Abstract
Glucuronidation is an important pathway for human drug metabolism. Four cloned and expressed human UDP-glucuronosyltransferases (UGT1A1, UGT1A6, UGT1A9, and UGT2B15) were used to screen a series of three potential drug substrates differing only in position of the phenol moiety. The meta and para phenols, UK-156,037 and UK-157,147, were found to be substrates for UGT1A1 withKm values of 256 and 105 μM, respectively. The ortho phenol UK-157,261 was glucuronidated predominantly by UGT1A9 with a Km of 45 μM. The latter Km compares favorably with the known UGT1A9 substrate propofol (Km= 200 μM). In a series of competition experiments, UK-157,261 was shown to inhibit the glucuronidation of propofol by UGT1A9 with aKi value of 65 μM. This result indicates that even the most potent of these compounds is extremely unlikely to interact in the clinic with the glucuronidation of propofol. This study shows the utility of the expressed human UDP-glucuronosyltransferases in determining substrate structure-activity relationships and potential drug-drug interactions.
Glucuronidation is an important route of drug metabolism in humans with a number of therapeutic agents cleared by this pathway (Clarke and Burchell, 1994). The UDP-glucuronosyltransferase isoforms in humans and common laboratory species are not identical, and it would be of great value to identify the specific human UGTs1involved where glucuronidation has been established as a pathway for drug detoxification in laboratory species (Burchell et al., 1995). Previously, evaluation of human drug metabolism has been studied by the use of microsomal preparations from human tissues but identification of the isoform(s) responsible is complicated by overlapping substrate specificity and the lack of specific UGT inhibitors. Further complications are caused by variability in the quality of tissue and the effects of environmental factors such as diet and prior drug therapy on the expression of drug-metabolizing enzymes in human tissue. Many forms of UDP-glucuronosyltransferase have now been cloned from a variety of species, including at least 19 human forms of the enzyme (Mackenzie et al., 1997).
The use of cloned and expressed human drug-metabolizing enzymes has become a widely used alternative for the study of human drug metabolism (Guengerich et al., 1997). Comparison of the kinetic parameters of substrate turnover can be used to assess the potential risk of interaction between substrates at the single enzyme level where sufficient information on the substrate specificities of the responsible drug-metabolizing isoforms is available. However there are many variables to consider when scaling from simple in vitro experiments to the in vivo situation (Remmel and Burchell, 1993). Simple rate based measurements for UGT turnover have been used in the past for a basic evaluation of the contribution of UGT isoforms to the glucuronidation of drugs (Wooster et al., 1993) This article presents a full in vitro kinetic treatment for substrate competition at the single enzyme level.
Four cloned and expressed human UDP-glucuronosyltransferases were used to screen three compounds from Pfizer Central Research, Sandwich, Kent, UK. The series of compounds known as UK-157,147, UK-156,037, and UK-157,261 are all structural isomers (Fig.1). The structure includes a phenol group and the difference between the compounds is the position of the phenolic hydroxyl group. UK-157,147 contains a meta phenol, UK-156,037 a para phenol, and UK-157,261 an orthophenol group. The human UGTs used were UGT1A1, UGT1A6, UGT1A9, and UGT2B15. The UGT isoforms were selected on the basis of previous reports that indicated that they were all capable of glucuronidating xenobiotic compounds, including 17α-ethinylestradiol (UGT1A1;Ebner et al., 1993), paracetamol (UGT1A6; Bock et al., 1993), propofol (UGT1A9; Ebner and Burchell, 1993) as well as a large number of other xenobiotics such as phenols, coumarins, and anthraquinones. UGT2B15 is also reported to glucuronidate simple phenolic compounds as well as endobiotic steroids (Green et al., 1994). Before this study, work completed at Pfizer Central Research in Sandwich had identified only a single phenolic glucuronide conjugate formed in human liver microsomes in vitro.
Materials and Methods
Tissue Culture.
