Structure-Property Relationships and Machine Learning Models for Addressing CYP3A4-Mediated Victim Drug-Drug Interaction Risk in Drug Discovery.

Among the FDA-approved small molecule drugs (2005-2016) that are primarily metabolized by cytochrome P450 (CYP), 64% are primarily metabolized by CYP3A4. As the proportion of an individual drug's fraction metabolized through CYP3A4 increases, the risk for the drug to be a victim of an interaction with CYP3A4 inhibitors or inducers increases. Therefore, it is important to assess the extent of involvement of individual CYP enzymes in the overall clearance for a scaffold early in discovery and mitigate the CYP3A4-mediated victim-drug-drug interaction (DDI) risk, if warranted by the desired clinical profile of the drug. To mitigate the CYP3A4-mediated victim DDI risk in discovery, we analyzed the physicochemical properties of the CYP3A4 substrates and found that molecular weight was the property that provided the best separation of the CYP3A4 substrates from other CYP substrates. In addition, neutral and basic compounds with MW ≥ 360 g/mol tend to be primarily metabolized by CYP3A4, whereas acidic compounds with MW < 360 g/mol are most likely to be primarily metabolized by other CYP enzymes. We then developed Support Vector Machine based on fingerprints (SVM-FP) and Deep-Learning (DL) models to predict if a molecule will be primarily metabolized by CYP3A4. Our models were trained on 2306 compounds, which is the largest training set among published models for this endpoint. Both models showed positive predictive values (PPV) > 80% in predicting a CYP3A4 substrate on a prospective testing set. Given the high PPV of the models, project teams can confidently deprioritize compounds predicted to be CYP3A4 substrates to avoid the potential liability of CYP3A4 victim DDI. Teams can then focus time and resources on synthesizing compounds that are predicted to have a lower dependency on CYP3A4 metabolism and confirm that experimentally. Through such iterative in silico-in vitro learning circles, drug discovery teams can decide if metabolism through non-CYP3A4 pathways could be achieved in the SAR of a chemical series to mitigate the CYP3A4 victim DDI risk.

Exploring Tunable Hyperparameters for Deep Neural Networks with Industrial ADME Data Sets.

Deep learning has drawn significant attention in different areas including drug discovery. It has been proposed that it could outperform other machine learning algorithms, especially with big data sets. In the field of pharmaceutical industry, machine learning models are built to understand quantitative structure-activity relationships (QSARs) and predict molecular activities, including absorption, distribution, metabolism, and excretion (ADME) properties, using only molecular structures. Previous reports have demonstrated the advantages of using deep neural networks (DNNs) for QSAR modeling. One of the challenges while building DNN models is identifying the hyperparameters that lead to better generalization of the models. In this study, we investigated several tunable hyperparameters of deep neural network models on 24 industrial ADME data sets. We analyzed the sensitivity and influence of five different hyperparameters including the learning rate, weight decay for L2 regularization, dropout rate, activation function, and the use of batch normalization. This paper focuses on strategies and practices for DNN model building. Further, the optimized model for each data set was built and compared with the benchmark models used in production. Based on our benchmarking results, we propose several practices for building DNN QSAR models.

In Silico and in Vitro Assessment of OATP1B1 Inhibition in Drug Discovery.

The organic anion-transporting polypeptide 1B1 transporter belongs to the solute carrier superfamily and is highly expressed at the basolateral membrane of hepatocytes. Several clinical studies show drug-drug interactions involving OATP1B1, thereby prompting the International Transporter Consortium to label OATP1B1 as a critical transporter that can influence a compound's disposition. To examine OATP1B1 inhibition early in the drug discovery process, we established a medium-throughput concentration-dependent OATP1B1 assay. To create an in silico OATP1B1 inhibition model, deliberate in vitro assay enrichment was performed with publically known OATP1B1 inhibitors, noninhibitors, and compounds from our own internal chemistry. To date, approximately 1200 compounds have been tested in the assay with 60:40 distribution between noninhibitors and inhibitors. Bagging, random forest, and support vector machine fingerprint (SVM-FP) quantitative structure-activity relationship classification models were created, and each method showed positive and negative predictive values >90%, sensitivity >80%, specificity >95%, and Matthews correlation coefficient >0.8 on a prospective test set indicating the ability to distinguish inhibitors from noninhibitors. A SVMF-FP regression model was also created that showed an R2 of 0.39, Spearman's rho equal to 0.76, and was capable of predicting 69% of the prospective test set within the experimental variability of the assay (3-fold). In addition to the in silico quantitative structure-activity relationship (QSAR) models, physicochemical trends were examined to provide structure activity relationship guidance to early discovery teams. A JMP partition tree analysis showed that among the compounds with calculated logP >3.5 and ≥1 negatively charged atom, 94% were identified as OATP1B1 inhibitors. The combination of the physicochemical trends along with an in silico QSAR model provides discovery project teams a valuable tool to identify and address drug-drug interaction liability due to OATP1B1 inhibition.

