Glucuronidation of 3'-azido-3'-deoxythymidine catalyzed by human liver UDP-glucuronosyltransferase. Significance of nucleoside hydrophobicity and inhibition by xenobiotics

The enzymatic glucuronidation of 3'-azido-3'-deoxythymidine (AZT) catalyzed by human liver microsomal UDP-glucuronosyltransferase (EC 2.4.1.17, UDPGT) was inhibited by a number of nucleoside analogs. The inhibitory potency of these nucleoside analogs correlated with their hydrophobicity (r2 = 0.90, N = 13). Since similar results were obtained with solubilized UDPGT (r2 = 0.87, N = 7), the affinity of the nucleosides for UDPGT was probably being assessed rather than the ability of the compounds to access the membrane-bound enzyme. Three homologous inhibitors, 3'-azido-2',3'-dideoxyuridine (AzddU), 5-ethyl-AzddU, and 5-propyl-AzddU, were also studied as substrates of UDPGT. The substrate efficiency (Vmax/Km) of these three compounds and AZT also correlated with their hydrophobicity (r2 = 0.94). Sixteen drugs that are structurally unrelated to nucleosides also inhibited the glucuronidation of AZT. The mechanism of inhibition was competitive for seven compounds tested. Ki values were estimated from Dixon plots for nine other less soluble inhibitors; their mechanism of inhibition was assumed to be competitive. Since the peak physiological drug concentrations of the tested inhibitors are considerably less than their Ki values, none of these compounds are expected to strongly inhibit AZT glucuronidation in humans. However, the rank order of these drugs with respect to their inhibitory potential is probenecid greater than chrloramphenicol greater than naproxen greater than phenylbutazone much greater than other drugs tested.     

Glucuronidation of 3'-azido-3'-deoxythymidine by rat and human liver microsomes

The glucuronidation of 3'-azido-3'-deoxythymidine (AZT) by rat and human liver microsomes has been studied in vitro. The AZT-glucuronide was preliminarily identified through specific hydrolysis by beta-glucuronidase and rigorous product identification was performed by high-field proton nuclear magnetic resonance and fast-atom-bombardment mass spectrometry. A beta-linked 5'-O-glucuronide was the exclusive product formed in liver microsomes. Rat and human liver microsomal uridine 5'-diphosphoglucuronyltransferase activities toward AZT were investigated. These studies revealed that AZT had a lower Km and a 5-6-fold higher relative catalytic efficiency for uridine 5'-diphosphoglucuronyltransferase in human as compared to rat liver microsomes which may play a role in the quantitative differences observed in the degree of AZT glucuronidation between rat and human.     

The effect of malaria infection on 3'-azido-3'-deoxythymidine and paracetamol glucuronidation in rat liver microsomes.

The effect of malaria infection on UDP-glucuronosyltransferase (UDPGT) activity was investigated in rat liver microsomes using 3'-azido-3'-deoxythymidine and paracetamol. The Michaelis-Menten parameters, Km and Vmax were calculated and intrinsic clearance values were estimated for normal and infected livers. The results show that malaria infection alters the activity of UDPGT.     

3'-azido-3'-deoxythymidine drug interactions. Screening for inhibitors in human liver microsomes.

Zidovudine is a widely used antiretroviral drug active against human immunodeficiency virus. The drug interactions of this compound, which are primarily eliminated as a glucuronide, have not yet been extensively studied. Because zidovudine is frequently combined with other drugs, complete knowledge of interactions is essential to optimize AIDS therapy. We therefore screened the effect of 55 molecules, representative of 20 different therapeutic classes, on 3'-azido-3'-deoxythymidine (AZT) glucuronidation by human liver microsomes. We demonstrate that many drugs caused more than 15% inhibition of AZT glucuronidation in vitro, whereas major antibiotics (ceftazidine, isoniazid, aminoglycosides, macrolides, and sulfamides), antivirals (2',3'-dideoxycytidine, 2',3'-dideoxyinosine, and acyclovir), flucytosine, metronidazole, acetaminophen, and ranitidine had no effect. For compounds that appeared to inhibit AZT glucuronidation, extrapolation to the clinical situation must take into account both the in vitro apparent Ki values and the usual expected plasma level for the coadministered drug. By considering these parameters, this work indicates that clinically relevant inhibition of AZT glucuronidation may be observed with the following drugs: cefoperazone, penicillin G, amoxicilin, piperacillin, chloramphenicol, vancomycin, miconazole, rifampicin, phenobarbital, carbamazepine, phenytoin, valproic acid, quinidine, phenylbutazone, ketoprofen, probenecid, and propofol. Complementary clinical and pharmacokinetic studies should be performed to validate these assumptions.     

