human plasma samples and analyzed by LC-MS, individually. The product mixtures from HLM incubations were used as standards. As shown in Figure 5C, the representative chromatogram from plasma extracts clearly showed the presence of 25OHD₃-3-glucuronide in human plasma, with a mean concentration from these 3 samples of approximately 2nM. In contrast, 25OHD₃-25-glucuronide and t-25OHD₃-25-glucuronide were not detectable under the conditions employed. In addition, 25OHD₃-3-glucuronide was also detected, for the first time, in human bile. After a series of extraction procedures, bile extracts containing putative 25OHD₃ glucuronides were derivatized with PTAD and analyzed by LC-MS/MS (Figure 6A). Standard 25OHD₃-3-glucuronide (Figure 6B) and 25OHD₃-25-glucuronide (Figure 6C) were subjected to PTAD derivatization in parallel. The identity of 3- or 25-O-glucuronide was based on the presence or absence of the characteristic fragment ion m/z 474 (Figure 3) and retention time. The representative chromatograms from bile extracts clearly showed the presence of 25OHD₃-3-glucuronide, but not 25OHD₃-25-glucuronide, in human bile, which is in agreement with the fragmentation pattern of standard 25OHD₃-3-glucuronide (Figure 6).

Figure 6.

The presence of 25OHD₃-3-glucuronide in human bile and its binding to DBP. Representative chromatograms from human bile (A), 25OHD₃-3-glucuronide (B), and 25OHD₃-25-glucuronide (C) after PTAD derivatization. Human bile (5 mL) was subjected to liquid-liquid extraction and solid-phase extraction using Waters Oasis WAX anion exchange cartridges as described in Materials and Methods. The bile extracts were derivatized with PTAD and analyzed by LC-MS/MS. Putative 25OHD₃ monoglucuronides were detected by multiple reaction monitoring (MRM) channels of m/z 734 → 298 and 734 → 474. Specifically, m/z 734 → 474 is a characteristic fragment from 25OHD₃-3-glucuronide. The standards 25OHD₃-3-glucuronide and 25OHD₃-25-glucuronide were also analyzed under the same conditions. D, Binding of 25OHD₃-3-glucuronide and 25OHD₃ to rat plasma DBP. The competitive binding assay was performed as described in Materials and Methods. Various concentrations of standards 25OHD₃ and 25OHD₃-3-glucuronide were used in the assay, and the percentage of bound ³H-25OHD₃ was calculated after competitive displacement of ³H-25OHD₃ by addition of the tested compounds.

Binding of 25OHD₃-3-glucuronide to rat plasma DBP

The selective persistence of 25OHD₃-3-glucuronide in plasma suggested the possibility that it might have significant affinity for the DBP, restricting its clearance by the kidney, as occurs for 25OHD₃. To test this hypothesis, the concentration-dependent binding of 25OHD₃-3-glucuronide and 25OHD₃ to rat plasma DBP was measured by radioligand binding assay (28). As seen in Figure 6D, 25OHD₃-3-glucuronide exhibited high affinity for DBP that was essentially identical to that of 25OHD₃.

Discussion

25OHD₃ is the major circulating form of vitamin D₃ in humans. At steady state, blood concentrations represent a balance between its formation rate and clearance by several oxidative and conjugative processes. Thus, interindividual differences in the efficiency of any of these processes, including direct glucuronidation, may contribute to variation in circulating plasma concentrations of 25OHD₃ and, hence, hormone action. Although glucuronide conjugates of 25OHD₃ have been described for different model animals (13, 16), there is far less known about their identity and formation in humans. In this report, we describe the identification of 3 different monoglucuronide conjugates of 25OHD₃, their formation by UGT1A4 and UGT1A3, as well as a genetic and environmental source of interindividual variation.

Specifically, we demonstrated that 3 different glucuronide isomers of 25OHD₃ are produced by the human liver. As expected, conjugation occurred at both C-25 and C-3 positions of 25OHD₃. The kinetic constants (K_m and V_max) associated with these reactions were similar to that seen for other endogenous UGT substrates, including estradiol, bilirubin, and 1α,25(OH)₂D₃ (17, 31, 32). Although the K_m for 25OHD₃ conjugation reactions catalyzed by recombinant UGT1A4 was lower than that seen for HLMs, the need to use alamethicin to permeabilize the microsomal membrane and impaired access of cofactor and efflux of UGT product may explain the relatively modest difference in parameter estimates. Based on a simple comparison of intrinsic clearances, UGT1A4 is the principal catalyst of 25OHD₃ glucuronidation in human liver and in vivo.

Interestingly, our results also suggested formation of 5,6-trans-25OHD₃-25-glucuronide. Whether this process is enzymatic or nonenzymatic is still unknown. Because 25OHD₃ is not likely to be converted spontaneously to 5,6-trans-25OHD₃ (30), t-25OHD₃-25-glucuronide may derive through isomerization of a precursor conjugate (ie, 25-O-glucuronide). Interestingly, the putative 5,6-trans-25OHD₃-25-glucuronide was the minor product in the 2- and 4-hour incubations but increased disproportionately to become the dominant product after 24 hours. This would suggest that it arises secondarily from one of the other glucuronide isomers through an enzyme-catalyzed isomerization process. However, it also raises the question as to whether other secondary metabolites not detected by our analytical methods are produced over time, resulting in an underestimate of the initial rate of product formation in hepatocytes. This could potentially explain the absence of evidence of induction of total hepatocyte activity after 4 and 24 hours of incubation with 25OHD₃.

