A Laboratory-Specific Scaling Factor to Predict the In Vivo Human Clearance of Aldehyde Oxidase Substrates

Improved clearance predictions for aldehyde oxidase substrates using a novel triculture human hepatocyte model

Over the last several decades, efforts in medicinal chemistry have aimed to reduce the extent of CYP metabolism of new chemical entities. This approach, however, has led to increased susceptibility to metabolism by non-CYP-mediated pathways, particularly involving other phase I enzymes such as aldehyde oxidase (AO). Commonly used in vitro models, such as suspended or cocultured primary human hepatocytes, have limitations in evaluating the disposition of compounds metabolized by AO due to low or variable levels of enzyme activity. Thus, an in vitro model that exhibits high to moderate levels of AO activity that can better predict the contribution of AO to drug metabolism and its impact on drug clearance is needed. A novel, 2D+ primary human hepatocyte model, TruVivo, was evaluated for its potential utility to improve hepatic clearance (CLh) predictions and determine the contribution of AO to drug metabolism in humans. TruVivo demonstrated stable levels of AO activity for at least 2 weeks that were higher than levels in other hepatocyte models. CLh predictions generated using TruVivo for the reference compounds carbazeran, zoniporide, zaleplon, and O6-benzylguanine were within 2-fold of reported in vivo CL values. Furthermore, the estimated fraction metabolized by AO for zaleplon and zoniporide was within 25% of reported in vivo values, whereas that for carbazeran and O6-benzylguanine was similar to those generated in other systems. These findings suggest TruVivo may offer a novel means to assess CLh of AO substrates more accurately, even over extended incubation times for low clearance compounds. SIGNIFICANCE STATEMENT: The TruVivo in vitro primary human hepatocyte model maintains high levels of aldehyde oxidase (AO) activity for at least 14 days in culture, making the system suitable for evaluating slowly metabolized compounds, particularly those metabolized by AO. This novel system may therefore be useful for improving human clearance predictions for AO substrates.

Prediction of hepatic metabolic clearance in rats and dogs using long-term cocultured hepatocytes

Knowledge of the accuracy of in vitro to in vivo extrapolation (IVIVE) in preclinical species pharmacokinetics is often used to assess the reliability of physiologically based modeling to predict clinical pharmacokinetics of new chemical entities. In this study, the in vitro hepatic clearance was evaluated using suspended hepatocytes and long-term cocultured hepatocytes for 27 internal drug candidates, and 20 (dogs) or 24 (rats) marketed drugs. Prospective identification of rate-determining step(s) was performed through application of the Extended Clearance Classification System. The HepatoPac, suitable for the assessment of low metabolic turnover in humans, did not demonstrate significant superiority of the metabolic clearance prediction in rats and dogs relative to suspended hepatocytes using a set of the internal drug candidates. A 3.6- to 5.9-fold underestimation of the upscaled unbound intrinsic clearance (CLint,h,u) for Extended Clearance Classification System 1a/2 compounds was determined in rats and dogs throughout the study. Limited predictability of the active transport clearance using both liver preparations was confirmed with a 9.8- to 213-fold error. Nonetheless, for compounds with CLint,h ≤ 3 μL/min/mg in suspended hepatocytes, reassessment of CLint,h using HepatoPac can be still considered to achieve a reasonable IVIVE. For marketed drugs, the in vitro CLint,h,u values obtained from hepatocyte suspensions and HepatoPac were consistent with the in vivo clearance data in rats and dogs with a few exceptions. Species differences in some drug-metabolizing enzymes were highlighted as potential hindrance to clearance IVIVE in preclinical species and humans. SIGNIFICANCE STATEMENT: The current findings provide insights into the differences and similarities between in vitro clearance measurements using hepatocyte suspension assays and HepatoPac, and their translation to in vivo clearance values in rats and dogs. In addition, a gap analysis was performed discussing species differences of the drug-metabolizing enzyme isoforms to complement the current strategy of predicting human pharmacokinetics prior to phase 1 with physiologically based pharmacokinetic modeling.