V79 and recombinant cell lines were grown up in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 100 units/ml penicillin, and 0.1 mg/ml streptomycin. Cell cultures were grown in 75-cm2 flasks (CoStar, Cambridge, MA) fitted with vented caps in humidified incubators at 37°C with the atmosphere maintained at 5% CO2. V79 cells heterologously expressing human UGTs were maintained under optimized constant selection concentrations of geneticin (G418; GIBCO, Paisley, Scotland). These were V79/UGT1A1, 1 mg/ml; V79/UGT1A6, 100 μg/ml; and V79/UGT1A9, 200 μg/ml.
HEK293 cells expressing UGT2B15, used by kind permission of Dr. T. Tephly, University of Iowa, Iowa City, IA, were grown up in the above-described media supplemented with 10 mM HEPES (pH 7.4) and 700 μg/ml geneticin. The cloning and expression of these UGT isoforms are reported elsewhere (Fournel-Gigleux et al., 1990; Wooster et al., 1991; Sutherland et al., 1992; Green et al., 1994).
Standardized UGT Activation by Sonication.
Cells were disrupted by a standard optimized sonication method. Pellets containing cells harvested from two 75-cm2 tissue culture flasks were thawed before assaying and resuspended in 200 μl of phosphate-buffered saline, pH 7.4. Each 200-μl suspension of cells was sonicated for four 5-s bursts (Microson ultrasonic cell disruptor; Heat Technologies, Farmingdale, NY), allowing at least 1 min on ice between bursts. Aliquots of cells prepared by this method were pooled before addition to the assays.
Activation of the human liver microsomes was also achieved by sonication, which was optimized to the number of 5-s bursts that yielded optimal activity. As with the cellular sonication, this was found to be four 5-s bursts with a minute on ice between sonication treatments.
UGT Assays.
UGT assays were performed by an adaptation of a previously published method (Ethell et al., 1998). Assays were incubated in 100 mM Tris/maleate buffer (pH 7.4) containing 5 mM MgCl2, typically 500 μM substrate, and 350 to 500 μg of cellular sonicate in a total volume of 100 μl. Screening assays were incubated for 60 min at 37°C at two concentrations of UDPGA: 50 μM and 2 mM (containing 0.1 μCi [14C]UDPGA; NEN DuPont, Stevenage, Hertfordshire, UK). UDPGA (50 μM) assays allow higher levels of radiolabel incorporation into the glucuronide, which enhances assay sensitivity making this a useful tool for screening. Kinetic determinations were performed at 2 mM UDPGA or higher. The compounds screened were UK-157,147, UK-156,037, and UK-157,261 (Pfizer Central Research). Each batch of screening assays was accompanied by positive controls for the cell lines: UGT1A9, 500 μM propofol (Aldrich, Gillingham, Dorset, UK); UGT1A6, 500 μM 1-naphthol (Fluka, Gillingham, Dorset, UK); UGT1A1, 250 μM 17α-ethinylestradiol (Sigma, Gillingham, Dorset, UK); and UGT2B15, 500 μM 8-hydroxyquinoline (BDH, Poole, Dorset, UK). The unknown compounds were also assayed with human liver microsomes to establish a retention time for the radiolabeled conjugate on the radiochemical HPLC system.
Incubations were terminated by the addition of 100 μl of methanol that had been prechilled to −20°C. The precipitated proteins were removed by centrifugation at 1000g. The resulting supernatant was then transferred to an HPLC vial and 150 μl of this volume directly injected onto a Shimadzu HPLC, which consisted of two LC-6A pumps, and an SCL6B system controller/autosampler (donated by Pfizer Central Research). Data were collected using JCL6000 for Windows software (Jones Chromatography, Hengoed, Wales, UK). The gradient conditions were changed from the original method. The linear gradient from 0 to 100% acetonitrile in 0.05 M ammonium acetate was reduced from 15 to 13 min to reduce the total cycle time between injections. The column used was changed to a Techsphere 5ODS2 (HPLC Technology, Macclesfield, Cheshire, UK), which gave comparable separation to the Spherisorb 5ODS2 column (HPLC Technology) used previously.
