Elucidating the Drug-Drug Interaction Potential of Hydroxychloroquine Metabolites due to Direct and Time-Dependent Inhibition of CYP3A

Abstract ID 93442 Poster Board 037 Aims: Hydroxychloroquine (HCQ) is a safe and effective treatment of rheumatoid arthritis and systemic lupus erythematosus. HCQ is metabolized into three metabolites: desethylhydroxychloroquine (DHCQ), desethylchloroquine (DCQ), and didesethylchloroquine (DDCQ). Recently, we showed that HCQ metabolites are cytochrome P450 (CYP) 3A direct and time-dependent inhibitors. Therefore, this study aimed to determine the in vitro inhibition constant (Ki), maximal inactivation rate (kinact), and inhibitor concentration at half-maximal inactivation rate (KI) parameters of CYP3A inhibition by HCQ metabolites to predict their drug-drug interaction (DDI) potential with CYP3A substrates. Methods: In Ki experiments, HCQ metabolites were incubated with human liver microsomes (HLM; 0.05 mg/ml), NADPH and midazolam for 2 min. In KI-kinact experiments, HCQ metabolites were preincubated with HLM (0.25 mg/ml) and NADPH for up to 10 min. At predetermined time points, aliquots of the preincubation mixture were diluted 20-fold into mixtures containing NADPH and midazolam (20 μM) and incubated for an additional 2 min. Non-specific binding to HLM was accounted for. The combined effect of HCQ metabolites on the plasma exposure of the CYP3A substrate midazolam was predicted using a static mechanistic model for direct and time-dependent inhibition. Predictions were carried out at clinically relevant concentrations in plasma, blood, and the liver: HCQ metabolite concentrations were estimated based on HCQ concentrations in blood (Somer et al. Br J Clin Pharmacol. 2000;49:549-554), metabolic ratios, blood to plasma ratios, tissue to blood concentration coefficients (Kp) and unbound fractions in plasma. Results: Ki values for CYP3A direct inhibition were 91.7, 132 and 27.7 μM for DHCQ, DCQ and DDCQ. KI and kinact values for CYP3A time-dependent inhibition were 548, 67.3 and 32.0 μM, and 0.60, 0.14 and 0.18 min⁻¹ for DHCQ, DCQ and DDCQ, respectively. The predicted fold increase in midazolam plasma area under the concentration-time curve (AUCR) following CYP3A inhibition by HCQ metabolites was 1.6-1.9 at estimated clinically relevant total plasma concentrations. Because HCQ and its metabolites accumulate extensively in blood and tissues, DDI predictions were also carried out using estimated blood and liver concentrations: midazolam AUCR was predicted to be ≥5.3 and ≥15.5, respectively. Conclusion: Our static predictions suggest that strong DDIs (AUCR > 5) may occur between HCQ and CYP3A substrates.