Machine learning-driven identification of drugs inhibiting cytochrome P450 2C9

Improving the accuracy of prediction models for small datasets of Cytochrome P450 inhibition with deep learning

The cytochrome P450 (CYP) superfamily metabolises a wide range of compounds; however, drug-induced CYP inhibition can lead to adverse interactions. Identifying potential CYP inhibitors is crucial for safe drug administration. This study investigated the application of deep learning techniques to the prediction of CYP inhibition, focusing on the challenges posed by limited datasets for CYP2B6 and CYP2C8 isoforms. To tackle these limitations, we leveraged larger datasets for related CYP isoforms, compiling comprehensive data from public databases containing IC50 values for 12,369 compounds that target seven CYP isoforms. We constructed single-task, fine-tuning, multitask, and multitask models incorporating data imputation on the missing values. Notably, the multitask models with data imputation demonstrated significant improvement in CYP inhibition prediction over the single-task models. Using the most accurate prediction models, we evaluated the inhibitory activity of approved drugs against CYP2B6 and CYP2C8. Among the 1,808 approved drugs analysed, our multitask models with data imputation identified 161 and 154 potential inhibitors of CYP2B6 and CYP2C8, respectively. This study underscores the significant potential of multitask deep learning, particularly when utilising a graph convolutional network with data imputation, to enhance the accuracy of CYP inhibition predictions under the conditions of limited data availability.Scientific contributionThis study demonstrates that even with small datasets, accurate prediction models can be constructed by utilising related data effectively. Also, our imputation techniques on the missing values improved the prediction accuracy of CYP2B6 and CYP2C8 inhibition significantly.

Employing Automated Machine Learning (AutoML) Methods to Facilitate the In Silico ADMET Properties Prediction

The rationale for using ADMET prediction tools in the early drug discovery paradigm is to guide the design of new compounds with favorable ADMET properties and ultimately minimize the attrition rates of drug failures. Artificial intelligence (AI) in in silico ADMET modeling has gained momentum due to its high-throughput and low-cost attributes. In this study, we developed a machine learning model capable of predicting 11 ADMET properties of chemical compounds. Each model was constructed by combining one of 40 classification algorithms including random forest (RF), extreme gradient boosting (XGB), support vector machine (SVM), and gradient boosting (GB) with one of three predefined hyperparameter configurations. This process can be efficiently performed using automated machine learning (AutoML) methods, which automatically search for the best combination of model algorithms and optimized hyperparameters. We developed optimal predictive models for 11 different ADMET properties using the Hyperopt-sklearn AutoML method. All of the developed models depicted an area under the ROC curve (AUC) >0.8. Furthermore, our developed models outperformed most of the ADMET properties and showed comparable performance in other properties when evaluated on external data sets and compared with published predictive models. Our results support the applicability of AutoML in ADMET prediction and will be helpful for ADMET prediction in early-stage drug discovery.

Insight into the Relationships Between Chemical, Protein and Functional Variables in the PBP/GOBP Family in Moths Based on Machine Learning

During their lives, insects must cope with a plethora of chemicals, of which a few will have an impact at the behavioral level. To detect these chemicals, insects use several protein families located in their main olfactory organs, the antennae. Inside the antennae, odorant-binding proteins (OBPs), as the most studied protein family, bind volatile chemicals to transport them. Pheromone-binding proteins (PBPs) and general-odorant-binding proteins (GOPBs) are two subclasses of OBPs and have evolved in moths with a putative olfactory role. Predictions for OBP-chemical interactions have remained limited, and functional data collected over the years unused. In this study, chemical, protein and functional data were curated, and related datasets were created with descriptors. Regression algorithms were implemented and their performance evaluated. Our results indicate that XGBoostRegressor exhibits the best performance (R2 of 0.76, RMSE of 0.28 and MAE of 0.20), followed by GradientBoostingRegressor and LightGBMRegressor. To the best of our knowledge, this is the first study showing a correlation among chemical, protein and functional data, particularly in the context of the PBP/GOBP family of proteins in moths.