Synthesis and biological evaluation of aryl-oxadiazoles as inhibitors of Mycobacterium tuberculosis.

Despite increased research efforts to find new treatments for tuberculosis in recent decades, compounds with novel mechanisms of action are still required. We previously identified a series of novel aryl-oxadiazoles with anti-tubercular activity specific for bacteria using butyrate as a carbon source. We explored the structure activity relationship of this series. Structural modifications were performed in all domains to improve potency and physico-chemical properties. A number of compounds displayed sub-micromolar activity against M. tuberculosis utilizing butyrate, but not glucose as the carbon source. Compounds showed no or low cytotoxicity against eukaryotic cells. Three compounds were profiled in mouse pharmacokinetic studies. Plasma clearance was low to moderate but oral exposure suggested solubility-limited drug absorption in addition to first pass metabolism. The presence of a basic nitrogen in the linker slightly increased solubility, and salt formation optimized aqueous solubility. Our findings suggest that the 1,3,4-oxadiazoles are useful tools and warrant further investigation.

QSAR Model of Unbound Brain-to-Plasma Partition Coefficient, Kp,uu,brain: Incorporating P-glycoprotein Efflux as a Variable.

We report development and prospective validation of a QSAR model of the unbound brain-to-plasma partition coefficient, Kp,uu,brain, based on the in-house data set of ∼1000 compounds. We discuss effects of experimental variability, explore the applicability of both regression and classification approaches, and evaluate a novel, model-within-a-model approach of including P-glycoprotein efflux prediction as an additional variable. When tested on an independent test set of 91 internal compounds, incorporation of P-glycoprotein efflux information significantly improves the model performance resulting in an R2 of 0.53, RMSE of 0.57, Spearman's Rho correlation coefficient of 0.73, and qualitative prediction accuracy of 0.8 (kappa = 0.6). In addition to improving the performance, one of the key advantages of this approach is the larger chemical space coverage provided indirectly through incorporation of the in vitro, higher throughput data set that is 4 times larger than the in vivo data set.

Integration of in silico and in vitro tools for scaffold optimization during drug discovery: predicting P-glycoprotein efflux.

In silico tools are regularly utilized for designing and prioritizing compounds to address challenges related to drug metabolism and pharmacokinetics (DMPK) during the process of drug discovery. P-Glycoprotein (P-gp) is a member of the ATP-binding cassette (ABC) transporters with broad substrate specificity that plays a significant role in absorption and distribution of drugs that are P-gp substrates. As a result, screening for P-gp transport has now become routine in the drug discovery process. Typically, bidirectional permeability assays are employed to assess in vitro P-gp efflux. In this article, we use P-gp as an example to illustrate a well-validated methodology to effectively integrate in silico and in vitro tools to identify and resolve key barriers during the early stages of drug discovery. A detailed account of development and application of in silico tools such as simple guidelines based on physicochemical properties and more complex quantitative structure-activity relationship (QSAR) models is provided. The tools were developed based on structurally diverse data for more than 2000 compounds generated using a robust P-gp substrate assay over the past several years. Analysis of physicochemical properties revealed a significantly lower proportion (<10%) of P-gp substrates among the compounds with topological polar surface area (TPSA) <60 Å(2) and the most basic cpKa <8. In contrast, this proportion of substrates was greater than 75% for compounds with TPSA >60 Å(2) and the most basic cpKa >8. Among the various QSAR models evaluated to predict P-gp efflux, the Bagging model provided optimum prediction performance for prospective validation based on chronological test sets. Four sequential versions of the model were built with increasing numbers of compounds to train the models as new data became available. Except for the first version with the smallest training set, the QSAR models exhibited robust prediction profiles with positive prediction values (PPV) and negative prediction values (NPV) exceeding 80%. The QSAR model demonstrated better concordance with the manual P-gp substrate assay than an automated P-gp substrate screen. The in silico and the in vitro tools have been effectively integrated during early stages of drug discovery to resolve P-gp-related challenges exemplified by several case studies. Key learning based on our experience with P-gp can be widely applicable across other DMPK-related challenges.

