Prediction of Oral Bioavailability in Rats: Transferring Insights from in Vitro Correlations to (Deep) Machine Learning Models Using in Silico Model Outputs and Chemical Structure Parameters.

Oral administration of drug products is a strict requirement in many medical indications. Therefore, bioavailability prediction models are of high importance for prioritization of compound candidates in the drug discovery process. However, oral exposure and bioavailability are difficult to predict, as they are the result of various highly complex factors and/or processes influenced by the physicochemical properties of a compound, such as solubility, lipophilicity, or charge state, as well as by interactions with the organism, for instance, metabolism or membrane permeation. In this study, we assess whether it is possible to predict intravenous (iv) or oral drug exposure and oral bioavailability in rats. As input parameters, we use (i) six experimentally determined in vitro and physicochemical endpoints, namely, membrane permeation, free fraction, metabolic stability, solubility, pKa value, and lipophilicity; (ii) the outputs of six in silico absorption, distribution, metabolism, and excretion models trained on the same endpoints, or (iii) the chemical structure encoded as fingerprints or simplified molecular input line entry system strings. The underlying data set for the models is an unprecedented collection of almost 1900 data points with high-quality in vivo experiments performed in rats. We find that drug exposure after iv administration can be predicted similarly well using hybrid models with in vitro- or in silico-predicted endpoints as inputs, with fold change errors (FCE) of 2.28 and 2.08, respectively. The FCEs for exposure after oral administration are higher, and here, the prediction from in vitro inputs performs significantly better in comparison to in silico-based models with FCEs of 3.49 and 2.40, respectively, most probably reflecting the higher complexity of oral bioavailability. Simplifying the prediction task to a binary alert for low oral bioavailability, based only on chemical structure, we achieve accuracy and precision close to 70%.

Reliable and Performant Identification of Low-Energy Conformers in the Gas Phase and Water.

Prediction of compound properties from structure via quantitative structure-activity relationship and machine-learning approaches is an important computational chemistry task in small-molecule drug research. Though many such properties are dependent on three-dimensional structures or even conformer ensembles, the majority of models are based on descriptors derived from two-dimensional structures. Here we present results from a thorough benchmark study of force field, semiempirical, and density functional methods for the calculation of conformer energies in the gas phase and water solvation as a foundation for the correct identification of relevant low-energy conformers. We find that the tight-binding ansatz GFN-xTB shows the lowest error metrics and highest correlation to the benchmark PBE0-D3(BJ)/def2-TZVP in the gas phase for the computationally fast methods and that in solvent OPLS3 becomes comparable in performance. MMFF94, AM1, and DFTB+ perform worse, whereas the performance-optimized but far more expensive functional PBEh-3c yields energies almost perfectly correlated to the benchmark and should be used whenever affordable. On the basis of our findings, we have implemented a reliable and fast protocol for the identification of low-energy conformers of drug-like molecules in water that can be used for the quantification of strain energy and entropy contributions to target binding as well as for the derivation of conformer-ensemble-dependent molecular descriptors.

Finerenone Impedes Aldosterone-dependent Nuclear Import of the Mineralocorticoid Receptor and Prevents Genomic Recruitment of Steroid Receptor Coactivator-1.

Aldosterone regulates sodium homeostasis by activating the mineralocorticoid receptor (MR), a member of the nuclear receptor superfamily. Hyperaldosteronism leads todeleterious effects on the kidney, blood vessels, and heart. Although steroidal antagonists such as spironolactone and eplerenone are clinically useful for the treatment of cardiovascular diseases, they are associated with several side effects. Finerenone, a novel nonsteroidal MR antagonist, is presently being evaluated in two clinical phase IIb trials. Here, we characterized the molecular mechanisms of action of finerenone and spironolactone at several key steps of the MR signaling pathway. Molecular modeling and mutagenesis approaches allowed identification of Ser-810 and Ala-773 as key residues for the high MR selectivity of finerenone. Moreover, we showed that, in contrast to spironolactone, which activates the S810L mutant MR responsible for a severe form of early onset hypertension, finerenone displays strict antagonistic properties. Aldosterone-dependent phosphorylation and degradation of MR are inhibited by both finerenone and spironolactone. However, automated quantification of MR subcellular distribution demonstrated that finerenone delays aldosterone-induced nuclear accumulation of MR more efficiently than spironolactone. Finally, chromatin immunoprecipitation assays revealed that, as opposed to spironolactone, finerenone inhibits MR, steroid receptor coactivator-1, and RNA polymerase II binding at the regulatory sequence of the SCNN1A gene and also remarkably reduces basal MR and steroid receptor coactivator-1 recruitment, unraveling a specific and unrecognized inactivating mechanism on MR signaling. Overall, our data demonstrate that the highly potent and selective MR antagonist finerenone specifically impairs several critical steps of the MR signaling pathway and therefore represents a promising new generation MR antagonist.

