WhichP450: a multi-class categorical model to predict the major metabolising CYP450 isoform for a compound

Investigation of in silico studies for cytochrome P450 isoforms specificity

Cytochrome P450 (CYP450) enzymes comprise a highly diverse superfamily of heme-thiolate proteins that responsible for catalyzing over 90 % of enzymatic reactions associated with xenobiotic metabolism in humans. Accurately predicting whether chemicals are substrates or inhibitors of different CYP450 isoforms can aid in pre-selecting hit compounds for the drug discovery process, chemical toxicology studies, and patients treatment planning. In this work, we investigated in silico studies on CYP450s specificity over past twenty years, categorizing these studies into structure-based and ligand-based approaches. Subsequently, we utilized 100 of the most frequently prescribed drugs to test eleven machine learning-based prediction models which were published between 2015 and 2024. We analyzed various aspects of the evaluated models, such as their datasets, algorithms, and performance. This will give readers with a comprehensive overview of these prediction models and help them choose the most suitable one to do prediction. We also provide our insights for future research trend in both structure-based and ligand-based approaches in this field.

Assessment of the in vitro metabolic stability of CEP-37440, a selective FAK/ALK inhibitor, in HLMs using fast UPLC-MS/MS method: in silico metabolic lability and DEREK alerts screening

Introduction

CEP-37440 was synthesized and supplied by the research and development division of Teva Branded Pharmaceutical Products (West Chester, PA, United States). CEP-37440 represents a newly developed compound that exhibits selectivity inhibition of Focal Adhesion Kinase and Anaplastic Lymphoma Kinase FAK/ALK receptors, demonstrating novel characteristics as an orally active inhibitor. The simultaneous inhibition of ALK and FAK can effectively address resistance and enhance the therapeutic efficacy against tumors through a synergistic mechanism.

Methods

The objective of this research was to create an LC-MS/MS method that is precise, efficient, environmentally friendly, and possesses a high level of sensitivity for the quantification of CEP-37440 in human liver microsomes (HLMs). The aforementioned approach was subsequently employed to evaluate the metabolic stability of CEP-37440 in HLMs in an in vitro setting. The validation procedures for the LC-MS/MS analytical method in the HLMs were performed following the bio-analytical method validation guidelines set out by the US-FDA. The AGREE program was utilized to assess the ecological impacts of the current LC-MS/MS methodology.

Results and Discussion

The calibration curve linearity was seen in the range of 1-3000 ng/mL. The inter-day accuracy (% RE) exhibited a range of -2.33% to 3.22%, whilst the intra-day accuracy demonstrated a range of -4.33% to 1.39%. The inter-day precision (% RSD) exhibited a range of 0.38% to 3.60%, whilst the intra-day precision demonstrated a range of 0.16% to 6.28%. The determination of the in vitro half-life (t1/2) and moderate intrinsic clearance (Clint) of CEP-37440 yielded values of 23.24 min and 34.74 mL/min/kg, respectively. The current manuscript is considered the first analytical study for CEP-37440 quantification with the application to metabolic stability assessment. These results suggest that CEP-37440 can be categorized as a pharmaceutical agent with a moderate extraction ratio. Consequently, it is postulated that the administration of CEP-37440 to patients may not lead to the accrual of dosages within the human organs. According to in silico P450 metabolic and DEREK software, minor structural alterations to the ethanolamine moiety or substitution of the group in drug design have the potential to enhance the metabolic stability and safety profile of novel derivatives in comparison to CEP-37440.

Prediction of Cytochrome P450 Substrates Using the Explainable Multitask Deep Learning Models

Cytochromes P450 (P450s or CYPs) are the most important phase I metabolic enzymes in the human body and are responsible for metabolizing ∼75% of the clinically used drugs. P450-mediated metabolism is also closely associated with the formation of toxic metabolites and drug-drug interactions. Therefore, it is of high importance to predict if a compound is the substrate of a given P450 in the early stage of drug development. In this study, we built the multitask learning models to simultaneously predict the substrates of five major drug-metabolizing P450 enzymes, namely, CYP3A4, 2C9, 2C19, 2D6, and 1A2, based on the collected substrate data sets. Compared to the single-task model and conventional machine learning models, the multitask fingerprints and graph neural networks model achieved superior performance with the average AUC values of 90.8% on the test set. Notably, the multitask model demonstrated its good performance on the small amount of substrate data sets such as CYP1A2, 2C9, and 2C19. In addition, the Shapley additive explanation and the attention mechanism were used to reveal specific substructures associated with P450 substrates, which were further confirmed and complemented by the substructure mining tool and the literature.

An ultra-fast ultra-high-performance liquid chromatography-tandem mass spectrometry method for estimating the in vitro metabolic stability of palbociclib in human liver microsomes: In silico study for metabolic lability, absorption, distribution, metabolism, and excretion features, and DEREK alerts screening

Palbociclib (Ibrance; Pfizer) was approved for the management of metastatic breast cancer characterized by hormone receptor-positive/human epidermal growth factor receptor 2 negative status. The objective of this study was to create a fast, precise, environmentally friendly, and highly sensitive ultra-high-performance liquid chromatography-tandem mass spectrometry approach for quantifying palbociclib (PAB) in human liver microsomes with the application for assessing metabolic stability. The validation features were performed in agreement with the bioanalytical method validation standards outlined by the US Food and Drug Administration. The StarDrop software (WhichP450 and DEREK modules) was used in screening the metabolic lability and structural alerts of PAB. The separation of PAB and encorafenib (as an internal standard) was achieved on a C8 column, employing an isocratic mobile phase. The inter-day and intra-day accuracy and precision ranged from -6.00% to 4.64% and from -2.33% to 3.13%, respectively. The constructed calibration curve displayed a linearity in the range of 1-3000 ng/mL. The sensitivity of the established approach was proven by the lower limit of quantification of 0.73 ng/mL. The Analytical GREEness calculator results revealed the high level of greenness of the developed method. The PAB's metabolic stability (t1/2 of 18.5 min and a moderate clearance (Clint) of 44.8 mL/min/kg) suggests a high extraction ratio medication that matched the WhichP450 software results.