Repurposing COVID-19 Compounds (via MMV COVID Box): Almitrine and Bortezomib Induce Programmed Cell Death in Trypanosoma cruzi

Chagas disease, caused by the protozoan Trypanosoma cruzi, affects millions globally, with limited treatment options available. Current therapies, such as benznidazole and nifurtimox, present challenges, including their toxicity, side effects, and inefficacy in the chronic phase. This study explores the potential of drug repurposing as a strategy to identify new treatments for T. cruzi, focusing on compounds from the Medicines for Malaria Venture (MMV) COVID Box. An initial screening of 160 compounds identified eight with trypanocidal activity, with almitrine and bortezomib showing the highest efficacy. Both compounds demonstrated significant activity against the epimastigote and amastigote stages of the parasite and showed no cytotoxicity in murine macrophage cells. Key features of programmed cell death (PCD), such as chromatin condensation, mitochondrial membrane potential disruption, and reactive oxygen species accumulation, were observed in T. cruzi treated with these compounds. The potential to induce controlled cell death of these two compounds in T. cruzi suggests they are promising candidates for further research. This study reinforces drug repurposing as a viable approach to discovering novel treatments for neglected tropical diseases like Chagas disease.

In vitro assessment of inhibitory effects of kinase inhibitors on CYP2C9, 3A and 1A2: Prediction of drug-drug interaction risk with warfarin and direct oral anticoagulants

OBJECTIVE

This study aimed to evaluate the cytochrome P450 (CYP)-mediated drug-drug interaction (DDI) potential of kinase inhibitors with warfarin and direct oral anticoagulants (DOACs).

METHODS

An in vitro CYP probe substrate cocktail assay was used to study the inhibitory effects of fifteen kinase inhibitors on CYP2C9, 3A, and 1A2. Then, DDI predictions were performed using both mechanistic static and physiologically-based pharmacokinetic (PBPK) models.

RESULTS

Linsitinib, masitinib, regorafenib, tozasertib, trametinib, and vatalanib were identified as competitive CYP2C9 inhibitors (Ki = 1.4, 1.0, 1.1, 3.8, 0.5, and 0.1 μM, respectively). Masitinib and vatalanib were competitive CYP3A inhibitors (Ki = 1.3 and 0.2 μM), and vatalanib noncompetitively inhibited CYP1A2 (Ki = 2.0 μM). Moreover, linsitinib and tozasertib were CYP3A time-dependent inhibitors (KI = 26.5 and 400.3 μM, kinact = 0.060 and 0.026 min-1, respectively). Only linsitinib showed time-dependent inhibition of CYP1A2 (KI = 13.9 μM, kinact = 0.018 min-1). Mechanistic static models identified possible DDI risks for linsitinib and vatalanib with (S)-/(R)-warfarin, and for masitinib with (S)-warfarin. PBPK simulations further confirmed that vatalanib may increase (S)- and (R)-warfarin exposure by 4.37- and 1.80-fold, respectively, and that linsitinib may increase (R)-warfarin exposure by 3.10-fold. Mechanistic static models predicted a smaller risk of DDIs between kinase inhibitors and apixaban or rivaroxaban. The greatest AUC increases (1.50-1.74) were predicted for erlotinib in combination with apixaban and rivaroxaban. Linsitinib, masitinib, and vatalanib were predicted to have a smaller effect on apixaban and rivaroxaban AUCs (AUCR 1.22-1.53). No kinase inhibitor was predicted to increase edoxaban exposure.

CONCLUSIONS

Our results suggest that several kinase inhibitors, including vatalanib and linsitinib, can cause CYP-mediated drug-drug interactions with warfarin and, to a lesser extent, with apixaban and rivaroxaban. The work provides mechanistic insights into the risk of DDIs between kinase inhibitors and anticoagulants, which can be used to avoid preventable DDIs in the clinic.

