Early Stages Of Drug Discovery Research Articles

Development of a high-throughput screening platform to identify new therapeutic agents for Medulloblastoma Group 3

Pediatric brain tumors (PBTs) represent about 25 % of all pediatric cancers and are the most common solid tumors in children and adolescents. Medulloblastoma (MB) is the most frequently occurring malignant PBT, accounting for almost 10 % of all pediatric cancer deaths. MB Group 3 (MB G3) accounts for 25–30 % of all MB cases and has the worst outcome, particularly when associated with MYC amplification. However, no targeted treatments for this group have been developed so far.Here we describe a unique high throughput screening (HTS) platform specifically designed to identify new therapies for MB G3. The platform incorporates optimized and validated 2D and 3D efficacy and toxicity models, that account for tumor heterogenicity, limited efficacy and unacceptable toxicity from the very early stage of drug discovery. The platform has been validated by conducting a pilot HTS campaign with a 1280 lead-like compound library. Results showed 8 active compounds, targeting MB reported targets and several are currently approved or in clinical trials for pediatric patients with PBTs, including MB. Moreover, hits were combined to avoid tumor resistance, identifying 3 synergistic pairs, one of which is currently under clinical study for recurrent MB and other PBTs.

In silico prediction of ocular toxicity of compounds using explainable machine learning and deep learning approaches.

The accurate identification of chemicals with ocular toxicity is of paramount importance in health hazard assessment. In contemporary chemical toxicology, there is a growing emphasis on refining, reducing, and replacing animal testing in safety evaluations. Therefore, the development of robust computational tools is crucial for regulatory applications. The performance of predictive models is heavily reliant on the quality and quantity of data. In this investigation, we amalgamated the most extensive dataset (4901 compounds) sourced from governmental GHS-compliant databases and literature to develop binary classification models of chemical ocular toxicity. We employed 12 molecular representations in conjunction with six machine learning algorithms and two deep learning algorithms to create a series of binary classification models. The findings indicated that the deep learning method GCN outperformed the machine learning models in cross-validation, achieving an impressive AUC of 0.915. However, the top-performing machine learning model (RF-Descriptor) demonstrated excellent performance with an AUC of 0.869 on the test set and was therefore selected as the best model. To enhance model interpretability, we conducted the SHAP method and attention weights analysis. The two approaches offered visual depictions of the relevance of key descriptors and substructures in predicting ocular toxicity of chemicals. Thus, we successfully struck a delicate balance between data quality and model interpretability, rendering our model valuable for predicting and comprehending potential ocular-toxic compounds in the early stages of drug discovery.

A Comprehensive Review of the Advancement in Omic Technologies in the Field of Drug Discovery and Development

Abstract: Discovering new drugs is time-consuming and expensive and involves many different tools from various domains. Numerous omic technologies, such as genomics, transcriptomics, proteomics, and metabolomics, have been created to speed up the process. Leveraging genetic and genomic insights, these methodologies play a pivotal role. Genetic insights aid in target identification, prioritization, and the prediction of drug outcomes. Gene expression data informs drug discovery, while proteomics uncovers targets and facilitates high-throughput profiling. Enhancing drug efficacy necessitates mechanistic insights into downstream effects, enabling side effects and resistance prediction. Early-stage drug discovery now extensively employs diverse metabolomics platforms. This review underscores the recent strides of omic technologies in drug discovery, affirming their role in enhancing drug viability and regulatory approval. The emphasis lies on the latest advancements in genomics, transcriptomics, proteomics, and metabolomics, collectively fortifying drug development.