Characterization of the in vitro metabolic profile of nazartinib in HLMs using UPLC-MS/MS method: In silico metabolic lability and DEREK structural alerts screening using StarDrop software

The orally given, irreversible, third-generation inhibitor of the epidermal growth factor receptor (EGFR), known as Nazartinib (EGF816), is now undergoing investigation in Phase II clinical trials conducted by Novartis for Non-Small Cell Lung Cancer. The primary aim of the current research was to establish a rapid, specific, environmentally friendly, and highly versatile UPLC-MS/MS methodology for the determination of nazartinib (NZT) levels in human liver microsomes (HLMs). Subsequently, same approach was used to examine the metabolic stability of NZT. The UPLC-MS/MS method employed in HLMs was validated as stated in the bioanalytical method validation criteria outlined by the US- FDA. The evaluation of the metabolic stability of NZT and the identification of potentially structural alarms were performed using the StarDrop software package that includes the P450 and DEREK software. The calibration curve for NZT showed a linearity in the range from 1 to 3000 ng/mL. The inter-day accuracy and precision exhibited a range of values between -4.33 % and 4.43 %, whereas the intra-day accuracy and precision shown a range of values between -2.78 % and 7.10 %. The sensitivity of the developed approach was verified through the determination of a LLOQ of 0.39 ng/mL. The intrinsic clearance and in vitro half-life of NZT were assessed to be 46.48 mL/min/kg and 17.44 min, respectively. In our preceding inquiry, we have effectively discerned the bioactivation center, denoted by the carbon atom between the unsaturated conjugated system and aliphatic linear tertiary amine. In the context of computational software, making minor adjustments or substituting the dimethylamino-butenoyl moiety throughout the drug design process may increase the metabolic stability and safety properties of new synthesized derivatives. The efficiency of utilizing different in silico software approaches to conserve resources and reduce effort was proved by the outcomes attained from in vitro incubation experiments and the use of NZT in silico software.

Ultra-fast UPLC-MS/MS approach for estimating X-376 in human liver microsomes: Evaluation of metabolic stability via in silico software and in vitro analysis

X-376 is a novel anaplastic lymphoma kinase (ALK) inhibitor that is capable of penetrating the blood brain barrier. This makes it suitable for use in patients with ALK-positive non-small cell lung cancer (NSCLC) who have metastases in the central nervous system. This study developed a highly sensitive, fast, eco-friendly, and reliable UPLC-MS/MS approach to quantify X-376 in human liver microsomes (HLMs). This approach was used to evaluate X-376's metabolic stability in HLMs in vitro. The UPLC-MS/MS analytical technique validation followed US-FDA bio-analytical method validation guidelines. StarDrop software, containing P450 metabolic and DEREK modules, was utilized to scan X-376's chemical structure for metabolic lability and hazardous warnings. X-376 and Encorafenib (ENF as internal standard) were resoluted on the Eclipse Plus C18 column utilizing an isocratic mobile phase method. The X-376 calibration curve was linear from 1 to 3000 ng/mL. The precision and accuracy of this study's UPLC-MS/MS approach were tested for intra- and inter-day measurements. Inter-day accuracy was -1.32% to 9.36% while intra-day accuracy was -1.5% to 10.00%. The intrinsic clearance (Clint) and in vitro half-life (t1/2) of X-376 were 59.77 mL/min/kg and 13.56 min. The high extraction ratio of X-376 supports the 50 mg twice-daily dose for ALK-positive NSCLC and CNS metastases patients. In silico software suggests that simple structural changes to the piperazine ring or group substitution in drug design may improve metabolic stability and safety compared to X-376.

Predicting routes of phase I and II metabolism based on quantum mechanics and machine learning

Unexpected metabolism could lead to the failure of many late-stage drug candidates or even the withdrawal of approved drugs. Thus, it is critical to predict and study the dominant routes of metabolism in the early stages of research.We describe the development and validation of a 'WhichEnzyme' model that accurately predicts the enzyme families most likely to be responsible for a drug-like molecule's metabolism. Furthermore, we combine this model with our previously published regioselectivity models for Cytochromes P450, Aldehyde Oxidases, Flavin-containing Monooxygenases, UDP-glucuronosyltransferases and Sulfotransferases - the most important Phase I and Phase II drug metabolising enzymes - and a 'WhichP450' model that predicts the Cytochrome P450 isoform(s) responsible for a compound's metabolism.The regioselectivity models are based on a mechanistic understanding of these enzymes' actions and use quantum mechanical simulations with machine learning methods to accurately predict sites of metabolism and the resulting metabolites. We train heuristics based on the outputs of the 'WhichEnzyme', 'WhichP450', and regioselectivity models to determine the most likely routes of metabolism and metabolites to be observed experimentally.Finally, we demonstrate that this combination delivers high sensitivity in identifying experimentally reported metabolites and higher precision than other methods for predicting in vivo metabolite profiles.

Selective and brain-penetrant ACSS2 inhibitors target breast cancer brain metastatic cells

Breast cancer brain metastasis (BCBM) typically results in an end-stage diagnosis and is hindered by a lack of brain-penetrant drugs. Tumors in the brain rely on the conversion of acetate to acetyl-CoA by the enzyme acetyl-CoA synthetase 2 (ACSS2), a key regulator of fatty acid synthesis and protein acetylation. Here, we used a computational pipeline to identify novel brain-penetrant ACSS2 inhibitors combining pharmacophore-based shape screen methodology with absorption, distribution, metabolism, and excretion (ADME) property predictions. We identified compounds AD-5584 and AD-8007 that were validated for specific binding affinity to ACSS2. Treatment of BCBM cells with AD-5584 and AD-8007 leads to a significant reduction in colony formation, lipid storage, acetyl-CoA levels and cell survival in vitro. In an ex vivo brain-tumor slice model, treatment with AD-8007 and AD-5584 reduced pre-formed tumors and synergized with irradiation in blocking BCBM tumor growth. Treatment with AD-8007 reduced tumor burden and extended survival in vivo. This study identifies selective brain-penetrant ACSS2 inhibitors with efficacy towards breast cancer brain metastasis.

DeepP450: Predicting Human P450 Activities of Small Molecules by Integrating Pretrained Protein Language Model and Molecular Representation

Cytochrome P450 enzymes (CYPs) play a crucial role in Phase I drug metabolism in the human body, and CYP activity toward compounds can significantly affect druggability, making early prediction of CYP activity and substrate identification essential for therapeutic development. Here, we established a deep learning model for assessing potential CYP substrates, DeepP450, by fine-tuning protein and molecule pretrained models through feature integration with cross-attention and self-attention layers. This model exhibited high prediction accuracy (0.92) on the test set, with area under the receiver operating characteristic curve (AUROC) values ranging from 0.89 to 0.98 in substrate/nonsubstrate predictions across the nine major human CYPs, surpassing current benchmarks for CYP activity prediction. Notably, DeepP450 uses only one model to predict substrates/nonsubstrates for any of the nine CYPs and exhibits certain generalizability on novel compounds and different categories of human CYPs, which could greatly facilitate early stage drug design by avoiding CYP-reactive compounds.