Metabolic-intermediate complex formation with cytochrome P450: theoretical studies in elucidating the reaction pathway for the generation of reactive nitroso intermediate.

Mechanism-based inhibition (MBI) of cytochrome P450 (CYP) can lead to drug-drug interactions and often to toxicity. Some aliphatic and aromatic amines can undergo biotransformation reactions to form reactive metabolites such as nitrosoalkanes, leading to MBI of CYPs. It has been proposed that the nitrosoalkanes coordinate with the heme iron, forming metabolic-intermediate complex (MIC), resulting in the quasi-irreversible inhibition of CYPs. Limited mechanistic details regarding the formation of reactive nitroso intermediate and its coordination with heme-iron have been reported. A quantum chemical analysis was performed to elucidate potential reaction pathways for the generation of nitroso intermediate and the formation of MIC. Elucidation of the energy profile along the reaction path, identification of three-dimensional structures of reactive intermediates and transition states, as well as charge and spin density analyses, were performed using the density functional B3LYP method. The study was performed using Cpd I [iron (IV-oxo] heme porphine with SH(-) as the axial ligand) to represent the catalytic domain of CYP, simulating the biotransformation process. Three pathways: (i) N-oxidation followed by proton shuttle, (ii) N-oxidation followed by 1,2-H shift, and (iii) H-abstraction followed by rebound mechanism, were studied. It was observed that the proton shuttle pathway was more favorable over the whole reaction leading to reactive nitroso intermediate. This study revealed that the MIC formation from a primary amine is a favorable exothermic process, involving eight different steps and preferably takes place on the doublet spin surface of Cpd I. The rate-determining step was identified to be the first N-oxidation of primary amine.

How hydrogen bonds impact P-glycoprotein transport and permeability.

The requirement to cross a biological membrane can be a complex process especially if multidrug transporters such as P-gp must be considered. Drug partitioning into the lipid membrane and efflux by P-gp are tightly coupled processes wherein H-bonding interactions play a key role. All H-bond donors and acceptors are not equal in terms of the strength of the H-bonds that they form, hence it is important to consider their relative strength. Using various examples from literature, we illustrate the benefits of considering the relative strengths of individual H-bonds and introducing intramolecular H-bonds to increase membrane permeability and/or decrease P-gp efflux.

High-throughput hydrogen bond strength calculation and its applications in optimizing drug ADME properties.

AIM: Modifying the molecule's intrinsic hydrogen bond strength (HBS) is a useful approach in optimizing its permeability and P-glycoprotein (P-gp) efflux. Quantum mechanics (QM) based computation has been utilized to estimate the molecular intrinsic HBS. Despite its usefulness, the computation is time consuming for a large set of molecules. METHODOLOGY/RESULTS: We introduced a fragment-based high-throughput HBS calculation method and validated it with internal and external datasets. Examples have been presented where the P-gp efflux and permeability can be optimized by modulating calculated HBS. CONCLUSION: The results will enable medicinal chemists to calculate HBS in a high-throughput manner while optimizing permeability and P-gp efflux. This will further improve the efficiency of balancing multiple properties during drug discovery process.

Plasticity of the Mycobacterium tuberculosis respiratory chain and its impact on tuberculosis drug development.