Establishing a physiologically based pharmacokinetic framework for aldehyde oxidase and dual aldehyde oxidase-CYP substrates

Aldehyde oxidase (AO) contributes to the clearance of many approved and investigational small molecule drugs, which are often dual substrates of AO and drug-metabolizing enzymes such as cytochrome P450s (CYPs). As such, the lack of established framework for quantitative translation of the clinical pharmacologic correlates of AO-mediated clearance represents an unmet need. This study aimed to evaluate the utility of physiologically based pharmacokinetic (PBPK) modeling in the development of AO and dual AO-CYP substrates. PBPK models were developed for capmatinib, idelalisib, lenvatinib, zaleplon, ziprasidone, and zoniporide, incorporating in vitro functional data from human liver subcellular fractions and human hepatocytes. Prediction of metabolic elimination with/without the additional empirical scaling factors (ESFs) was assessed. Clinical pharmacokinetics, human mass balance, and drug-drug interaction (DDI) studies with CYP3A4 modulators, where available, were used to refine/verify the models. Due to the lack of clinically significant AO-DDIs with known AO inhibitors, the fraction metabolized by AO (fmAO) was verified indirectly. Clearance predictions were improved by using ESFs (GMFE ≤1.4-fold versus up to fivefold with physiologically-based scaling only). Observed fmi from mass balance studies were crucial for model verification/refinement, as illustrated by capmatinib, where the fmAO (40%) was otherwise underpredicted up to fourfold. Subsequently, independent DDI studies with ketoconazole, itraconazole, rifampicin, and carbamazepine verified the fmCYP3A4, with predicted ratios of the area under the concentration-time curve (AUCR) within 1.5-fold of the observations. In conclusion, this study provides a novel PBPK-based framework for predicting AO-mediated pharmacokinetics and quantitative assessment of clinical DDI risks for dual AO-CYP substrates within a totality-of-evidence approach.

Evaluation and Optimization of a Microcavity Plate-Based Human Hepatocyte Spheroid Model for Predicting Clearance of Slowly Metabolized Drug Candidates

In vitro clearance assays are routinely conducted in drug discovery to predict in vivo clearance, but low metabolic turnover compounds are often difficult to evaluate. Hepatocyte spheroids can be cultured for days, achieving higher drug turnover, but have been hindered by limitations on cell number per well. Corning Elplasia microcavity 96-well microplates enable the culture of 79 hepatocyte spheroids per well. In this study, microcavity spheroid properties (size, hepatocyte function, longevity, culturing techniques) were assessed and optimized for clearance assays, which were then compared with microsomes, hepatocyte suspensions, two-dimensional-plated hepatocytes, and macrowell spheroids cultured as one per well. Higher enzyme activity coupled with greater hepatocyte concentrations in microcavity spheroids enabled measurable turnover of all 17 test compounds, unlike the other models that exhibited less drug turnover. Microcavity spheroids also predicted intrinsic clearance (CLint) and blood clearance (CLb) within threefold for 53% [9/17; average absolute fold error (AAFE), 3.9] and 82% (14/17; AAFE, 2.6) of compounds using a linear regression correction model, respectively. An alternate method incorporating mechanistic modeling that accounts for mass transport (permeability and diffusion) within spheroids demonstrated improved predictivity for CLint (12/17; AAFE, 4.0) and CLb (14/17; AAFE, 2.1) without the need for empirical scaling factors. The estimated fraction of drug metabolized by cytochrome P450 3A4 (fm,CYP3A4) using 3 μM itraconazole was within 25% of observed values for 6 of 8 compounds, with 5 of 8 compounds within 10%. In sum, spheroid cultures in microcavity plates permit the ability to test and predict clearance as well as fm,CYP3A4 of low metabolic turnover compounds and represent a valuable complement to conventional in vitro clearance assays. SIGNIFICANCE STATEMENT: Culturing multiple spheroids in ultralow attachment microcavities permits accurate quantitation of metabolically stable compounds in substrate depletion assays, overcoming limitations with singly cultured spheroids. In turn, this permits robust estimates of intrinsic clearance, which is improved with the consideration of mass transport within the spheroid. Incubations with 3 μM itraconazole enabled assessments of CYP3A4 involvement in hepatic clearance.

Challenges and Opportunities for In Vitro-In Vivo Extrapolation of Aldehyde Oxidase-Mediated Clearance: Toward a Roadmap for Quantitative Translation