Discovery of novel brain permeable human ACSS2 inhibitors for blocking breast cancer brain metastatic growth

Breast-cancer brain metastasis (BCBM) poses a significant clinical challenge, resulting in an end-stage diagnosis and hindered by limited therapeutic options. The blood-brain barrier (BBB) acts as an anatomical and physiological hurdle for therapeutic compounds, restricting the effective delivery of therapies to the brain. In order to grow and survive in a nutrient-poor environment, tumors in the brain must adapt to their metabolic needs, becoming highly dependent on acetate. These tumors rely on the conversion of acetate to acetyl-CoA by the enzyme Acetyl-CoA synthetase 2 (ACSS2), a key metabolic enzyme involved in regulating fatty acid synthesis and protein acetylation in tumor cells. ACSS2 has emerged as a crucial enzyme required for the growth of tumors in the brain. Here, we utilized a computational pipeline, combining pharmacophore-based shape screen methodology with ADME property predictions to identify novel brain-permeable ACSS2 inhibitors. From a small molecule library, this approach identified 30 potential ACSS2 binders, from which two candidates, AD-5584 and AD-8007, were validated for their binding affinity, predicted metabolic stability, and, notably, their ability to traverse the BBB. We show that treatment of BCBM cells, MDA-MB-231BR, with AD-5584 and AD-8007 leads to a significant reduction in lipid storage, reduction in colony formation, and increase in cell death in vitro . Utilizing an ex vivo orthotopic brain-slice tumor model, we show that treatment with AD-8007 and AD-5584 significantly reduces tumor size and synergizes with radiation in blocking BCBM tumor growth ex vivo. Importantly, we show that following intraperitoneal injections with AD-5584 and AD-8007, we can detect these compounds in the brain, confirming their BBB permeability. Thus, we have identified and validated novel ACSS2 inhibitor candidates for further drug development and optimization as agents for treating patients with breast cancer brain metastasis.

Discovery, Crystallographic Studies, and Mechanistic Investigations of Novel Phenylalanine Derivatives Bearing a Quinazolin-4-one Scaffold as Potent HIV Capsid Modulators

Optimization of compound 11L led to the identification of novel HIV capsid modulators, quinazolin-4-one-bearing phenylalanine derivatives, displaying potent antiviral activities against both HIV-1 and HIV-2. Notably, derivatives 12a2 and 21a2 showed significant improvements, with 2.5-fold over 11L and 7.3-fold over PF74 for HIV-1, and approximately 40-fold over PF74 for HIV-2. The X-ray co-crystal structures confirmed the multiple pocket occupation of 12a2 and 21a2 in the binding site. Mechanistic studies revealed a dual-stage inhibition profile, where the compounds disrupted capsid-host factor interactions at the early stage and promoted capsid misassembly at the late stage. Remarkably, 12a2 and 21a2 significantly promoted capsid misassembly, outperforming 11L, PF74, and LEN. The substitution of easily metabolized amide bond with quinolin-4-one marginally enhanced the stability of 12a2 in human liver microsomes compared to controls. Overall, 12a2 and 21a2 highlight their potential as potent HIV capsid modulators, paving the way for future advancements in anti-HIV drug design.

An Ultrafast UPLC-MS/MS Method for Characterizing the In Vitro Metabolic Stability of Acalabrutinib

Acalabrutinib, commercially known as Calquence®, is a pharmacological molecule that has robust inhibitory activity against Bruton tyrosine kinase. The medicine in question was carefully developed by the esteemed pharmaceutical company AstraZeneca. The FDA granted authorization on 21 November 2019 for the utilization of acalabrutinib (ACB) in the treatment of small lymphocytic lymphoma (SLL) or chronic lymphocytic leukemia (CLL) in adult patients. The aim of this study was to develop a UPLC-MS/MS method that is effective, accurate, environmentally sustainable, and has a high degree of sensitivity. The methodology was specifically developed with the intention of quantifying ACB in human liver microsomes (HLMs). The methodology described above was subsequently utilized to assess the metabolic stability of ACB in HLMs in an in vitro environment. The validation procedures for the UPLC-MS/MS method in the HLMs were conducted in accordance with the bioanalytical method validation criteria established by the U.S.- DA. The utilization of the StarDrop software (version 6.6), which integrates the P450 metabolic module and DEREK software (KB 2018 1.1), was employed for the purpose of evaluating the metabolic stability and identifying potential hazardous alarms associated with the chemical structure of ACB. The calibration curve, as established by the ACB, demonstrated a linear correlation across the concentration range of 1 to 3000 ng/mL in the matrix of HLMs. The present study conducted an assessment of the accuracy and precision of the UPLC-MS/MS method in quantifying inter-day and intra-day fluctuations. The inter-day accuracy demonstrated a spectrum of values ranging from -1.00% to 8.36%, whilst the intra-day accuracy presented a range of values spanning from -2.87% to 4.11%. The t1/2 and intrinsic clearance (Clint) of ACB were determined through in vitro testing to be 20.45 min and 39.65 mL/min/kg, respectively. The analysis concluded that the extraction ratio of ACB demonstrated a moderate level, thus supporting the recommended dosage of ACB (100 mg) to be administered twice daily for the therapeutic treatment of persons suffering from B-cell malignancies. Several computational tools have suggested that introducing minor structural alterations to the butynoyl group, particularly the alpha, beta-unsaturated amide moiety, or substituting this group during the drug design procedure, could potentially enhance the metabolic stability and safety properties of novel derivatives in comparison to ACB.

Evaluation of Alectinib Metabolic Stability in HLMs Using Fast LC-MS/MS Method: In Silico ADME Profile, P450 Metabolic Lability, and Toxic Alerts Screening

