Kathleen M. Knights

The prediction of drug-glucuronidation parameters in humans: UDP-glucuronosyltransferase enzyme-selective substrate and inhibitor probes for reaction phenotyping and in vitro–in vivo extrapolation of drug clearance and drug-drug interaction potential

Major advances in the characterization of uridine diphosphate (UDP)-glucuronosyltransferase (UGT) enzyme substrate and inhibitor selectivities and the development of experimental paradigms to investigate xenobiotic glucuronidation in vitro now permit the prediction of a range of drug-glucuronidation parameters in humans. In particular, the availability of substrate and inhibitor “probes” for the major hepatic drug metabolizing UGTs together with batteries of recombinant enzymes allow the reaction phenotyping of drug glucuronidation reactions. Additionally, in vitro experimental approaches and scaling strategies have been successfully applied to the quantitative prediction of in vivo clearance via glucuronidation and drug-drug interaction potential.

The “Albumin Effect” and Drug Glucuronidation: Bovine Serum Albumin and Fatty Acid-Free Human Serum Albumin Enhance the Glucuronidation of UDP-Glucuronosyltransferase (UGT) 1A9 Substrates but Not UGT1A1 and UGT1A6 Activities

Bovine serum albumin (BSA) and fatty acid-free human serum albumin (HSAFAF) reduce the Km values for UGT2B7 substrates by sequestering inhibitory long-chain fatty acids released by incubations of human liver microsomes (HLM) and HEK293 cells expressing this enzyme. However, the scope of the “albumin effect” is unknown. In this investigation we characterized the effects of albumin on the kinetics of 4-methylumbelliferone (4MU) glucuronidation by UDP-glucuronosyltransferase (UGT) 1A1, 1A6, and 1A9, and propofol (PRO) glucuronidation by UGT1A9 and HLM. BSA and HSAFAF, but not human serum albumin, reduced the Km values for 4MU and PRO glucuronidation by UGT1A9. For example, HSAFAF (2%) reduced the Km values for 4MU and PRO glucuronidation from 13.4 to 2.9 and 41 to 7.2 μM, respectively. Similarly, HSAFAF (2%) reduced the Km for PRO glucuronidation by HLM from 127 to 10.6 μM. Arachidonic, linoleic, and oleic acids and a mixture of these decreased the rates of 4MU and PRO glucuronidation by UGT1A9. Km values for these reactions were increased 3- to 6-fold by the fatty acid mixture. Inhibition was reversed by the addition of BSA (2%). Extrapolation of kinetic constants for PRO glucuronidation by HLM in the presence of HSAFAF predicted in vivo hepatic clearance within 15%. Fatty acids had no effect on 4MU glucuronidation by UGT1A1 and UGT1A6 but, paradoxically, all forms of albumin altered the kinetic model for 4MU glucuronidation by UGT1A6 (from Michaelis-Menten to two-site). Only BSA caused a similar effect on 4MU glucuronidation by UGT1A1. It is concluded that BSA and HSAFAF reduce the Km values of only those enzymes inhibited by long-chain unsaturated fatty acids.

Binding of Inhibitory Fatty Acids Is Responsible for the Enhancement of UDP-Glucuronosyltransferase 2B7 Activity by Albumin: Implications for in Vitro-in Vivo Extrapolation

Studies were performed to elucidate the mechanism responsible for the reduction in Km values of UDP-glucuronosyltransferase 2B7 (UGT2B7) substrates observed for incubations conducted in the presence of albumin. Addition of bovine serum albumin (BSA) and fatty acid-free human serum albumin (HSA-FAF), but not “crude” HSA, resulted in an approximate 90% reduction in the Km values for the glucuronidation of zidovudine (AZT) by human liver microsomes (HLM) and UGT2B7 and a 50 to 75% reduction in the S50 for 4-methylumbelliferone (4MU) glucuronidation by UGT2B7, without affecting Vmax. Oleic, linoleic, and arachidonic acids were shown to be the most abundant unsaturated long-chain fatty acids present in crude HSA and in the membranes of HLM and human embryonic kidney (HEK)293 cells, and it was demonstrated that these and other unsaturated long-chain fatty acids were UGT2B7 substrates. Glucuronides with Rf (retention factor) values corresponding to the glucuronides of linoleic and arachidonic acid were detected when HLM and HEK293 cell lysates were incubated with radiolabeled cofactor, and the intensity of the bands was modulated by the presence of crude HSA (increased) and BSA or HSA-FAF (decreased). Oleic, linoleic, and arachidonic acid inhibited AZT and 4MU glucuronidation by HLM and/or UGT2B7, due to an increase in Km/S50 without a change in Vmax. Addition of BSA and HSA-FAF reversed the inhibition. Likewise, coexpression of UGT2B7 and HSA in HEK293 cells reduced the Km/S50 values of these substrates. It is postulated that BSA and HSA-FAF sequester inhibitory fatty acids released during incubations, and the apparent high Km values observed for UGT2B7 substrates arise from the presence of these endogenous inhibitors.

