The Artificial Intelligence-Powered New Era in Pharmaceutical Research and Development: A Review.

Currently, artificial intelligence (AI), machine learning (ML), and deep learning (DL) are gaining increased interest in many fields, particularly in pharmaceutical research and development, where they assist in decision-making in complex situations. Numerous research studies and advancements have demonstrated how these computational technologies are used in various pharmaceutical research and development aspects, including drug discovery, personalized medicine, drug formulation, optimization, predictions, drug interactions, pharmacokinetics/ pharmacodynamics, quality control/quality assurance, and manufacturing processes. Using advanced modeling techniques, these computational technologies can enhance efficiency and accuracy, handle complex data, and facilitate novel discoveries within minutes. Furthermore, these technologies offer several advantages over conventional statistics. They allow for pattern recognition from complex datasets, and the models, typically developed from data-driven algorithms, can predict a given outcome (model output) from a set of features (model inputs). Additionally, this review discusses emerging trends and provides perspectives on the application of AI with quality by design (QbD) and the future role of AI in this field. Ethical and regulatory considerations associated with integrating AI into pharmaceutical technology were also examined. This review aims to offer insights to researchers, professionals, and others on the current state of AI applications in pharmaceutical research and development and their potential role in the future of research and the era of pharmaceutical Industry 4.0 and 5.0.

Discovery of CW-3308 as a Potent, Selective, and Orally Efficacious PROTAC Degrader of BRD9.

The bromodomain-containing protein BRD9 has emerged as an attractive therapeutic target. In the present study, we successfully identified a number of highly potent BRD9 degraders by using two different cereblon ligands developed in our laboratory. Further optimization led to the discovery of CW-3308 as a potent, selective, and orally bioavailable BRD9 degrader. It displayed degradation potency (DC50) < 10 nM and efficiency (Dmax) > 90% against BRD9 in the G401 rhabdoid tumor and HS-SY-II synovial sarcoma cell lines and had a high degradation selectivity over BRD7 and BRD4 proteins. CW-3308 achieved 91% of oral bioavailability in mice. A single oral dose efficiently reduced the BRD9 protein by >90% in the synovial sarcoma HS-SY-II xenograft tumor tissue. Oral administration effectively inhibited HS-SY-II xenograft tumor growth in mice. CW-3308 is a promising lead compound for further optimization and extensive evaluation for the treatment of synovial sarcoma, rhabdoid tumor, and other BRD9-dependent human diseases.

Holistic in silico developability assessment of novel classes of small proteins using publicly available sequence-based predictors.

The development of novel therapeutic proteins is a lengthy and costly process, with an average attrition rate of 91% (Thomas et al. Clinical Development Success Rates and Contributing Factors 2011-2020, 2021). To increase the probability of success and ensure robust drug supply beyond approval, it is essential to assess the developability profile of new potential drug candidates as early and broadly as possible in development (Jain et al. MAbs, 2023. https://doi.org/10.1016/j.copbio.2011.06.002 ). Predicting these properties in silico is expected to be the next leap in innovation as it would enable significantly reduced development timelines combined with broader screens at lower costs. However, developing predictive algorithms typically requires substantial datasets generated under very defined conditions, a limiting factor especially for new classes of therapeutic proteins that hold immense clinical promise. Here we describe a strategy for assessing the developability of a novel class of small therapeutic Anticalin® proteins using machine learning in conjunction with a knowledge-driven approach. The knowledge-driven approach considers developability attributes such as aggregation propensity, charge variants, immunogenicity, specificity, thermal stability, hydrophobicity, and potential post-translational modifications, to calculate a holistic developability score. Based on sequence-derived descriptors as input parameters we established novel statistical models designed to predict the developability scores for Anticalin proteins. The best models yielded low root mean square errors across the entire dataset and were further validated by removing input data from individual screening campaigns and predicting developability scores for those drug candidates. The adoption of the described workflow will enable significantly streamlined preclinical development of Anticalin drug candidates and could potentially be applied to other therapeutic protein scaffolds.

Gene regulatory networks reveal sex difference in lung adenocarcinoma.

