ADMET-PrInt: Evaluation of ADMET Properties: Prediction and Interpretation.

Great progress in the development of computational strategies for drug design applications has revolutionized the process of searching for new drugs. Although the focus of in silico strategies is still put on the provision of the desired activity of a compound to the considered target, characterization of a compound in terms of its physicochemical and ADMET properties becomes an indispensable element of computer-aided drug design protocols. In the study, an online application ADMET-PrInt for in silico assessment of selected compound features: cardiotoxicity, solubility, genotoxicity, membrane permeability, and plasma protein binding was prepared. In addition to the prediction of particular property, ADMET-PrInt enables also the identification of compound features influencing this property thanks to the application of two explainability approaches: local interpretabile model-agnostic explanations and counterfactual analysis. It is an important factor for medicinal chemists, as it greatly facilitates the process of optimization of the compound structure in terms of the evaluated properties. The intuitive webpage, available at admet.if-pan.krakow.pl, allows making use of all predictive and interpretability models also by nonexperts and nonprogrammers.

Generative Models Should at Least Be Able to Design Molecules That Dock Well: A New Benchmark.

Designing compounds with desired properties is a key element of the drug discovery process. However, measuring progress in the field has been challenging due to the lack of realistic retrospective benchmarks, and the large cost of prospective validation. To close this gap, we propose a benchmark based on docking, a widely used computational method for assessing molecule binding to a protein. Concretely, the goal is to generate drug-like molecules that are scored highly by SMINA, a popular docking software. We observe that various graph-based generative models fail to propose molecules with a high docking score when trained using a realistically sized training set. This suggests a limitation of the current incarnation of models for de novo drug design. Finally, we also include simpler tasks in the benchmark based on a simpler scoring function. We release the benchmark as an easy to use package available at https://github.com/cieplinski-tobiasz/smina-docking-benchmark. We hope that our benchmark will serve as a stepping stone toward the goal of automatically generating promising drug candidates.

The Subtype Selectivity in Search of Potent Hypotensive Agents among 5,5-Dimethylhydantoin Derived α1-Adrenoceptors Antagonists.

In order to find new hypotensive drugs possessing higher activity and better selectivity, a new series of fifteen 5,5-dimethylhydantoin derivatives (1-15) was designed. Three-step syntheses, consisting of N-alkylations using standard procedures as well as microwaves, were carried out. Crystal structures were determined for compounds 7-9. All of the synthesized 5,5-dimethylhydantoins were tested for their affinity to α1-adrenergic receptors (α1-AR) using both in vitro and in silico methods. Most of them displayed higher affinity (Ki < 127.9 nM) to α1-adrenoceptor than urapidil in radioligand binding assay. Docking to two subtypes of adrenergic receptors, α1A and α1B, was conducted. Selected compounds were tested for their activity towards two α1-AR subtypes. All of them showed intrinsic antagonistic activity. Moreover, for two compounds (1 and 5), which possess o-methoxyphenylpiperazine fragments, strong activity (IC50 < 100 nM) was observed. Some representatives (3 and 5), which contain alkyl linker, proved selectivity towards α1A-AR, while two compounds with 2-hydroxypropyl linker (11 and 13) to α1B-AR. Finally, hypotensive activity was examined in rats. The most active compound (5) proved not only a lower effective dose than urapidil but also a stronger effect than prazosin.

Prediction of probability distributions of molecular properties: towards more efficient virtual screening and better understanding of compound representations.

Various in silico approaches to predict activity and properties of chemical compounds constitute nowadays the basis of computer-aided drug design. While there is a general focus on the predictions of values, mathematically more appropriate is the prognosis of probability distributions, which offers additional possibilities, such as the evaluation of uncertainty, higher moments, and quantiles. In this study, we applied the Hierarchical Correlation Reconstruction approach to assess several ADMET properties of chemical compounds. It uses multiple linear regression to independently assess multiple moments, which are then finally combined into predicted probability distribution. The method enables inexpensive selection of compounds with properties nearly certain to fall into the particular range during virtual screening and automatic rejection of predictions characterized by high rate of uncertainty; however, unlike to the currently used virtual screening methods, it focuses on the prediction of the property distribution, not its actual value. Moreover, the presented protocol enables detection of structural features, which should be carefully considered when optimizing compounds towards particular property, as well as it provides deeper understanding of the examined compound representations.

