Methods to assess pulmonary metabolism

Drug metabolism involves the biochemical modification of pharmaceutical substances or xenobiotics by living organisms, usually through specialized enzymes. This process normally converts lipophilic chemicals into less potent, and more hydrophilic products that facilitates their elimination from the body (Mittal et al., 2015). However, these processes may also convert the drug into more lipophilic, more potent or even toxic metabolites (Macherey and Dansette, 2015). In order to design effective and safe dosage regimens, the pharmacology, toxicology, and drug-drug interactions of the drug and its metabolites should be thoroughly understood (Tillement and Tremblay, 2007). As a result, the study of drug metabolism is vital to the pharmaceutical industry. Drug metabolism is generally divided into two distinct phases. Phase I reactions include oxidation, reduction and hydrolysis which are mediated by enzymes such as cytochrome P450 (CYPs), flavin-containing monooxygenases (FMOs), aldehyde oxidase (AO) and various hydrolases (Shehata, 2010). In phase II metabolism, enzymes such as uridine 5’-diphosphoglucuronosyltransferase (UGT), sulfotransferase, glutathione S-transferase (GST) and N-acetyl transferase (NAT) catalyze the conjugation of drugs with endogenous molecules (Testa and Clement, 2015).
Although the liver is the primary site of drug metabolism, other organs including skin, lungs, kidneys, and intestine also possess considerable metabolic capacity (Krishna and Klotz, 1994). Furthermore, drug-metabolizing enzymes are known to be expressed in the lungs, albeit to a lesser extent than in liver (Somers et al., 2007). Recent reports have highlighted the metabolic role of lungs, a highly perfused organ that is in direct contact with inhaled xenobiotics and drugs (Borghardt et al., 2018). In addition, phase I enzymes such as CYP1A1 may play an important role in bioactivation of inhaled procarcinogens, such as those present in tobacco smoke (Anttila et al., 2011). From the perspective of pharmacotherapy, pulmonary drug metabolism may cause a first-pass effect for inhaled medicines, as well as contribute to overall clearance of systemically administered drugs (Winkler et al., 2004). Frequent exposure of the lungs to environmental xenobiotics may also lead to induction of drug-metabolizing enzymes via pregnane X receptors (PXR), constitutive androstane receptors (CAR), or aryl hydrocarbon receptors (AhR), a phenomenon that can significantly alter the rate of drug metabolism (Tolson and Wang, 2010).
Despite the potential importance of lung metabolism for respiratory therapies, relatively little is known about the actual activity and protein abundance of drug-metabolizing enzymes in lung tissue (Hukkanen et al., 2002). The lack of robustness and consistency of existing experimental models of lung metabolism leads to considerable difficulties in the interpretation and prediction of drug clearance. This project was designed to address these challenges and establish a robust and predictive experimental model for rat and human pulmonary drug metabolism. Therefore the aim of this thesis was 1) to investigate the further development of a precision-cut lung slicing (PCLS) model to accurately estimate pulmonary drug clearance in rat, 2) to examine the pulmonary metabolic activity of rat and human phase I and phase II enzymes using this model, and 3) to compare the PCLS model with currently available in vitro and in vivo experimental models in order to better understand the contribution of pulmonary metabolism to drug elimination.
1) Establish a PCLS model to accurately estimate pulmonary drug clearance in rat
PCLS technology is a 3D organotypic tissue model which reflects the natural and relevant microanatomy and the metabolic function of the lung (Neuhaus et al., 2017). Although the use of PCLS is becoming accepted as a research tool to investigate pulmonary drug metabolism, the protocols applied vary between laboratories and there is still opportunity to improve and standardize the methodology. Therefore, some of the key experimental factors used in the PCLS procedure were optimized with the aim of reducing variability and tissue damage and retaining lung metabolic activity. Due to the limited availability of fresh human lung tissue, method optimization was performed using rat lung tissue (under the assumption that rat is a suitable surrogate for human) and referring to a well-known CYP1A1 substrate, mavoglurant (AFQ056). The choice of mavoglurant as a test compound was based on two factors; 1) it is a CYP1A1 substrate and this enzyme is located primarily in extra-hepatic organs such as lung, kidney, brain and intestine (Drahushuk et al., 1998, Cheung et al., 1999, Paine et al., 1999, Smith et al., 2001) and 2) metabolism of mavoglurant by CYP1A1 produces a specific metabolite, CBJ474, that serves as a marker for CYP1A1 metabolic activity (Walles et al., 2013). During the optimization process of the PCLS model it was possible to achieve higher mavoglurant turnover by performing incubations on dynamic organ culture system. This investigation demonstrated the importance of optimization and standardization of PCLS conditions.
2) Investigate the activity of phase I and phase II metabolic enzymes in rat and human lungs using PCLS model
Preclinical species such as rats and mice are commonly used for the optimization of pharmacokinetic (PK) properties and for testing in vivo efficacy of new chemical entities. The PK data from these preclinical species is also often used for the prediction of human PK, and therefore, it is desirable that drug metabolism in these species is representative of that in human. For the comparison of enzyme activity in rat and human lungs, a selection of phase I and phase II probe substrates (please refer to chapter 3.1, Table 3 for a list of the probe substrates tested) were incubated using the optimized PCLS model. The results showed that there are remarkable differences in pulmonary metabolic activity between rat and human, reflecting species dependent expression of drug-metabolizing enzymes. CYP-mediated metabolic activity was relatively low in both species, whereas phase II enzyme activities appeared to be more significant in rat than in human. Therefore, care should be taken when extrapolating metabolism data from animal models to humans.
3) Comparison of PCLS model with established in vitro and in vivo experimental models to be able to understand the contribution of pulmonary metabolism to drug elimination
A variety of in vitro and ex vivo lung models such as cell culture, sub-cellular fractions, tissue slices and isolated perfused lung models have been used to investigate pulmonary metabolism. Each tool has advantages and disadvantages and can be used to answer specific scientific questions. However, due to the diversity of the cells in the lung it is difficult to obtain quantitative data. In this research project, mavoglurant and benzydamine (well-characterized FMO substrate) were incubated using rat lung microsomes (in vitro) and rat PCLS (ex vivo) and also administered intravenously (i.v.) and intra-arterially (i.a.) to rats (in vivo). Previous human in vivo studies had indicated that extra-hepatic metabolism contributes to the elimination of mavoglurant (Novartis internal data). The goal was to use these models to understand if the lung makes a significant contribution to total body clearance. Using the well-stirred organ model, lung clearance (CLlung) of mavoglurant was estimated from microsomal and PCLS data and compared to the in vivo data obtained from i.v. and i.a. dosing to rats. The data generated from these three experimental models were comparable and the data suggested that the contribution of pulmonary metabolism to the elimination of mavoglurant is negligible. The same experiments were also performed using benzydamine. Interestingly, calculated lung clearance from microsomal data were 8-fold higher than lung clearance estimates from the PCLS model. Hence, similar to the PCLS-derived predictions, in vivo data indicated very low pulmonary clearance.