Alectinib, also known as Alecensa®, is prescribed for the therapeutic treatment of individuals diagnosed with metastatic non-small cell lung cancer (NSCLC) who have a specific genetic mutation referred to as anaplastic lymphoma kinase (ALK) positivity. The Food and Drug Administration granted regular approval to alectinib, a drug developed by Hoffmann-La Roche, Inc. (Basel, Switzerland)/Genentech, Inc. (South San Francisco, CA, USA), on 6 November 2017. The screening of the metabolic stability and identification of hazardous alarms within the chemical structure of ALC was conducted using the StarDrop software package (version 6.6), which incorporates the P450 metabolic module and DEREK software (KB 2018 1.1). The primary aim of this investigation was to develop a high-throughput and accurate LC-MS/MS technique for the quantification of ALC in the metabolic matrix (human liver microsomes; HLMs). The aforementioned methodology was subsequently employed to assess the metabolic stability of ALC in HLMs through in vitro tests, with the obtained results further validated using in silico software. The calibration curve of the ALC showed a linear correlation that exists within the concentration range from 1 to 3000 ng/mL. The LC-MS/MS approach that was recommended exhibited accuracy and precision levels for both inter-day and intra-day measurements. Specifically, the accuracy values ranged from -2.56% to 3.45%, while the precision values ranged from -3.78% to 4.33%. The sensitivity of the established approach was proved by its ability to adhere to an LLOQ of 0.82 ng/mL. The half-life (t1/2) and intrinsic clearance (Clint) of ALC were estimated to be 22.28 min and 36.37 mL/min/kg, correspondingly, using in vitro experiments. The ALC exhibited a moderate extraction ratio. The metabolic stability and safety properties of newly created derivatives can be enhanced by making modest adjustments to the morpholine and piperidine rings or by substituting the substituent, as per computational software. In in silico ADME prediction, ALC was shown to have poor water solubility and high gastrointestinal absorption along with inhibition of some cytochrome P450s (CYP2C19 and CYP2C9) without inhibition of others (CYP1A2, CYP3A4, and CYP2D6) and P-glycoprotein substrate. The study design that involves using both laboratory experiments and different in silico software demonstrates a novel and groundbreaking approach in the establishment and uniformization of LC-MS/MS techniques for the estimation of ALC concentrations, identifying structural alerts and the assessment of its metabolic stability. The utilization of this study strategy has the potential to be employed in the screening and optimization of prospective compounds during the drug creation process. This strategy may also facilitate the development of novel derivatives of the medicine that maintain the same biological action by targeted structural modifications, based on an understanding of the structural alerts included within the chemical structure of ALC.

Design, synthesis, and mechanistic study of 2-piperazineone-bearing peptidomimetics as novel HIV capsid modulators

HIV-1 capsid (CA) is an attractive target for its indispensable roles in the viral life cycle. We report the design, synthesis, and mechanistic study of a novel series of 2-piperazineone peptidomimetics as HIV capsid modulators by mimicking the structure of host factors binding to CA. F-Id-3o was the most potent compound from the synthesized series, with an anti-HIV-1 EC50 value of 6.0 μM. However, this series of compounds showed a preference for HIV-2 inhibitory activity, in which Id-3o revealed an EC50 value of 2.5 μM (anti-HIV-2 potency), an improvement over PF74. Interestingly, F-Id-3o did bind HIV-1 CA monomers and hexamers with comparable affinity, unlike PF74, consequently showing antiviral activity in the early and late stages of the HIV-1 lifecycle. Molecular dynamics simulations shed light on F-Id-3o and Id-3o binding modes within the HIV-1/2 CA protein and provide a possible explanation for the increased anti-HIV-2 potency. Metabolic stability assays in human plasma and human liver microsomes indicated that although F-Id-3o has enhanced metabolic stability over PF74, further optimization is necessary. Moreover, we utilized computational prediction of drug-like properties and metabolic stability of F-Id-3o and PF74, which correlated well with experimentally derived metabolic stability, providing an efficient computational pipeline for future preselection based on metabolic stability prediction. Overall, the 2-piperazineone-bearing peptidomimetics are a promising new chemotype in the CA modulators class with considerable optimization potential.

Artificial Intelligence in Drug Toxicity Prediction: Recent Advances, Challenges, and Future Perspectives

Toxicity prediction is a critical step in the drug discovery process that helps identify and prioritize compounds with the greatest potential for safe and effective use in humans, while also reducing the risk of costly late-stage failures. It is estimated that over 30% of drug candidates are discarded owing to toxicity. Recently, artificial intelligence (AI) has been used to improve drug toxicity prediction as it provides more accurate and efficient methods for identifying the potentially toxic effects of new compounds before they are tested in human clinical trials, thus saving time and money. In this review, we present an overview of recent advances in AI-based drug toxicity prediction, including the use of various machine learning algorithms and deep learning architectures, of six major toxicity properties and Tox21 assay end points. Additionally, we provide a list of public data sources and useful toxicity prediction tools for the research community and highlight the challenges that must be addressed to enhance model performance. Finally, we discuss future perspectives for AI-based drug toxicity prediction. This review can aid researchers in understanding toxicity prediction and pave the way for new methods of drug discovery.

Assessment of In Silico and In Vitro Selpercatinib Metabolic Stability in Human Liver Microsomes Using a Validated LC-MS/MS Method

Selpercatinib (SLP; brand name Retevmo®) is a selective and potent RE arranged during transfection (RET) inhibitor. On 21 September 2022, the FDA granted regular approval to SLP (Retevmo, Eli Lilly, and Company). It is considered the only and first RET inhibitor for adults with metastatic or locally advanced solid tumors with RET gene fusion. In the current experiment, a highly specific, sensitive, and fast liquid chromatography tandem mass spectrometry (LC-MS/MS) method for quantifying SLP in human liver microsomes (HLMs) was developed and applied to the metabolic stability evaluation of SLP. The LC-MS/MS method was validated following the bioanalytical methodology validation guidelines outlined by the FDA (linearity, selectivity, matrix effect, accuracy, precision, carryover, and extraction recovery). SLP was detected by a triple quadrupole detector (TQD) using a positive ESI source and multiple reaction monitoring (MRM) mode for mass spectrometric analysis and estimation of analytes ions. The IS-normalized matrix effect and extraction recovery were acceptable according to the FDA guidelines for the bioanalysis of SLP. The SLP calibration standards were linear from 1 to 3000 ng/mL HLMs matrix, with a regression equation (y = 1.7298x + 3.62941) and coefficient of variation (r2 = 0.9949). The intra-batch and inter-batch precision and accuracy of the developed LC-MS/MS method were −6.56–5.22% and 5.08–3.15%, respectively. SLP and filgotinib (FLG) (internal standard; IS) were chromatographically separated using a Luna 3 µm PFP (2) stationary phase (150 × 4.6 mm) with an isocratic mobile phase at 23 ± 1 °C. The limit of quantification (LOQ) was 0.78 ng/mL, revealing the LC-MS/MS method sensitivity. The intrinsic clearance and in vitro t1/2 (metabolic stability) of SLP in the HLMs matrix were 34 mL/min/kg and 23.82 min, respectively, which proposed an intermediate metabolic clearance rate of SLP, confirming the great value of this type of kinetic experiment for more accurate metabolic stability predictions. The literature review approved that the established LC-MS/MS method is the first developed and reported method for quantifying SLP metabolic stability.