The viability of Mycobacterium tuberculosis (Mtb) depends on energy generated by its respiratory chain. Cytochrome bc1-aa3 oxidase and type-2 NADH dehydrogenase (NDH-2) are respiratory chain components predicted to be essential, and are currently targeted for drug development. Here we demonstrate that an Mtb cytochrome bc1-aa3 oxidase deletion mutant is viable and only partially attenuated in mice. Moreover, treatment of Mtb-infected marmosets with a cytochrome bc1-aa3 oxidase inhibitor controls disease progression and reduces lesion-associated inflammation, but most lesions become cavitary. Deletion of both NDH-2 encoding genes (Δndh-2 mutant) reveals that the essentiality of NDH-2 as shown in standard growth media is due to the presence of fatty acids. The Δndh-2 mutant is only mildly attenuated in mice and not differently susceptible to clofazimine, a drug in clinical use proposed to engage NDH-2. These results demonstrate the intrinsic plasticity of Mtb's respiratory chain, and highlight the challenges associated with targeting the pathogen's respiratory enzymes for tuberculosis drug development.

Mollamides B and C, Cyclic hexapeptides from the indonesian tunicate Didemnum molle.

Two new cyclic hexapeptides, mollamides B (1) and C (2), were isolated from the Indonesian tunicate Didemnum molle along with the known peptide keenamide A (3). The structures were established using 1D and 2D NMR experiments. The relative configuration of mollamide B at the thiazoline moiety was determined using molecular modeling coupled with NMR-derived restraints. Their absolute configuration was determined using Marfey's method. The new peptides have been evaluated for their antimicrobial, antimalarial, anticancer, anti-HIV-1, anti-Mtb, and anti-inflammatory activities. Keenamide A and mollamide B show cytotoxicity against several cancer cell lines.

In Silico Absorption, Distribution, Metabolism, Excretion, and Pharmacokinetics (ADME-PK): Utility and Best Practices. An Industry Perspective from the International Consortium for Innovation through Quality in Pharmaceutical Development.

In silico tools to investigate absorption, distribution, metabolism, excretion, and pharmacokinetics (ADME-PK) properties of new chemical entities are an integral part of the current industrial drug discovery paradigm. While many companies are active in the field, scientists engaged in this area do not necessarily share the same background and have limited resources when seeking guidance on how to initiate and maintain an in silico ADME-PK infrastructure in an industrial setting. This work summarizes the views of a group of industrial in silico and experimental ADME scientists, participating in the In Silico ADME Working Group, a subgroup of the International Consortium for Innovation through Quality in Pharmaceutical Development (IQ) Drug Metabolism Leadership Group. This overview on the benefits, caveats, and impact of in silico ADME-PK should serve as a resource for medicinal chemists, computational chemists, and DMPK scientists working in drug design to increase their knowledge in the area.

Ligand-supported homology modeling of the human angiotensin II type 1 (AT(1)) receptor: insights into the molecular determinants of telmisartan binding.

Angiotensin II type 1 (AT(1)) receptor belongs to the super-family of G-protein-coupled receptors, and antagonists of the AT(1) receptor are effectively used in the treatment of hypertension. To understand the molecular interactions of these antagonists, such as losartan and telmisartan, with the AT(1) receptor, a homology model of the human AT(1) (hAT(1)) receptor with all connecting loops was constructed from the 2.6 A resolution crystal structure (PDB i.d., 1L9H) of bovine rhodopsin. The initial model generated by MODELLER was subjected to a stepwise ligand-supported model refinement. This protocol involved initial docking of non-peptide AT(1) antagonists in the putative binding site, followed by several rounds of iterative energy minimizations and molecular dynamics simulations. The final model was validated based on its correlation with several structure-activity relationships and site-directed mutagenesis data. The final model was also found to be in agreement with a previously reported AT(1) antagonist pharmacophore model. Docking studies were performed for a series of non-peptide AT(1) receptor antagonists in the active site of the final hAT(1) receptor model. The docking was able to identify key molecular interactions for all the AT(1) antagonists studied. Reasonable correlation was observed between the interaction energy values and the corresponding binding affinities of these ligands, providing further validation for the model. In addition, an extensive unrestrained molecular dynamics simulation showed that the docking-derived bound pose of telmisartan is energetically stable. Knowledge gained from the present studies can be used in structure-based drug design for developing novel ligands for the AT(1) receptor.