High-throughput assay to simultaneously evaluate activation of CYP3A and the direct and time-dependent inhibition of CYP3A, CYP2C9, and CYP2D6 using liquid chromatography-tandem mass spectrometry

In the early stages of drug discovery, adequate evaluation of the potential drug–drug interactions (DDIs) of drug candidates is important. Several CYP3A activators are known to lead to underestimation of DDIs. These compounds affect midazolam 1'-hydroxylation but not midazolam 4-hydroxylation. We used both metabolic reactions of midazolam to evaluate the activation and inhibition of CYP3A activators simultaneously. For our CYP inhibition assay using cocktail probe substrates, simultaneous liquid chromatography-tandem mass spectrometry monitoring of 1'-hydroxymidazolam and 4-hydroxymidazolam for CYP3A was established in addition to monitoring of 4-hydroxydiclofenac and 1'-hydroxybufuralol for CYP2C9 and CYP2D6. The results of our cocktail inhibition assay were well correlated with those of a single inhibition assay, as were the estimated inhibition parameters for typical CYP3A inhibitors. In our assay, a proprietary compound that activated midazolam 1'-hydroxylation and tended to inhibit 4-hydroxylation was evaluated along with known CYP3A activators. All compounds were well characterised by comparison of the results of midazolam 1'- and 4-hydroxylation. In conclusion, our CYP cocktail inhibition assay can detect CYP3A activation and assess the direct and time-dependent inhibition potentials for CYP3A, CYP2C9, and CYP2D6. This method is expected to be very efficient in the early stages of drug discovery.

Developing a GNN-based AI model to predict mitochondrial toxicity using the bagging method.

Mitochondrial toxicity has been implicated in the development of various toxicities, including hepatotoxicity. Therefore, mitochondrial toxicity has become a major screening factor in the early discovery phase of drug development. Several models have been developed to predict mitochondrial toxicity based on chemical structures. However, they only provide a binary classification of positive or negative results and do not provide the substructures that contribute to a positive decision. Therefore, we developed an artificial intelligence (AI) model to predict mitochondrial toxicity and visualize structural alerts. To construct the model, we used the open-source software library kMoL, which employs a graph neural network approach that allows learning from chemical structure data. We also utilized the integrated gradient method, which enables the visualization of substructures that contribute to positive results. The dataset used to construct the AI model exhibited a significant imbalance, with significantly more negative than positive data. To address this, we employed the bagging method, which resulted in a model with high predictive performance, as evidenced by an F1 score of 0.839. This model can also be used to visualize substructures that contribute to mitochondrial toxicity using the integrated gradient method. Our AI model predicts mitochondrial toxicity based on chemical structures and may contribute to screening mitochondrial toxicity in the early stages of drug discovery.

A gravity-driven tissue chip to study the efficacy and toxicity of cancer therapeutics.

Tissue chip and organs-on-chip technologies have emerged as promising tools in preclinical studies. In oncology, this is driven by the high failure rates of candidate drugs in clinical trials mainly due to inadequate efficacy or intolerable toxicity and the need for better predictive preclinical models than those traditionally used. However, the intricate design, fabrication, operation, and limited compatibility with automation limit the utility of tissue chips. To tackle these issues, we designed a novel 32-unit tissue chip in the format of standard 96-well plates to streamline automation, fabricated it using 3D printing, and leveraged gravity-driven flow to bypass the need for external flow devices. Each unit includes three interconnected tissue compartments that model liver, tumor, and bone marrow stroma. The focus on liver and bone marrow stroma was due to their respective roles in drug metabolism and disturbances to the bone marrow niche from off-target toxicity of chemotherapies. We analyzed flow patterns, mixing, and oxygen transport among and within the compartments through finite element simulations and demonstrated the utility of the tissue chip to study the efficacy of commonly-used cytotoxic cancer drugs against tumor cells and their toxicity toward liver and bone marrow cells. The ability to simultaneously study drug efficacy and toxicity in high throughput can help select promising therapeutics in early stages of drug discovery in preclinical studies.

Recent developments and applications of ambient mass spectrometry imaging in pharmaceutical research: an overview.