Epirubicin glucuronidation is catalyzed by human UDP-glucuronosyltransferase 2B7.

Epirubicin is one of the most active agents for breast cancer. The formation of epirubicin glucuronide by liver UDP-glucuronosyltransferase (UGT) is its main inactivating pathway. This study aimed to investigate epirubicin glucuronidation in human liver microsomes, to identify the specific UGT isoform for this reaction, and to correlate epirubicin glucuronidation with other UGT substrates. Microsomes from human livers were used. UGTs specifically expressed in cellular systems, as well as two UGT2B7 variants, were screened for epirubicin glucuronidation. Epirubicin, morphine, and SN-38 glucuronides were measured by high-pressure liquid chromatography. The mean +/- S.D. formation rate of epirubicin glucuronide in human liver microsomes (n = 47) was 138 +/- 37 pmol/min/mg (coefficient of variation, 24%). This phenotype was normally distributed. We screened commercially available UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, UGT2B7, and UGT2B15 for epirubicin glucuronidation. Only UGT2B7 converted epirubicin to its glucuronide. No differences in epirubicin glucuronidation were found in HK293 cells expressing the two UGT2B7 variants at position 268. Catalytic efficiency (V(max)/K(m)) of epirubicin glucuronidation was 1.4 microl/min/mg, a value higher than that observed for morphine, a substrate of UGT2B7. Formation of epirubicin glucuronide was significantly related to that of morphine-3-glucuronide (r = 0.76, p < 0.001) and morphine-6-glucuronide (r = 0.73, p < 0.001). No correlation was found with SN-38, a substrate of UGT1A1 (r = 0.04). UGT2B7 is the major human UGT catalyzing epirubicin glucuronidation, and UGT2B7 is the candidate gene for this phenotype. The reported tyrosine to histidine polymorphism in UGT2B7 does not alter the formation rate of epirubicin glucuronide, and undiscovered genetic polymorphisms in UGT2B7 might change the metabolic fate of this important anticancer agent.     

Glucuronidation of 3'-azido-3'-deoxythymidine: human and rat enzyme specificity

Since preclinical studies indicated that 3'-azido-3'-deoxythymidine (AZT, zidovudine, Retrovir, BW A509U), a potent anti-HIV agent, is not metabolized extensively in rats, rabbits, mice, guinea pigs, cats, or dogs, the extensive biotransformation of AZT observed in humans was not expected. On average, approximately 75% of an oral AZT dose is recovered in human urine as a single metabolite while only 14-18% of the dose is recovered unchanged. Ultraviolet, infrared, nuclear magnetic resonance, and mass spectra and enzymatic degradation characterized the isolated major metabolite as a 5'-O-glucuronide (3'-azido-3'-deoxy-5'-beta-D-glucopyranuronosylthymidine, GAZT), a very unique nucleoside metabolite. These observations suggest that UDP-glucuronosyltransferase (UDPGT), EC2.4.1.17, mediates the in vivo biotransformation of AZT to GAZT. Since glucuronidation is one of the major conjugation reactions involved in the metabolic conversion of xenobiotics to more polar, water-soluble metabolites, it is an important detoxification pathway in humans. Therefore, it is important to understand the enzymatic basis for the discrepancy between metabolism of AZT in laboratory mammals and humans. This is especially relevant in light of the use of laboratory mammals to predict the metabolism of novel pharmaceutical agents in humans. The study presented herein confirms that liver UDPGT does catalyze the glucuronidation of AZT and that the higher substrate efficiency of AZT with human enzyme compared to rodent enzyme may account for metabolic differences observed in vivo.     