BACKGROUND: Lung adenocarcinoma (LUAD) has been observed to have significant sex differences in incidence, prognosis, and response to therapy. However, the molecular mechanisms responsible for these disparities have not been investigated extensively. METHODS: Sample-specific gene regulatory network methods were used to analyze RNA sequencing data from non-cancerous human lung samples from The Genotype Tissue Expression Project (GTEx) and lung adenocarcinoma primary tumor samples from The Cancer Genome Atlas (TCGA); results were validated on independent data. RESULTS: We found that genes associated with key biological pathways including cell proliferation, immune response and drug metabolism are differentially regulated between males and females in both healthy lung tissue and tumor, and that these regulatory differences are further perturbed by tobacco smoking. We also discovered significant sex bias in transcription factor targeting patterns of clinically actionable oncogenes and tumor suppressor genes, including AKT2 and KRAS. Using differentially regulated genes between healthy and tumor samples in conjunction with a drug repurposing tool, we identified several small-molecule drugs that might have sex-biased efficacy as cancer therapeutics and further validated this observation using an independent cell line database. CONCLUSIONS: These findings underscore the importance of including sex as a biological variable and considering gene regulatory processes in developing strategies for disease prevention and management.

Lung adenocarcinoma (LUAD) is a disease that affects males and females differently. Biological sex not only influences chances of developing the disease, but also how the disease progresses and how effective various therapies may be. We analyzed sex-specific gene regulatory networks consisting of transcription factors and the genes they regulate in both healthy lung tissue and in LUAD and identified sex-biased differences. We found that genes associated with cell proliferation, immune response, and drug metabolism are differentially targeted by transcription factors between males and females. We also found that several genes that are drug targets in LUAD, are also regulated differently between males and females. Importantly, these differences are also influenced by an individual's smoking history. Extending our analysis using a drug repurposing tool, we found candidate drugs with evidence that they might work better for one sex or the other. These results demonstrate that considering the differences in gene regulation between males and females will be essential if we are to develop precision medicine strategies for preventing and treating LUAD.

Attribute-guided prototype network for few-shot molecular property prediction.

The molecular property prediction (MPP) plays a crucial role in the drug discovery process, providing valuable insights for molecule evaluation and screening. Although deep learning has achieved numerous advances in this area, its success often depends on the availability of substantial labeled data. The few-shot MPP is a more challenging scenario, which aims to identify unseen property with only few available molecules. In this paper, we propose an attribute-guided prototype network (APN) to address the challenge. APN first introduces an molecular attribute extractor, which can not only extract three different types of fingerprint attributes (single fingerprint attributes, dual fingerprint attributes, triplet fingerprint attributes) by considering seven circular-based, five path-based, and two substructure-based fingerprints, but also automatically extract deep attributes from self-supervised learning methods. Furthermore, APN designs the Attribute-Guided Dual-channel Attention module to learn the relationship between the molecular graphs and attributes and refine the local and global representation of the molecules. Compared with existing works, APN leverages high-level human-defined attributes and helps the model to explicitly generalize knowledge in molecular graphs. Experiments on benchmark datasets show that APN can achieve state-of-the-art performance in most cases and demonstrate that the attributes are effective for improving few-shot MPP performance. In addition, the strong generalization ability of APN is verified by conducting experiments on data from different domains.

FitScore: a fast machine learning-based score for 3D virtual screening enrichment.

Enhancing virtual screening enrichment has become an urgent problem in computational chemistry, driven by increasingly large databases of commercially available compounds, without a commensurate drop in in vitro screening costs. Docking these large databases is possible with cloud-scale computing. However, rapid docking necessitates compromises in scoring, often leading to poor enrichment and an abundance of false positives in docking results. This work describes a new scoring function composed of two parts - a knowledge-based component that predicts the probability of a particular atom type being in a particular receptor environment, and a tunable weight matrix that converts the probability predictions into a dimensionless score suitable for virtual screening enrichment. This score, the FitScore, represents the compatibility between the ligand and the binding site and is capable of a high degree of enrichment across standardized docking test sets.

Repurposing Anidulafungin for Alzheimer's Disease via Fragment-Based Drug Discovery.

The misfolding and aggregation of beta-amyloid (Aß) peptides have been implicated as key pathogenic events in the early stages of Alzheimer's disease (AD). Inhibiting Aß aggregation represents a potential disease-modifying therapeutic approach to AD treatment. Previous studies have identified various molecules that inhibit Aß aggregation, some of which share common chemical substructures (fragments) that may be key to their inhibitory activity. Employing fragment-based drug discovery (FBDD) methods may facilitate the identification of these fragments, which can subsequently be used to screen new inhibitors and provide leads for further drug development. In this study, we used an in silico FBDD approach to identify 17 fragment clusters that are significantly enriched among Aß aggregation inhibitors. These fragments were then used to screen anti-infective agents, a promising drug class for repurposing against amyloid aggregation. This screening process identified 16 anti-infective drugs, 5 of which were chosen for further investigation. Among the 5 candidates, anidulafungin, an antifungal compound, showed high efficacy in inhibiting Aß aggregation in vitro. Kinetic analysis revealed that anidulafungin selectively blocks the primary nucleation step of Aß aggregation, substantially delaying Aß fibril formation. Cell viability assays demonstrated that anidulafungin can reduce the toxicity of oligomeric Aß on BV2 microglia cells. Molecular docking simulations predicted that anidulafungin interacted with various Aß species, including monomers, oligomers, and fibrils, potentially explaining its activity against Aß aggregation and toxicity. This study suggests that anidulafungin is a potential drug to be repurposed for AD, and FBDD is a promising approach for discovering drugs to combat Aß aggregation.