The application of ambient mass spectrometry imaging "MSI" is expanding in the areas of fundamental research on drug delivery and multiple phases of the process of identifying and developing drugs. Precise monitoring of a drug's pharmacological workflows, such as intake, distribution, metabolism, and discharge, is made easier by MSI's ability to determine the concentrations of the initiating drug and its metabolites across dosed samples without losing spatial data. Lipids, glycans, and proteins are just a few of the many phenotypes that MSI may be used to concurrently examine. Each of these substances has a particular distribution pattern and biological function throughout the body. MSI offers the perfect analytical tool for examining a drug's pharmacological features, especially in vitro and in vivo effectiveness, security, probable toxic effects, and putative molecular pathways, because of its high responsiveness in chemical and physical environments. The utilization of MSI in the field of pharmacy has further extended from the traditional tissue examination to the early stages of drug discovery and development, including examining the structure-function connection, high-throughput capabilities in vitro examination, and ex vivo research on individual cells or tumor spheroids. Additionally, an enormous array of endogenous substances that may function as tissue diagnostics can be scanned simultaneously, giving the specimen a highly thorough characterization. Ambient MSI techniques are soft enough to allow for easy examination of the native sample to gather data on exterior chemical compositions. This paper provides a scientific and methodological overview of ambient MSI utilization in research on pharmaceuticals.

Exploring the Evolution of Analytical Method Validation: Current Perspectives

In this review, we delve into the critical process of analytical method development and validation within the realm of pharmaceutical drug development. The creation of sound analytical methods stands as a cornerstone from the early stages of drug discovery through to the eventual marketing approval. Our objective is to meticulously examine the methodology behind developing, optimizing, and validating analytical methods, particularly focusing on their application to drug products – from the formulation's developmental stage to the production of commercial batches. We begin by elucidating the significance of method development tailored to the specific components of finished drug products or in-process tests. Additionally, we delve into the intricacies of sample preparation techniques essential for accurate analysis. Central to our discussion are the practical approaches employed to ascertain critical parameters including selectivity, specificity, limit of detection, limit of quantitation, linearity, range accuracy, precision, recovery, solution stability, ruggedness, and robustness of liquid chromatographic methods. These parameters are indispensable for supporting routine, in-process, and stability analyses crucial for ensuring the safety, efficacy, and quality of pharmaceutical products. By dissecting the intricacies of method development and validation, this review aims to provide pharmaceutical researchers and practitioners with valuable insights and practical guidelines to navigate the complex landscape of analytical method validation in drug development.

Three-dimensional in vitro model of bone metastases of neuroblastoma as a tool for pharmacological evaluations.

In vitro metastatic models are foreseen to introduce a breakthrough in the field of preclinical screening of more functional small-molecule pharmaceuticals and biologics. To achieve this goal, the complexity of current in vitro systems requests an appropriate upgrade to approach the three-dimensional (3D) in vivo metastatic disease. Here, we explored the potential of our 3D β-tricalcium phosphate (β-TCP) model of neuroblastoma bone metastasis for drug toxicity assessment. Tailor-made scaffolds with interconnected channels were produced by combining 3D printing and slip casting method. The organization of neuroblastoma cells into a mesenchymal stromal cell (MSC) network, cultured under bioactive conditions provided by β-TCP, was monitored by two-photon microscopy. Deposition of extracellular matrix protein Collagen I by MSCs and persistent growth of tumor cells confirmed the cell-supportive performance of our 3D model. When different neuroblastoma cells were treated with conventional chemotherapeutics, the β-TCP model provided the necessary reproducibility and accuracy of experimental readouts. Drug efficacy evaluation was done for 3D and 2D cell cultures, highlighting the need for a higher dose of chemotherapeutics under 3D conditions to achieve the expected cytotoxicity in tumor cells. Our results confirm the importance of 3D geometry in driving native connectivity between nonmalignant and tumor cells and sustain β-TCP scaffolds as a reliable and affordable drug screening platform for use in the early stages of drug discovery.

QSPRmodeler - An open source application for molecular predictive analytics.

The drug design process can be successfully supported using a variety of in silico methods. Some of these are oriented toward molecular property prediction, which is a key step in the early drug discovery stage. Before experimental validation, drug candidates are usually compared with known experimental data. Technically, this can be achieved using machine learning approaches, in which selected experimental data are used to train the predictive models. The proposed Python software is designed for this purpose. It supports the entire workflow of molecular data processing, starting from raw data preparation followed by molecular descriptor creation and machine learning model training. The predictive capabilities of the resulting models were carefully validated internally and externally. These models can be easily applied to new compounds, including within more complex workflows involving generative approaches.