Development and Validation of a Rapid LC-MS/MS Method for Quantifying Alvocidib: In Silico and In Vitro Metabolic Stability Estimation in Human Liver Microsomes

Alvocidib (AVC; flavopiridol) is a potent cyclin-dependent kinase inhibitor used in patients with acute myeloid leukemia (AML). The FDA has approved orphan drug designation to AVC for treating patients with AML. In the current work, the in silico calculation of AVC metabolic lability was done using the P450 metabolism module of the StarDrop software package, that is expressed as a composite site lability (CSL). This was followed by establishing an LC-MS/MS analytical method for AVC estimation in human liver microsomes (HLMs) to assess metabolic stability. AVC and glasdegib (GSB), used as internal standards (IS), were separated utilizing a C18 column (reversed chromatography) with an isocratic mobile phase. The lower limit of quantification (LLOQ) was 5.0 ng/mL, revealing the sensitivity of the established LC-MS/MS analytical method that exhibited a linearity in the range 5-500 ng/mL in the HLMs matrix with correlation coefficient (R² = 0.9995). The interday and intraday accuracy and precision of the established LC-MS/MS analytical method were -1.4% to 6.7% and -0.8% to 6.4%, respectively, confirming the reproducibility of the LC-MS/MS analytical method. The calculated metabolic stability parameters were intrinsic clearance (CL_int) and in vitro half-life (t_1/2) of AVC at 26.9 µL/min/mg and 25.8 min, respectively. The in silico results from the P450 metabolism model matched the results generated from in vitro metabolic incubations; therefore, the in silico software can be used to predict the metabolic stability of the drugs, saving time and resources. AVC exhibits a moderate extraction ratio, indicating reasonable in vivo bioavailability. The established chromatographic methodology was the first LC-MS/MS method designed for AVC estimation in HLMs matrix that was applied for AVC metabolic stability estimation.

Development of an LC-MS/MS Method for Quantification of Sapitinib in Human Liver Microsomes: In Silico and In Vitro Metabolic Stability Evaluation

Sapitinib (AZD8931, SPT) is a tyrosine kinase inhibitor of the epidermal growth factor receptor (EGFR) family (pan-erbB). In multiple tumor cell lines, STP has been shown to be a much more potent inhibitor of EGF-driven cellular proliferation than gefitinib. In the current study, a highly sensitive, rapid, and specific LC-MS/MS analytical method for the estimation of SPT in human liver microsomes (HLMs) was established with application to metabolic stability assessment. The LC-MS/MS analytical method was validated in terms of linearity, selectivity, precision, accuracy, matrix effect, extraction recovery, carryover, and stability following the FDA guidelines for bioanalytical method validation. SPT was detected using electrospray ionization (ESI) as an ionization source under multiple reaction monitoring (MRM) in the positive ion mode. The IS-normalized matrix factor and extraction recovery were acceptable for the bioanalysis of SPT. The SPT calibration curve was linear, from 1 ng/mL to 3000 ng/mL HLM matrix samples, with a linear regression equation of y = 1.7298x + 3.62941 (r² = 0.9949). The intraday and interday accuracy and precision values of the LC-MS/MS method were -1.45-7.25% and 0.29-6.31%, respectively. SPT and filgotinib (FGT) (internal standard; IS) were separated through the use of an isocratic mobile phase system with a Luna 3 µm PFP(2) column (150 × 4.6 mm) stationary phase column. The limit of quantification (LOQ) was 0.88 ng/mL, confirming the LC-MS/MS method sensitivity. The intrinsic clearance and in vitro half-life of STP were 38.48 mL/min/kg and 21.07 min, respectively. STP exhibited a moderate extraction ratio that revealed good bioavailability. The literature review demonstrated that the current analytical method is the first developed LC-MS/MS method for the quantification of SPT in an HLM matrix with application to SPT metabolic stability evaluation.

Rapid LC-MS/MS Bosutinib Quantification with Applications in Metabolic Stability Estimation

Bosutinib (BOS) is FDA approved drug for the treatment of chronic phase (CP) Philadelphia chromosome-positive (Ph+) chronic myelogenous leukemia (CML). We report a fast, sensitive, and simple LC-MS/MS method, validated for the determination of BOS in human liver microsomes, utilizing tofacitinib (TOF) as the internal standard. The separation of BOS and TOF was done using a 1.8 μm C18 column (2.1 × 50 mm) at room temperature using the isocratic elution system of acetonitrile-water (30:70, v/v) containing 0.1 M formic acid at a flow rate of 0.15 mL/min, and a triple-quadrupole tandem mass spectrometer (TQD-MS) with an electrospray ionization (ESI) source that was operated in the positive ion mode. The method was validated according to the European Medicines Agency, and the rapid and specific quantification of BOS in human liver microsomes was achieved in the range of 5-200 ng/mL, with a determination coefficient of 0.999. Intra- and inter-day accuracy and precision values were <4% in all cases. The procedure is rapid, specific, reliable, and can be applied in metabolic stability evaluations since it is the first LC-MS/MS method specific to BOS quantification. The metabolic stability assessment of BOS showed high CL_int (34.3 µL/min/mg) and short in vitro t_1/2 values of 20.21 min, indicating that BOS may be rapidly eliminated from the blood by the liver.

Discovery and Mechanistic Investigation of Piperazinone Phenylalanine Derivatives with Terminal Indole or Benzene Ring as Novel HIV-1 Capsid Modulators

HIV-1 capsid (CA) performs multiple roles in the viral life cycle and is a promising target for antiviral development. In this work, we describe the design, synthesis, assessment of antiviral activity, and mechanistic investigation of 20 piperazinone phenylalanine derivatives with a terminal indole or benzene ring. Among them, F₂-7f exhibited moderate anti-HIV-1 activity with an EC₅₀ value of 5.89 μM, which was slightly weaker than the lead compound PF74 (EC₅₀ = 0.75 μM). Interestingly, several compounds showed a preference for HIV-2 inhibitory activity, represented by 7f with an HIV-2 EC₅₀ value of 4.52 μM and nearly 5-fold increased potency over anti-HIV-1 (EC₅₀ = 21.81 μM), equivalent to PF74 (EC₅₀ = 4.16 μM). Furthermore, F₂-7f preferred to bind to the CA hexamer rather than to the monomer, similar to PF74, according to surface plasmon resonance results. Molecular dynamics simulation indicated that F₂-7f and PF74 bound at the same site. Additionally, we computationally analyzed the ADMET properties for 7f and F₂-7f. Based on this analysis, 7f and F₂-7f were predicted to have improved drug-like properties and metabolic stability over PF74, and no toxicities were predicted based on the chemotype of 7f and F₂-7f. Finally, the experimental metabolic stability results of F₂-7f in human liver microsomes and human plasma moderately correlated with our computational prediction. Our findings show that F₂-7f is a promising small molecule targeting the HIV-1 CA protein with considerable development potential.