Underestimation of aldehyde oxidase (AO)-mediated clearance by current in vitro assays leads to uncertainty in human dose projections, thereby reducing the likelihood of success in drug development. In the present study we first evaluated the current drug development practices for AO substrates. Next, the overall predictive performance of in vitro-in vivo extrapolation of unbound hepatic intrinsic clearance (CLint,u) and unbound hepatic intrinsic clearance by AO (CLint,u,AO) was assessed using a comprehensive literature database of in vitro (human cytosol/S9/hepatocytes) and in vivo (intravenous/oral) data collated for 22 AO substrates (total of 100 datapoints from multiple studies). Correction for unbound fraction in the incubation was done by experimental data or in silico predictions. The fraction metabolized by AO (fmAO) determined via in vitro/in vivo approaches was found to be highly variable. The geometric mean fold errors (gmfe) for scaled CLint,u (mL/min/kg) were 10.4 for human hepatocytes, 5.6 for human liver cytosols, and 5.0 for human liver S9, respectively. Application of these gmfe's as empirical scaling factors improved predictions (45%-57% within twofold of observed) compared with no correction (11%-27% within twofold), with the scaling factors qualified by leave-one-out cross-validation. A road map for quantitative translation was then proposed following a critical evaluation on the in vitro and clinical methodology to estimate in vivo fmAO In conclusion, the study provides the most robust system-specific empirical scaling factors to date as a pragmatic approach for the prediction of in vivo CLint,u,AO in the early stages of drug development. SIGNIFICANCE STATEMENT: Confidence remains low when predicting in vivo clearance of AO substrates using in vitro systems, leading to de-prioritization of AO substrates from the drug development pipeline to mitigate risk of unexpected and costly in vivo impact. The current study establishes a set of empirical scaling factors as a pragmatic tool to improve predictability of in vivo AO clearance. Developing clinical pharmacology strategies for AO substrates by utilizing mass balance/clinical drug-drug interaction data will help build confidence in fmAO.

PBPK Modelling for Drugs Cleared by Non-CYP Enzymes: State-of-the-Art and Future Perspectives

Physiologically-based pharmacokinetic (PBPK) modeling has become the established method for predicting human pharmacokinetics (PK) and drug-drug interactions (DDI). The number of drugs cleared by non-CYP enzyme metabolism has increased steadily and to date, there is no consolidated overview of PBPK modeling for drugs cleared by non-CYP enzymes. This review aims to describe the state-of-the-art for PBPK modeling for drugs cleared via non-CYP enzymes, to identify successful strategies, to describe gaps and to provide suggestion to overcome them. To this end, we conducted a detailed literature search and found 58 articles published before the 1st of January 2023 containing 95 examples of clinical PBPK models for 62 non-CYP enzyme substrates. Reviewed articles covered the drug clearance by uridine 5'-diphospho-glucuronosyltransferases (UGTs), aldehyde oxidase (AO), flavin-containing monooxygenases (FMOs), sulfotransferases (SULTs) and carboxylesterases (CES), with UGT2B7, UGT1A9, CES1, FMO3 and AO being the enzymes most frequently involved. In vitro-in vivo extrapolation (IVIVE) of intrinsic clearance and the bottom-up PBPK modeling involving non-CYP enzymes remains challenging. We observed that the middle-out modeling approach was applied in 80% of the cases, with metabolism parameters optimized in 73% of the models. Our review could not identify a standardized approach used for model optimization based on clinical data, with manual optimization employed most frequently. Successful development of models for UGT2B7, UGT1A9, CES1, and FMO3 substrates provides a foundation for other drugs metabolized by these enzymes and guides the way forward in creating PBPK models for other enzymes in these families. Significance Statement Our review charts the rise of PBPK modeling for drugs cleared by non-CYP enzymes. Analyzing 58 articles and 62 non-CYP enzyme substrates, we found that UGTs, AO, FMOs, SULTs, and CES were the main enzyme families involved and that UGT2B7, UGT1A9, CES1, FMO3 and AO are the individual enzymes with the strongest PBPK modeling precedents. Approaches established for these enzymes can now be extended to additional substrates and to drugs metabolized by enzymes that are similarly well characterized.

Dissecting Parameters Contributing to the Underprediction of Aldehyde Oxidase-Mediated Metabolic Clearance of Drugs