Machine Learning-Based In Silico Prediction of the Inhibitory Activity of Chemical Substances Against Rat and Human Cytochrome P450s

The prediction of cytochrome P450 inhibition by a computational (quantitative) structure-activity relationship approach using chemical structure information and machine learning would be useful for toxicity research as a simple and rapid in silico tool. However, there are few in silico models focusing on the species differences between rat and human in the P450s inhibition. This study aimed to establish in silico models to classify chemical substances as inhibitors or non-inhibitors of various rat and human P450s, using only molecular descriptors. Using the in-house test results from our in vitro experiments, we used 326 substances for model construction and internal validation data. Apart from the 326 substances, 60 substances were used as external validation data set. We focused on seven rat P450s (CYP1A1, CYP1A2, CYP2B1, CYP2C6, CYP2D1, CYP2E1, and CYP3A2) and 11 human P450s (CYP1A1, CYP1A2, CYP1B1, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4). Most of the models established using XGBoost showed an area under the receiver operating characteristic curve (ROC-AUC) of 0.8 or more in the internal validation. When we set an applicability domain for the models and confirmed their generalization performance through external validation, most of the models showed an ROC-AUC of 0.7 or more. Interestingly, for CYP1A1 and CYP1A2, we discovered that a human P450 inhibitory activity model can predict rat P450 inhibitory activity and vice versa. These models are the first attempts to predict inhibitory activity against a wide variety of P450s in both rats and humans using chemical structure information. Our experimental results and in silico models would be helpful to support information for species similarities and differences in chemical-induced toxicity.

GTransCYPs: an improved graph transformer neural network with attention pooling for reliably predicting CYP450 inhibitors

State‑of‑the‑art medical studies proved that predicting CYP450 enzyme inhibitors is beneficial in the early stage of drug discovery. However, accurate machine learning-based (ML) in silico methods for predicting CYP450 inhibitors remains challenging. Here, we introduce GTransCYPs, an improved graph neural network (GNN) with a transformer mechanism for predicting CYP450 inhibitors. This model significantly enhances the discrimination between inhibitors and non-inhibitors for five major CYP450 isozymes: 1A2, 2C9, 2C19, 2D6, and 3A4. GTransCYPs learns information patterns from molecular graphs by aggregating node and edge representations using a transformer. The GTransCYPs model utilizes transformer convolution layers to process features, followed by a global attention-pooling technique to synthesize the graph-level information. This information is then fed through successive linear layers for final output generation. Experimental results demonstrate that the GTransCYPs model achieved high performance, outperforming other state-of-the-art methods in CYP450 prediction.Scientific contributionThe prediction of CYP450 inhibition via computational techniques utilizing biological information has emerged as a cost-effective and highly efficient approach. Here, we presented a deep learning (DL) architecture based on GNN with transformer mechanism and attention pooling (GTransCYPs) to predict CYP450 inhibitors. Four GTransCYPs of different pooling technique were tested on an experimental tasks on the CYP450 prediction problem for the first time. Graph transformer with attention pooling algorithm achieved the best performances. Comparative and ablation experiments provide evidence of the efficacy of our proposed method in predicting CYP450 inhibitors. The source code is publicly available at https://github.com/zonwoo/GTransCYPs .

Machine Learning and Deep Learning Models for Predicting Noncovalent Inhibitors of AmpC β-Lactamase

Continuous use of antibiotics leads to the ability of bacteria to adapt by developing complex antibiotic resistance (AR) mechanisms. The synthesis of β-lactamases is a widely observed AR mechanism. The class C β-lactamase (AmpC) causes significant resistance toward β-lactam antibiotics, and new treatments are urgently needed. Noncovalent inhibitors have been developed against a broad spectrum of β-lactamases. In this study, we developed robust and accurate models for predicting noncovalent inhibitors of AmpC using large compound data sets and machine/deep learning modeling. We created support vector machine (SVM), random forest (RF), and feed-forward neural network (FFNN) classification models. The cross-validation (CV) accuracies varied between 80 and 82%, as combined models reached an accuracy of 83%. We analyzed the physicochemical characteristics of the noncovalent inhibitors and predicted the binding modes for some of them. Such models are helpful for identifying new noncovalent inhibitors in order to establish novel solutions against the growing resistance to standard β-lactam inhibitors. The best RF, SVM, and FFNN models for predicting noncovalent inhibitors of AmpC β-lactamase are available in the GitHub repository, https://github.com/UPCmctr/ML-DL-AmpC-B-lactamase.