Homology modeling of G-protein-coupled receptors and implications in drug design.

G-protein-coupled receptors (GPCRs) are considered therapeutically important due to their involvement in a variety of processes governing several cellular functions, and their tractability as drug targets. A large percentage of drugs on the market, and in development stages, target the super family of the GPCRs. The enormous interest in GPCR drug design is, however, limited by the scarcity of structural information. The only GPCR for which a three dimensional (3D) structure is reported is bovine rhodopsin and it belongs to class A of the GPCR family. As a result, there has been considerable interest in alternative techniques, for example, homology modeling of GPCRs, in order to derive useful three dimensional models of other proteins for use in structure-based drug design. However, homology modeling of GPCRs is not straightforward, and encounters several problems, owing to the availability of a single structural template, as well as the low degree of sequence homology between the template and target sequences. There are several key issues which need to be considered during every stage of GPCR homology modeling, in order to derive reasonable 3D models. Homology modeling of GPCRs has been utilized increasingly in the past few years and has been successful, not only in furthering the understanding of ligand-protein interactions, but also in the identification of new and potent ligands. Thus, with the lessons learned from past experiences and new developments, homology modeling in case of GPCRs can be harnessed for developing more reliable three dimensional models. This, in turn, will provide better tools to use in structure-based drug design leading to the identification of novel and potent GPCR ligands for several therapeutic indications.

Identification of novel parasitic cysteine protease inhibitors by use of virtual screening. 2. The available chemical directory.

The incidence of parasitic infections such as malaria, leishmaniasis, and trypanosomiasis has been steadily increasing. Since the existing chemotherapy of these diseases suffers from lack of safe and effective drugs and/or the presence of widespread drug resistance, there is an urgent need for development of potent, mechanism-based antiparasitic agents against these diseases. Cysteine proteases have been established as valid targets for this purpose. The Available Chemical Directory consisting of nearly 355,000 compounds was screened in silico against the homology models of plasmodial cysteine proteases, falcipain-2, and falcipain-3, to identify structurally diverse non-peptide inhibitors. The study led to identification of 22 inhibitors of parasitic cysteine proteases out of which 18 compounds were active against falcipain-2 and falcipain-3. Eight compounds exhibited dual activity against both enzymes. Additionally, four compounds were found to inhibit L. donovani cysteine protease. While one of the cysteine protease inhibitors also exhibited in vitro antiplasmodial activity with an IC50 value of 9.5 microM, others did not show noticeable antiplasmodial activity up to 20 microM. A model identifying important pharmacophoric features common to the structurally diverse falcipain-2 inhibitors has also been developed. Very few potent non-peptide inhibitors of the parasitic cysteine proteases have been reported so far, and identification of these novel and chemically diverse inhibitors should provide leads to be optimized into candidates to treat protozoal infections.

Design and synthesis of the first generation of dithiolane thiazolidinedione- and phenylacetic acid-based PPARgamma agonists.

A series of novel derivatives of potent antioxidant vitamin, alpha-lipoic acid, and related analogues were designed, synthesized, and evaluated for their PPARgamma agonist activities. Compounds 9a and the water soluble analogue11e were found to be potent PPARgamma agonists. Compound 9a appeared to have a significant role in improving insulin sensitivity and reducing triglyceride levels in fa/fa rats as well as inhibited proliferation of a variety of normal and neoplastic cultured human cell types. These novel compounds may prove efficacious not only in the treatment of Type 2 diabetes, but also atherosclerosis, prevention of vascular restenosis, and inflammatory skin diseases.

Stereochemistry Balances Cell Permeability and Solubility in the Naturally Derived Phepropeptin Cyclic Peptides.