Multitargeting the Action of 5-HT6 Serotonin Receptor Ligands by Additional Modulation of Kinases in the Search for a New Therapy for Alzheimer's Disease: Can It Work from a Molecular Point of View?

In view of the unsatisfactory treatment of cognitive disorders, in particular Alzheimer's disease (AD), the aim of this review was to perform a computer-aided analysis of the state of the art that will help in the search for innovative polypharmacology-based therapeutic approaches to fight against AD. Apart from 20-year unrenewed cholinesterase- or NMDA-based AD therapy, the hope of effectively treating Alzheimer's disease has been placed on serotonin 5-HT6 receptor (5-HT6R), due to its proven, both for agonists and antagonists, beneficial procognitive effects in animal models; however, research into this treatment has so far not been successfully translated to human patients. Recent lines of evidence strongly emphasize the role of kinases, in particular microtubule affinity-regulating kinase 4 (MARK4), Rho-associated coiled-coil-containing protein kinase I/II (ROCKI/II) and cyclin-dependent kinase 5 (CDK5) in the etiology of AD, pointing to the therapeutic potential of their inhibitors not only against the symptoms, but also the causes of this disease. Thus, finding a drug that acts simultaneously on both 5-HT6R and one of those kinases will provide a potential breakthrough in AD treatment. The pharmacophore- and docking-based comprehensive literature analysis performed herein serves to answer the question of whether the design of these kind of dual agents is possible, and the conclusions turned out to be highly promising.

The relationship between stereochemical and both, pharmacological and ADME-Tox, properties of the potent hydantoin 5-HT7R antagonist MF-8.

This study concerns synthesis and evaluation of pharmacodynamic and pharmacokinetic profile for all four stereoisomers of MF-8 (5-(4-fluorophenyl)-3-(2-hydroxy-3-(4-(2-methoxyphenyl)piperazin-1-yl)propyl)-5-methylimidazolidine-2,4-dione), the previously described, highly potent 5-HT7R ligand with antidepressant activity on mice. The combination of DFT calculations of 1H NMR chemical shifts with docking and dynamic simulations, in comparison to experimental screening results, provided prediction of the configuration for one of two present stereogenic centers. The experimental data for stereoisomers (MF-8A-MF-8D) confirmed the significant impact of stereochemistry on both, 5-HT7R affinity and antagonistic action, with Ki and Kb values in the range of 3-366 nM and 0.024-99 µM, respectively. We also indicated the stereochemistry-dependent influence of the tested compounds on P-glycoprotein efflux, absorption in Caco-2 model, metabolic pathway as well as CYP3A4 and CYP2C9 activities.

Low Basicity as a Characteristic for Atypical Ligands of Serotonin Receptor 5-HT2.

Serotonin receptors are extensively examined by academic and industrial researchers, due to their vital roles, which they play in the organism and constituting therefore important drug targets. Up to very recently, it was assumed that the basic nitrogen in compound structure is a necessary component to make it active within this receptor system. Such nitrogen interacts in its protonated form with the aspartic acid from the third transmembrane helix (D3x32) forming a hydrogen bond tightly fitting the ligand in the protein binding site. However, there are several recent studies that report strong serotonin receptor affinity also for compounds without a basic moiety in their structures. In the study, we carried out a comprehensive in silico analysis of the low-basicity phenomenon of the selected serotonin receptor ligands. We focused on the crystallized representatives of the proteins of 5-HT1B, 5-HT2A, 5-HT2B, and 5-HT2C receptors, and examined the problem both from the ligand- and structure-based perspectives. The study was performed for the native proteins, and for D3x32A mutants. The investigation resulted in the determination of nonstandard structural requirements for activity towards serotonin receptors, which can be used in the design of new nonbasic ligands.