Design, Synthesis, and Mechanistic Study of 2-Pyridone-Bearing Phenylalanine Derivatives as Novel HIV Capsid Modulators

The AIDS pandemic is still of importance. HIV-1 and HIV-2 are the causative agents of this pandemic, and in the absence of a viable vaccine, drugs are continually required to provide quality of life for infected patients. The HIV capsid (CA) protein performs critical functions in the life cycle of HIV-1 and HIV-2, is broadly conserved across major strains and subtypes, and is underexploited. Therefore, it has become a therapeutic target of interest. Here, we report a novel series of 2-pyridone-bearing phenylalanine derivatives as HIV capsid modulators. Compound FTC-2 is the most potent anti-HIV-1 compound in the new series of compounds, with acceptable cytotoxicity in MT-4 cells (selectivity index HIV-1 > 49.57; HIV-2 > 17.08). However, compound TD-1a has the lowest EC50 in the anti-HIV-2 assays (EC50 = 4.86 ± 1.71 μM; CC50= 86.54 ± 29.24 μM). A water solubility test found that TD-1a showed a moderately increased water solubility compared with PF74, while the water solubility of FTC-2 was improved hundreds of times. Furthermore, we use molecular simulation studies to provide insight into the molecular contacts between the new compounds and HIV CA. We also computationally predict drug-like properties and metabolic stability for FTC-2 and TD-1a. Based on this analysis, TD-1a is predicted to have improved drug-like properties and metabolic stability over PF74. This study increases the repertoire of CA modulators and has important implications for developing anti-HIV agents with novel mechanisms, especially those that inhibit the often overlooked HIV-2.

Predicting Regioselectivity of AO, CYP, FMO, and UGT Metabolism Using Quantum Mechanical Simulations and Machine Learning

Unexpected metabolism in modification and conjugation phases can lead to the failure of many late-stage drug candidates or even withdrawal of approved drugs. Thus, it is critical to predict the sites of metabolism (SoM) for enzymes, which interact with drug-like molecules, in the early stages of the research. This study presents methods for predicting the isoform-specific metabolism for human AOs, FMOs, and UGTs and general CYP metabolism for preclinical species. The models use semi-empirical quantum mechanical simulations, validated using experimentally obtained data and DFT calculations, to estimate the reactivity of each SoM in the context of the whole molecule. Ligand-based models, trained and tested using high-quality regioselectivity data, combine the reactivity of the potential SoM with the orientation and steric effects of the binding pockets of the different enzyme isoforms. The resulting models achieve κ values of up to 0.94 and AUC of up to 0.92.

A validated LC-MS/MS analytical method for the quantification of pemigatinib: metabolic stability evaluation in human liver microsomes

Pemigatinib (PMB) is a small molecule inhibitor of fibroblast growth factor receptor 1 (FGFR1), FGFR2 and FGFR3. On April 17, 2020, the US Food and Drug Administration granted accelerated approval for PMB for the treatment of adults with previously treated, unresectable metastatic or locally advanced cholangiocarcinoma with a fibroblast growth factor receptor 2 (FGFR2) fusion or other rearrangement. PMB is considered the first targeted treatment for cholangiocarcinoma approved in the US. In this study, in silico prediction of PMB metabolic stability was done using the WhichP450 module of the StarDrop software package. Further, an LC-MS/MS analytical method was developed for PMB quantification in human liver microsomes (HLM) to experimentally assess metabolic stability. PMB and flavopiridol (FVL), used as an internal standard IS, were resolved using an isocratic mobile phase and a C18 stationary phase. The LC-MS/MS method showed linearity in the range of 5 to 500 ng mL⁻¹ in an HLM matrix (R² = 0.9995). The lower limit of quantification (LLOQ) was 5 ng mL⁻¹, indicating sensitivity. The inter- and intra-day accuracy and precision were within a variability of 10, confirming the reproducibility of the method. The measured in vitro half-life and intrinsic clearance of PMB were 27.29 min and 25.40 μL min⁻¹ mg⁻¹, respectively. PMB showed a moderate extraction ratio suggesting good bioavailability. The developed analytical method is the first LC-MS/MS method specific for PMB quantification with application to metabolic stability assessment.

PMB showed a moderate extraction ratio suggesting good bioavailability. The developed analytical method is the first LC-MS/MS method specific for PMB quantification with application to metabolic stability assessment.

Profiling of in vivo, in vitro and reactive zorifertinib metabolites using liquid chromatography ion trap mass spectrometry

Zorifertinib (AZD-3759; ZFB) is a potent, novel, oral, small molecule used for the treatment of non-small cell lung cancer (NSCLC). ZFB is Epidermal Growth Factor Receptor (EGFR) inhibitor that is characterized by good permeability of the blood–brain barrier for (NSCLC) patients with EGFR mutations. The present research reports the profiling of in vitro, in vivo and reactive metabolites of ZFB. Prediction of vulnerable metabolic sites and reactivity pathways (cyanide and GSH) of ZFB were performed by WhichP450™ module (StarDrop software package) and XenoSite reactivity model (XenoSite Web Predictor-Home), respectively. ZFB in vitro metabolites were done by incubation with isolated perfused rat liver hepatocytes and rat liver microsomes (RLMs). Extraction of ZFB and its related metabolites from the incubation matrix was done by protein precipitation. In vivo metabolism was performed by giving ZFB (10 mg kg⁻¹) through oral gavage to Sprague Dawley rats that were housed in metabolic cages. Urine was collected at specific time intervals (0, 6, 12, 18, 24, 48, 72, 96 and 120 h) from ZFB dosing. The collected urine samples were filtered then stored at −70 °C. N-Methyl piperazine ring of ZFB undergoes phase I metabolism forming iminium intermediates that were stabilized using potassium cyanide as a trapping agent. Incubation of ZFB with RLMs were performed in the presence of 1.0 mM KCN and 1.0 mM glutathione to check reactive intermediates as it is may be responsible for toxicities associated with ZFB usage. For in vitro metabolites there were six in vitro phase I metabolites, three in vitro phase II metabolites, seven reactive intermediates (four GSH conjugates and three cyano adducts) of ZFB were detected by LC-IT-MS. For in vivo metabolites there were six in vivo phase I and three in vivo phase II metabolites of ZFB were detected by LC-IT-MS. In vitro and in vivo phase I metabolic pathways were N-demethylation, O-demethylation, hydroxylation, reduction, defluorination and dechlorination. In vivo phase II metabolic reaction was direct sulphate and glucuronic acid conjugation with ZFB.