Enzyme-mediated drug-drug interactions: a review of in vivo and in vitro methodologies, regulatory guidance, and translation to the clinic

Enzyme-mediated pharmacokinetic drug-drug interactions can be caused by altered activity of drug metabolizing enzymes in the presence of a perpetrator drug, mostly via inhibition or induction. We identified a gap in the literature for a state-of-the art detailed overview assessing this type of DDI risk in the context of drug development. This manuscript discusses in vitro and in vivo methodologies employed during the drug discovery and development process to predict clinical enzyme-mediated DDIs, including the determination of clearance pathways, metabolic enzyme contribution, and the mechanisms and kinetics of enzyme inhibition and induction. We discuss regulatory guidance and highlight the utility of in silico physiologically-based pharmacokinetic modeling, an approach that continues to gain application and traction in support of regulatory filings. Looking to the future, we consider DDI risk assessment for targeted protein degraders, an emerging small molecule modality, which does not have recommended guidelines for DDI evaluation. Our goal in writing this report was to provide early-career researchers with a comprehensive view of the enzyme-mediated pharmacokinetic DDI landscape to aid their drug development efforts.

Designing cytochrome P450 enzymes for use in cancer gene therapy

Cancer is a significant global socioeconomic burden, as millions of new cases and deaths occur annually. In 2020, almost 10 million cancer deaths were recorded worldwide. Advancements in cancer gene therapy have revolutionized the landscape of cancer treatment. An approach with promising potential for cancer gene therapy is introducing genes to cancer cells that encode for chemotherapy prodrug metabolizing enzymes, such as Cytochrome P450 (CYP) enzymes, which can contribute to the effective elimination of cancer cells. This can be achieved through gene-directed enzyme prodrug therapy (GDEPT). CYP enzymes can be genetically engineered to improve anticancer prodrug conversion to its active metabolites and to minimize chemotherapy side effects by reducing the prodrug dosage. Rational design, directed evolution, and phylogenetic methods are some approaches to developing tailored CYP enzymes for cancer therapy. Here, we provide a compilation of genetic modifications performed on CYP enzymes aiming to build highly efficient therapeutic genes capable of bio-activating different chemotherapeutic prodrugs. Additionally, this review summarizes promising preclinical and clinical trials highlighting engineered CYP enzymes' potential in GDEPT. Finally, the challenges, limitations, and future directions of using CYP enzymes for GDEPT in cancer gene therapy are discussed.

Marine Prostanoids with Cytotoxic Activity from Octocoral Clavularia spp

Octocoral of the genus Clavularia is a kind of marine invertebrate possessing abundant cytotoxic secondary metabolites, such as prostanoids and dolabellanes. In our continuous natural product study of C. spp., two previously undescribed prostanoids [clavulone I-15-one (1) and 12-O-deacetylclavulone I (2)] and eleven known analogs (3-13) were identified. The structures of these new compounds were elucidated based on analysis of their 1D and 2D NMR, HRESIMS, and IR data. Additionally, all tested prostanoids (1 and 3-13) showed potent cytotoxic activities against the human oral cancer cell line (Ca9-22). The major compound 3 showed cytotoxic activity against the Ca9-22 cells with the IC50 value of 2.11 ± 0.03 μg/mL, which echoes the cytotoxic effect of the coral extract. In addition, in silico tools were used to predict the possible effects of isolated compounds on human tumor cell lines and nitric oxide production, as well as the pharmacological potentials.

Targeting PIM kinases in cancer therapy: An update on pharmacological small-molecule inhibitors

PIM kinases, a serine/threonine kinase family with three isoforms, has been well-known to participate in multiple physiological processes by phosphorylating various downstream targets. Accumulating evidence has recently unveiled that aberrant upregulation of PIM kinases (PIM1, PIM2, and PIM3) are closely associated with tumor cell proliferation, migration, survival, and even resistance. Inhibiting or silencing of PIM kinases has been reported have remarkable antitumor effects, such as anti-proliferation, pro-apoptosis and resensitivity, indicating the therapeutic potential of PIM kinases as potential druggable targets in many types of human cancers. More recently, several pharmacological small-molecule inhibitors have been preclinically and clinically evaluated and showed their therapeutic potential; however, none of them has been approved for clinical application so far. Thus, in this perspective, we focus on summarizing the oncogenic roles of PIM kinases, key signaling network, and pharmacological small-molecule inhibitors, which will provide a new clue on discovering more candidate antitumor drugs targeting PIM kinases in the future.

