ANDROMEDA by Prosilico and log D outperform human hepatocytes for the prediction of intrinsic hepatic metabolic clearance of carboxylic acids

Abstract

AbstractIntroductionExtrahepatic metabolism/conjugation, deconjugation of their metabolites, and low and varying unbound fraction in plasma (fu), is characteristic for carboxylic drugs. Thus, it is comparably difficult to estimate theirin vivointrinsic hepatic metabolic clearance (CLint) and hepatic CL (CLH) and to predict theirin vivoCLint, CLHand CL. One objective was to investigate the laboratory variability of fuand CLintfor carboxylic acids. Another objective was to compare human hepatocytes, measured log D and the software ANDROMEDA with regards to prediction of humanin vivoCLintof carboxylic acids.Materials and MethodsMeasured unbound hepatocyte CLint, non-renal CL (surrogate for CLH), non-renal CLint(surrogate for hepatic metabolic CLint), log D and fudata were taken from studies in the literature. ANDROMEDA (by Prosilico; version 1.0) prediction software was used forin silicopredictions of CLintfor carboxylic acids not used in the training set of its CLint-model.Results and DiscussionMean and maximum differences between highest and lowest reportedin vivoCLintpredicted from hepatocyte CLintwere 209- and 1,476-fold (n=8), respectively. Corresponding estimates forin vitrofuwere 17- and 50-fold, respectively. The data set with the apparently highest number of carboxylic acids contains 39 carboxylic acids within vitroCLintand log D (both measured at the same laboratory),in vivoCLintandin vitrofu. 18 carboxylic acids were excluded as theirin vitroCLintwas below the limit of quantification. The correlation coefficient (R2) for log hepatocyte predictedin vivoCLintvslogin vivoCLintwas 0.34. The corresponding R2for log Dvslogin vivoCLintwas 0.40 (0.47 for 64 carboxylic acids). The Q2(forward-looking R2) forin silico(ANDROMEDA) predicted and measured login vivoCLintfor 12 carboxylic acids was 0.86. The corresponding R2for hepatocytes and log D were 0.67 and 0.66, respectively. ANDROMEDA produced a lower maximum prediction error compared to hepatocytes and also predicted thein vivoCLintfor all carboxylic acids out of reach for the hepatocyte assay.ConclusionVery large interlaboratory variability was demonstrated for plasma protein binding and hepatocyte assays. Log D, and especially ANDROMEDA, outperformed the hepatocyte assay for the prediction of CLintof carboxylic acidsin vivoin man.