Best of both worlds: combining pharma data and state of the art modeling technology to improve in Silico pKa prediction.

In a unique collaboration between a software company and a pharmaceutical company, we were able to develop a new in silico pKa prediction tool with outstanding prediction quality. An existing pKa prediction method from Simulations Plus based on artificial neural network ensembles (ANNE), microstates analysis, and literature data was retrained with a large homogeneous data set of drug-like molecules from Bayer. The new model was thus built with curated sets of ∼14,000 literature pKa values (∼11,000 compounds, representing literature chemical space) and ∼19,500 pKa values experimentally determined at Bayer Pharma (∼16,000 compounds, representing industry chemical space). Model validation was performed with several test sets consisting of a total of ∼31,000 new pKa values measured at Bayer. For the largest and most difficult test set with >16,000 pKa values that were not used for training, the original model achieved a mean absolute error (MAE) of 0.72, root-mean-square error (RMSE) of 0.94, and squared correlation coefficient (R(2)) of 0.87. The new model achieves significantly improved prediction statistics, with MAE = 0.50, RMSE = 0.67, and R(2) = 0.93. It is commercially available as part of the Simulations Plus ADMET Predictor release 7.0. Good predictions are only of value when delivered effectively to those who can use them. The new pKa prediction model has been integrated into Pipeline Pilot and the PharmacophorInformatics (PIx) platform used by scientists at Bayer Pharma. Different output formats allow customized application by medicinal chemists, physical chemists, and computational chemists.

Structures of synaptic vesicle protein 2A and 2B bound to anticonvulsants.

Epilepsy is a common neurological disorder characterized by abnormal activity of neuronal networks, leading to seizures. The racetam class of anti-seizure medications bind specifically to a membrane protein found in the synaptic vesicles of neurons called synaptic vesicle protein 2 (SV2) A (SV2A). SV2A belongs to an orphan subfamily of the solute carrier 22 organic ion transporter family that also includes SV2B and SV2C. The molecular basis for how anti-seizure medications act on SV2s remains unknown. Here we report cryo-electron microscopy structures of SV2A and SV2B captured in a luminal-occluded conformation complexed with anticonvulsant ligands. The conformation bound by anticonvulsants resembles an inhibited transporter with closed luminal and intracellular gates. Anticonvulsants bind to a highly conserved central site in SV2s. These structures provide blueprints for future drug design and will facilitate future investigations into the biological function of SV2s.

Cyclobutane-bearing restricted anchoring residues enabled geometry-specific hydrocarbon peptide stapling.

Stapled peptides are regarded as the promising next-generation therapeutics because of their improved secondary structure, membrane permeability and metabolic stability as compared with the prototype linear peptides. Usually, stapled peptides are obtained by a hydrocarbon stapling technique, anchoring from paired olefin-terminated unnatural amino acids and the consequent ring-closing metathesis (RCM). To investigate the adaptability of the rigid cyclobutane structure in RCM and expand the chemical diversity of hydrocarbon peptide stapling, we herein described the rational design and efficient synthesis of cyclobutane-based conformationally constrained amino acids, termed (E)-1-amino-3-(but-3-en-1-yl)cyclobutane-1-carboxylic acid (E7) and (Z)-1-amino-3-(but-3-en-1-yl)cyclobutane-1-carboxylic acid (Z7). All four combinations including E7-E7, E7-Z7, Z7-Z7 and Z7-E7 were proven to be applicable in RCM-mediated peptide stapling to afford the corresponding geometry-specific stapled peptides. With the aid of the combined quantum and molecular mechanics, the E7-E7 combination was proven to be optimal in both the RCM reaction and helical stabilization. With the spike protein of SARS-CoV-2 as the target, a series of cyclobutane-bearing stapled peptides were obtained. Among them, E7-E7 geometry-specific stapled peptides indeed exhibit higher α-helicity and thus stronger biological activity than canonical hydrocarbon stapled peptides. We believe that this methodology possesses great potential to expand the scope of the existing peptide stapling strategy. These cyclobutane-bearing restricted anchoring residues served as effective supplements for the existing olefin-terminated unnatural amino acids and the resultant geometry-specific hydrocarbon peptide stapling provided more potential for peptide therapeutics.