Characterization of Niflumic Acid as a Selective Inhibitor of Human Liver Microsomal UDP-Glucuronosyltransferase 1A9: Application to the Reaction Phenotyping of Acetaminophen Glucuronidation

Enzyme selective inhibitors represent the most valuable experimental tool for reaction phenotyping. However, only a limited number of UDP-glucuronosyltransferase (UGT) enzyme-selective inhibitors have been identified to date. This study characterized the UGT enzyme selectivity of niflumic acid (NFA). It was demonstrated that 2.5 μM NFA is a highly selective inhibitor of recombinant and human liver microsomal UGT1A9 activity. Higher NFA concentrations (50–100 μM) inhibited UGT1A1 and UGT2B15 but had little effect on the activities of UGT1A3, UGT1A4, UGT1A6, UGT2B4, UGT2B7, and UGT2B17. NFA inhibited 4-methylumbelliferone and propofol (PRO) glucuronidation by recombinant UGT1A9 and PRO glucuronidation by human liver microsomes (HLM) according to a mixed (competitive-noncompetitive) mechanism, with Ki values ranging from 0.10 to 0.40 μM. Likewise, NFA was a mixed or noncompetitive inhibitor of recombinant and human liver microsomal UGT1A1 (Ki range 14–18 μM), whereas competitive inhibition (Ki 62 μM) was observed with UGT2B15. NFA was subsequently applied to the reaction phenotyping of human liver microsomal acetaminophen (APAP) glucuronidation. Consistent with previous reports, APAP was glucuronidated by recombinant UGT1A1, UGT1A6, UGT1A9, and UGT2B15. NFA concentrations in the range of 2.5 to 100 μM inhibited APAP glucuronidation by UGT1A1, UGT1A9, and UGT2B15 but not by UGT1A6. The mean Vmax for APAP glucuronidation by HLM was reduced by 20, 35, and 40%, respectively, in the presence of 2.5, 50, and 100 μM NFA. Mean Km values decreased in parallel with Vmax, although the magnitude of the decrease was smaller. Taken together, the NFA inhibition data suggest that UGT1A6 is the major enzyme involved in APAP glucuronidation.

Renal drug metabolism in humans: the potential for drug-endobiotic interactions involving cytochrome P450 (CYP) and UDP-glucuronosyltransferase (UGT).

Although knowledge of human renal cytochrome P450 (CYP) and UDP‐glucuronosyltransferase (UGT) enzymes and their role in xenobiotic and endobiotic metabolism is limited compared with hepatic drug and chemical metabolism, accumulating evidence indicates that human kidney has significant metabolic capacity. Of the drug metabolizing P450s in families 1 to 3, there is definitive evidence for only CYP 2B6 and 3A5 expression in human kidney. CYP 1A1, 1A2, 1B1, 2A6, 2C19, 2D6 and 2E1 are not expressed in human kidney, while data for CYP 2C8, 2C9 and 3A4 expression are equivocal. It is further known that several P450 enzymes involved in the metabolism of arachidonic acid and eicosanoids are expressed in human kidney, CYP 4A11, 4F2, 4F8, 4F11 and 4F12. With the current limited evidence of drug substrates for human renal P450s drug–endobiotic interactions arising from inhibition of renal P450s, particularly effects on arachidonic acid metabolism, appear unlikely. With respect to the UGTs, 1A5, 1A6, 1A7, 1A9, 2B4, 2B7 and 2B17 are expressed in human kidney, whereas UGT 1A1, 1A3, 1A4, 1A8, 1A10, 2B10, 2B11 and 2B15 are not. The most abundantly expressed renal UGTs are 1A9 and 2B7, which play a significant role in the glucuronidation of drugs, arachidonic acid, prostaglandins, leukotrienes and P450 derived arachidonic acid metabolites. Modulation by drug substrates (e.g. NSAIDs) of the intrarenal activity of UGT1A9 and UGT2B7 has the potential to perturb the metabolism of renal mediators including aldosterone, prostaglandins and 20‐hydroxyeicosatetraenoic acid, thus disrupting renal homeostasis.

The stereospecific incorporation of fenoprofen into rat hepatocyte and adipocyte triacylglycerols

The formation of triacylglycerols containing fenoprofen was studied in rat isolated adipocytes and hepatocytes incubated with [3H]glycerol and R or S fenoprofen. In both hepatocytes and adipocytes there was a high-affinity enzymatic process for the synthesis of triacylglycerol containing fenoprofen which was stereospecific for the R enantiomer. The apparent Km values for R fenoprofen were 1.0 microM in adipocytes and 2.8 microM in hepatocytes. These results are consistent with the proposed stereospecific formation of R-2-arylpropionyl-CoA thioesters resulting in the stereospecific formation of R-tri-acylglycerol at clinically relevant unbound fenoprofen concentrations. In isolated hepatocytes, but not adipocytes, a second low-affinity enzymatic process for the synthesis of triacylglycerol containing fenoprofen was also observed. However, this process (Km = 3780 microM) occurred at concentrations much higher than those found in man with usual doses.

