Abstracts Biological Sciences

Roles of cytochromes P450 and microsomal epoxide hydrolase in drug-drug interactions involving valporic acid and its analogues

by Susan I. Hurst

Valproic acid (VPA) therapy is complicated by inhibition-based drug-drug interactions with other anticonvulsants. It was hypothesized that VPA may inhibit one or more cytochrome P450 isoforms. VPA was incubated with selective cytochrome P450 (1A2, 2A6, 2C9, 2C19, 2E1, and 3A4) probes in microsomes and the only isoform that exhibited pronounced inhibition (57%, 3000 m M) was CYP2C9. VPA competitively inhibited CYP2C9 with a Ki value of 1209 +/- 167 m M. It also inhibited the in vitro metabolism of phenytoin and phenobarbital. Neither non-specific microsomal protein-binding, time-dependent inactivation, nor liver/plasma partition ratios (rat model) could be used to reconcile measured in vitro Ki values with the observed clinical interactions (e.g. phenytoin and phenobarbital) and the extensive plasma protein-binding of VPA.Several VPA derivatives were investigated for their ability to inhibit microsomal epoxide hydrolase (mEH). Neither valproylglycinamide nor its metabolite valproylglycine were potent inhibitors of mEH compared to the antiepileptic drugs VPA and valpromide. Also, N-methyl-tetramethylcyclopropane carboxamide was not a potent mEH inhibitor. Its metabolite, tetramethylcyclopropane carboxamide, exhibited greater inhibition of mEH, but with IC50s at least 20-fold larger than the IC50 for valpromide.The clinical interaction between dezinamide and carbamazepine was investigated via the effects of dezinamide on CYP3A4 and mEH. Dezinamide had either no effect (probe R-warfarin) or it increased the activity of CYP3A4 (probe carbamazepine). Dezinamide inhibited mEH via mixed inhibition at concentrations that are likely to be obtained in vivo (Ki of 18.8 to 20.8 m g/ml using various probe mEH substrates). Thus the overall predicted effect would be an increase in the major metabolite of carbamazepine, carbamazepine-10,11-epoxide.There was no effect of mEH allelic variance (H113/H113 or R139/R139 mutations) on valpromide inhibition of mEH. As valpromide is primarily a competitive inhibitor of mEH, the findings suggest these mutations do not affect inhibitor binding at the active site. If they can be generalized to other primarily competitive mEH inhibitors, they also suggest that individuals homozygous for these mutations are not more susceptible to mEH inhibition. However, the present findings do not preclude effects of these mutations on another enzyme site, relevant to the binding of a noncompetitive inhibitor.