Pharmacological and physiological roles of the carboxylesterase 1 enzyme complex

Abstract

The mammalian carboxylesterase 1 (Ces1/CES1) proteins are enzymes that break down a large number of (pro)drugs. The activity of these enzymes in the body can influence how effective a certain (pro)drug is. Ces1/CES1 enzymes are also known to affect lipid metabolism, thus affecting susceptibility to obesity-related metabolic disease. We generated and characterized a Ces1 cluster knockout mouse model, and a “humanized” CES1 strain. These genetically modified mice were subjected to pharmacokinetic and toxicity testing, looking at the anticancer (pro)drugs irinotecan and capecitabine, and how their behavior is affected by the Ces1/CES1 modifications. In parallel, we studied the phenotypic effects of full mouse Ces1 cluster deficiency on lipid/glucose homeostasis, and assessed the reversal ability of hepatic human CES1 in TgCES1 mice under medium-fat diet conditions. We identified and characterized an FVB(NKI) mouse substrain (previously considered wild-type in the Netherlands Cancer Institute) in which the Ces1c/d/e genes were naturally mutated. The in vivo pharmacokinetics of the anti-cancer prodrugs irinotecan and capecitabine was clearly altered in this substrain. Ces1c/d/e deficiency further disrupted lipid homeostasis, and Ces1c/d/e-/- mice were prone to developing obesity-related metabolic disease, associated with an activated hepatic acute phase response under high-fat diet conditions. We observed that hepatic human CES1 exerted the predicted protein-dose dependent effects on hydrolysis of irinotecan to SN-38 in liver, and on capecitabine plasma exposure. Unexpectedly, however, whereas TgCES1-low mice were capable of mostly re-establishing the lipid homeostasis in adipose tissues, which was disrupted in Ces1-/- mice, TgCES1-high mice failed to rescue the disrupted lipid homeostasis, both on a medium- and high-fat diet. This expression threshold-dependent regulation of lipid homeostasis by hepatic CES1 is novel, but as of yet not fully understood. It therefore raises many questions that should be further investigated. We also investigated the involvement of ABCB1 and ABCG2 in transport of the BRAFV600E inhibitor encorafenib in vitro and in vivo. It was efficiently transported by canine and human ABCB1 and ABCG2 and by mouse Abcg2 in vitro. In vivo data indicate that both Abcb1a/1b and Abcg2 are involved in restricting the net accumulation of encorafenib in the brain. The strikingly low brain concentration of encorafenib, even in the absence of Abcb1a/1b and Abcg2, suggests that encorafenib may not have optimal access to malignancies positioned behind a functional BBB. If this would also apply to the human brain, this might perhaps turn out to be a limitation for encorafenib in targeting malignancies positioned behind a functional BBB. The second-generation FLT3 inhibitor quizartinib was moderately transported by human ABCB1 and slightly by mouse Abcg2 in vitro. Our study is the first to show that the brain penetration of quizartinib, but not its oral availability, can be markedly limited by both mouse Abcb1 and mouse Abcg2. Furthermore, our in vivo experiments did not point to a prominent role of CYP3A in limiting the oral availability or tissue distribution of quizartinib, at least in mice. Finally, the oral bioavailability of quizartinib in mice was quite high, in spite of a modest rate of net absorption.