Six in vitro phase I metabolites, three in vitro phase II metabolites, seven reactive intermediates (four GSH conjugates and three cyano adducts), six in vivo phase I and three in vivo phase II metabolites of ZFB were detected by LC-IT-MS.

Artificial Intelligence, Machine Learning, and Deep Learning in Real-Life Drug Design Cases

The discovery and development of drugs is a long and expensive process with a high attrition rate. Computational drug discovery contributes to ligand discovery and optimization, by using models that describe the properties of ligands and their interactions with biological targets. In recent years, artificial intelligence (AI) has made remarkable modeling progress, driven by new algorithms and by the increase in computing power and storage capacities, which allow the processing of large amounts of data in a short time. This review provides the current state of the art of AI methods applied to drug discovery, with a focus on structure- and ligand-based virtual screening, library design and high-throughput analysis, drug repurposing and drug sensitivity, de novo design, chemical reactions and synthetic accessibility, ADMET, and quantum mechanics.

LC-MS/MS Estimation of Rociletinib Levels in Human Liver Microsomes: Application to Metabolic Stability Estimation

BACKGROUND

Rociletinib (CO-1686; RLC) is a new, small molecule that is orally administered to inhibit mutant-selective covalent inhibitor of most epidermal growth factor receptor (EGFR)-mutated forms, including T790M, L858R, and exon 19 deletions, but not exon 20 insertions. Non-small-cell lung cancer (NSCLC) with a gene mutation that encodes EGFR is sensitive to approved EGFR inhibitors, but usually resistance develops, which is frequently mediated by T790M EGFR mutation. RLC is an EGFR inhibitor found to be active in preclinical models of EGFR-mutated NSCLC with or without T790M.

METHODS

In silico drug metabolism prediction of RLC was executed with the aid of the WhichP450 module (StarDrop software package) to verify its metabolic liability. Second, a fast, accurate, and competent LC-MS/MS assay was developed for RLC quantification to determine its metabolic stability. RLC and bosutinib (BOS) (internal standard; IS) were separated using an isocratic elution system with a C18 column (reversed stationary phase).

RESULTS

The developed LC-MS/MS analytical method showed linearity of 5-500 ng/mL with r2 ≥ 0.9998 in the human liver microsomes (HLMs) matrix. A limit of quantification of 4.6 ng/mL revealed the sensitivity of the analytical method, while the acquired inter- and intra-day accuracy and precision values below 4.63% inferred the method reproducibility. RLC metabolic stability estimation was calculated using intrinsic clearance (20.15 µL/min/mg) and in vitro half-life (34.39 min) values.

CONCLUSION

RLC exhibited a moderate extraction ratio indicative of good bioavailability. The developed analytical method herein is the first LC-MS/MS assay for RLC metabolic stability.

Cytochrome P450 enzymes: a review on drug metabolizing enzyme inhibition studies in drug discovery and development

Assessment of drug candidate's potential to inhibit cytochrome P450 (CYP) enzymes remains crucial in pharmaceutical drug discovery and development. Both direct and time-dependent inhibition of drug metabolizing CYP enzymes by the concomitant administered drug is the leading cause of drug-drug interactions (DDIs), resulting in the increased toxicity of the victim drug. In this context, pharmaceutical companies have grown increasingly diligent in limiting CYP inhibition liabilities of drug candidates in the early stages and examining risk assessments throughout the drug development process. This review discusses different strategies and decision-making processes for assessing the drug-drug interaction risks by enzyme inhibition and lays particular emphasis on in vitro study designs and interpretation of CYP inhibition data in a stage-appropriate context.

CYPstrate: A Set of Machine Learning Models for the Accurate Classification of Cytochrome P450 Enzyme Substrates and Non-Substrates

The interaction of small organic molecules such as drugs, agrochemicals, and cosmetics with cytochrome P450 enzymes (CYPs) can lead to substantial changes in the bioavailability of active substances and hence consequences with respect to pharmacological efficacy and toxicity. Therefore, efficient means of predicting the interactions of small organic molecules with CYPs are of high importance to a host of different industries. In this work, we present a new set of machine learning models for the classification of xenobiotics into substrates and non-substrates of nine human CYP isozymes: CYPs 1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, and 3A4. The models are trained on an extended, high-quality collection of known substrates and non-substrates and have been subjected to thorough validation. Our results show that the models yield competitive performance and are favorable for the detection of CYP substrates. In particular, a new consensus model reached high performance, with Matthews correlation coefficients (MCCs) between 0.45 (CYP2C8) and 0.85 (CYP3A4), although at the cost of coverage. The best models presented in this work are accessible free of charge via the "CYPstrate" module of the New E-Resource for Drug Discovery (NERDD).

Rapid Optimization of the Metabolic Stability of a Human Immunodeficiency Virus Type-1 Capsid Inhibitor Using a Multistep Computational Workflow

Poor metabolic stability of the human immunodeficiency virus type-1 (HIV-1) capsid (CA) inhibitor PF-74 is a major concern in its development toward clinical use. To improve on the metabolic stability, we employed a novel multistep computationally driven workflow, which facilitated the rapid design of improved PF-74 analogs in an efficient manner. Using this workflow, we designed three compounds that interact specifically with the CA interprotomer pocket, inhibit HIV-1 infection, and demonstrate enantiomeric preference. Moreover, using this workflow, we were able to increase the metabolic stability 204-fold in comparison to PF-74 in only three analog steps. These results demonstrate our ability to rapidly design CA compounds using a novel computational workflow that has improved metabolic stability over the parental compound. This workflow can be further applied to the redesign of PF-74 and other promising inhibitors with a stability shortfall.

A comprehensive comparison of molecular feature representations for use in predictive modeling

Machine learning methods are commonly used for predicting molecular properties to accelerate material and drug design. An important part of this process is deciding how to represent the molecules. Typically, machine learning methods expect examples represented by vectors of values, and many methods for calculating molecular feature representations have been proposed. In this paper, we perform a comprehensive comparison of different molecular features, including traditional methods such as fingerprints and molecular descriptors, and recently proposed learnable representations based on neural networks. Feature representations are evaluated on 11 benchmark datasets, used for predicting properties and measures such as mutagenicity, melting points, activity, solubility, and IC50. Our experiments show that several molecular features work similarly well over all benchmark datasets. The ones that stand out most are Spectrophores, which give significantly worse performance than other features on most datasets. Molecular descriptors from the PaDEL library seem very well suited for predicting physical properties of molecules. Despite their simplicity, MACCS fingerprints performed very well overall. The results show that learnable representations achieve competitive performance compared to expert based representations. However, task-specific representations (graph convolutions and Weave methods) rarely offer any benefits, even though they are computationally more demanding. Lastly, combining different molecular feature representations typically does not give a noticeable improvement in performance compared to individual feature representations.