Novel N-Substituted 3-Aryl-4-(diethoxyphosphoryl)azetidin-2-ones as Antibiotic Enhancers and Antiviral Agents in Search for a Successful Treatment of Complex Infections.

A novel series of N-substituted cis- and trans-3-aryl-4-(diethoxyphosphoryl)azetidin-2-ones were synthesized by the Kinugasa reaction of N-methyl- or N-benzyl-(diethyoxyphosphoryl)nitrone and selected aryl alkynes. Stereochemistry of diastereoisomeric adducts was established based on vicinal H3-H4 coupling constants in azetidin-2-one ring. All the obtained azetidin-2-ones were evaluated for the antiviral activity against a broad range of DNA and RNA viruses. Azetidin-2-one trans-11f showed moderate inhibitory activity against human coronavirus (229E) with EC50 = 45 µM. The other isomer cis-11f was active against influenza A virus H1N1 subtype (EC50 = 12 µM by visual CPE score; EC50 = 8.3 µM by TMS score; MCC > 100 µM, CC50 = 39.9 µM). Several azetidin-2-ones 10 and 11 were tested for their cytostatic activity toward nine cancerous cell lines and several of them appeared slightly active for Capan-1, Hap1 and HCT-116 cells values of IC50 in the range 14.5-97.9 µM. Compound trans-11f was identified as adjuvant of oxacillin with significant ability to enhance the efficacy of this antibiotic toward the highly resistant S. aureus strain HEMSA 5. Docking and molecular dynamics simulations showed that enantiomer (3R,4S)-11f can be responsible for the promising activity due to the potency in displacing oxacillin at ß-lactamase, thus protecting the antibiotic from undesirable biotransformation.

Molecular Insights into an Antibiotic Enhancer Action of New Morpholine-Containing 5-Arylideneimidazolones in the Fight against MDR Bacteria.

In the search for an effective strategy to overcome antimicrobial resistance, a series of new morpholine-containing 5-arylideneimidazolones differing within either the amine moiety or at position five of imidazolones was explored as potential antibiotic adjuvants against Gram-positive and Gram-negative bacteria. Compounds (7-23) were tested for oxacillin adjuvant properties in the Methicillin-susceptible S. aureus (MSSA) strain ATCC 25923 and Methicillin-resistant S. aureus MRSA 19449. Compounds 14-16 were tested additionally in combination with various antibiotics. Molecular modelling was performed to assess potential mechanism of action. Microdilution and real-time efflux (RTE) assays were carried out in strains of K. aerogenes to determine the potential of compounds 7-23 to block the multidrug efflux pump AcrAB-TolC. Drug-like properties were determined experimentally. Two compounds (10, 15) containing non-condensed aromatic rings, significantly reduced oxacillin MICs in MRSA 19449, while 15 additionally enhanced the effectiveness of ampicillin. Results of molecular modelling confirmed the interaction with the allosteric site of PBP2a as a probable MDR-reversing mechanism. In RTE, the compounds inhibited AcrAB-TolC even to 90% (19). The 4-phenylbenzylidene derivative (15) demonstrated significant MDR-reversal "dual action" for ß-lactam antibiotics in MRSA and inhibited AcrAB-TolC in K. aerogenes. 15 displayed also satisfied solubility and safety towards CYP3A4 in vitro.

Mutual Support of Ligand- and Structure-Based Approaches-To What Extent We Can Optimize the Power of Predictive Model? Case Study of Opioid Receptors.