Accurate prediction of Kp,uu,brain based on experimental measurement of Kp,brain and computed physicochemical properties of candidate compounds in CNS drug discovery

A mathematical equation model was developed by building the relationship between the fu,b/fu,p ratio and the computed physicochemical properties of candidate compounds, thereby predicting Kp,uu,brain based on a single experimentally measured Kp,brain value. A total of 256 compounds and 36 marketed published drugs including acidic, basic, neutral, zwitterionic, CNS-penetrant, and non-CNS penetrant compounds with diverse structures and physicochemical properties were involved in this study. A strong correlation was demonstrated between the fu,b/fu,p ratio and physicochemical parameters (CLogP and ionized fraction). The model showed good performance in both internal and external validations. The percentages of compounds with Kp,uu,brain predictions within 2-fold variability were 80.0 %–83.3 %, and more than 90 % were within a 3-fold variability. Meanwhile, “black box” QSAR models constructed by machine learning approaches for predicting fu,b/fu,p ratio based on the chemical descriptors are also presented, and the ANN model displayed the highest accuracy with an RMSE value of 0.27 and 86.7 % of the test set drugs fell within a 2-fold window of linear regression. These models demonstrated strong predictive power and could be helpful tools for evaluating the Kp,uu,brain by a single measurement parameter of Kp,brain during lead optimization for CNS penetration evaluation and ranking CNS drug candidate molecules in the early stages of CNS drug discovery.

3D printed porous membrane integrated devices to study the chemoattractant induced behavioural response of aquatic organisms.

Biological models with genetic similarities to humans are used for exploratory research to develop behavioral screening tools and understand sensory-motor interactions. Their small, often mm-sized appearance raises challenges in the straightforward quantification of their subtle behavioral responses and calls for new, customisable research tools. 3D printing provides an attractive approach for the manufacture of custom designs at low cost; however, challenges remain in the integration of functional materials like porous membranes. Nanoporous membranes have been integrated with resin exchange using purpose-designed resins by digital light projection 3D printing to yield functionally integrated devices using a simple, economical and semi-automated process. Here, the impact of the layer thickness and layer number on the porous properties - parameters unique for 3D printing - are investigated, showing decreases in mean pore diameter and porosity with increasing layer height and layer number. From the same resin formulation, materials with average pore size between 200 and 600 nm and porosity between 45% and 61% were printed. Membrane-integrated devices were used to study the chemoattractant induced behavioural response of zebrafish embryos and planarians, both demonstrating a predominant behavioral response towards the chemoattractant, spending >85% of experiment time in the attractant side of the observation chamber. The presented 3D printing method can be used for printing custom designed membrane-integrated devices using affordable 3D printers and enable fine-tuning of porous properties through adjustment of layer height and number. This accessible approach is expected to be adopted for applications including behavioural studies, early-stage pre-clinical drug discovery and (environmental) toxicology.

DrugPose: benchmarking 3D generative methods for early stage drug discovery

Molecule generation in 3D space has gained attention in the past few years.

PharmacoNet: deep learning-guided pharmacophore modeling for ultra-large-scale virtual screening

As ultra-large-scale virtual screening becomes critical for early-stage drug discovery, highly efficient screening methods are gaining prominence. Deep-learning-based approaches which directly estimate binding affinities without binding conformation have attracted great...

FRAGSITE2: A structure and fragment-based approach for virtual ligand screening.

Protein function annotation and drug discovery often involve finding small molecule binders. In the early stages of drug discovery, virtual ligand screening (VLS) is frequently applied to identify possible hits before experimental testing. While our recent ligand homology modeling (LHM)-machine learning VLS method FRAGSITE outperformed approaches that combined traditional docking to generate protein-ligand poses and deep learning scoring functions to rank ligands, a more robust approach that could identify a more diverse set of binding ligands is needed. Here, we describe FRAGSITE2 that shows significant improvement on protein targets lacking known small molecule binders and no confident LHM identified template ligands when benchmarked on two commonly used VLS datasets: For both the DUD-E set and DEKOIS2.0 set and ligands having a Tanimoto coefficient (TC) < 0.7 to the template ligands, the 1% enrichment factor (EF1% ) of FRAGSITE2 is significantly better than those for FINDSITEcomb2.0 , an earlier LHM algorithm. For the DUD-E set, FRAGSITE2 also shows better ROC enrichment factor and AUPR (area under the precision-recall curve) than the deep learning DenseFS scoring function. Comparison with the RF-score-VS on the 76 target subset of DEKOIS2.0 and a TC < 0.99 to training DUD-E ligands, FRAGSITE2 has double the EF1% . Its boosted tree regression method provides for more robust performance than a deep learning multiple layer perceptron method. When compared with the pretrained language model for protein target features, FRAGSITE2 also shows much better performance. Thus, FRAGSITE2 is a promising approach that can discover novel hits for protein targets. FRAGSITE2's web service is freely available to academic users at http://sites.gatech.edu/cssb/FRAGSITE2.