Benzimidazole-Oxadiazole Hybrids-Development in Medicinal Chemistry: An Overview.

To increase the success rate of drug discovery, one practical strategy is to begin molecular hybridisation. The presence of two or more pharmacophores in a single unit leads to a pharmacological potency greater than the sum of each individual moiety's potency. Heterocyclic compounds are very widely distributed in nature and are essential for life activities. Benzimidazole and oxadiazole are privileged structures in medicinal chemistry and are widely used in drug discovery and development due to their vast biological properties. The drug-like properties (like pharmacokinetics and pharmacodynamics) of the individual scaffolds can be improved by benzimidazole-oxadiazole chimeric molecules via a molecular hybridisation approach. Benzimidazole and oxadiazole cores can either be fused or incorporated using either functional groups/bonds. Over the last few decades, drug discovery scientists have predicted that these moieties could be interconnected to yield a novel or modified hybrid compound. Benzimidazole and oxadiazole hybrids were identified as the most potent anticancer, antimicrobial, anti-inflammatory, antioxidant, anticonvulsant, antidepressant, antihypertensive and antitubercular agents. In this context, the present review describes the biological properties of benzimidazole-oxadiazole (1,3,4 and 1,2,4) hybrids, their possible structure-activity relationship and the mechanism of action studies presented. This review article is intended to stimulate fresh ideas in the search for rational designs of more active and less toxic benzimidazole-oxadiazole hybrid prospective therapeutic candidates, as well as more effective diagnostic agents and pathologic probes.

Unveiling the potential of phytochemicals to inhibit nuclear receptor binding SET domain protein 2 for cancer: Pharmacophore screening, molecular docking, ADME properties, and molecular dynamics simulation investigations.

Nuclear receptor binding SET domain protein 2 (NSD2) significantly contributes to the development of cancer, making it a promising target for cancer drug discovery. This research explores natural compounds as potential selective inhibitors for NSD2 in cancer treatment. Employing a comprehensive in silico approach, the study utilized pharmacophore modeling, molecular docking, pharmacokinetic profiling, and molecular dynamics simulations. An e-pharmacophore model-based screening using the first selective and potent ligand bound to NSD2 identified 49,248 natural compounds from the SuperNatural 3.0 database (containing 449,008 molecules) with acceptable alignment with the developed pharmacophore hypotheses. Subsequently, molecular docking was executed to assess the standout compounds which led to the selection of ten candidates that surpassed the reference inhibitor in accordance w the binding affinity expressed as a G score. Ligand-residue interaction analyses of the top three hits (SN0450102, SN0410255, and SN0142336) revealed diverse crucial interactions with the NSD2 active site, including hydrogen bonds, pi-pi stacking, and hydrophobic contacts with key amino acid residues in the NSD2-PWWP1 domain. Pharmacokinetic profiling confirmed the drug-likability for the refined hits, indicating good cellular permeability and minimal blood-brain barrier penetration. Molecular dynamics simulations for 200 nanoseconds affirmed the stability of protein-ligand complexes, with minimal fluctuations in root mean square deviation and root mean square fluctuation analyses. Overall, this study identified promising natural compounds as potential pharmaceutical agents in the treatment of NSD2-associated cancers.

Discovery of Triazolopyrimidine Derivatives as Selective P2X3 Receptor Antagonists Binding to an Unprecedented Allosteric Site as Evidenced by Cryo-Electron Microscopy.

The P2X3 receptor (P2X3R), an ATP-gated cation channel predominantly expressed in C- and Aδ-primary afferent neurons, has been proposed as a drug target for neurological inflammatory diseases, e.g., neuropathic pain, and chronic cough. Aiming to develop novel, selective P2X3R antagonists, tetrazolopyrimidine-based hit compound 9 was optimized through structure-activity relationship studies by modifying the tetrazole core as well as side chain substituents. The optimized antagonist 26a, featuring a cyclopropane-substituted triazolopyrimidine core, displayed potent P2X3R-antagonistic activity (IC50 = 54.9 nM), 20-fold selectivity versus the heteromeric P2X2/3R, and high selectivity versus other P2XR subtypes. Noncompetitive P2X3R blockade was experimentally confirmed by calcium influx assays. Cryo-electron microscopy revealed that 26a stabilizes the P2X3R in its desensitized state, acting as a molecular barrier to prevent ions from accessing the central pore. In vivo studies in a rat neuropathic pain model (spinal nerve ligation) showed dose-dependent antiallodynic effects of 26a, thus presenting a novel, promising lead structure.

