Characterization of Divergent Metabolic Pathways in Elucidating an Unexpected, Slow-Forming, and Long Half-Life Major Metabolite of Iclepertin.

PURPOSE: After single oral dosing of the glycine reuptake transporter (GlyT1) inhibitor, iclepertin (BI 425809), a single major circulating metabolite, M530a, was identified. However, upon multiple dosing, a second major metabolite, M232, was observed with exposure levels ~ twofold higher than M530a. Studies were conducted to characterize the metabolic pathways and enzymes responsible for formation of both major human metabolites. METHODS: In vitro studies were conducted with human and recombinant enzyme sources and enzyme-selective inhibitors. The production of iclepertin metabolites was monitored by LC-MS/MS. RESULTS: Iclepertin undergoes rapid oxidation to a putative carbinolamide that spontaneously opens to an aldehyde, M528, which then undergoes reduction by carbonyl reductase to the primary alcohol, M530a. However, the carbinolamide can also undergo a much slower oxidation by CYP3A to form an unstable imide metabolite, M526, that is subsequently hydrolyzed by a plasma amidase to form M232. This difference in rate of metabolism of the carbinolamine explains why high levels of the M232 metabolite were not observed in vitro and in single dose studies in humans, but were observed in longer-term multiple dose studies. CONCLUSIONS: The long half-life iclepertin metabolite M232 is formed from a common carbinolamine intermediate, that is also a precursor of M530a. However, the formation of M232 occurs much more slowly, likely contributing to its extensive exposure in vivo. These results highlight the need to employ adequate clinical study sampling periods and rigorous characterization of unexpected metabolites, especially when such metabolites are categorized as major, thus requiring safety assessment.

Predictive In Vitro-In Vivo Extrapolation for Time Dependent Inhibition of CYP1A2, CYP2C8, CYP2C9, CYP2C19, and CYP2D6 Using Pooled Human Hepatocytes, Human Liver Microsomes, and a Simple Mechanistic Static Model.

Inactivation of Cytochrome P450 (CYP450) enzymes can lead to significant increases in exposure of comedicants. The majority of reported in vitro to in vivo extrapolation (IVIVE) data have historically focused on CYP3A, leaving the assessment of other CYP isoforms insubstantial. To this end, the utility of human hepatocytes (HHEP) and human liver microsomes (HLM) to predict clinically relevant drug-drug interactions was investigated with a focus on CYP1A2, CYP2C8, CYP2C9, CYP2C19, and CYP2D6. Evaluation of IVIVE for CYP2B6 was limited to only weak inhibition. A search of the University of Washington Drug-Drug Interaction Database was conducted to identify a clinically relevant weak, moderate, and strong inhibitor for selective substrates of CYP1A2, CYP2C8, CYP2C9, CYP2C19, and CYP2D6, resulting in 18 inhibitors for in vitro characterization against 119 clinical interaction studies. Pooled human hepatocytes and HLM were preincubated with increasing concentrations of inhibitors for designated timepoints. Time dependent inhibition was detected in HLM for four moderate/strong inhibitors, suggesting that some optimization of incubation conditions (i.e., lower protein concentrations) is needed to capture weak inhibition. Clinical risk assessment was conducted by incorporating the in vitro derived kinetic parameters maximal rate of enzyme inactivation (min-1) (kinact) and concentration of inhibitor resulting in 50% of the maximum enzyme inactivation (KI) into static equations recommended by regulatory authorities. Significant overprediction was observed when applying the basic models recommended by regulatory agencies. Mechanistic static models, which consider the fraction of metabolism through the impacted enzyme, using the unbound hepatic inlet concentration lead to the best overall prediction accuracy with 92% and 85% of data from HHEPs and HLM, respectively, within twofold of the observed value. SIGNIFICANCE STATEMENT: Coupling time-dependent inactivation parameters derived from pooled human hepatocytes and human liver microsomes (HLM) with a mechanistic static model provides an easy and quantitatively accurate means to determine clinical drug-drug interaction risk from in vitro data. Optimization is needed to evaluate time-dependent inhibition (TDI) for weak and moderate inhibitors using HLM. Recommendations are made with respect to input parameters for in vitro to in vivo extrapolation (IVIVE) of TDI with non-CYP3A enzymes using available data from HLM and human hepatocytes.

HLM-beads: Rapid Assessment of Nonspecific Binding to Human Liver Microsomes Using Magnetizable Beads.

