In vitro and in vivo glucuronidation of midazolam in humans

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT center dot Midazolam is extensively metabolized and is eliminated in urine predominantly as conjugates of 1'-hydroxymidazolam. center dot Midazolam is known to undergo N-glucuronidation in vitro by UGT1A4. center dot In vitro 1'-hydroxymidazolam is O-glucuronidated by UGT2B4 and UGT2B7, and N-glucuronidated by UGT1A4. WHAT THIS STUDY ADDS center dot N-glucuronide of midazolam has been quantified in human urine, indicating for the first time that this route of metabolism occurs in vivo. center dot Metabolism of 4-hydroxymidazolam has been compared with that of 1'-hydroxymidazolam in vitro. center dot This study provides further evidence, in vitro and in vivo, of the importance of N-glucuronidation in the metabolism of midazolam and its metabolites. Midazolam (MDZ) is a benzodiazepine used as a CYP3A4 probe in clinical and in vitro studies. A glucuronide metabolite of MDZ has been identified in vitro in human liver microsome (HLM) incubations. The primary aim of this study was to understand the in vivo relevance of this pathway. An authentic standard of N-glucuronide was generated from microsomal incubations and isolated using solid-phase extraction. The structure was confirmed using proton nuclear magnetic resonance (NMR) and (1)H-(13)C long range correlation experiments. The metabolite was quantified in vivo in human urine samples. Enzyme kinetic behaviour of the pathway was investigated in HLM and recombinant UGT (rUGT) enzymes. Additionally, preliminary experiments were performed with 1'-OH midazolam (1'-OH MDZ) and 4-OH-midazolam (4-OH MDZ) to investigate N-glucuronidation. NMR data confirmed conjugation of midazolam N-glucuronide (MDZG) standard to be on the alpha-nitrogen of the imidazole ring. In vivo, MDZG in the urine accounted for 1-2% of the administered dose. In vitro incubations confirmed UGT1A4 as the enzyme of interest. The pathway exhibited atypical kinetics and a substrate inhibitory cooperative binding model was applied to determine K(m) (46 mu M, 64 mu M), V(max) (445 pmol min(-1) mg(-1), 427 pmol min(-1) mg(-1)) and K(i) (58 mu M, 79 mu M) in HLM and rUGT1A4, respectively. From incubations with HLM and rUGT enzymes, N-glucuronidation of 1'-OH MDZ and 4-OH MDZ is also inferred. A more complete picture of MDZ metabolism and the enzymes involved has been elucidated. Direct N-glucuronidation of MDZ occurs in vivo. Pharmacokinetic modelling using Simcyp (TM) illustrates an increased role for UGT1A4 under CYP3A inhibited conditions.