We investigated the effect of variability and instability in aldehyde oxidase (AO) content and activity on the scaling of in vitro metabolism data. AO content and activity in human liver cytosol (HLC) and five recombinant human AO preparations (rAO) were determined using targeted proteomics and carbazeran oxidation assay, respectively. AO content was highly variable as indicated by the relative expression factor (REF; i.e., HLC to rAO content) ranging from 0.001 to 1.7 across different in vitro systems. The activity of AO in HLC degrades at a 10-fold higher rate in the presence of the substrate as compared with the activity performed after preincubation without substrate. To scale the metabolic activity from rAO to HLC, a protein-normalized activity factor (pnAF) was proposed wherein the activity was corrected by AO content, which revealed up to sixfold higher AO activity in HLC versus rAO systems. A similar value of pnAF was observed for another substrate, ripasudil. Physiologically based pharmacokinetic (PBPK) modeling revealed a significant additional clearance (CL; 66%), which allowed for the successful prediction of in vivo CL of four other substrates, i.e., O-benzyl guanine, BIBX1382, zaleplon, and zoniporide. For carbazeran, the metabolite identification study showed that the direct glucuronidation may be contributing to around 12% elimination. Taken together, this study identified differential protein content, instability of in vitro activity, role of additional AO clearance, and unaccounted metabolic pathways as plausible reasons for the underprediction of AO-mediated drug metabolism. Consideration of these factors and integration of REF and pnAF in PBPK models will allow better prediction of AO metabolism. SIGNIFICANCE STATEMENT: This study elucidated the plausible reasons for the underprediction of aldehyde oxidase (AO)-mediated drug metabolism and provided recommendations to address them. It demonstrated that integrating protein content and activity differences and accounting for the loss of AO activity, as well as consideration of extrahepatic clearance and additional pathways, would improve the in vitro to in vivo extrapolation of AO-mediated drug metabolism using physiologically based pharmacokinetic modeling.

The Disconnect in Intrinsic Clearance Determined in Human Hepatocytes and Liver Microsomes Results from Divergent Cytochrome P450 Activities

Candidate drugs may exhibit higher unbound intrinsic clearances (CLint,u) in human liver microsomes (HLMs) relative to human hepatocytes (HHs), posing a challenge as to which value is more predictive of in vivo clearance (CL). This work was aimed at better understanding the mechanism(s) underlying this 'HLM:HH disconnect' via examination of previous explanations, including passive permeability limited CL or cofactor exhaustion in hepatocytes. A series of structurally related, passively permeable (Papps > 5 × 10-6 cm/s), 5-azaquinazolines were studied in different liver fractions, and metabolic rates and routes were determined. A subset of these compounds demonstrated a significant HLM:HH (CLint,u ratio 2-26) disconnect. Compounds were metabolized via combinations of liver cytosol aldehyde oxidase (AO), microsomal cytochrome P450 (CYP) and flavin monooxygenase (FMO). For this series, the lack of concordance between CLint,u determined in HLM and HH contrasted with an excellent correlation of AO dependent CLint,u determined in human liver cytosol[Formula: see text], r2 = 0.95, P < 0.0001). The HLM:HH disconnect for both 5-azaquinazolines and midazolam was as a result of significantly higher CYP activity in HLM and lysed HH fortified with exogenous NADPH relative to intact HH. Moreover, for the 5-azaquinazolines, the maintenance of cytosolic AO and NADPH-dependent FMO activity in HH, relative to CYP, supports the conclusion that neither substrate permeability nor intracellular NADPH for hepatocytes were limiting CLint,u Further studies are required to identify the underlying cause of the lower CYP activities in HH relative to HLM and lysed hepatocytes in the presence of exogenous NADPH. SIGNIFICANCE STATEMENT: Candidate drugs may exhibit higher intrinsic clearance in human liver microsomes relative to human hepatocytes, posing a challenge as to which value is predictive of in vivo clearance. This work demonstrates that the difference in activity determined in liver fractions results from divergent cytochrome P450 but not aldehyde oxidase or flavin monooxygenase activity. This is inconsistent with explanations including substrate permeability limitations or cofactor exhaustion and should inform the focus of further studies to understand this cytochrome P450 specific disconnect phenomenon.

Inactivation of Human Aldehyde Oxidase by Small Sulfhydryl-Containing Reducing Agents

Human aldehyde oxidase (hAOX1) is a molybdoflavoenzyme that belongs to the xanthine oxidase (XO) family. hAOX1 is involved in phase I drug metabolism, but its physiologic role is not fully understood to date, and preclinical studies consistently underestimated hAOX1 clearance. In the present work, we report an unexpected effect of the common sulfhydryl-containing reducing agents, e.g., dithiothreitol (DTT), on the activity of hAOX1 and mouse aldehyde oxidases. We demonstrate that this effect is due to the reactivity of the sulfido ligand bound at the molybdenum cofactor with the sulfhydryl groups. The sulfido ligand coordinated to the Mo atom in the XO family of enzymes plays a crucial role in the catalytic cycle and its removal results in the total inactivation of these enzymes. Because liver cytosols, S9 fractions, and hepatocytes are commonly used to screen the drug candidates for hAOX1, our study suggests that DTT treatment of these samples should be avoided, otherwise false negative results by an inactivated hAOX1 might be obtained. SIGNIFICANCE STATEMENT: This work characterizes the inactivation of human aldehyde oxidase (hAOX1) by sulfhydryl-containing agents and identifies the site of inactivation. The role of dithiothreitol in the inhibition of hAOX1 should be considered for the preparation of hAOX1-containing fractions for pharmacological studies on drug metabolism and drug clearance.