Discovery of gem-Dimethyl-hydroxymethylpyridine Derivatives as Potent Non-nucleoside RSV Polymerase Inhibitors.

Respiratory syncytial virus (RSV) is an RNA virus infecting the upper and lower respiratory tract and is recognized as a major respiratory health threat, particularly to older adults, immunocompromised individuals, and young children. Around 64 million children and adults are infected every year worldwide. Despite two vaccines and a new generation monoclonal antibody recently approved, no effective antiviral treatment is available. In this manuscript, we present the medicinal chemistry efforts resulting in the identification of compound 28 (JNJ-8003), a novel RSV non-nucleoside inhibitor displaying subnanomolar activity in vitro as well as prominent efficacy in mice and a neonatal lamb models.

Discovery of macrocyclic covalent inhibitors for severe acute respiratory syndrome coronavirus 2 3CL protease.

The coronavirus disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been spread worldwide for more than 3 years. Although the hospitalization rate and mortality have decreased dramatically due to wide vaccination effort and improved treatment options, the disease is still a global health issue due to constant viral mutations, causing negative impact on social and economic activities. In addition, long COVID and complications arising from COVID-19 weeks after infection have become a concern for public health experts. Therefore, better treatments for COVID-19 are still needed. Herein, we describe a class of macrocyclic peptidomimetic compounds that are potent inhibitors of SARS-Cov-2 3CL protease (3CLpro). Significantly, some of the compounds showed a higher stability against human liver microsomes (HLM t1/2 > 180 min) and may be suitable for oral administration without the need for a pharmacokinetic (PK) boosting agent such as ritonavir.

Discovery of a Potent and Orally Bioavailable Xanthine Oxidase/Urate Transporter 1 Dual Inhibitor as a Potential Treatment for Hyperuricemia and Gout.

The main uric acid-lowering agents in clinical use for hyperuricemia and gout are xanthine oxidase (XO) inhibitors or urate transporter 1 (URAT1) inhibitors. While these therapies can partially control the disease, they have various limitations. The development of XO/URAT1 dual inhibitors offers the potential to enhance therapeutic potency and reduce toxicity compared with single-target inhibitors. Through scaffold hopping from the XO inhibitor febuxostat (2) and the URAT1 inhibitor probenecid (3), followed by structure-activity relationship (SAR) studies, we identified compound 27 as a potent dual inhibitor of XO and URAT1. Compound 27 demonstrated significant dual inhibition in vitro (XO IC50 = 35 nM; URAT1 IC50 = 31 nM) and exhibited favorable pharmacology and pharmacokinetic (PK) profiles in multiple species including monkeys. Furthermore, toxicity studies in rats and monkeys revealed general safety profiles, supporting that compound 27 emerges as a promising novel drug candidate with potent XO/URAT1 dual inhibition for the treatment of gout.

Discovery of a potent, selective, and orally available EGFR C797S mutant inhibitor (DS06652923) with in vivo antitumor activity.

The C797S mutation is one of the major factors behind resistance to the third-generation epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs). Herein, we describe the discovery of DS06652923, a novel, potent, and orally available EGFR-triple-mutant inhibitor. Through scaffold hopping from the previously reported nicotinamide derivative, a novel biaryl scaffold was obtained. The potency was successfully enhanced by the introduction of basic substituents based on analysis of the docking study results. In addition, the difluoromethoxy group on the pyrazole ring improved the kinase selectivity by inducing steric clash with the other kinases. The most optimized compound, DS06652923, achieved tumor regression in the Ba/F3 allograft model upon its oral administration.

Discovery Small-Molecule p300 Inhibitors Derived from a Newly Developed Indazolone-Focused DNA-Encoded Library.

The DNA-encoded library (DEL) is a robust tool for chemical biology and drug discovery. In this study, we developed a DNA-compatible light-promoted reaction that is highly efficient and plate-compatible for DEL construction based on the formation of the indazolone scaffold. Employing this high-efficiency approach, we constructed a DEL featuring an indazolone core, which enabled the identification of a novel series of ligands specifically targeting E1A-binding protein (p300) after DEL selection. Taken together, our findings underscore the feasibility of light-promoted reactions in DEL synthesis and unveil promising avenues for developing p300-targeting inhibitors.