Radioactive UDPGA and glucuronides were detected using model 9701 radioactivity monitors (Reeve Analytical, Glasgow, Scotland, UK) fitted with a 200-μl flow cell packed with silanized cerium activated lithium glass as scintillant (GS1/TSX; Reeve Analytical).
Kinetic parameters were determined using the same assay conditions with a shorter incubation time of 40 min. The range of substrate concentrations used was typically 10, 50, 100, 250, 500, and 1000 μM (assays performed in duplicate). Kinetic parameters were calculated using a nonlinear regression method (Prism 2; GraphPad Software, San Diego, CA).
Inhibition of Propofol Glucuronidation by UK-157,261.
Propofol glucuronidation has been found to be carried out by UGT1A9 (Ebner and Burchell, 1993). The effect of increasing concentrations of UK-157,261 (0–250 μM) on this reaction was studied. The propofol concentration range studied was 50 to 750 μM. Incubations were carried out 37°C for 40 min in the presence of 2 mM [14C]UDPGA (approximately 0.1 μCi). Incubations were terminated as described above. However, propofol and UK-157,261 glucuronides were found to coelute on the gradient system and an isocratic HPLC system was developed to separate them. The chromatography conditions were acetonitrile/0.05 M ammonium acetate pH 5.2 (25:75, v/v) at a flow rate of 1 ml/min. [14C]UDPGA and glucuronide conjugates were detected as in the gradient HPLC assays. The propofol glucuronide eluted at 6.4 min and the UK-157,261 glucuronide eluted at 5.3 min. The total run time for each injection was 8 min. All assays were performed in duplicate.
Effect of Saccharic Acid 1,4-Lactone.
It was not know whether β-glucuronidase was present in the cell lines used for the screening and kinetic assays. Assays were performed as previously described in the presence and absence of 10 mM saccharic acid 1,4-lactone, an inhibitor of β-glucuronidase (Sigma). Comparing the activities from assays that contain saccharic acid 1,4-lactone with those that do not allows an assessment of whether β-glucuronidase (if present) can affect the rate of glucuronidation measured in these assays.
Protein Determination.
Protein concentrations were measured using the method of Lowry et al. (1951) using bovine serum albumin as standard.
Results
Preliminary rates of glucuronidation of UK-157,147, UK-156,037, and UK-157,261 in cell lines expressing UGT1A1, UGT1A6, UGT1A9, UGT2B15, and human liver microsomes are shown in Table1. Cell line incubations were performed at 50 μM and 2 mM UDPGA concentration. All three UK compounds were glucuronidated in human liver microsomal incubations with the rate of glucuronide production being highest for UK-156,037. Incubations with human liver microsomes only produced a single conjugate, which eluted at the same time as the most abundant glucuronide conjugate detected in assays with expressed UGT1A1. The phenolic glucuronide had previously been identified as the only product in assays with human liver microsomes (data not shown). This is perhaps not surprising considering that the phenolic position is less sterically hindered and is more nucleophilic than the secondary aliphatic hydroxyl group. Because the major glucuronide conjugate that eluted in the UGT1A1 assays elutes with the previously identified phenol glucuronide from assays with human liver microsomes, it was assumed that the earlier eluting peak was indeed the phenol glucuronide and the later eluting peak was most likely to be the glucuronide of the secondary aliphatic alcohol.
Assays that contained saccharic acid 1,4-lactone gave no significant difference in activity from those that did not. Comparing assays with each cell line in the presence and absence of inhibitor the differences were insignificant: UGT1A1 and UGT1A9 showed less activity in assays that contained saccharic acid 1,4-lactone (−1 and −4%, respectively) than those that did not. UGT2B15 expressed in HEK293 cells showed a slight increase in the presence of the inhibitor (+1.6%). All the differences in activity are < ±5%. Variation of experimental results within these limits is expected and as such it appears unlikely that β-glucuronidase (if present in these cell lines) has any effect on the assays themselves.