Design and Preclinical Characterization Program toward Asundexian (BAY 2433334), an Oral Factor XIa Inhibitor for the Prevention and Treatment of Thromboembolic Disorders.

Activated coagulation factor XI (FXIa) is a highly attractive antithrombotic target as it contributes to the development and progression of thrombosis but is thought to play only a minor role in hemostasis so that its inhibition may allow for decoupling of antithrombotic efficacy and bleeding time prolongation. Herein, we report our major efforts to identify an orally bioavailable, reversible FXIa inhibitor. Using a protein structure-based de novo design approach, we identified a novel micromolar hit with attractive physicochemical properties. During lead modification, a critical problem was balancing potency and absorption by focusing on the most important interactions of the lead series with FXIa while simultaneously seeking to improve metabolic stability and the cytochrome P450 interaction profile. In clinical trials, the resulting compound from our extensive research program, asundexian (BAY 2433334), proved to possess the desired DMPK properties for once-daily oral dosing, and even more importantly, the initial pharmacological hypothesis was confirmed.

Target 2035 - an update on private sector contributions.

Target 2035, an international federation of biomedical scientists from the public and private sectors, is leveraging 'open' principles to develop a pharmacological tool for every human protein. These tools are important reagents for scientists studying human health and disease and will facilitate the development of new medicines. It is therefore not surprising that pharmaceutical companies are joining Target 2035, contributing both knowledge and reagents to study novel proteins. Here, we present a brief progress update on Target 2035 and highlight some of industry's contributions.

Machine Learning Applied to the Modeling of Pharmacological and ADMET Endpoints.

The well-known concept of quantitative structure-activity relationships (QSAR) has been gaining significant interest in the recent years. Data, descriptors, and algorithms are the main pillars to build useful models that support more efficient drug discovery processes with in silico methods. Significant advances in all three areas are the reason for the regained interest in these models. In this book chapter we review various machine learning (ML) approaches that make use of measured in vitro/in vivo data of many compounds. We put these in context with other digital drug discovery methods and present some application examples.

CACHE (Critical Assessment of Computational Hit-finding Experiments): A public-private partnership benchmarking initiative to enable the development of computational methods for hit-finding.

One aspirational goal of computational chemistry is to predict potent and drug-like binders for any protein, such that only those that bind are synthesized. In this Roadmap, we describe the launch of Critical Assessment of Computational Hit-finding Experiments (CACHE), a public benchmarking project to compare and improve small molecule hit-finding algorithms through cycles of prediction and experimental testing. Participants will predict small molecule binders for new and biologically relevant protein targets representing different prediction scenarios. Predicted compounds will be tested rigorously in an experimental hub, and all predicted binders as well as all experimental screening data, including the chemical structures of experimentally tested compounds, will be made publicly available, and not subject to any intellectual property restrictions. The ability of a range of computational approaches to find novel binders will be evaluated, compared, and openly published. CACHE will launch 3 new benchmarking exercises every year. The outcomes will be better prediction methods, new small molecule binders for target proteins of importance for fundamental biology or drug discovery, and a major technological step towards achieving the goal of Target 2035, a global initiative to identify pharmacological probes for all human proteins.

Target 2035 - update on the quest for a probe for every protein.