In humans, dozens of UGTs catalyze conjugation of both endogenous and xenobiotic molecules. Glucuronide conjugation of any single substrate is often catalyzed by multiple UGT isozymes, due to their broad substrate specificity. However, we found marked enzyme selectivity, with production of the 3 isomeric 25OHD₃ glucuronide conjugates catalyzed exclusively by UGT1A4 (primarily) and UGT1A3 (secondarily). Previously, we showed that glucuronidation of 1α,25(OH)₂D₃ is catalyzed by UGT1A4, UGT2B4, and UGT2B7, with the greatest efficiency shown by UGT1A4. The same enzymatic preference was seen with 25OHD₃ conjugation, and thus, interindividual variation in UGT1A4 activity could affect the elimination of both hormone species and have perhaps a synergistic effect on vitamin D-mediated biological responses. One example where this could occur is with individuals who are homozygous for a “gain-of-function” allelic variant of UGT1A4. Previous investigators in the field have shown that the UGT1A4*3 (Leu48Val) variant confers enhanced clearance of the drug olanzapine (27, 33), through increased N-glucuronidation of the drug (26), a reaction known to be catalyzed by UGT1A4. Based on these findings and our demonstration that the UGT1A4*3 allele is associated with enhanced 25OHD₃ glucuronidation activity, one might anticipate lower circulating levels of 25OHD₃ in the blood of individuals carrying the variant allele, compared with those homozygous for the wild-type allele, all other factors being equal. This prediction can be readily tested in a large enough population.

Relatively little is known about the regulation of UGT1A4 and UGT1A3 in humans. However, there is clear evidence that UGT1A4 activity is increased substantially during pregnancy (34, 35), possibly through the action of elevated estrogen concentrations (36). Our demonstration that UGT1A4 is the principal catalyst of 25OHD₃ glucuronidation raises the question of whether or not there is enhanced 25OHD₃ conjugation during pregnancy. Interestingly, total plasma 1α,25(OH)₂D₃ concentrations increase during pregnancy (37) as a result in part of enhanced renal 1α-hydroxylase activity. However, 25OHD₃ deficiency during pregnancy remains a clinical problem and is associated with adverse maternal and fetal outcomes (38–40). Thus, homozygosity for the UGT1A4*3 allele might be a significant risk factor for vitamin D deficiency in pregnancy.

Exposure to xenobiotics has also been shown to affect UGT1A4 expression and thus might influence 25OHD₃ homeostasis. Of note, agonists of the aryl hydrocarbon receptor, such as dioxin, can enhance gene transcription (34, 41). There is also evidence to suggest that PXR or CAR agonists up-regulate UGT1A4 expression (42). In our investigation, we found that treatment of human hepatocytes with rifampin, phenobarbital, carbamazepine, or hyperforin all resulted in elevated UGT1A4 mRNA content, suggesting an enhancement of gene transcription through a nuclear receptor-mediated pathway. However, as noted above, full concordance between this effect and increased 25OHD₃ glucuronidation was not evident. Confounding by time-dependent sequential metabolism of the initial 3- or 25-O-glucuronide conjugates, or discordance between UGT mRNA levels and protein contents/activities, may have contributed to this discrepancy. Further work is needed to resolve the issue.

The detection of significant concentrations of endogenous 25OHD₃-3-glucuronide in human plasma and bile suggests that glucuronidation may be an important pathway of 25OHD₃ clearance in vivo. Excretion of the conjugates across that basolateral membrane of hepatocytes by perhaps one of the multidrug resistance-associated protein transporters after their intracellular formation would explain their presence in human blood. There is strong evidence for the formation and biliary excretion of 25OHD₃-glucuronide(s) from studies that followed the disposition of radiolabeled iv 25OHD₃ in humans (12). Our results demonstrate the presence of 25OHD₃-3-glucuronide, but not 25OHD₃-25-glucuronide and t-25OHD₃-25-glucuronide, in human bile. Detection of only 25OHD₃-3-glucuronide may reflect differential secretion of 25OHD₃ glucuronides across basolateral and canalicular membranes and conversion of 25OHD₃-25-glucuronide to t-25OHD₃-25-glucuronide. It may also be the indirect result of high-affinity binding of 25OHD₃-3-glucuronide to DBP, which would be expected to extend its half-life in blood and perhaps permit hepatic reuptake and biliary clearance. The absence of 25OHD₃-25-glucuronide and t-25OHD₃-25-glucuronide in bile may also be due to metabolic instability in vivo via an unknown catalytic pathway, which needs further investigation. More importantly, we postulate that the secretion of 25OHD₃-3-glucuronide into bile may be the first step in a paracrine signaling loop that might help regulate intestinal VDR gene targets, such as TRPV6 and CYP3A4 (21). These findings raise the possibility that interindividual differences in the efficiency of 25OHD₃ glucuronidation by UGT1A4 might contribute to variability in the efficiency of intestinal calcium absorption and CYP3A4-dependent drug metabolism.

In summary, we have identified UGT1A4 and UGT1A3 as the principal catalysts of 25OHD₃ glucuronidation in human liver. The enzymes generate 3 monoglucuronide regioisomers: 25OHD₃-25-glucronide, 25OHD₃-3-glucuronide, and 5,6-trans-25OHD₃-25-glucuronide. Formation of these metabolites is affected by a common gain-of-function polymorphism (UGT1A4*3) and may be inducible by PXR/CAR agonists, such as rifampin, carbamazepine, and phenobarbital. Moreover, the presence of the dominant glucuronide product in human plasma and bile suggests that the metabolic pathway is significant and may contribute to homeostatic control of vitamin D effects in the body.