In vitro glucuronidation of thyroxine and triiodothyronine by liver microsomes and recombinant human UDP-glucuronosyltransferases.

Glucuronidation, which may take place on the phenolic hydroxyl and carboxyl groups, is a major pathway of metabolism for thyroxine (T4) and triiodothyronine (T3). In this study, a liquid chromatography/mass spectrometry (LC/MS) method was developed to separate phenolic and acyl glucuronides of T4 and T3. The method was used to collect the phenolic glucuronide of T4 for definitive characterization by NMR and to determine effects of incubation pH, species differences, and human UDP-glucuronosyltransferases (UGTs) involved in the formation of the glucuronides. Formation of T4 phenolic glucuronide was favored at pH 7.4, whereas formation of T4 acyl glucuronide was favored at pH 6.8. All the UGTs examined catalyzed the formation of T4 phenolic glucuronide except UGT1A4; the highest activity was detected with UGT1A3, UGT1A8, and UGT1A10, followed by UGT1A1 and UGT2B4. Formation of T3 phenolic glucuronide was observed in the order of UGT1A8 > UGT1A10 > UGT1A3 > UGT1A1; trace activity was observed with UGT1A6 and UGT1A9. UGT1A3 was the major isoform catalyzing the formation of T4 and T3 acyl glucuronides. In liver microsomes, phenolic glucuronidation was the highest in mice for T4 and in rats for T3 and lowest in monkeys for both T4 and T3. Acyl glucuronidation was highest in humans and lowest in mice for T4 and T3. Phenolic glucuronidation was higher than acyl glucuronidation for T4 in humans; in contrast, the acyl glucuronidation was slightly higher than phenolic glucuronidation for T3. UGT activities were lower toward T3 than T4 in all the species. The LC/MS method was a useful tool in studying glucuronidation of T4 and T3.     

Acyl glucuronidation of fluoroquinolone antibiotics by the UDP-glucuronosyltransferase 1A subfamily in human liver microsomes.

Acyl glucuronidation is an important metabolic pathway for fluoroquinolone antibiotics. However, it is unclear which human UDP-glucuronosyltransferase (UGT) enzymes are involved in the glucuronidation of the fluoroquinolones. The in vitro formation of levofloxacin (LVFX), grepafloxacin (GPFX), moxifloxacin (MFLX), and sitafloxacin (STFX) glucuronides was investigated in human liver microsomes and cDNA-expressed recombinant human UGT enzymes. The apparent Km values for human liver microsomes ranged from 1.9 to 10.0 mM, and the intrinsic clearance values (calculated as Vmax/Km) had a rank order of MFLX > GPFX > STFX > > LVFX. In a bank of human liver microsomes (n = 14), the glucuronidation activities of LVFX, MFLX, and STFX correlated highly with UGT1A1-selective beta-estradiol 3-glucuronidation activity, whereas the glucuronidation activity of GPFX correlated highly with UGT1A9-selective propofol glucuronidation activity. Among 12 recombinant UGT enzymes, UGT1A1, 1A3, 1A7, and 1A9 catalyzed the glucuronidation of these fluoroquinolones. Results of enzyme kinetics studies using the recombinant UGT enzymes indicated that UGT1A1 most efficiently glucuronidates MFLX, and UGT1A9 most efficiently glucuronidates GPFX. In addition, the glucuronidation activities of MFLX and STFX in human liver microsomes were potently inhibited by bilirubin with IC50 values of 4.9 microM and 4.7 microM, respectively; in contrast, the glucuronidation activity of GPFX was inhibited by mefenamic acid with an IC50 value of 9.8 microM. These results demonstrate that UGT1A1, 1A3, and 1A9 enzymes are involved in the glucuronidation of LVFX, GPFX, MFLX, and STFX in human liver microsomes, and that MFLX and STFX are predominantly glucuronidated by UGT1A1, whereas GPFX is mainly glucuronidated by UGT1A9.     

The UDP-glucuronosyltransferase 2B7 isozyme is responsible for gemfibrozil glucuronidation in the human liver.

