Blinded Predictions and Post Hoc Analysis of the Second Solubility Challenge Data: Exploring Training Data and Feature Set Selection for Machine and Deep Learning Models.

Accurate methods to predict solubility from molecular structure are highly sought after in the chemical sciences. To assess the state of the art, the American Chemical Society organized a "Second Solubility Challenge" in 2019, in which competitors were invited to submit blinded predictions of the solubilities of 132 drug-like molecules. In the first part of this article, we describe the development of two models that were submitted to the Blind Challenge in 2019 but which have not previously been reported. These models were based on computationally inexpensive molecular descriptors and traditional machine learning algorithms and were trained on a relatively small data set of 300 molecules. In the second part of the article, to test the hypothesis that predictions would improve with more advanced algorithms and higher volumes of training data, we compare these original predictions with those made after the deadline using deep learning models trained on larger solubility data sets consisting of 2999 and 5697 molecules. The results show that there are several algorithms that are able to obtain near state-of-the-art performance on the solubility challenge data sets, with the best model, a graph convolutional neural network, resulting in an RMSE of 0.86 log units. Critical analysis of the models reveals systematic differences between the performance of models using certain feature sets and training data sets. The results suggest that careful selection of high quality training data from relevant regions of chemical space is critical for prediction accuracy but that other methodological issues remain problematic for machine learning solubility models, such as the difficulty in modeling complex chemical spaces from sparse training data sets.

Membrane Permeability in Cyclic Peptides is Modulated by Core Conformations.

Cyclic peptides have the potential to bind to challenging targets, which are undruggable with small molecules, but their application is limited by low membrane permeability. Here, using a series of cyclic pentapeptides, we showed that established physicochemical criteria of permeable peptides are heavily violated. We revealed that a dominant core conformation, stabilized by amides' shielding pattern, could guide the design of novel compounds. As a result, counter-intuitive strategies, such as incorporation of polar residues, can be beneficial for permeability. We further find that core globularity is a promising descriptor, which can extend the capability of standard predictive models.

Findings of the Second Challenge to Predict Aqueous Solubility.

Ten years ago, we issued an open prediction challenge to the cheminformatics community: would participants be able to predict the equilibrium intrinsic solubilities of 32 druglike molecules using only a high-precision (CheqSol instrument, performed in one laboratory) set of 100 compounds as a training set? The "solubility challenge" was a widely recognized success and spurred many discussions about the prediction methods and quality of data. We revisited the competition a second time recently and challenged the community to a different challenge, not a blind test this time but using a larger test set of molecules, gathered and curated from published sources (mostly "gold standard" saturation shake-flask measurements), where the average interlaboratory reproducibility for the molecules was estimated to be ∼0.17 log unit. Also, a second test set was included, comprising "contentious" molecules, the reported (mostly shake-flask) solubility of which had higher average uncertainty, ∼0.62 log unit. In the second competition, the participants were invited to use their own training sets, provided that the training sets did not contain any of the test set molecules. We were motivated to revisit the competition to (1) examine to what extent computational methods had improved in 10 years, (2) verify that data quality may not be the main limiting factor in the accuracy of the prediction method, and (3) attempt to seek a relationship between the makeup of the training set data and the prediction outcome.

Multisolvent Models for Solvation Free Energy Predictions Using 3D-RISM Hydration Thermodynamic Descriptors.

The potential to predict solvation free energies (SFEs) in any solvent using a machine learning (ML) model based on thermodynamic output, extracted exclusively from 3D-RISM simulations in water is investigated. The models on multiple solvents take into account both the solute and solvent description and offer the possibility to predict SFEs of any solute in any solvent with root mean squared errors less than 1 kcal/mol. Validations that involve exclusion of fractions or clusters of the solutes or solvents exemplify the model's capability to predict SFEs of novel solutes and solvents with diverse chemical profiles. In addition to being predictive, our models can identify the solute and solvent features that influence SFE predictions. Furthermore, using 3D-RISM hydration thermodynamic output to predict SFEs in any organic solvent reduces the need to run 3D-RISM simulations in all these solvents. Altogether, our multisolvent models for SFE predictions that take advantage of the solvation effects are expected to have an impact in the property prediction space.

Solubility Challenge Revisited after Ten Years, with Multilab Shake-Flask Data, Using Tight (SD ∼ 0.17 log) and Loose (SD ∼ 0.62 log) Test Sets.