Twenty years after the publication of the first draft of the human genome, our knowledge of the human proteome is still fragmented. The challenge of translating the wealth of new knowledge from genomics into new medicines is that proteins, and not genes, are the primary executers of biological function. Therefore, much of how biology works in health and disease must be understood through the lens of protein function. Accordingly, a subset of human proteins has been at the heart of research interests of scientists over the centuries, and we have accumulated varying degrees of knowledge about approximately 65% of the human proteome. Nevertheless, a large proportion of proteins in the human proteome (∼35%) remains uncharacterized, and less than 5% of the human proteome has been successfully targeted for drug discovery. This highlights the profound disconnect between our abilities to obtain genetic information and subsequent development of effective medicines. Target 2035 is an international federation of biomedical scientists from the public and private sectors, which aims to address this gap by developing and applying new technologies to create by year 2035 chemogenomic libraries, chemical probes, and/or biological probes for the entire human proteome.

A new mode of mineralocorticoid receptor antagonism by a potent and selective nonsteroidal molecule.

Limitations of current steroidal mineralocorticoid receptor (MR) antagonists have stimulated the search for a new generation of molecules. We screened for novel nonsteroidal compounds and identified MR antagonists derived from the chemical class of dihydropyridines. Chemical optimization resulted in BR-4628, which displays high in vitro and in vivo MR potency as well as selectivity with respect to the other steroid hormone receptors and the L-type calcium channel. Biochemical studies demonstrated that BR-4628 forms complexes with MR that do not promote the recruitment of transcriptional co-regulators. Docking experiments, using the crystal structure of the MR ligand-binding domain in an agonist conformation, revealed that BR-4628 accommodates in the MR ligand-binding cavity differently in comparison with the classical steroidal MR antagonists. An alanine scanning mutagenesis approach, based on BR-4628 docking, allowed identifying its anchoring mode within the ligand-binding cavity. Altogether, we propose that BR-4628 is a bulky antagonist that inactivates MR through a passive mechanism. It represents the prototype of a new class of MR antagonists.

Oral, direct thrombin and factor†Xa inhibitors: the replacement for warfarin, leeches, and pig intestines?

To prevent thromboses after surgery, patients have until now had to inject themselves daily with heparin. For stroke prophylaxis in atrial fibrillation, patients take vitamin K antagonists of the coumarin type, which have a narrow therapeutic window and whose dosage must be regularly monitored. In order to improve the standard of therapy in thromboembolic diseases such as deep-vein thrombosis, pulmonary embolism, and stroke in atrial fibrillation, intensive research has been carried out over the last decade in the search for new, orally active thrombin and factor Xa inhibitors. A number of these compounds are already on the market or are in advanced clinical development; they could revolutionize the anticoagulant market.

Druggability Assessment for Selected Serine Proteases in a Pharmaceutical Industry Setting.

Target druggability assessment is an integral part of the early target characterization and selection process in pharmaceutical industry. Here, we investigate a set of five different serine proteases from the blood coagulation cascade. The aim of this study is twofold. Firstly, leveraging the wealth of available in-house high-throughput screening (HTS) data, we analyze HTS hit rates and discuss their predictive value for the development of small molecule (SMOL) candidates. Purely structure-activity relationship (SAR) based druggability ratings are compared with computational protein-structure based druggability assessments. Secondly, we evaluate the impact of using conformational ensembles from molecular dynamics (MD) simulations instead of single static crystal structures as basis for computational druggability assessments. Based on this study, we recommend incorporating molecular dynamics routinely into the early target characterization process, especially if only a single X-ray structure is available.

Bayer's in silico ADMET platform: a journey of machine learning over the past two decades.

Over the past two decades, an in silico absorption, distribution, metabolism, and excretion (ADMET) platform has been created at Bayer Pharma with the goal to generate models for a variety of pharmacokinetic and physicochemical endpoints in early drug discovery. These tools are accessible to all scientists within the company and can be a useful in assisting with the selection and design of novel leads, as well as the process of lead optimization. Here. we discuss the development of machine-learning (ML) approaches with special emphasis on data, descriptors, and algorithms. We show that high company internal data quality and tailored descriptors, as well as a thorough understanding of the experimental endpoints, are essential to the utility of our models. We discuss the recent impact of deep neural networks and show selected application examples.

Design, Synthesis, and Pharmacological Characterization of a Neutral, Non-Prodrug Thrombin Inhibitor with Good Oral Pharmacokinetics.