Gemfibrozil, a fibrate hypolipidemic agent, is eliminated in humans by glucuronidation. A gemfibrozil glucuronide has been reported to show time-dependent inhibition of cytochrome P450 2C8. Comprehensive assessment of the drug interaction between gemfibrozil and cytochrome P450 2C8 substrates requires a clear understanding of gemfibrozil glucuronidation. However, the primary UDP-glucuronosyltransferase (UGT) isozymes responsible for gemfibrozil glucuronidation remain to be determined. Here, we identified the main UGT isozymes involved in gemfibrozil glucuronidation. Evaluation of 12 recombinant human UGT isozymes shows gemfibrozil glucuronidation activity in UGT1A1, UGT1A3, UGT1A9, UGT2B4, UGT2B7, and UGT2B17, with UGT2B7 showing the highest activity. The kinetics of gemfibrozil glucuronidation in pooled human liver microsomes (HLMs) follows Michaelis-Menten kinetics with high and low affinity components. The high affinity K(m) value was 2.5 microM, which is similar to the K(m) value of gemfibrozil glucuronidation in recombinant UGT2B7 (2.2 microM). In 16 HLMs, a significant correlation was observed between gemfibrozil glucuronidation and both morphine 3-OH glucuronidation (r = 0.966, p < 0.0001) and flurbiprofen glucuronidation (r = 0.937, p < 0.0001), two reactions mainly catalyzed by UGT2B7, whereas no significant correlation was observed between gemfibrozil glucuronidation and either estradiol 3beta-glucuronidation and propofol glucuronidation, two reactions catalyzed by UGT1A1 and UGT1A9, respectively. Flurbiprofen and mefenamic acid inhibited gemfibrozil glucuronidation in HLMs with similar IC(50) values to those reported in recombinant UGT2B7. These results suggest that UGT2B7 is the main isozyme responsible for gemfibrozil glucuronidation in humans.     

Metabolism of ketotifen by human liver microsomes. In vitro characterization of a tertiary amine glucuronidation

Biotransformation of ketotifen was investigated in vitro using human liver microsomes. Three of the four metabolic pathways observed in vivo in man were exhibited under the conditions of incubation, namely demethylation, N-oxidation, and N-glucuronidation, the absent route being the ketoreduction, which probably has a cytosolic localization. The kinetic parameters of the N-glucuronidation (KM for ketotifen and UDPGA and Vmax) were determined with native and detergent-treated microsomes. Treatment by Triton X-100 increased by about 3-fold the conjugation reaction. No sex difference was observed and N-glucuronidation did not seem to be inhibited either by bilirubin or by 4-nitrophenol. Thus, human liver microsomes are a useful and suitable in vitro model for studying metabolic routes, specific for man, as in the case of ketotifen. Obviously, the results obtained can only reflect partially the multiplicity of in vivo events and interpretation has to be complemented by investigations with other models.     

In vitro glucuronidation using human liver microsomes and the pore-forming peptide alamethicin.

The UDP-glucuronosyltransferases (UGTs) are a superfamily of membrane-bound enzymes whose active site is localized inside the endoplasmic reticulum. Glucuronidation using human liver microsomes has traditionally involved disruption of the membrane barrier, usually by detergent treatment, to attain maximal enzyme activity. The goals of the current work were to develop a universal method to glucuronidate xenobiotic substrates using microsomes, and to apply this method to sequential oxidation-glucuronidation reactions. Three assays of UGT catalytic activity estradiol-3-glucuronidation, acetaminophen-O-glucuronidation, and morphine-3-glucuronidation, which are relatively selective probes for human UGT1A1, 1A6, and 2B7 isoforms, respectively, were developed. Treatment of microsomes with the pore-forming peptide alamethicin (50 microg/mg protein) resulted in conjugation rates 2 to 3 times the rates observed with untreated microsomes. Addition of physiological concentrations of Mg(2+) to the alamethicin-treated microsomes yielded rates that were 4 to 7 times the rates with untreated microsomes. Optimized assay conditions were found not to detrimentally affect cytochrome P450 activity as determined by effects on testosterone 6beta-hydroxylation and 7-ethoxycoumarin deethylation. Formation of estradiol-3-glucuronide displayed atypical kinetics, and data best fit the Hill equation, yielding apparent kinetic parameters of K(m)(app) = 0.017 mM, V(max)(app) = 0.4 nmol/mg/min, and n = 1.8. Formation of acetaminophen-O-glucuronide also best fit the Hill equation, with K(m)(app) = 4 mM, V(max)(app) = 1.5 nmol/mg/min, and n = 1.4. Alternatively, morphine-3-glucuronide formation displayed Michaelis-Menten kinetics, with K(m)(app) = 2 mM and V(max)(app) = 2. 5 nmol/mg/min. Finally, alamethicin treatment of microsomes was found to be effective in facilitating the sequential oxidation-glucuronidation of 7-ethoxycoumarin.     