In early drug development, drug-drug interaction risk is routinely assessed using human liver microsomes (HLMs). Nonspecific binding of drugs to HLMs can affect the determination of accurate enzyme parameters (Km, Ki, KI). Previously, we described a novel in vitro model consisting of HLMs bound to magnetizable beads [HLM-magnetizable-beads system (HLM-beads)]. The HLM-beads enable rapid separation of HLMs from incubation media by applying a magnetic field. Here, HLM-beads were further characterized and evaluated as a tool to assess HLM nonspecific binding of small molecules. The free fractions (fu,mic) of 13 compounds (chosen based on their pKa values) were determined using HLM-beads under three HLM concentrations (0.025, 0.50, and 1.0 mg/ml) and compared with those determined by equilibrium dialysis. Most fu,mic values obtained using HLM-beads were within 0.5- to 2-fold of the values determined using equilibrium dialysis. The highest fold difference were observed for high binders itraconazole and BIRT2584 (1.9- to 2.9-fold), as the pronounced adsorption of these compounds to the equilibrium dialysis apparatus interfered with their fu,mic determination. Correlation and linear regression analysis of the fu,mic values generated using HLM-beads and equilibrium dialysis was conducted. Overall, a good correlation of fu,mic values obtained by the two methods were observed, as the r and R2 values from correlational analysis and linear regression analysis were >0.9 and >0.89, respectively. These studies demonstrate that HLM-beads can produce comparable fu,mic values as determined by equilibrium dialysis while reducing the time required for this type of study from hours to only 10 minutes and compound apparatus adsorption. SIGNIFICANCE STATEMENT: This work introduces a new method of rapidly assessing nonspecific microsomal binding using human liver microsomes bound to magnetizable beads.

Contribution of Major Metabolites toward Complex Drug-Drug Interactions of Deleobuvir: In Vitro Predictions and In Vivo Outcomes.

The drug-drug interaction (DDI) potential of deleobuvir, an hepatitis C virus (HCV) polymerase inhibitor, and its two major metabolites, CD 6168 (formed via reduction by gut bacteria) and deleobuvir-acyl glucuronide (AG), was assessed in vitro. Area-under-the-curve (AUC) ratios (AUCi/AUC) were predicted using a static model and compared with actual AUC ratios for probe substrates in a P450 cocktail of caffeine (CYP1A2), tolbutamide (CYP2C9), and midazolam (CYP3A4), administered before and after 8 days of deleobuvir administration to HCV-infected patients. In vitro studies assessed inhibition, inactivation and induction of P450s. Induction was assessed in a short-incubation (10 hours) hepatocyte assay, validated using positive controls, to circumvent cytotoxicity seen with deleobuvir and its metabolites. Overall, P450 isoforms were differentially affected by deleobuvir and its two metabolites. Of note was more potent CYP2C8 inactivation by deleobuvir-AG than deleobuvir and P450 induction by CD 6168 but not by deleobuvir. The predicted net AUC ratios for probe substrates were 2.92 (CYP1A2), 0.45 (CYP2C9), and 0.97 (CYP3A4) compared with clinically observed ratios of 1.64 (CYP1A2), 0.86 (CYP2C9), and 1.23 (CYP3A4). Predictions of DDI using deleobuvir alone would have significantly over-predicted the DDI potential for CYP3A4 inhibition (AUC ratio of 6.15). Including metabolite data brought the predicted net effect close to the observed DDI. However, the static model over-predicted the induction of CYP2C9 and inhibition/inactivation of CYP1A2. This multiple-perpetrator DDI scenario highlights the application of the static model for predicting complex DDI for CYP3A4 and exemplifies the importance of including key metabolites in an overall DDI assessment.

Defining the Role of Gut Bacteria in the Metabolism of Deleobuvir: In Vitro and In Vivo Studies.