Artificial Intelligence in Drug Discovery: A Comprehensive Review of Data-driven and Machine Learning Approaches

As expenditure on drug development increases exponentially, the overall drug discovery process requires a sustainable revolution. Since artificial intelligence (AI) is leading the fourth industrial revolution, AI can be considered as a viable solution for unstable drug research and development. Generally, AI is applied to fields with sufficient data such as computer vision and natural language processing, but there are many efforts to revolutionize the existing drug discovery process by applying AI. This review provides a comprehensive, organized summary of the recent research trends in AI-guided drug discovery process including target identification, hit identification, ADMET prediction, lead optimization, and drug repositioning. The main data sources in each field are also summarized in this review. In addition, an in-depth analysis of the remaining challenges and limitations will be provided, and proposals for promising future directions in each of the aforementioned areas.

Practical Applications of Deep Learning To Impute Heterogeneous Drug Discovery Data

Contemporary deep learning approaches still struggle to bring a useful improvement in the field of drug discovery because of the challenges of sparse, noisy, and heterogeneous data that are typically encountered in this context. We use a state-of-the-art deep learning method, Alchemite, to impute data from drug discovery projects, including multitarget biochemical activities, phenotypic activities in cell-based assays, and a variety of absorption, distribution, metabolism, and excretion (ADME) endpoints. The resulting model gives excellent predictions for activity and ADME endpoints, offering an average increase in R² of 0.22 versus quantitative structure-activity relationship methods. The model accuracy is robust to combining data across uncorrelated endpoints and projects with different chemical spaces, enabling a single model to be trained for all compounds and endpoints. We demonstrate improvements in accuracy on the latest chemistry and data when updating models with new data as an ongoing medicinal chemistry project progresses.

Advances in the study of drug metabolism - symposium report of the 12th Meeting of the International Society for the Study of Xenobiotics (ISSX)

The 12th International Society for the Study of Xenobiotics (ISSX) meeting, held in Portland, OR, USA from July 28 to 31, 2019, was attended by diverse members of the pharmaceutical sciences community. The ISSX New Investigators Group provides learning and professional growth opportunities for student and early career members of ISSX. To share meeting content with those who were unable to attend, the ISSX New Investigators herein elected to highlight the "Advances in the Study of Drug Metabolism" symposium, as it engaged attendees with diverse backgrounds. This session covered a wide range of current topics in drug metabolism research including predicting sites and routes of metabolism, metabolite identification, ligand docking, and medicinal and natural products chemistry, and highlighted approaches complemented by computational modeling. In silico tools have been increasingly applied in both academic and industrial settings, alongside traditional and evolving in vitro techniques, to strengthen and streamline pharmaceutical research. Approaches such as quantum mechanics simulations facilitate understanding of reaction energetics toward prediction of routes and sites of drug metabolism. Furthermore, in tandem with crystallographic and orthogonal wet lab techniques for structural validation of drug metabolizing enzymes, in silico models can aid understanding of substrate recognition by particular enzymes, identify metabolic soft spots and predict toxic metabolites for improved molecular design. Of note, integration of chemical synthesis and biosynthesis using natural products remains an important approach for identifying new chemical scaffolds in drug discovery. These subjects, compiled by the symposium organizers, presenters, and the ISSX New Investigators Group, are discussed in this review.

Identification and in silico prediction of metabolites of tebufenozide derivatives by major human cytochrome P450 isoforms

Cytochrome P450 (CYP) enzymes constitute a superfamily of heme-containing monooxygenases. CYPs are involved in the metabolism of many chemicals such as drugs and agrochemicals. Therefore, examining the metabolic reactions by each CYP isoform is important to elucidate their substrate recognition mechanisms. The clarification of these mechanisms may be useful not only for the development of new drugs and agrochemicals, but also for risk assessment of chemicals. In our previous study, we identified the metabolites of tebufenozide, an insect growth regulator, formed by two human CYP isoforms: CYP3A4 and CYP2C19. The accessibility of each site of tebufenozide to the reaction center of CYP enzymes and the susceptibility of each hydrogen atom for metabolism by CYP enzymes were evaluated by a docking simulation and hydrogen atom abstraction energy estimation at the density functional theory level, respectively. In this study, the same in silico prediction method was applied to the metabolites of tebufenozide derivatives by major human CYPs (CYP1A2, 2C9, 2C19, 2D6, and 3A4). In addition, the production rate of the metabolites by CYP3A4 was quantitively analyzed by frequency based on docking simulation and hydrogen atom abstraction energy using the classical QSAR approach. Then, the obtained QSAR model was applied to predict the sites of metabolism and the metabolite production order by each CYP isoform.

Composition and Orientation of the Core Region of Novel HIV-1 Entry Inhibitors Influences Metabolic Stability

Fostemsavir/temsavir is an investigational HIV-1 entry inhibitor currently in late-stage clinical trials. Although it holds promise to be a first-in-class Env-targeted entry inhibitor for the clinic, issues with bioavailability relegate its use to salvage therapies only. As such, the development of a small molecule HIV-1 entry inhibitor that can be used in standard combination antiretroviral therapy (cART) remains a longstanding goal for the field. We previously demonstrated the ability of extending the chemotypes available to this class of inhibitor as the first step towards this overarching goal. In addition to poor solubility, metabolic stability is a crucial determinant of bioavailability. Therefore, in this short communication, we assess the metabolic stabilities of five of our novel chemotype entry inhibitors. We found that changing the piperazine core region of temsavir alters the stability of the compound in human liver microsome assays. Moreover, we identified an entry inhibitor with more than twice the metabolic stability of temsavir and demonstrated that the orientation of the core replacement is critical for this increase. This work further demonstrates the feasibility of our long-term goal—to design an entry inhibitor with improved drug-like qualities—and warrants expanded studies to achieve this.

Computational methods and tools to predict cytochrome P450 metabolism for drug discovery

In this review, we present important, recent developments in the computational prediction of cytochrome P450 (CYP) metabolism in the context of drug discovery. We discuss in silico models for the various aspects of CYP metabolism prediction, including CYP substrate and inhibitor predictors, site of metabolism predictors (i.e., metabolically labile sites within potential substrates) and metabolite structure predictors. We summarize the different approaches taken by these models, such as rule‐based methods, machine learning, data mining, quantum chemical methods, molecular interaction fields, and docking. We highlight the scope and limitations of each method and discuss future implications for the field of metabolism prediction in drug discovery.