Ten years ago we issued, in conjunction with the Journal of Chemical Information and Modeling, an open prediction challenge to the cheminformatics community. Would they be able to predict the intrinsic solubilities of 32 druglike compounds using only a high-precision set of 100 compounds as a training set? The "Solubility Challenge" was a widely recognized success and spurred many discussions about the prediction methods and quality of data. Regardless of the obvious limitations of the challenge, the conclusions were somewhat unexpected. Despite contestants employing the entire spectrum of approaches available then to predict aqueous solubility and disposing of an extremely tight data set, it was not possible to identify the best methods at predicting aqueous solubility, a variety of methods and combinations all performed equally well (or badly). Several authors have suggested since then that it is not the poor quality of the solubility data which limits the accuracy of the predictions, but the deficient methods used. Now, ten years after the original Solubility Challenge, we revisit it and challenge the community to a new test with a much larger database with estimates of interlaboratory reproducibility.

Fast and General Method To Predict the Physicochemical Properties of Druglike Molecules Using the Integral Equation Theory of Molecular Liquids.

We report a method to predict physicochemical properties of druglike molecules using a classical statistical mechanics based solvent model combined with machine learning. The RISM-MOL-INF method introduced here provides an accurate technique to characterize solvation and desolvation processes based on solute-solvent correlation functions computed by the 1D reference interaction site model of the integral equation theory of molecular liquids. These functions can be obtained in a matter of minutes for most small organic and druglike molecules using existing software (RISM-MOL) (Sergiievskyi, V. P.; Hackbusch, W.; Fedorov, M. V. J. Comput. Chem. 2011, 32, 1982-1992). Predictions of caco-2 cell permeability and hydration free energy obtained using the RISM-MOL-INF method are shown to be more accurate than the state-of-the-art tools for benchmark data sets. Due to the importance of solvation and desolvation effects in biological systems, it is anticipated that the RISM-MOL-INF approach will find many applications in biophysical and biomedical property prediction.

Case study: cremophor EL-based liquid formulations as simple substitutes for amorphous solid dispersions in early preclinical in vivo studies.

OBJECTIVES: The objective of the present case study was to increase the exposure of the poorly soluble crystalline compound A. METHODS: Mice received 10 mg/kg of crystalline compound A formulated in eight different cosolvent, oil, and cyclodextrin mixtures. KEY FINDINGS: In all cases, AUC0-24h and maximum blood/plasma concentration (Cmax) were in the range of 6-16 µM × h and <1.4 µm, respectively, with a bioavailability below 18%. When 6% cremophor (CrEL) was added to three selected vehicles, AUC0-24h and Cmax increased ~5-10 times. The obtained pharmacokinetic profile of the most improved formulation using CrEL was possible to superimpose on the one obtained after administration of a CrEL-free amorphous solid dispersion (ASD, HPMC-AS:drug, 80:20) suspension of compound A. CONCLUSIONS: It is crucial to find an optimal screen vehicle as early as possible for a poorly water-soluble lead series and then avoid time and resource-consuming vehicle testing of multiple compounds in vivo. An ASD approach is more suited for clinical development when more time and resources are allocated to the project. In this case study, some preclinical formulations were used to maximize exposure but also as preindicators for ASDs later in the development chain.

Discovery of N-[[1-[2-(tert-butylcarbamoylamino)ethyl]-4-(hydroxymethyl)-4-piperidyl]methyl]-3,5-dichloro-benzamide as a selective T-type calcium channel (Cav3.2) inhibitor.

The T-type calcium channel inhibitor Mibefradil was reported to protect the heart from atrial remodeling, a key process involved in the development of atrial fibrillation and arrhythmias. Mibefradil is not a selective T-type calcium channel inhibitor and also affects the function of different ion channels. Our aim was to develop a selective T-type calcium channel inhibitor to validate the importance of T-type-related pharmacology in atrial fibrillation. Structural optimisation of a previously disclosed hit series focussed on minimising exposure to the central nervous system and improving pharmacokinetic properties, while maintain adequate potency and selectivity. This resulted in the design of N-[[1-[2-(tert-butylcarbamoylamino)ethyl]-4-(hydroxymethyl)-4-piperidyl]methyl]-3,5-dichloro-benzamide, a novel, selective, peripherally restricted chemical probe to verify the role of T-type calcium channel inhibition on atrial fibrillation protection.

Selective and Bioavailable HDAC6 2-(Difluoromethyl)-1,3,4-oxadiazole Substrate Inhibitors and Modeling of Their Bioactivation Mechanism.