Despite extensive research on small molecule thrombin inhibitors for oral application in the past decades, only a single double prodrug with very modest oral bioavailability has reached human therapy as a marketed drug. We have undertaken major efforts to identify neutral, non-prodrug inhibitors. Using a holistic analysis of all available internal data, we were able to build computational models and apply these for the selection of a lead series with the highest possibility of achieving oral bioavailability. In our design, we relied on protein structure knowledge to address potency and identified a small window of favorable physicochemical properties to balance absorption and metabolic stability. Protein structure information on the pregnane X receptor helped in overcoming a persistent cytochrome P450 3A4 induction problem. The selected compound series was optimized to a highly potent, neutral, non-prodrug thrombin inhibitor by designing, synthesizing, and testing derivatives. The resulting optimized compound, BAY1217224, has reached first clinical trials, which have confirmed the desired pharmacokinetic properties.

Structure-Permeability Relationship of Semipeptidic Macrocycles-Understanding and Optimizing Passive Permeability and Efflux Ratio.

We herein report the first thorough analysis of the structure-permeability relationship of semipeptidic macrocycles. In total, 47 macrocycles were synthesized using a hybrid solid-phase/solution strategy, and then their passive and cellular permeability was assessed using the parallel artificial membrane permeability assay (PAMPA) and Caco-2 assay, respectively. The results indicate that semipeptidic macrocycles generally possess high passive permeability based on the PAMPA, yet their cellular permeability is governed by efflux, as reported in the Caco-2 assay. Structural variations led to tractable structure-permeability and structure-efflux relationships, wherein the linker length, stereoinversion, N-methylation, and peptoids site-specifically impact the permeability and efflux. Extensive nuclear magnetic resonance, molecular dynamics, and ensemble-based three-dimensional polar surface area (3D-PSA) studies showed that ensemble-based 3D-PSA is a good predictor of passive permeability.

Structure-based design, synthesis and in vitro characterization of potent 17beta-hydroxysteroid dehydrogenase type 1 inhibitors based on 2-substitutions of estrone and D-homo-estrone.

In search for specific drugs against steroid-dependent cancers we have developed a novel set of potent inhibitors of steroidogenic human 17beta-hydroxysteroid dehydrogenase type 1 (17beta-HSD 1). The X-ray structure of 17beta-HSD 1 in complex with estradiol served as basis for the design of the inhibitors. 2-Substituted estrone and D-homo-estrone derivatives were synthesized and tested for 17beta-HSD 1 inhibition. The best 17beta-HSD 1 inhibitor, 2-phenethyl-D-homo-estrone, revealed an IC(50) of 15 nM in vitro. The inhibitory potency of compounds is comparable or better to that of previously described inhibitors. An interaction within the cofactor binding site is not necessary to obtain this high binding affinity for substances developed.

An approach towards enhancement of a screening library: The Next Generation Library Initiative (NGLI) at Bayer - against all odds?

Pharmaceutical companies often refer to 'screening their library' when performing high-throughput screening (HTS) on a corporate compound collection to identify lead structures for small-molecule drug discovery programs. Characteristics of such a library, including the size, chemical space covered, and physicochemical properties, often determine the success of a screening campaign. Therefore, strategies to maintain and enhance the overall quality of screening collections are crucial to stay competitive and to cope with the 'novelty erosion' that is observed gradually. The Next Generation Library Initiative (NGLI), the enhancement of Bayer's HTS collection by 500000 newly designed compounds within 5 years, is addressing exactly this challenge. Here, we describe this collaborative project, which involves all internal medicinal chemists in a crowd-sourcing approach, as well as selected external partners, to reach this ambitious goal.

Improving the hit-to-lead process: data-driven assessment of drug-like and lead-like screening hits.

Drug-like and lead-like hits derived from HTS campaigns provide good starting points for lead optimization. However, too strong emphasis on potency as hit-selection parameter might hamper the success of such projects. A detailed absorption, distribution, metabolism, excretion and toxicology (ADME-Tox) profiling is needed to help identify hits with a minimum number of (known) liabilities. This is particularly true for drug-like hits. Herein, we describe how to break down large numbers of screening hits and we provide a comprehensive overview of the strengths and weaknesses for each structural class. The overall profile (e.g. ligand efficiency, selectivity and ADME-Tox) is the distinctive feature that will define the priority for follow-up.