Aldehyde oxidase mediated drug metabolism: an underpredicted obstacle in drug discovery and development

Aldehyde oxidase (AO) has garnered curiosity as a non-CYP metabolizing enzyme in drug development due to unexpected consequences such as toxic metabolite generation and high metabolic clearance resulting in the clinical failure of new drugs. Therefore, poor AO mediated clearance prediction in preclinical nonhuman species remains a significant obstacle in developing novel drugs. Various isoforms of AO, such as AOX1, AOX3, AOX3L1, and AOX4 exist across species, and different AO activity among humans influences the AO mediated drug metabolism. Therefore, carefully considering the unique challenges is essential in developing successful AO substrate drugs. The in vitro to in vivo extrapolation underpredicts AO mediated drug clearance due to the lack of reliable representative animal models, substrate-specific activity, and the discrepancy between absolute concentration and activity. An in vitro tool to extrapolate in vivo clearance using a yard-stick approach is provided to address the underprediction of AO mediated drug clearance. This approach uses a range of well-known AO drug substrates as calibrators for qualitative scaling new drugs into low, medium, or high clearance category drugs. So far, in vivo investigations on chimeric mice with humanized livers (humanized mice) have predicted AO mediated metabolism to the best extent. This review addresses the critical aspects of the drug discovery stage for AO metabolism studies, challenges faced in drug development, approaches to tackle AO mediated drug clearance's underprediction, and strategies to decrease the AO metabolism of drugs.

Contribution of Extrahepatic Aldehyde Oxidase Activity to Human Clearance

Aldehyde oxidase (AOX) is a soluble, cytosolic enzyme that metabolizes various N-heterocyclic compounds and organic aldehydes. It has wide tissue distribution with highest levels found in liver, kidney, and lung. Human clearance projections of AOX substrates by in vitro assessments in isolated liver fractions (cytosol, S9) and even hepatocytes have been largely underpredictive of clinical outcomes. Various hypotheses have been suggested as to why this is the case. One explanation is that extrahepatic AOX expression contributes measurably to AOX clearance and is at least partially responsible for the often observed underpredictions. Although AOX expression has been confirmed in several extrahepatic tissues, activities therein and potential contribution to overall human clearance have not been thoroughly studied. In this work, the AOX enzyme activity using the S9 fractions of select extrahepatic human tissues (kidney, lung, vasculature, and intestine) were measured using carbazeran as a probe substrate. Measured activities were scaled to a whole-body clearance using best-available parameters and compared with liver S9 fractions. Here, the combined scaled AOX clearance obtained from the kidney, lung, vasculature, and intestine is very low and amounted to <1% of liver. This work suggests that AOX metabolism from extrahepatic sources plays little role in the underprediction of activity in human. One of the notable outcomes of this work has been the first direct demonstration of AOX activity in human vasculature. SIGNIFICANCE STATEMENT: This work demonstrates aldehyde oxidase (AOX) activity is measurable in a variety of extrahepatic human tissues, including vasculature, yet activities and potential contributions to human clearance are relatively low and insignificant when compared with the liver. Additionally, the modeling of the tissue-specific in vitro kinetic data suggests that AOX may be influenced by the tissue it resides in and thus show different affinity, activity, and modified activity over time.

Recent developments in predicting CYP-independent metabolism

As lead optimization efforts have successfully reduced metabolic liabilities due to cytochrome P450 (CYP)-mediated metabolism, there has been an increase in the frequency of involvement of non-CYP enzymes in the metabolism of investigational compounds. Although there have been numerous notable advancements in the characterization of non-CYP enzymes with respect to their localization, reaction mechanisms, species differences and identification of typical substrates, accurate prediction of non-CYP-mediated clearance, with a particular emphasis with the difficulties in accounting for any extrahepatic contributions, remains a challenge. The current manuscript comprehensively summarizes the recent advancements in the prediction of drug metabolism and the in vitro to in vitro extrapolation of clearance for substrates of non-CYP drug metabolizing enzymes.