Histone deacetylase 6 (HDAC6) is a unique member of the HDAC family mainly targeting cytosolic nonhistone substrates, such as α-tubulin, cortactin, and heat shock protein 90 to regulate cell proliferation, metastasis, invasion, and mitosis in tumors. We describe the identification and characterization of a series of 2-(difluoromethyl)-1,3,4-oxadiazoles (DFMOs) as selective nonhydroxamic acid HDAC6 inhibitors. By comparing structure-activity relationships and performing quantum mechanical calculations of the HDAC6 catalytic mechanism, we show that potent oxadiazoles are electrophilic substrates of HDAC6 and propose a mechanism for the bioactivation. We also observe that the inherent electrophilicity of the oxadiazoles makes them prone to degradation in water solution and the generation of potentially toxic products cannot be ruled out, limiting the developability for chronic diseases. However, the oxadiazoles demonstrate high oral bioavailability and low in vivo clearance and are excellent tools for studying the role of HDAC6 in vitro and in vivo in rats and mice.

The Solid State Landscape of the Sildenafil Drug.

Sildenafil, the active ingredient of the drug developed by Pfizer for the treatment of erectile dysfunction was firstly synthesized in 1989 in the United Kingdom and since then it has become one of the most prescribed drugs for sexual performance in the western world with more than 2.7 million prescriptions in the US in 2021. Since its discovery, this drug compound has attracted the interest of formulators and crystallographers, with a high number of crystal forms of sildenafil being found and characterized, including polymorphs, hydrates, solvates, salts and cocrystals, converting it in one of the most promiscuous multicomponent crystal former drugs in the pharmaceutical sciences arena. In this minireview, the polymorph, pseudopolymorph and multicomponent solid forms landscape of sildenafil is presented through a comprehensive compilation of their 42 solid forms reported in literature.

Discovery of N-(1-adamantyl)-2-(4-alkylpiperazin-1-yl)acetamide derivatives as T-type calcium channel (Cav3.2) inhibitors.

Chemical evolution of a HTS-based fragment hit resulted in the identification of N-(1-adamantyl)-2-[4-(2-tetrahydropyran-4-ylethyl)piperazin-1-yl]acetamide, a novel, selective T-type calcium channel (Ca(v)3.2) inhibitor with in vivo antihypertensive effect in rats.

AZD0284, a Potent, Selective, and Orally Bioavailable Inverse Agonist of Retinoic Acid Receptor-Related Orphan Receptor C2.

Inverse agonists of the nuclear receptor RORC2 have been widely pursued as a potential treatment for a variety of autoimmune diseases. We have discovered a novel series of isoindoline-based inverse agonists of the nuclear receptor RORC2, derived from our recently disclosed RORC2 inverse agonist 2. Extensive structure-activity relationship (SAR) studies resulted in AZD0284 (20), which combined potent inhibition of IL-17A secretion from primary human TH17 cells with excellent metabolic stability and good PK in preclinical species. In two preclinical in vivo studies, compound 20 reduced thymocyte numbers in mice and showed dose-dependent reduction of IL-17A containing γδ-T cells and of IL-17A and IL-22 RNA in the imiquimod induced inflammation model. Based on these data and a favorable safety profile, 20 was progressed to phase 1 clinical studies.

Identification and Optimization of Pyrrolidine Derivatives as Highly Potent Ghrelin Receptor Full Agonists.

Muscle atrophy and cachexia are common comorbidities among patients suffering from cancer, chronic obstructive pulmonary disease, and several other chronic diseases. The peptide hormone ghrelin exerts pleiotropic effects including the stimulation of growth hormone secretion and subsequent increase of insulin-like growth factor-1 levels, an important mediator of muscle growth and repair. Ghrelin also acts on inflammation, appetite, and adipogenesis and therefore has been considered a promising therapeutic target for catabolic conditions. We previously reported on the synthesis and properties of an indane based series of ghrelin receptor full agonists which led to a sustained increase of insulin-like growth factor-1 in a dog pharmacodynamic study. Herein we report on the identification of a series of pyrrolidine or piperidine based full agonists and attempted optimization to give compounds with profiles suitable for progression as clinical candidates.

Discovery of Potent and Orally Bioavailable Inverse Agonists of the Retinoic Acid Receptor-Related Orphan Receptor C2.