Deleobuvir is a potent inhibitor of the hepatitis C virus nonstructural protein 5B polymerase. In humans, deleobuvir underwent extensive reduction to form CD 6168. This metabolite was not formed in vitro in aerobic incubations with human liver microsomes or cytosol. Anaerobic incubations of deleobuvir with rat and human fecal homogenates produced CD 6168. Using these in vitro formation rates, a retrospective analysis was conducted to assess whether the fecal formation of CD 6168 could account for the in vivo levels of this metabolite. The formation of CD 6168 was also investigated using a pseudo-germ free (pGF) rat model, in which gut microbiota were largely eradicated by antibiotic treatment. Plasma exposure (area under the curve from 0 to ∞) of CD 6168 was approximately 9-fold lower in pGF rats (146 ± 64 ng·h/ml) compared with control rats (1,312 ± 649 ng·h/ml). Similarly, in pGF rats, lower levels of CD 6168 (1.5% of the deleobuvir dose) were excreted in feces compared with control rats (42% of the deleobuvir dose). In agreement with these findings, in pGF rats, approximately all of the deleobuvir dose was excreted as deleobuvir into feces (105% of dose), whereas only 26% of the deleobuvir dose was excreted as deleobuvir in control rats. These differences in plasma and excretion profiles between pGF and control rats confirm the role of gut bacteria in the formation of CD 6168. These results underline the importance of evaluating metabolism by gut bacteria and highlight experimental approaches for nonclinical assessment of bacterial metabolism in drug development.

A numerical method for analysis of in vitro time-dependent inhibition data. Part 2. Application to experimental data.

Time-dependent inhibition (TDI) of cytochrome P450 enzymes is an important cause of drug-drug interactions. The standard approach to characterize the kinetics of TDI is to determine the rate of enzyme loss, kobs, at various inhibitor concentrations, [I], and replot the kobs versus [I] to obtain the key kinetic parameters, KI and kinact. In our companion manuscript (Part 1; Nagar et al., 2014) in this issue of Drug Metabolism and Disposition, we used simulated datasets to develop and test a new numerical method to analyze in vitro TDI data. Here, we have applied this numerical method to five TDI datasets. Experimental datasets include the inactivation of CYP2B6, CYP2C8, and CYP3A4. None of the datasets exhibited Michaelis-Menten-only kinetics, and the numerical method allowed use of more complex models to fit each dataset. Quasi-irreversible as well as partial inhibition kinetics were observed and parameterized. Three datasets required the use of a multiple-inhibitor binding model. The mechanistic and clinical implications provided by these analyses are discussed. Together with the results in Part 1, we have developed and applied a new numerical method for analysis of in vitro TDI data. This method appears to be generally applicable to model in vitro TDI data with atypical and complex kinetic schemes.

Aminobenzotriazole inhibits and induces several key drug metabolizing enzymes complicating its utility as a pan CYP inhibitor for reaction phenotyping.

Aminobenzotriazole (ABT) is commonly used as a non-selective inhibitor of cytochrome P450 (CYP) enzymes to assign contributions of CYP versus non-CYP pathways to the metabolism of new chemical entities. Despite widespread use, a systematic review of the drug-drug interaction (DDI) potential for ABT has not been published nor have the implications for using it in plated hepatocyte models for low clearance reaction phenotyping. The goal being to investigate the utility of ABT as a pan-CYP inhibitor for reaction phenotyping of low clearance compounds by evaluating stability over the incubation period, inhibition potential against UGT and sulfotransferase enzymes, and interaction with nuclear receptors involved in the regulation of drug metabolizing enzymes and transporters. Induction potential for additional inhibitors used to ascribe fraction metabolism (fm ), pathway including erythromycin, ketoconazole, azamulin, atipamezole, ZY12201, and quinidine was also investigated. ABT significantly inhibited the clearance of a non-selective UGT substrate 4-methylumbelliferone, with several UGTs shown to be inhibited using selective probe substrates in human hepatocytes and rUGTs. The inhibitors screened in the induction assay were shown to induce enzymes regulated through Aryl Hydrocarbon Receptor, Constitutive Androstane Receptor, and Pregnane X Receptor. Lastly, a case study identifying the mechanisms of a clinical DDI between Palbociclib and ARV-471 is provided as an example of the potential consequences of using ABT to derive fm . This work demonstrates that ABT is not an ideal pan-CYP inhibitor for reaction phenotyping of low clearance compounds and establishes a workflow that can be used to enable robust characterization of other prospective inhibitors.

Comparison of Commonly Used and New Methods to Determine Small Molecule Non-Specific Binding to Human Liver Microsomes.