Discovery of Selective Inhibitors for the Lysosomal Parkinson's Disease Channel TMEM175.

TMEM175 is a lysosomal potassium and proton channel that is associated with the development of Parkinson's disease. Advances in understanding the physiological roles of TMEM175 have been hampered by the absence of selective inhibitors, and studies involving genetic perturbations have yielded conflicting results. Here, we report the discovery and characterization of the first reported TMEM175-selective inhibitors, 2-phenylpyridin-4-ylamine (2-PPA), and AP-6. Cryo-EM structures of human TMEM175 bound by 2-PPA and AP-6 reveal that they act as pore blockers, binding at distinct sites in the pore and occluding the ion permeation pathway. Acute inhibition of TMEM175 by 2-PPA or AP-6 increases the level of lysosomal macromolecule catabolism, thereby accelerating macropinocytosis and other digestive processes. These inhibitors may serve as valuable tools to study the roles of TMEM175 in regulating lysosomal function and provide useful templates for future therapeutic development in Parkinson's disease.

Perspectives on the use of machine learning for ADME prediction at AstraZeneca.

A drug's pharmacokinetic (PK) profile will determine its dose and the frequency of administration as well as the likelihood of observing any adverse drug reactions.It is important to understand these PK properties as early as possible in the drug discovery process, ideally, to accurately predict these prior to synthesising the molecule leading to significant improvements in efficiency.In this paper, we describe the approaches used within AstraZeneca to improve our ability of predicting the preclinical and human pharmacokinetic profiles of novel molecules using machine learning and artificial intelligence.We will show how combining chemical structure-based approaches with experimentally derived properties enables improved predictions of in vivo pharmacokinetics and can be extended to molecules that go beyond the classical Lipinski's rule-of-five space.We will also discuss how combining these in vitro and in vivo predictive models could ultimately improve our ability to predict the human outcome at the point of chemical design.

Computational Methods for the Discovery and Optimization of TAAR1 and TAAR5 Ligands.

G-protein-coupled receptors (GPCRs) represent a family of druggable targets when treating several diseases and continue to be a leading part of the drug discovery process. Trace amine-associated receptors (TAARs) are GPCRs involved in many physiological functions with TAAR1 having important roles within the central nervous system (CNS). By using homology modeling methods, the responsiveness of TAAR1 to endogenous and synthetic ligands has been explored. In addition, the discovery of different chemo-types as selective murine and/or human TAAR1 ligands has helped in the understanding of the species-specificity preferences. The availability of TAAR1-ligand complexes sheds light on how different ligands bind TAAR1. TAAR5 is considered an olfactory receptor but has specific involvement in some brain functions. In this case, the drug discovery effort has been limited. Here, we review the successful computational efforts developed in the search for novel TAAR1 and TAAR5 ligands. A specific focus on applying structure-based and/or ligand-based methods has been done. We also give a perspective of the experimental data available to guide the future drug design of new ligands, probing species-specificity preferences towards more selective ligands. Hints for applying repositioning approaches are also discussed.

Developing a Semi-Supervised Approach Using a PU-Learning-Based Data Augmentation Strategy for Multitarget Drug Discovery.

Multifactorial diseases demand therapeutics that can modulate multiple targets for enhanced safety and efficacy, yet the clinical approval of multitarget drugs remains rare. The integration of machine learning (ML) and deep learning (DL) in drug discovery has revolutionized virtual screening. This study investigates the synergy between ML/DL methodologies, molecular representations, and data augmentation strategies. Notably, we found that SVM can match or even surpass the performance of state-of-the-art DL methods. However, conventional data augmentation often involves a trade-off between the true positive rate and false positive rate. To address this, we introduce Negative-Augmented PU-bagging (NAPU-bagging) SVM, a novel semi-supervised learning framework. By leveraging ensemble SVM classifiers trained on resampled bags containing positive, negative, and unlabeled data, our approach is capable of managing false positive rates while maintaining high recall rates. We applied this method to the identification of multitarget-directed ligands (MTDLs), where high recall rates are critical for compiling a list of interaction candidate compounds. Case studies demonstrate that NAPU-bagging SVM can identify structurally novel MTDL hits for ALK-EGFR with favorable docking scores and binding modes, as well as pan-agonists for dopamine receptors. The NAPU-bagging SVM methodology should serve as a promising avenue to virtual screening, especially for the discovery of MTDLs.