Tuning Potential Functions to Host-Guest Binding Data.

Software to more rapidly and accurately predict protein-ligand binding affinities is of high interest for early-stage drug discovery, and physics-based methods are among the most widely used technologies for this purpose. The accuracy of these methods depends critically on the accuracy of the potential functions that they use. Potential functions are typically trained against a combination of quantum chemical and experimental data. However, although binding affinities are among the most important quantities to predict, experimental binding affinities have not to date been integrated into the experimental data set used to train potential functions. In recent years, the use of host-guest complexes as simple and tractable models of binding thermodynamics has gained popularity due to their small size and simplicity, relative to protein-ligand systems. Host-guest complexes can also avoid ambiguities that arise in protein-ligand systems such as uncertain protonation states. Thus, experimental host-guest binding data are an appealing additional data type to integrate into the experimental data set used to optimize potential functions. Here, we report the extension of the Open Force Field Evaluator framework to enable the systematic calculation of host-guest binding free energies and their gradients with respect to force field parameters, coupled with the curation of 126 host-guest complexes with available experimental binding free energies. As an initial application of this novel infrastructure, we optimized generalized Born (GB) cavity radii for the OBC2 GB implicit solvent model against experimental data for 36 host-guest systems. This refitting led to a dramatic improvement in accuracy for both the training set and a separate test set with 90 additional host-guest systems. The optimized radii also showed encouraging transferability from host-guest systems to 59 protein-ligand systems. However, the new radii are significantly smaller than the baseline radii and lead to excessively favorable hydration free energies (HFEs). Thus, users of the OBC2 GB model currently may choose between GB cavity radii that yield more accurate binding affinities and GB cavity radii that yield more accurate HFEs. We suspect that achieving good accuracy on both will require more far-reaching adjustments to the GB model. We note that binding free-energy calculations using the OBC2 model in OpenMM gain about a 10× speedup relative to corresponding explicit solvent calculations, suggesting a future role for implicit solvent absolute binding free-energy (ABFE) calculations in virtual compound screening. This study proves the principle of using host-guest systems to train potential functions that are transferrable to protein-ligand systems and provides an infrastructure that enables a range of applications.

Modification of gradient HPLC method for determination of small molecules' affinity to human serum albumin under column safety conditions: Robustness and chemometrics study

In the early stages of drug discovery, beyond the biological activity screening, determining the physicochemical properties that affect the distribution of molecules in the human body is an essential step. Plasma protein binding (PPB) is one of the most important investigated endpoints. Nevertheless, the methodology for measuring %PPB is significantly less popular and standardized than other physicochemical properties, like lipophilicity. Here, we proposed how to modify protocols presented by Valko into column safety conditions and evaluated their robustness using fractional factorial design. For robustness testing, four factors were selected: column temperature, mobile phase flow rate, maximum isopropanol concentration in the mobile phase, and buffer pH. Elaborate methods have been applied for the analysis of HSA affinity for three groups of antibiotic-oriented substances that vary in chemical structure: fluoroquinolones, sulfonamides, and tetrazole derivatives. Furthermore, based on the reversed-phase chromatography the workflow of pilot studies was proposed to select molecules that have high affinity to HSA and can not be eluted from the HSA column using the concentration of organic modifier recommended by the column manufacturer.