Quaternary ammonium-linked glucuronidation of tamoxifen by human liver microsomes and UDP-glucuronosyltransferase 1A4

Tamoxifen (TAM), a nonsteroidal antiestrogen, is the most widely used drug for chemotherapy of hormone-dependent breast cancer in women. In the present study, we found a new potential metabolic pathway of TAM via N-linked glucuronic acid conjugation for excretion in humans. TAM N(+)-glucuronide was isolated from a reaction mixture consisting of TAM and human liver microsomes fortified with UDP-glucuronic acid (UDPGA) and identified with a synthetic specimen by high-performance liquid chromatography-electrospray ionization-mass spectrometry. However, no TAM-glucuronidating activity was detected in microsomes from rat, mouse, monkey, dog, and guinea pig livers. A strong correlation (r(2) =0.92 ) was observed between N-glucuronidating activities toward TAM and trifluoperazine, a probe substrate for human UDP-glucuronosyltransferase (UGT) 1A4, in human liver microsomes from eight donors (five females, three males). However, no correlation ( (r(2) =0.02 )) was observed in the activities between 7-hydroxy-4-(trifluoromethyl)coumarin and TAM. Only UGT1A4 catalyzed the N-linked glucuronidation of TAM among recombinant UGTs (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, UGT2B4, UGT2B7, UGT2B15, and UGT2B17) expressed in insect cells. Apparent K(m) values for TAM N-glucuronidation by human liver microsomes and recombinant UGT1A4 were 35.8 and 32.4 microM, respectively. These results strongly suggested that UGT1A4 could play a role in metabolism and excretion of TAM without Phase I metabolism in human liver. TAM N(+)-glucuronide still had binding affinity similar to TAM itself for human estrogen receptors, ERalpha and ERbeta, suggesting that TAM N(+)-glucuronide might contribute to the biological activity of TAM in vivo.     

Doses and time-dependent effects of 3'-azido-3'-deoxythymidine on T47D human breast cancer cells in vitro

The objective of the present work was to study the effects of 3'-azido-3'-deoxythymidine (azidothymidine, Zidovudine) on human breast cancer cells by using, as a model, the T47D cell line (typified as oestrogen-dependent and p53-mutated). Low azidothymidine doses (3.125 microM) increase the percentage of cells in S-phase, with the effect reversing after 24 hr of incubation; as azidothymidine doses increase, the magnitude and duration of its effect increase proportionally, although, even with the highest concentrations (50-100 microM) the effects decline after 48 hr of incubation. If media (containing azidothymidine or vehicle) are daily renewed, the azidothymidine effects (accumulation of cells in S-phase) are higher and decline later than when media and drug are not changed during the whole culture period, thereby suggesting that the reversion of azidothymidine effects could be related with a degradation of the drug or accumulation in media of substances which counteract its effects. Azidothymidine inhibits T47D cell proliferation at concentrations higher than 50 microM. The exposure to 50 or 100 microM azidothymidine induced cell apoptosis after 48 hr or more of incubation. We conclude that: a) azidothymidine, with appropriate doses and duration of treatment, synchronizes cells in S-phase, inhibits proliferation, and induces apoptosis, b) the discontinuous application of the drug rather than continuous exposure to it increases its efficiency to synchronize the T47D cell cycle. This in vitro anti-breast cancer activity suggests that a possible clinical usefulness of azidothymidine, either alone or associated with other drugs with cycle-specific antitumoural activity circumscribed to the S-phase of cell cycle, is worthy of investigation.     

