Marc Vrana , University of Washington, Seattle, WA
Xu Meijuan , Department of Clinical Pharmacology, Affiliated Hospital of Nanjing University of Traditional Chinese Medicine, Nanjing, China
Haeyoung Zhang , University of Washington, Seattle, WA
Bhagwat Prasad , University of Washington, Seattle, WA
In the early stages of drug development, drug metabolism data are generated using human in vitro models such as liver microsomes, cytosol, and S9. These subcellular fractions are isolated from liver tissue by a combination of low- and high-speed centrifugation steps which enrich drug metabolizing enzymes (DMEs). While the localizations of traditional hepatic DMEs (CYPs, UGTs) are well characterized, localization of other important DMEs including carboxylesterases (CESs), aldehyde oxidase (AOX), paraoxonases (PONs), etc., is not well known. This knowledge gap is a major limitation in using the data for in vitro to in vivo extrapolation (IVIVE). Therefore, the aims of this study were to characterize localization of non-CYP and non-UGT DMEs in human liver subcellular fractions, and to evaluate the impact of processing variables on the apparent localization of these DMEs.
Using samples procured from the University of Washington liver bank (n=3), we followed a standard procedure for the isolation of microsomes, changing key variables (centrifugation speeds) in the fractionation process in order to assess their impact on the a) enrichment and b) recovery of 28 important DMEs, including CYPs, UGTs, CESs, PONs, ALDH, cholinesterases (CHes), arylacetamide deacetylase (AADAC), sulfotransferases (SULTs), AOXs, and EPHXs. The protein quantification of these DMEs was performed in each fraction using validated LC-MS/MS proteomics methods [1]. The protein expression was normalized to the total protein concentration in each fraction and the enrichment and recovery were calculated.
CES1 and CES2 were found to be significantly enriched in both the cytosol and microsomes (166±15% and 77±10%). Other esterases, including PONs, ACHes, and AADAC were much more highly enriched in the microsomes, demonstrating localization to the endoplasmic reticulum. The soluble enzymes likewise showed a predictable pattern of recovery, with ADH1A, 1B, 1C, and ALDH1A1 having an average enrichment of 185±9% and 14.6±2% in the cytosol and microsomes, respectively. As expected CYPs and UGTs were found at low relative concentrations in the cytosol fraction as contamination, and were highly enriched in the microsomes when compared to the homogenate (0.2 vs. 4.2 fold respectively). Using a lower centrifugation speed in the initial separation of bigger organelles (6000 vs. 15000xg) generally produced better recovery on average, but also produced much greater variability. Further, a significant and variable loss of microsomal DMEs was observed in the mitochondrial pellet as well as the final microsomal wash step.
LC-MS/MS proteomics is an accurate and high-throughput technique to characterize localization of DMEs. CESs were confirmed at similar levels in both microsomes and cytosol, while other esterases (CHes, AADAC, PONs) were highly enriched in the microsomal fraction. Additionally, it was determined that the initial low speed centrifugation has a marked effect on the recovery of proteins. Further, a significant loss of microsomal DMEs occurred during the processing that must be mitigated or otherwise accounted for in IVIVE calculations to ensure physiological relevance.

1. Vrana, M., et al., CPT Pharmacometrics Syst Pharmacol, 2017. 6(4): p. 267-276.