Incubation of known substrates for each UGT isoform (17 α-ethinylestradiol for UGT1A1, 1-naphthol for UGT1A6, propofol for UGT1A9, and 8-hydroxyquinoline for UGT2B15) showed that the preparations were all capable of glucuronidation. However, there was a marked isoform specificity for the glucuronidation of the three Pfizer compounds. UGT1A6 did not glucuronidate any of the UK compounds tested. UGT2B15 did not glucuronidate UK-157,147 or UK-157,261 but was shown to produce the glucuronide of UK-156,037 at a very low rate. UGT1A9 also produced glucuronides of UK-157,147 and UK-156, 037 at low rates but only when 50 μM UDPGA was used. At 2 mM UDPGA, no glucuronides could be detected for these two compounds. This is most likely due to the dilution effect of the [14C]UDPGA with unlabeled UDPGA, rendering detection of low-level glucuronides difficult. In contrast to UK-157,147 and UK-156,037, UK-157,261 was glucuronidated at a significant rate by UGT1A9. UGT1A1 did not glucuronidate UK-157,261 but experiments with UK-157,147 and UK-156,037 showed the presence of two glucuronides, both produced at significant rates by UGT1A1. The significance of these two glucuronides is not known although there are several potential sites of glucuronidation on these molecules (both aliphatic hydroxyl and phenol). The likelihood is that UGT1A1 is not present in human liver microsomes at sufficient concentrations to allow the production of both glucuronides as seen with expressed UGT1A1.
Overall, these preliminary rate experiments showed that UGT1A1 was capable of glucuronidating UK-157,147 and UK-156,037, whereas the isoform most likely to be responsible for the metabolism of UK-157,261 is UGT1A9. These combinations were taken forward for full kinetic evaluation.
The kinetic parameters for UK-157,147 and UK-156,037 with UGT1A1 and for UK-156,261 with UGT1A9 are shown in Table2. Representative chromatograms are shown in Fig. 2. ApparentKm values for UK-157,147 and UK-156,037 against UGT1A1 were determined as the sum of the two glucuronide peaks detected and were 105 and 256 μM, respectively. TheKm value for UK-157,261 versus UGT1A9 was 45 μM. The ratioVmax/Km for each substrate gives an indication of the intrinsic clearance of that substrate by that isoform. This is a useful parameter for extrapolating from the in vitro to the in vivo situation. The intrinsic clearance of UK-157,261 by UGT1A9 was more than 3-fold that for UK-157,147 and UK-156,037 by UGT1A1.
Following on from the above data, literature evidence was reviewed to examine the potential for these UK compounds to interact with the glucuronidation of standard substrates. Two standard substrates for UGT1A1 are bilirubin and 17 α-ethinylestradiol, which are reported to have Km values with UGT1A1 of 25 μM (Senafi et al., 1994) and 55 μM (Ethell et al., 1998). Because theKm values for UK-157,147 and UK-156,037 versus UGT1A1 are in excess of these values, it was thought unlikely that they would reduce the glucuronidation of these standard substrates and no further work was carried out with these compounds. Propofol is a standard substrate for UGT1A9 with a reportedKm of 200 μM (Ebner and Burchell, 1993). Because the Km for UK-157,261 for UGT1A9 was 45 μM, a reduction of the glucuronidation of propofol by UK-157,261 was possible. Therefore, competition experiments were carried out and typical chromatograms are shown in Fig.3. The chromatogram show glucuronide formed from assays using a fixed concentration of propofol and increasing concentrations of the competing UK-157,261. The peak that elutes first is the 14C-labeled UK-157,261 and the later eluting peak is the 14C-labeled propofol. These chromatograms clearly indicate a marked decrease in the propofol glucuronidation as UK-157,261 concentration was increased, coupled to a corresponding increase in the levels of UK-157,261 glucuronide present.