Cyclic peptide (CP) natural products provide useful model systems for mapping "beyond-Rule-of-5" (bRo5) space. We identified the phepropeptins as natural product CPs with potential cell permeability. Synthesis of the phepropeptins and epimeric analogues revealed much more rapid cellular permeability for the natural stereochemical pattern. Despite being more cell permeable, the natural compounds exhibited similar aqueous solubility as the corresponding epimers, a phenomenon explained by solvent-dependent conformational flexibility among the natural compounds. When analyzing the polarity of the solution structures we found that neither the number of hydrogen bonds nor the total polar surface area accurately represents the solvation energies of the high and low dielectric conformations. This work adds to a growing number of natural CPs whose solvent-dependent conformational behavior allows for a balance between aqueous solubility and cell permeability, highlighting structural flexibility as an important consideration in the design of molecules in bRo5 chemical space.

Discovery and Characterization of 2-Acylaminoimidazole Microsomal Prostaglandin E Synthase-1 Inhibitors.

As part of a program aimed at the discovery of antinociceptive therapy for inflammatory conditions, a screening hit was found to inhibit microsomal prostaglandin E synthase-1 (mPGES-1) with an IC50 of 17.4 µM. Structural information was used to improve enzyme potency by over 1000-fold. Addition of an appropriate substituent alleviated time-dependent cytochrome P450 3A4 (CYP3A4) inhibition. Further structure-activity relationship (SAR) studies led to 8, which had desirable potency (IC50 = 12 nM in an ex vivo human whole blood (HWB) assay) and absorption, distribution, metabolism, and excretion (ADME) properties. Studies on the formulation of 8 identified 8·H3PO4 as suitable for clinical development. Omission of a lipophilic portion of the compound led to 26, a readily orally bioavailable inhibitor with potency in HWB comparable to celecoxib. Furthermore, 26 was selective for mPGES-1 inhibition versus other mechanisms in the prostanoid pathway. These factors led to the selection of 26 as a second clinical candidate.

Discovery of the First α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic Acid (AMPA) Receptor Antagonist Dependent upon Transmembrane AMPA Receptor Regulatory Protein (TARP) γ-8.

Transmembrane AMPA receptor regulatory proteins (TARPs) are a family of scaffolding proteins that regulate AMPA receptor trafficking and function. TARP γ-8 is one member of this family and is highly expressed within the hippocampus relative to the cerebellum. A selective TARP γ-8-dependent AMPA receptor antagonist (TDAA) is an innovative approach to modulate AMPA receptors in specific brain regions to potentially increase the therapeutic index relative to known non-TARP-dependent AMPA antagonists. We describe here, for the first time, the discovery of a noncompetitive AMPA receptor antagonist that is dependent on the presence of TARP γ-8. Three major iteration cycles were employed to improve upon potency, CYP1A2-dependent challenges, and in vivo clearance. An optimized molecule, compound (-)-25 (LY3130481), was fully protective against pentylenetetrazole-induced convulsions in rats without the motor impairment associated with non-TARP-dependent AMPA receptor antagonists. Compound (-)-25 could be utilized to provide proof of concept for antiepileptic efficacy with reduced motor side effects in patients.

Identification of Phenoxyalkylbenzimidazoles with Antitubercular Activity.

We conducted an evaluation of the phenoxyalkylbenzimidazole series based on the exemplar 2-ethyl-1-(3-phenoxypropyl)-1H-benzo[d]imidazole for its antitubercular activity. Four segments of the molecule were examined systematically to define a structure-activity relationship with respect to biological activity. Compounds had submicromolar activity against Mycobacterium tuberculosis; the most potent compound had a minimum inhibitory concentration (MIC) of 52 nM and was not cytotoxic against eukaryotic cells (selectivity index = 523). Compounds were selective for M. tuberculosis over other bacterial species, including the closely related Mycobacterium smegmatis. Compounds had a bacteriostatic effect against aerobically grown, replicating M. tuberculosis, but were bactericidal against nonreplicating bacteria. Representative compounds had moderate to high permeability in MDCK cells, but were rapidly metabolized in rodents and human liver microsomes, suggesting the possibility of rapid in vivo hepatic clearance mediated by oxidative metabolism. These results indicate that the readily synthesized phenoxyalkylbenzimidazoles are a promising class of potent and selective antitubercular agents, if the metabolic liability can be solved.