Purpose: Over the last few decades, chemists have become increasingly adept at designing compounds that are able to circumvent cytochrome P450-mediated clearance. This has led to the emergence of other drug metabolizing enzymes as significant contributors to drug clearance, one of which is aldehyde oxidase (AO). AO belongs to a family of molybdenum-containing enzymes and exhibits an extraordinary degree of amino acid sequence similarity with another enzyme in the same family, i.e., xanthine oxidase (XO). AO is present in the cytosol and is known to have proclivity toward oxidizing aza-heterocyclic compounds. Since ongoing projects at NCATS have AO-mediated liabilities, we have collected cytosolic stability data for more than 1400 compounds in mouse (CD-1 male) and human cytosol fractions. Our goal is to understand this dataset and provide chemists with a set of guidelines which can be used to rationally modulate compounds and attenuate AO liability.
Methods: Cytosol stability assay was performed using the substrate depletion method. Briefly, each reaction mixture (110 μL) consisted of a test article (1 μM), human or mouse cytosol fractions (2mg/mL) in phosphate buffer at pH 7.4. Samples were incubated in 384-well plates at 37°C for 0, 5, 10, 15, 30 and 60 min. Sample analysis and half-life calculations were performed using a previously described method (Shah et al., 2016).
Results: Similar to reports in literature, we observed marked species differences in AO-mediated clearance. Herein we account our efforts to systematically identify structural features that affect AO metabolism within and across species. Through a combination of matched molecular pair and sub-structural fingerprint analysis, we identified a set of rules that were deemed important for AO-mediated metabolic stability from the 1400 compound training set (Table 1). For validation, we plan to apply these rules prospectively to a set of 370 chemically diverse compounds.
Conclusion: The validated rules will enhance lead optimization by guiding structural modifications. The code and the data will be incorporated in the publicly available NCATS web services (https://predictor.ncats.io).
Dac-Trung Nguyen– Employee, NCATS
Alexey Zakharov– Employee, NCATS
Khalida Shamim– Employee, NCATS
Jian-Kang Jiang– Employee, NCATS
Hao Li– Employee, NCATS
Wenwei Huang– Employee, NCATS
Xin Xu– Employee, NCATS, Maryland