The data from these chromatograms and from the competition experiments at all other propofol concentrations are indicated in Fig.4. These graphs clearly indicate that UK-157,261 glucuronidation can interact significantly with propofol glucuronidation. At the two lowest propofol concentrations (50 and 100 μM) UK-157,261 glucuronide becomes the major conjugate. As expected the level of inhibition increases with the concentration of UK-157,261. At 50 μM UK-157,261 the mean rate of propofol glucuronidation was reduced to 70 ± 5.4% of the control rate, whereas at 100 μM the mean rate was 59 ± 5.7% and at 200 μM 38 ± 9.3%. The errors are expressed as the standard deviation of the mean reduction.
A Dixon plot of the inhibition of propofol glucuronidation by the competing UK-157,261 is shown in Fig. 5. The convergence of the lines indicates a Kivalue of approximately 65 μM, which is consistent with theKm values determined independently from the competition study. The form of the graph is characteristic of simple competitive inhibition as predicted by the glucuronidation of both of these compounds by UGT1A9.
Discussion
The UGT isoforms responsible for the glucuronidation of three Pfizer compounds have been determined. The meta andpara phenols (UK-157,147 and UK-156,037, respectively) were shown to be substrates for UGT1A1, whereas the ortho phenol (UK-157,261) was not a substrate for this isoform. UK-157,261 was shown to be glucuronidated by UGT1A9 with the former compounds poor substrates for this isoform. Thus, the switch of the phenol from the less sterically hindered meta and para positions to the more sterically hindered ortho position has the effect of switching the isoform from UGT1A1 to UGT1A9. Interestingly, the presence of bulky groups adjacent to the phenol is also a feature of the propofol structure (2,6-diisopropylphenol) and marks this compound out as a specific substrate of UGT1A9.
Comparison of the activities toward these compounds in the cell lines with those in human liver microsomes shows that the activities for UK-157,147 and UK-156,037 with UGT1A1 are consistent with the activities seen in human liver microsomes. However the UGT1A9 activity toward UK-157,261 is much greater in the cell line assays than those measured in assays with human liver microsomes. The reason for this may be the relative expression levels of UGT1A9 in various tissues.Sutherland et al. (1993) have shown that UGT1A9 mRNA levels in human kidney are up to 3-fold that observed in human liver. In addition, glucuronidation of propofol (a UGT1A9 substrate that is specific to UGT1A9 in human liver and kidney) has been observed in microsomes prepared from human kidney (Sutherland et al,. 1993; McGurk et al., 1998). The lower activity toward UK-157,261 being due to lower levels of expressed UGT1A9 in human liver microsomes is supported by comparing the glucuronidation rates of propofol in the two UGT sources. Propofol activity in the microsomes is 4.5 times lower in human liver microsomes than in expressed 1A9 (HLM, 0.20 nmol/min/mg; UGT1A9, 0.89 nmol/min/mg) This compares remarkably well to the difference in glucuronidation rate for UK-157,261 of 4.4 times lower (HLM, 0.331 nmol/min/mg; UGT1A9, 0.075 nmol/min/mg) in HLM than expressed UGT1A9. These observations have implications for the evaluation of glucuronidation using in vitro systems because the utility of microsomal incubations will depend upon the substrate specificity of the drug studied and the tissue from which the microsomes are prepared. Such tissue differences in the expression of UGT isoforms highlights the potential utility of the individually expressed isoforms used in this investigation. This article describes the screening of three compounds across a series of UGT isoforms and the rapid identification of the major isoforms responsible for the glucuronidation of each compound. Reports on the substrate specificities of other cloned and expressed human UGTs indicate that other isoforms are capable of glucuronidating xenobiotics such as UGT1A3 (Mojarabbi et al., 1996), UGT1A4 (Green et al., 1995), UGT1A10 (Mojarabbi and MacKenzie, 1997), and UGT2B7 (Jin et al., 1993;Coffman et al., 1997). Ideally, a more comprehensive set of UGT isoenzymes could be used for this screening process to eliminate the possibility of contribution of other UGT isoforms to the metabolism of drugs.