The process of modern drug design would not exist in the current form without computational methods. They are part of every stage of the drug design pipeline, supporting the search and optimization of new bioactive substances. Nevertheless, despite the great help that is offered by in silico strategies, the power of computational methods strongly depends on the input data supplied at the stage of the predictive model construction. The studies on the efficiency of the computational protocols most often focus on global efficiency. They use general parameters that refer to the whole dataset, such as accuracy, precision, mean squared error, etc. In the study, we examined machine learning predictions obtained for opioid receptors (mu, kappa, delta) and focused on cases for which the predictions were the most accurate and the least accurate. Moreover, by using docking, we tried to explain prediction errors. We attempted to develop a rule of thumb, which can help in the prediction of compound activity towards opioid receptors via docking, especially those that have been incorrectly predicted by machine learning. We found out that although the combination of ligand- and structure-based path can be beneficial for the prediction accuracy, there still remain cases that cannot be reliably predicted by any available modeling method. In addition to challenging ligand- and structure-based predictions, we also examined the role of the application of machine-learning methods in comparison to simple statistical methods for both standard ligand-based representations (molecular fingerprints) and interaction fingerprints. All approaches were confronted in both classification (where compounds were assigned to the group of active and inactive group constructed on the basis of Ki values) and regression (where exact Ki value was predicted) experiments.

The Structural Determinants for α1-Adrenergic/Serotonin Receptors Activity among Phenylpiperazine-Hydantoin Derivatives.

Several studies confirmed the reciprocal interactions between adrenergic and serotoninergic systems and the influence of these phenomena on the pathogenesis of anxiety. Hence, searching for chemical agents with a multifunctional pharmacodynamic profile may bring highly effective therapy for CNS disorders. This study presents a deep structural insight into the hydantoin-arylpiperazine group and their serotonin/α-adrenergic activity. The newly synthesized compounds were tested in the radioligand binding assay and the intrinsic activity was evaluated for the selected derivatives. The computer-aided SAR analysis enabled us to answer questions about the influence of particular structural fragments on selective vs. multifunctional activity. As a result of the performed investigations, there were two leading structures: (a) compound 12 with multifunctional adrenergic-serotonin activity, which is a promising candidate to be an effective anxiolytic agent; (b) compound 14 with high α1A/α1D affinity and selectivity towards α1B, which is recommended due to the elimination of probable cardiotoxic effect. The structural conclusions of this work provide significant support for future lead optimization in order to achieve the desired pharmacodynamic profile in searching for new CNS-modulating agents.

In silico and in vitro studies on interaction of novel non-imidazole histamine H3R antagonists with CYP3A4.

The paper presents in silico study to explain differences in the influence of the series of non-imidazole histamine receptor H3 ligands on the activity of cytochrome P-450 3A4 isoform, which was verified in in vitro tests. The compounds appeared to induce broad range of effects - from significant inhibition (-61% reduction of CYP3A4 control activity) to extreme activation (+713% of control activity). Structure-activity relationship for examined compounds was analyzed, with special attention paid to the influence of substituent and the chain length. Docking, molecular dynamics studies, and their statistical analysis allowed to identify those interactions that can be responsible for determination of particular activity type of a compound toward CYP3A4 (activation/inhibition). It resulted in indication of several amino acid residues, which should be carefully analyzed during estimation of compound effects on CYP3A4 activity.

Emulating Docking Results Using a Deep Neural Network: A New Perspective for Virtual Screening.

Docking is one of the most important steps in virtual screening pipelines, and it is an established method for examining potential interactions between ligands and receptors. However, this method is computationally expensive, and it is often among the last steps of the process of compound libraries evaluation. In this work, we investigate the feasibility of learning a deep neural network to predict the docking output directly from a two-dimensional compound structure. The developed protocol is orders of magnitude faster than typical docking software, and it returns ligand-receptor complexes encoded in the form of the interaction fingerprint. Its speed and efficiency unlock the application possibilities, such as screening compound libraries of vast size on the basis of contact patterns or docking score (derived on the basis of predicted interaction schemes). We tested our approach on several G protein-coupled receptor targets and 4 CYP enzymes in retrospective virtual screening experiments, and a variant of graph convolutional network appeared to be most effective in emulating docking results. The method can be easily used by the community based on the code available in the Supporting Information.

How Sure Can We Be about ML Methods-Based Evaluation of Compound Activity: Incorporation of Information about Prediction Uncertainty Using Deep Learning Techniques.

