Physiologically based pharmacokinetic modeling of aldehyde oxidase drug-drug interactions mediated by erlotinib

The propensity for aldehyde oxidase (AO) substrates to be implicated in drug-drug interactions (DDI) has remained largely uninterrogated due to the lack of a precipitant drug which elicits potent inhibition of AO in vivo. Recently, we characterized the epidermal growth factor receptor inhibitor erlotinib as a clinical AO inhibitor and proposed, through mechanistic metabolism studies and static modeling, that AO inhibition was responsible for its observed DDI with the investigational drug OSI-930. However, as erlotinib also inhibits the organic anion transporting polypeptide 2B1 transporter in the liver and gut at clinically relevant concentrations, the potential contribution of transporter-mediated interactions to the observed clinical DDI with OSI-930 remained unclear. In this follow-up study, uptake studies in organic anion transporting polypeptide 2B1-transfected human embryonic kidney 293 cells confirmed that OSI-930 is not a substrate for this transporter. Physiologically based pharmacokinetic (PBPK) models for erlotinib and OSI-930, which incorporated competitive and time-dependent AO inhibition and an AO-mediated metabolism component, respectively, were iteratively developed, refined and verified using published clinical data. Our physiologically based pharmacokinetic model successfully recapitulated the in vivo AO-mediated DDI between erlotinib and OSI-930 occurring after multiple erlotinib doses, with predicted/observed Cmax and area under the curve ratios ranging from 0.84 to 1.06. Additional simulations were also conducted to investigate untested clinical DDI scenarios between erlotinib and other known AO substrates-O6-benzylguanine, zaleplon, ziprasidone, and zoniporide. These prospective simulations suggested a considerable DDI risk (ie, area under the curve ratio > 3-fold) for high fraction metabolized by AO (fm,AO) compounds codosed with clinical doses of erlotinib, thereby warranting further clinical investigation. SIGNIFICANCE STATEMENT: The contribution of OATP2B1 and AO inhibition to the observed erlotinib-OSI-930 DDI was investigated using in vitro experiments and PBPK modeling. Our study revealed that the DDI is due to inhibition of AO-mediated clearance of OSI-930 by erlotinib. Additionally, our erlotinib model also predicts clinically significant DDIs would occur when dosing erlotinib with high fm,AO AO substrates. These findings highlight the need to consider AO inhibition in DDI risk assessments and may help inform future drug development and regulatory strategies.