The kinetic data presented here allow likely interactions between substrate drugs to be predicted. The relativeKm values for UK-157,147 and UK-156,037 against UGT1A1 suggest that the likelihood of any interaction between the UGT1A1 substrates and bilirubin or 17 α-ethinylestradiol glucuronidation is low. Bilirubin has a reported apparentKm of 25 μM against UGT1A1 (Senafi et al., 1994). UGT1A1 also glucuronidates the synthetic steroid 17α-ethinylestradiol with a Km of 55 μM (Ethell et al., 1998). The Km values for UK-157,147 and UK-156,037 exceed those for bilirubin and 17α-ethinylestradiol and consequently an interaction is unlikely.
There is a potential for interaction between compound UK-157,261 and the anesthetic compound propofol. The Kmfor UK-157,261 against UGT1A9 is 45 μM and theKm for propofol has been shown to be around 200 μM (Ebner and Burchell, 1993). Thus, UK-157,261 has the potential to inhibit the glucuronidation of propofol in vitro. Additional information to illustrate this interaction was produced by the in vitro competition study with propofol. The steady-state plasma concentration of propofol after an infusion of 9 mg/kg/h has been estimated at 6 mg/l (34 μM) and 88% of the administered dose is excreted as a glucuronide conjugate in the urine (Langley and Heel, 1988). If the plasma concentrations of UK-157,261 reach a similar concentration this could lead to a significant reduction in the capacity of UGT1A9 to glucuronidate propofol and hence result in prolonged anesthesia. However, this is most unlikely to occur in the clinic because the therapeutic concentrations of UK-157,261 (low nM) are unlikely to approach those required for inhibition of propofol glucuronidation. Other interactions have been reported in vitro with human liver microsomes. It was shown that several drugs could inhibit propofol glucuronidation in vitro in experiments where the competing drug was present at either 0.5 or 5 mM. These drugs were oxazepam (18% of activity at 5 mM), ketoprofen (17% of activity at 5 mM), acetylsalicylic acid (16% of activity at 5 mM), and fentanyl (2% of activity at 5 mM) (Le Guellec et al., 1995). The concentrations of UK-157,261 in the competition studies only reached 250 μM and were capable of inhibiting propofol glucuronidation to 38 ± 9% of activity.
The present study has shown the potential of individually expressed isoforms of UGTs in the study of glucuronidation. It has proved possible to rapidly screen three compounds against a range of human isoforms to identify the isoforms responsible for the glucuronidation of these potential drug molecules. Once the isoforms responsible were identified, further kinetic evaluation was undertaken to assess the in vitro intrinsic clearance of the compounds studied. From this kinetic data the potential of these drugs to interact with other isoform substrates has been assessed and an in vitro interaction with propofol demonstrated. It is hoped that these cell line-expressed UGTs will be more widely used in future to examine structure-substrate specificity for glucuronidation.
Acknowledgments
This work was funded by an Medical Research Council collaborative studentship with Pfizer Central Research, Sandwich, Kent, UK, and the Wellcome Trust (to B.B.).
Footnotes
-
Send reprint requests to: Brian T. Ethell, Department of Molecular and Cellular Pathology, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY, Scotland. E-mail:b.ethell{at}dundee.ac.uk
- Abbreviations used are::
- UGT
- UDP-glucuronosyltransferase
- UDPGA
- UDP-glucuronic acid
- HPLC
- high-performance liquid chromatography
- HLM
- human liver microsome
- Received July 13, 2000.
- Accepted September 21, 2000.
- The American Society for Pharmacology and Experimental Therapeutics