The 'albumin effect' and in vitro - in vivo extrapolation: Sequestration of long chain unsaturated fatty acids enhances phenytoin hydroxylation by human liver microsomal and recombinant cytochrome P450 2C9

This study characterized the mechanism by which bovine serum albumin (BSA) reduces the Km for phenytoin (PHY) hydroxylation and the implications of the “albumin effect” for in vitro-in vivo extrapolation of kinetic data for CYP2C9 substrates. BSA and essentially fatty acid-free human serum albumin (HSA-FAF) reduced the Km values for PHY hydroxylation (based on unbound substrate concentration) by human liver microsomes (HLMs) and recombinant CYP2C9 by approximately 75%, with only a minor effect on Vmax. In contrast, crude human serum albumin increased the Km with both enzyme sources. Mass spectrometric analysis of incubations containing HLMs was consistent with the hypothesis that BSA sequesters long-chain unsaturated acids (arachidonic, linoleic, oleic) released from membranes. A mixture of arachidonic, linoleic and oleic acids, at a concentration corresponding to 1/20 of the content of HLMs, doubled the Km for PHY hydroxylation by CYP2C9, without affecting Vmax. This effect was reversed by addition of BSA to incubations. Ki values for arachidonic acid inhibition of human liver microsomal- and CYP2C9-catalyzed PHY hydroxylation were 3.8 and 1.6 μM, respectively. Similar effects were observed with heptadecanoic acid, the most abundant long-chain unsaturated acid present in Escherichia coli membranes. Extrapolation of intrinsic clearance (CLint) values for each enzyme source determined in the presence of BSA and HSA-FAF accurately predicted the known CLint for PHY hydroxylation in vivo. The results indicate that previously determined in vitro Km values for CYP2C9 substrates are almost certainly overestimates, and accurate in vitro-in vivo extrapolation of kinetic data for CYP2C9 substrates is achievable.

Amino acid conjugation: contribution to the metabolism and toxicity of xenobiotic carboxylic acids

Despite being the first conjugation reaction demonstrated in humans, amino acid conjugation as a route of metabolism of xenobiotic carboxylic acids is not well characterised. This is principally due to the small number and limited structural diversity of xenobiotic substrates for amino acid conjugation. Unlike CYP and uridine 5′-diphosphate glucuronosyltransferase, which are localised in the endoplasmic reticulum, the enzymes of amino acid conjugation reside in mitochondria. Unique among drug metabolism pathways, amino acid conjugation involves initial formation of a xenobiotic acyl-CoA thioester that is then conjugated principally with glycine in humans. However, formation of the xenobiotic acyl-CoA thioester does not always infer subsequent amino acid conjugation. Evidence is presented that in the absence of glycine conjugation substrates that form acyl-CoA thioesters perturb mitochondrial function. This review discusses literature on the enzymes involved and the concept that xenobiotic substrate selectivity provides a barrier to protect the metabolic integrity of the mitochondria.

ROLE OF HEPATIC FATTY ACID:COENZYME A LIGASES IN THE METABOLISM OF XENOBIOTIC CARBOXYLIC ACIDS

1. Formation of acyl‐coenzymes (Co)A occurs as an obligatory step in the metabolism of a variety of endogenous substrates, including fatty acids. The reaction is catalysed by ATP‐dependent acid:CoA ligases (EC 6.2.1.1‐2.1.3; AMP forming), classified on the basis of their ability to conjugate saturated fatty acids of differing chain lengths, short (C2‐C4), medium (C4‐C12) and long (C10‐C22). The enzymes are located in various cell compartments (cytosol, smooth endoplasmic reticulum, mitochondria and peroxisomes) and exhibit wide tissue distribution, with highest activity associated with liver and adipose tissue.

Defining the COX inhibitor selectivity of NSAIDs: implications for understanding toxicity.

The hypothesis that the anti-inflammatory activity of NSAIDs derives from COX inhibition is well established. It also underpins the accepted mechanism of the gastrointestinal and renal toxicity of NSAIDs. However, in terms of NSAID-induced cardiovascular toxicity, is COX inhibition then guilty by association? Multiple experimental models of COX-1/COX-2 inhibition have enabled ranking of the relative inhibitory activity of NSAIDs. Inhibition is expressed as an IC50 value and the index of COX selectivity as the ratio of the IC50 value for COX-2 and COX-1. These data informed the ‘imbalance hypothesis’ that the cardiovascular risk of NSAIDs results from an imbalance in the detrimental actions of COX-1-derived thromboxane A2 and the beneficial actions of COX-2-derived prostacyclin (PGI2). Data derived from in vitro models used to generate NSAID IC50 values are discussed in the context of the difficulties in defining COX selectivity and hence understanding the toxicity of NSAIDs in current clinical use.