The further optimization of a recently disclosed series of inverse agonists of the nuclear receptor RORC2 is described. Investigations into the left-hand side of compound 1, guided by X-ray crystal structures, led to the substitution of the 4-aryl-thiophenyl residue with the hexafluoro-2-phenyl-propan-2-ol moiety. This change resulted in to compound 28, which combined improved drug-like properties with good cell potency and a significantly lower dose, using an early dose to man prediction. Target engagement in vivo was demonstrated in the thymus of mice by a reduction in the number of double positive T cells after oral dosing.

Polymorph control: past, present and future.

The appearance and disappearance of polymorphs is no longer a mysterious and inexplicable process. Although methods for polymorph control are still imperfect, there is a large armoury of methods that can be used to tackle this important and challenging problem. We survey the methods and their successes over the last few years.

Identification and Pharmacological Profile of an Indane Based Series of Ghrelin Receptor Full Agonists.

Cachexia and muscle wasting are very common among patients suffering from cancer, chronic obstructive pulmonary disease, and other chronic diseases. Ghrelin stimulates growth hormone secretion via the ghrelin receptor, which subsequently leads to increase of IGF-1 plasma levels. The activation of the GH/IGF-1 axis leads to an increase of muscle mass and functional capacity. Ghrelin further acts on inflammation, appetite, and adipogenesis and for this reason was considered an important target to address catabolic conditions. We report the synthesis and properties of an indane based series of ghrelin receptor full agonists; they have been shown to generate a sustained increase of IGF-1 levels in dog and have been thoroughly investigated with respect to their functional activity.

Potent and Orally Bioavailable Inverse Agonists of RORγt Resulting from Structure-Based Design.

Retinoic acid receptor related orphan receptor γt (RORγt), has been identified as the master regulator of TH17-cell function and development, making it an attractive target for the treatment of autoimmune diseases by a small-molecule approach. Herein, we describe our investigations on a series of 4-aryl-thienyl acetamides, which were guided by insights from X-ray cocrystal structures. Efforts in targeting the cofactor-recruitment site from the 4-aryl group on the thiophene led to a series of potent binders with nanomolar activity in a primary human-TH17-cell assay. The observation of a DMSO molecule binding in a subpocket outside the LBD inspired the introduction of an acetamide into the benzylic position of these compounds. Hereby, a hydrogen-bond interaction of the introduced acetamide oxygen with the backbone amide of Glu379 was established. This greatly enhanced the cellular activity of previously weakly cell-active compounds. The best compounds combined potent inhibition of IL-17 release with favorable PK in rodents, with compound 32 representing a promising starting point for future investigations.

Diclofenac solubility: independent determination of the intrinsic solubility of three crystal forms.

The solubility in water of diclofenac ({2-[(2,6-dichlorophenyl)amino]phenyl}acetic acid), a potent nonsteroidal anti-inflammatory drug, has been investigated. The various solid forms have been characterized by thermogravimetric analysis, differential scanning calorimetry, and X-ray diffraction. The commercially available form of diclofenac is the anhydrous sodium salt. This was recrystallized from ethanol and precipitated as a hydrate containing four diclofenac anions, four sodium cations, and nineteen water molecules per unit cell. This crystal structure is similar to but different from an earlier report of the structure. Crystals of the acid form of diclofenac were anhydrous and corresponded to an earlier crystal structure. Separate solubility measurements on all three of these solid forms of diclofenac gave consistent results for the intrinsic solubility. The aqueous solubility values reported in the literature for diclofenac are spread over a large range, with a factor of 100 separating the largest and the smallest. Our value is at the smaller end of this range. It is the only one supported by three independent procedures and rigorous characterization of the solid forms. The experimental conditions were precisely controlled.

Discovery of AZD6642, an inhibitor of 5-lipoxygenase activating protein (FLAP) for the treatment of inflammatory diseases.

A drug discovery program in search of novel 5-lipoxygenase activating protein (FLAP) inhibitors focused on driving a reduction in lipophilicity with maintained or increased ligand lipophilic efficiency (LLE) compared to previously reported compounds led to the discovery of AZD6642 (15b). Introduction of a hydrophilic tetrahydrofuran (THF) ring at the stereogenic central carbon atom led to a significant shift in physicochemical property space. The structure-activity relationship exploration and optimization of DMPK properties leading to this compound are described in addition to pharmacokinetic analysis and an investigation of the pharmacokinetic (PK)-pharmacodynamic (PD) relationship based on ex vivo leukotriene B4 (LTB4) levels in dog. AZD6642 shows high specific potency and low lipophilicity, resulting in a selective and metabolically stable profile. On the basis of initial PK/PD relation measured, a low dose to human was predicted.