A great variety of computational approaches support drug design processes, helping in selection of new potentially active compounds, and optimization of their physicochemical and ADMET properties. Machine learning is a group of methods that are able to evaluate in relatively short time enormous amounts of data. However, the quality of machine-learning-based prediction depends on the data supplied for model training. In this study, we used deep neural networks for the task of compound activity prediction and developed dropout-based approaches for estimating prediction uncertainty. Several types of analyses were performed: the relationships between the prediction error, similarity to the training set, prediction uncertainty, number and standard deviation of activity values were examined. It was tested whether incorporation of information about prediction uncertainty influences compounds ranking based on predicted activity and prediction uncertainty was used to search for the potential errors in the ChEMBL database. The obtained outcome indicates that incorporation of information about uncertainty of compound activity prediction can be of great help during virtual screening experiments.

Molecular Modeling of µ Opioid Receptor Ligands with Various Functional Properties: PZM21, SR-17018, Morphine, and Fentanyl-Simulated Interaction Patterns Confronted with Experimental Data.

Molecular modeling approaches are an indispensable part of the drug design process. They not only support the process of searching for new ligands of a given receptor, but they also play an important role in explaining particular activity pathways of a compound. In this study, a comprehensive molecular modeling protocol was developed to explain the observed activity profiles of selected µ opioid receptor agents: two G protein-biased µ opioid receptor agonists(PZM21 and SR-17018), unbiased morphine, and the ß-arrestin-2-biased agonist,fentanyl. The study involved docking and molecular dynamics simulations carried out for three crystal structures of the target at a microsecond scale, followed by the statistical analysis of ligand-protein contacts. The interaction frequency between the modeled compounds and the subsequent residues of a protein during the simulation was also correlated with the output of in vitro and in vivo tests, resulting in the set of amino acids with the highest Pearson correlation coefficient values. Such indicated positions may serve as a guide for designing new G protein-biased ligands of the µ opioid receptor.

Development of New Methods Needs Proper Evaluation-Benchmarking Sets for Machine Learning Experiments for Class A GPCRs.

New computational approaches for virtual screening applications are constantly being developed. However, before a particular tool is used to search for new active compounds, its effectiveness in the type of task must be examined. In this study, we conducted a detailed analysis of various aspects of preparation of respective data sets for such an evaluation. We propose a protocol for fetching data from the ChEMBL database, examine various compound representations in terms of the possible bias resulting from the way they are generated, and define a new metric for comparing the structural similarity of compounds, which is in line with chemical intuition. The newly developed method is also used for the evaluation of various approaches for division of the data set into training and test set parts, which are also examined in detail in terms of being the source of possible results bias. Finally, machine learning methods are applied in cross-validation studies of data sets constructed within the paper, constituting benchmarks for the assessment of computational methods developed for virtual screening tasks. Additionally, analogous data sets for class A G protein-coupled receptors (100 targets with the highest number of records) were prepared. They are available at http://gmum.net/benchmarks/ , together with script enabling reproduction of all results available at https://github.com/lesniak43/ananas .

Three-dimensional descriptors for aminergic GPCRs: dependence on docking conformation and crystal structure.

Three-dimensional descriptors are often used to search for new biologically active compounds, in both ligand- and structure-based approaches, capturing the spatial orientation of molecules. They frequently constitute an input for machine learning-based predictions of compound activity or quantitative structure-activity relationship modeling; however, the distribution of their values and the accuracy of depicting compound orientations might have an impact on the power of the obtained predictive models. In this study, we analyzed the distribution of three-dimensional descriptors calculated for docking poses of active and inactive compounds for all aminergic G protein-coupled receptors with available crystal structures, focusing on the variation in conformations for different receptors and crystals. We demonstrated that the consistency in compound orientation in the binding site is rather not correlated with the affinity itself, but is more influenced by other factors, such as the number of rotatable bonds and crystal structure used for docking studies. The visualizations of the descriptors distributions were prepared and made available online at http://chem.gmum.net/vischem_stability , which enables the investigation of chemical structures referring to particular data points depicted in the figures. Moreover, the performed analysis can assist in choosing crystal structure for docking studies, helping in selection of conditions providing the best discrimination between active and inactive compounds in machine learning-based experiments.