Computational and artificial intelligence-based approaches for drug metabolism and transport prediction

Drug metabolism and transport, orchestrated by drug-metabolizing enzymes (DMEs) and drug transporters (DTs), are implicated in drug-drug interactions (DDIs) and adverse drug reactions (ADRs). Reliable and precise predictions of DDIs and ADRs are critical in the early stages of drug development to reduce the rate of drug candidate failure. A variety of experimental and computational technologies have been developed to predict DDIs and ADRs. Recent artificial intelligence (AI) approaches offer new opportunities for better predicting and understanding the complex processes related to drug metabolism and transport. We summarize the role of major DMEs and DTs, and provide an overview of current progress in computational approaches for the prediction of drug metabolism, transport, and DDIs, with an emphasis on AI including machine learning (ML) and deep learning (DL) modeling.

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

Drug metabolism and excretion play crucial roles in determining the efficacy and safety of drug candidates, and predicting these processes is an essential part of drug discovery and development. In recent years, artificial intelligence (AI) has emerged as a powerful tool for predicting drug metabolism and excretion, offering the potential to speed up drug development and improve clinical success rates. This review highlights recent advances in AI-based drug metabolism and excretion prediction, including deep learning and machine learning algorithms. We provide a list of public data sources and free prediction tools for the research community. We also discuss the challenges associated with the development of AI models for drug metabolism and excretion prediction and explore future perspectives in the field. We hope this will be a helpful resource for anyone who is researching in silico drug metabolism, excretion, and pharmacokinetic properties.

An Insight into the Metabolism of 2,5-Disubstituted Monotetrazole Bearing Bisphenol Structures: Emerging Bisphenol A Structural Congeners

The non-estrogenic 2,5-disubstituted tetrazole core-bearing bisphenol structures (TbB) are being researched as emerging structural congeners of Bisphenol A, an established industrial endocrine disruptor. However, there is no understanding of TbB's adverse effects elicited via metabolic activation. Therefore, the current study aimed to investigate the metabolism of TbB ligands, with in silico results serving as a guide for in vitro studies. The Cytochrome P450 enzymes (CYP) inhibitory assay of TbB ligands on the seven human liver CYP isoforms (i.e., 1A2, 2A6, 2D6, 2C9, 2C8, 2C19, and 3A4) using human liver microsomes (HLM) revealed TbB ligand 223-3 to have a 50% inhibitory effect on all the CYP isoforms at a 10 μM concentration, except 1A2. The TbB ligand 223-10 inhibited 2B6 and 2C8, whereas the TbB ligand 223-2 inhibited only 2C9. The first-order inactivity rate constant (Kobs) studies indicated TbB ligands 223-3, 223-10 to be time-dependent (TD) inhibitors, whereas the TbB 223-2 ligand did not show such a significant effect. The 223-3 exhibited a TD inhibition for 2C9, 2C19, and 1A2 with Kobs values of 0.0748, 0.0306, and 0.0333 min-1, respectively. On the other hand, the TbB ligand 223-10 inhibited 2C9 in a TD inhibition manner with Kobs value 0.0748 min-1. However, the TbB ligand 223-2 showed no significant TD inhibition effect on the CYPs. The 223-2 ligand biotransformation pathway by in vitro studies in cryopreserved human hepatocytes suggested the clearance via glucuronidation with the predominant detection of only 223-2 derived mono glucuronide as a potential inactive metabolite. The present study demonstrated that the 223-2 ligand did not elicit any significant adverse effect via metabolic activation, thus paving the way for its in vivo drug-drug interactions (DDI) studies.

DEEPCYPs: A deep learning platform for enhanced cytochrome P450 activity prediction

Cytochrome P450 (CYP) is a superfamily of heme-containing oxidizing enzymes involved in the metabolism of a wide range of medicines, xenobiotics, and endogenous compounds. Five of the CYPs (1A2, 2C9, 2C19, 2D6, and 3A4) are responsible for metabolizing the vast majority of approved drugs. Adverse drug-drug interactions, many of which are mediated by CYPs, are one of the important causes for the premature termination of drug development and drug withdrawal from the market. In this work, we reported in silicon classification models to predict the inhibitory activity of molecules against these five CYP isoforms using our recently developed FP-GNN deep learning method. The evaluation results showed that, to the best of our knowledge, the multi-task FP-GNN model achieved the best predictive performance with the highest average AUC (0.905), F1 (0.779), BA (0.819), and MCC (0.647) values for the test sets, even compared to advanced machine learning, deep learning, and existing models. Y-scrambling testing confirmed that the results of the multi-task FP-GNN model were not attributed to chance correlation. Furthermore, the interpretability of the multi-task FP-GNN model enables the discovery of critical structural fragments associated with CYPs inhibition. Finally, an online webserver called DEEPCYPs and its local version software were created based on the optimal multi-task FP-GNN model to detect whether compounds bear potential inhibitory activity against CYPs, thereby promoting the prediction of drug-drug interactions in clinical practice and could be used to rule out inappropriate compounds in the early stages of drug discovery and/or identify new CYPs inhibitors.