Structure activity relationship and target prediction for ABX464 analogues in Caenorhabditis elegans

Global challenges with treatment failures and/or widespread resistance in parasitic worms against commercially available anthelmintics lend impetus to the development of new anthelmintics with novel mechanism(s) of action. The free-living nematode Caenorhabditis elegans is an important model organism used for drug discovery, including the screening and structure-activity investigation of new compounds, and target deconvolution. Previously, we conducted a whole-organism phenotypic screen of the ‘Pandemic Response Box’ (from Medicines for Malaria Venture, MMV) and identified a hit compound, called ABX464, with activity against C. elegans and a related, parasitic nematode, Haemonchus contortus. Here, we tested a series of 44 synthesized analogues to explore the pharmacophore of activity on C. elegans and revealed five compounds whose potency was similar or greater than that of ABX464, but which were not toxic to human hepatoma (HepG2) cells. Subsequently, we employed thermal proteome profiling (TPP), protein structure prediction and an in silico-docking algorithm to predict ABX464-target candidates. Taken together, the findings from this study contribute significantly to the early-stage drug discovery of a new nematocide based on ABX464. Future work is aimed at validating the ABX464-protein interactions identified here, and at assessing ABX464 and associated analogues against a panel of parasitic nematodes, towards developing a new anthelmintic with a mechanism of action that is distinct from any of the compounds currently-available commercially.

Development of a Novel CYP3A4 Classifier Model via Site of Metabolism (SOM)-based Molecular Docking, Multivariate Analysis and Molecular Dynamics of Known Substrates and Inhibitors

CYP3A4 is a major hepatic enzyme essential for metabolizing diverse chemical entities. The development of CYP3A4 substrate and inhibitor classifier models is a valuable research strategy to prevent toxicokinetics. Currently, no molecular docking model is available to classify the substrates and inhibitors for a wide range of chemical entities. This study generated a CYP3A4 substrate-inhibitor classifier model using multivariate analysis of selected variables from site of metabolism (SOM)-based docking results and molecular descriptors. CYP3A4 SOM-based molecular docking experiment was performed on the substrates and inhibitors from the DrugBank database. The relevant descriptors were selected using the area under the receiving operating curve (AUROC) and boxplot analysis. The selected variables were used to generate a CYP3A4 substrate-inhibitor classifier model using multivariate analysis. Two complexes were selected for molecular dynamic simulations. A total of 326 substrates and 154 inhibitors were successfully docked. The SOM-based docking model predicted 78.5% SOM correctly of the substrates. The CDOCKER energy, improper dihedral energy, potential energy, initial RMS (root mean square) gradient, and RMS gradient demonstrated acceptable AUROC, sensitivity, specificity, and Youden’s index results. The selected variables were subjected to multivariate analysis, and the PLS-DA analysis classification model showed promising performance in predicting chemical entities’ behavior as substrates or inhibitors. This model classified the external samples correctly with over 70.0% prediction accuracy. Molecular dynamic simulations showed that the selected complexes obtained from the docking model produced stable complexes. The classifier model is applicable as a research tool for the early stages of drug discovery.

From Proteins to Ligands: Decoding Deep Learning Methods for Binding Affinity Prediction.

Accurate in silico prediction of protein-ligand binding affinity is important in the early stages of drug discovery. Deep learning-based methods exist but have yet to overtake more conventional methods such as giga-docking largely due to their lack of generalizability. To improve generalizability, we need to understand what these models learn from input protein and ligand data. We systematically investigated a sequence-based deep learning framework to assess the impact of protein and ligand encodings on predicting binding affinities for commonly used kinase data sets. The role of proteins is studied using convolutional neural network-based encodings obtained from sequences and graph neural network-based encodings enriched with structural information from contact maps. Ligand-based encodings are generated from graph-neural networks. We test different ligand perturbations by randomizing node and edge properties. For proteins, we make use of 3 different protein contact generation methods (AlphaFold2, Pconsc4, and ESM-1b) and compare these with a random control. Our investigation shows that protein encodings do not substantially impact the binding predictions, with no statistically significant difference in binding affinity for KIBA in the investigated metrics (concordance index, Pearson's R Spearman's Rank, and RMSE). Significant differences are seen for ligand encodings with random ligands and random ligand node properties, suggesting a much bigger reliance on ligand data for the learning tasks. Using different ways to combine protein and ligand encodings did not show a significant change in performance.