Synthesis and antiviral activity of several 2,5'-anhydro analogues of 3'-azido-3'-deoxythymidine, 3'-azido-2',3'-dideoxyuridine, 3'-azido-2',3'-dideoxy-5-halouridines, and 3'-deoxythymidine against human immunodeficiency virus and Rauscher-murine leukemia virus

Several 2,5'-anhydro analogues of 3'-azido-3'-deoxythymidine (AZT), 3'-azido-2'3'-dideoxyuridine (AZU), 3'-azido-2'3'-dideoxy-5-bromouridine, 3'-azido-2',3'-dideoxy-5-iodouridine, and 3'-deoxythymidine and the 3'-azido derivative of 5-methyl-2'-deoxyisocytidine have been synthesized for evaluation as potential anti-HIV (human immunodeficiency virus) agents. These 2,5'-anhydro derivatives, compounds 13-17, demonstrated significant anti-HIV-1 activity with IC50 values of 0.56, 4.95, 26.5, 27.1, and 48 microM, respectively. Compared to that of the parent compounds AZT and AZU, the respective 2,5'-anhydro analogues, compounds 13 and 14, were somewhat less active. Whereas AZT was cytotoxic with a TCID50 of 29 microM, the toxicity of the 2,5'-anhydro derivative of AZT, compound 13, was reduced considerably to a TCID50 value of greater than 100 microM. The 2,5'-anhydro analogue of 5-methyl-2'-deoxyisocytidine also demonstrated anti-HIV-1 activity with an IC50 value of 12 microM. These compounds were also evaluated against Rauscher-Murine leukemia virus (R-MuLV) in cell culture. Among them, AZT, 3'-azido-2',3'-dideoxy-5-iodouridine, 3'-azido-2',3'-dideoxy-5-bromouridine, and 2,5'-anhydro-3'-azido-3'-deoxythymidine (13) were found to be most active, with IC50 values of 0.023, 0.21, 0.23, and 0.27 microM, respectively.     

Predominant contribution of UDP-glucuronosyltransferase 2B7 in the glucuronidation of racemic flurbiprofen in the human liver.

Flurbiprofen is a nonsteroidal anti-inflammatory drug used as a racemic mixture. Although glucuronidation is one of its elimination pathways, the role of UDP-glucuronosyltransferase (UGT) in this process remains to be investigated. Thus, the kinetics of the stereoselective glucuronidation of racemic (R,S)-flurbiprofen by recombinant UGT isozymes and human liver microsomes (HLMs) were investigated, and the major human UGT isozymes involved were identified. UGT1A1, 1A3, 1A9, 2B4, and 2B7 showed glucuronidation activity for both (R)- and (S)-glucuronide, with UGT2B7 possessing the highest activity. UGT2B7 formed the (R)-glucuronide at a rate 2.8-fold higher than that for (S)-glucuronide, whereas the other UGTs had similar formation rates. The glucuronidation of racemic flurbiprofen by HLMs also resulted in the formation of (R)-glucuronide as the dominant form, which occurred to a degree similar to that by recombinant UGT2B7 (2.1 versus 2.8). The formation of (R)-glucuronide correlated significantly with morphine 3-OH glucuronidation (r = 0.96, p < 0.0001), morphine 6-OH glucuronidation (r = 0.91, p < 0.0001), and 3'-azido-3'-deoxythymidine glucuronidation (r = 0.85, p < 0.0001), a reaction catalyzed mainly by UGT2B7, in individual HLMs. In addition, the formation of both glucuronides correlated significantly (r = 0.99, p < 0.0001). Mefenamic acid inhibited the formation of both (R)- and (S)-glucuronide in HLMs with similar IC(50) values (2.0 and 1.7 muM, respectively), which are close to those in recombinant UGT2B7. In conclusion, these findings suggest that the formation of (R)- and (S)-glucuronide from racemic flurbiprofen is catalyzed by the same UGT isozyme, namely UGT2B7.