5-Arylideneimidazolones with Amine at Position 3 as Potential Antibiotic Adjuvants against Multidrug Resistant Bacteria.

Searching for new chemosensitizers of bacterial multidrug resistance (MDR), chemical modifications of (Z)-5-(4-chlorobenzylidene)-2-(4-methylpiperazin-1-yl)-3H-imidazol-4(5H)-one (6) were performed. New compounds (7⁻17), with fused aromatic rings at position 5, were designed and synthesized. Crystallographic X-ray analysis proved that the final compounds (7⁻17) were substituted with tertiary amine-propyl moiety at position 3 and primary amine group at 2 due to intramolecular Dimroth rearrangement. New compounds were evaluated on their antibiotic adjuvant properties in either Gram-positive or Gram-negative bacteria. Efflux pump inhibitor (EPI) properties towards the AcrAB-TolC pump in Enterobacter aerogenes (EA289) were investigated in the real-time efflux (RTE) assay. Docking and molecular dynamics were applied to estimate an interaction of compounds 6⁻17 with penicillin binding protein (PBP2a). In vitro ADME-Tox properties were evaluated for compound 9. Most of the tested compounds reduced significantly (4-32-fold) oxacillin MIC in highly resistant MRSA HEMSA 5 strain. The anthracene-morpholine derivative (16) was the most potent (32-fold reduction). The tested compounds displayed significant EPI properties during RTE assay (37⁻97%). The naphthyl-methylpiperazine derivative 9 showed the most potent "dual action" of both oxacillin adjuvant (MRSA) and EPI (E. aerogenes). Molecular modeling results suggested the allosteric mechanism of action of the imidazolones, which improved binding of oxacillin in the PBP2a active site in MRSA.

MetStabOn-Online Platform for Metabolic Stability Predictions.

Metabolic stability is an important parameter to be optimized during the complex process of designing new active compounds. Tuning this parameter with the simultaneous maintenance of a desired compound's activity is not an easy task due to the extreme complexity of metabolic pathways in living organisms. In this study, the platform for in silico qualitative evaluation of metabolic stability, expressed as half-lifetime and clearance was developed. The platform is based on the application of machine learning methods and separate models for human, rat and mouse data were constructed. The compounds' evaluation is qualitative and two types of experiments can be performed-regression, which is when the compound is assigned to one of the metabolic stability classes (low, medium, high) on the basis of numerical value of the predicted half-lifetime, and classification, in which the molecule is directly assessed as low, medium or high stability. The results show that the models have good predictive power, with accuracy values over 0.7 for all cases, for Sequential Minimal Optimization (SMO), k-nearest neighbor (IBk) and Random Forest algorithms. Additionally, for each of the analyzed compounds, 10 of the most similar structures from the training set (in terms of Tanimoto metric similarity) are identified and made available for download as separate files for more detailed manual inspection. The predictive power of the models was confronted with the external dataset, containing metabolic stability assessment via the GUSAR software, leading to good consistency of results for SMOreg and Naïve Bayes (~0.8 on average). The tool is available online.

Substructural Connectivity Fingerprint and Extreme Entropy Machines-A New Method of Compound Representation and Analysis.

Key-based substructural fingerprints are an important element of computer-aided drug design techniques. The usefulness of the fingerprints in filtering compound databases is invaluable, as they allow for the quick rejection of molecules with a low probability of being active. However, this method is flawed, as it does not consider the connections between substructures. After changing the connections between particular chemical moieties, the fingerprint representation of the compound remains the same, which leads to difficulties in distinguishing between active and inactive compounds. In this study, we present a new method of compound representation-substructural connectivity fingerprints (SCFP), providing information not only about the presence of particular substructures in the molecule but also additional data on substructure connections. Such representation was analyzed by the recently developed methodology-extreme entropy machines (EEM). The SCFP can be a valuable addition to virtual screening tools, as it represents compound structure with greater detail and more specificity, allowing for more accurate classification.