Machine learning and structure-based modeling for the prediction of UDP-glucuronosyltransferase inhibition

UDP-glucuronosyltransferases (UGTs) are responsible for 35% of the phase II drug metabolism. In this study, we focused on UGT1A1, which is a key UGT isoform. Strong inhibition of UGT1A1 may trigger adverse drug/herb-drug interactions, or result in disorders of endobiotic metabolism. Most of the current machine learning methods predicting the inhibition of drug metabolizing enzymes neglect protein structure and dynamics, both being essential for the recognition of various substrates and inhibitors. We performed molecular dynamics simulations on a homology model of the human UGT1A1 structure containing both the cofactor- (UDP-glucuronic acid) and substrate-binding domains to explore UGT conformational changes. Then, we created models for the prediction of UGT1A1 inhibitors by integrating information on UGT1A1 structure and dynamics, interactions with diverse ligands, and machine learning. These models can be helpful for further prediction of drug-drug interactions of drug candidates and safety treatments.

•

UGTs are responsible for 35% of the phase II drug metabolism reactions

•

We created machine learning models for prediction of UGT1A1 inhibitors

•

Our simulations suggested key residues of UGT1A1 involved in the substrate binding

A unified GCNN model for predicting CYP450 inhibitors by using graph convolutional neural networks with attention mechanism

Undesirable drug-drug interactions (DDIs) may lead to serious adverse side effects when more than two drugs are administered to a patient simultaneously. One of the most common DDIs is caused by unexpected inhibition of a specific human cytochrome P450 (CYP450), which plays a dominant role in the metabolism of the co-administered drugs. Therefore, a unified and reliable method for predicting the potential inhibitors of CYP450 family is extremely important in drug development. In this work, graph convolutional neural network (GCN) with attention mechanism and 1-D convolutional neural network (CNN) were used to extract the features of CYP ligands and the binding sites of CYP450 respectively, which were then combined to establish a unified GCN-CNN (GCNN) model for predicting the inhibitors of 5 dominant CYP isoforms, i.e., 1A2, 2C9, 2C19, 2D6, and 3A4. Overall, the established GCNN model showed good performances on the test samples and achieved better performances than the recently proposed iCYP-MFE model by using the same datasets. Based on the heat-map analysis of the resulting molecular graphs, the key structural determinants of the CYP inhibitors were further explored.

ABCG2/BCRP transport mechanism revealed through kinetically excited targeted molecular dynamics simulations

ABCG2/BCRP is an ABC transporter that plays an important role in tissue protection by exporting endogenous substrates and xenobiotics. ABCG2 is of major interest due to its involvement in multidrug resistance (MDR), and understanding its complex efflux mechanism is essential to preventing MDR and drug-drug interactions (DDI). ABCG2 export is characterized by two major conformational transitions between inward- and outward-facing states, the structures of which have been resolved. Yet, the entire transport cycle has not been characterized to date. Our study bridges the gap between the two extreme conformations by studying connecting pathways. We developed an innovative approach to enhance molecular dynamics simulations, ‘kinetically excited targeted molecular dynamics’, and successfully simulated the transitions between inward- and outward-facing states in both directions and the transport of the endogenous substrate estrone 3-sulfate. We discovered an additional pocket between the two substrate-binding cavities and found that the presence of the substrate in the first cavity is essential to couple the movements between the nucleotide-binding and transmembrane domains. Our study shed new light on the complex efflux mechanism, and we provided transition pathways that can help to identify novel substrates and inhibitors of ABCG2 and probe new drug candidates for MDR and DDI.