Accurate measurement of non-specific binding of a drug candidate to human liver microsomes (HLM) can be critical for the accurate determination of key enzyme kinetic parameters such as Michaelis-Menton (Km), reversible inhibition (Ki), or inactivation (KI) constants. Several methods have been developed to determine non-specific binding of small molecules to HLM, such as rapid equilibrium dialysis (RED), ultrafiltration (UF), HLM bound to magnetizable beads (HLM-beads), ultracentrifugation (UC), the linear extrapolation stability assay (LESA), and the Transil™ system. Despite various differences in methodology between these methods, it is generally presumed that similar free fraction values (fu,mic) should be generated. To evaluate this hypothesis, a test set of 9 compounds were selected, representing low (high fu,mic value) and significant (low fu,mic value) HLM binding, respectively, across HLM concentrations tested in this manuscript. The fu,mic values were determined using a single compound concentration (1.0 µM) and three HLM concentrations (0.025, 0.50, and 1.0 mg/mL). When the HLM non-specific binding event is not extensive resulting in high fu,mic values, all methods generated similar fu,mic values. However, fu,mic values varied markedly across assay formats when high binding to HLM occurred, where fu,mic values differed by up to 33-fold depending on the method used. Potential causes for such discrepancies across the various methods employed, practical implications related to conduct the different assays, and implications to clinical drug-drug interaction (DDI) predictions are discussed.

Metabolic switching of BILR 355 in the presence of ritonavir. II. Uncovering novel contributions by gut bacteria and aldehyde oxidase.

Ritonavir (RTV) was used as a boosting agent to increase the clinical exposure of 11-ethyl-5,11-dihydro-5-methyl-8-[2-[(1-oxido-4-quinolinyl)oxy]ethyl]-6H-dipyrido[3,2-b:2',3'-e][1,4]diazepin-6-one (BILR 355), an inhibitor of the human immunodeficiency virus, by inhibiting the CYP3A-mediated metabolism of BILR 355. However, although the levels of BILR 355 increased upon concomitant administration of RTV, a metabolite of BILR 355, BILR 516, which was not detected previously in humans dosed with BILR 355 alone, became a disproportionate human metabolite with levels exceeding the parent levels at steady state. This was an unusual finding based on the in vitro and in vivo metabolic profiles of BILR 355 available at that time. Our studies reveal that BILR 355 is reduced to an intermediate, BILR 402, by gut bacteria and the reduced metabolite (BILR 402) is then oxidized by aldehyde oxidase to form BILR 516, the disproportionate human metabolite. The role of aldehyde oxidase helped to explain the somewhat unique formation of BILR 516 in humans compared with preclinical animal species. This article underlines the increasing importance of two individually atypical enzymes in drug development, gut bacterial biotransformation and aldehyde oxidase, which in combination provided a unique metabolic pathway. In addition, this article clearly elucidates an example of novel metabolic switching and, it is hoped, raises awareness of the potential for metabolic switching in combination drug therapies.

Metabolic switching of BILR 355 in the presence of ritonavir. I. Identifying an unexpected disproportionate human metabolite.

11-Ethyl-5,11-dihydro-5-methyl-8-[2-[(1-oxido-4-quinolinyl)oxy] ethyl]-6H-dipyrido[3,2-b:2',3'-e][1,4]diazepin-6-one (BILR 355) is an inhibitor of the human immunodeficiency virus-1. BILR 355 exhibited a nonlinear pharmacokinetic profile and low exposure after oral administration to humans. This article describes the in vitro metabolism of BILR 355, which is correlated with the in vivo nonlinearity findings. Our in vitro studies had demonstrated that BILR 355 was extensively metabolized by cytochrome P450 3A. Thus, BILR 355 was concomitantly administered with ritonavir (RTV) in an attempt to boost systemic exposure, which did occur in humans. In addition, the expectation was that the overall metabolism of BILR 355 would be decreased with concomitant administration of RTV. Subsequent metabolite profiling was performed using human plasma samples obtained from clinical phase Ib studies with concomitant administration of BILR 355 and RTV. A total of 18 metabolites was observed. Their structures were proposed on the basis of high-performance liquid chromatography-tandem mass spectrometry technologies, and 10 metabolites were confirmed by comparison with synthetic standards. We were surprised to find that a disproportionate human metabolite, BILR 516, was uncovered during this metabolite profiling study and pharmacokinetic analysis of BILR 516 showed that it had a longer half-life and higher exposure than the parent compound at steady state. Of interest, BILR 516 was not detected in human plasma when BILR 355 was administered alone. Therefore, whereas RTV boosted the exposure of BILR 355, it resulted in a significant metabolic switching of BILR 355. Overall, this article demonstrates an unusual example of metabolic switching and raises concern about the consequence of metabolic switching during drug development.