Drug Discovery Applications Research Articles

STORMM: Structure and topology replica molecular mechanics for chemical simulations.

The Structure and TOpology Replica Molecular Mechanics (STORMM) code is a next-generation molecular simulation engine and associated libraries optimized for performance on fast, vectorized central processor units and graphics processing units (GPUs) with independent memory and tens of thousands of threads. STORMM is built to run thousands of independent molecular mechanical calculations on a single GPU with novel implementations that tune numerical precision, mathematical operations, and scarce on-chip memory resources to optimize throughput. The libraries are built around accessible classes with detailed documentation, supporting fine-grained parallelism and algorithm development as well as copying or swapping groups of systems on and off of the GPU. A primary intention of the STORMM libraries is to provide developers of atomic simulation methods with access to a high-performance molecular mechanics engine with extensive facilities to prototype and develop bespoke tools aimed toward drug discovery applications. In its present state, STORMM delivers molecular dynamics simulations of small molecules and small proteins in implicit solvent with tens to hundreds of times the throughput of conventional codes. The engineering paradigm transforms two of the most memory bandwidth-intensive aspects of condensed-phase dynamics, particle-mesh mapping, and valence interactions, into compute-bound problems for several times the scalability of existing programs. Numerical methods for compressing and streamlining the information present in stored coordinates and lookup tables are also presented, delivering improved accuracy over methods implemented in other molecular dynamics engines. The open-source code is released under the MIT license.

Untargeted metabolomic profiling of Prinsepia utilis Royle leaves by Ultra-Performance Liquid Chromatography-Quadrupole-Time of Flight-Mass Spectrometry.

Prinsepia utilis Royle, native to the Himalayas, is esteemed in Chinese and Indian folk medicine for its diverse medicinal benefits, targeting arthritis, pain relief, bone disorders, and joint discomfort. This study examined the 25% aqueous methanol extract of P.utilis leaves using UPLC-Q-TOF-MS/MS, identifying 78 metabolites, 76 of which were reported for the first time in P.utilis. These included 64 phenolics represented by 56 flavonoids, 5 phenolic acids, 3 phenolic glycosides, 4 terpenoids, 2 lignan glycosides, and 8 other compounds, expanding the knowledge of its chemical composition. These findings lay a foundation for further research, providing insights into potential bioactive compounds and opening avenues for applications in natural product drug discovery, traditional medicine, and nutraceutical development, leveraging the plant's established traditional uses.

Overcoming Challenges in O-Nitration: Selective Alcohol Nitration Deploying N,6-Dinitrosaccharin and Lewis Acid Catalysis.

Nitrate esters hold pivotal roles in pharmaceuticals, energetic materials, and atmospheric processes, motivating the development of efficient synthesis routes. Here, we present a novel catalytic method for the synthesis of nitrates via the direct O-nitration of alcohols, addressing limitations of current traditional methods. Leveraging bench-stable and recoverable N,6-dinitrosaccharin reagent, our catalytic strategy employs magnesium triflate to achieve mild and selective O-nitration of alcohols, offering broad substrate scope and unprecedentedly large functional group tolerance (e.g. alkenes, alkynes, carbonyls). DFT mechanistic studies reveal a dual role of the magnesium catalyst in the activation of both the nitrating reagent and the alcohol substrate. They also unveil a barrierless proton transfer upon formation of a widely-accepted - yet elusive in solution - nitrooxonium ion intermediate. Overall, our work contributes to the development of mild, selective, and sustainable approaches to nitrates synthesis, with potential applications in drug discovery, materials science, and environmental chemistry.

Overcoming Challenges in O‐Nitration: Selective Alcohol Nitration Deploying N,6‐Dinitrosaccharin and Lewis Acid Catalysis

Nitrate esters hold pivotal roles in pharmaceuticals, energetic materials, and atmospheric processes, motivating the development of efficient synthesis routes. Here, we present a novel catalytic method for the synthesis of nitrates via the direct O‐nitration of alcohols, addressing limitations of current traditional methods. Leveraging bench‐stable and recoverable N,6‐dinitrosaccharin reagent, our catalytic strategy employs magnesium triflate to achieve mild and selective O‐nitration of alcohols, offering broad substrate scope and unprecedented large functional group tolerance (e.g. alkenes, alkynes, carbonyls). DFT mechanistic studies reveal a dual role of the magnesium catalyst in the activation of both the nitrating reagent and the alcohol substrate. They also unveil a barrierless proton transfer upon formation of a widely‐accepted — yet elusive in solution — nitrooxonium ion intermediate. Overall, our work contributes to the development of mild, selective, and sustainable approaches to nitrates synthesis, with potential applications in drug discovery, materials science, and environmental chemistry.

Predicting blood–brain barrier permeability of molecules with a large language model and machine learning

Predicting the blood–brain barrier (BBB) permeability of small-molecule compounds using a novel artificial intelligence platform is necessary for drug discovery. Machine learning and a large language model on artificial intelligence (AI) tools improve the accuracy and shorten the time for new drug development. The primary goal of this research is to develop artificial intelligence (AI) computing models and novel deep learning architectures capable of predicting whether molecules can permeate the human blood–brain barrier (BBB). The in silico (computational) and in vitro (experimental) results were validated by the Natural Products Research Laboratories (NPRL) at China Medical University Hospital (CMUH). The transformer-based MegaMolBART was used as the simplified molecular input line entry system (SMILES) encoder with an XGBoost classifier as an in silico method to check if a molecule could cross through the BBB. We used Morgan or Circular fingerprints to apply the Morgan algorithm to a set of atomic invariants as a baseline encoder also with an XGBoost classifier to compare the results. BBB permeability was assessed in vitro using three-dimensional (3D) human BBB spheroids (human brain microvascular endothelial cells, brain vascular pericytes, and astrocytes). Using multiple BBB databases, the results of the final in silico transformer and XGBoost model achieved an area under the receiver operating characteristic curve of 0.88 on the held-out test dataset. Temozolomide (TMZ) and 21 randomly selected BBB permeable compounds (Pred scores = 1, indicating BBB-permeable) from the NPRL penetrated human BBB spheroid cells. No evidence suggests that ferulic acid or five BBB-impermeable compounds (Pred scores < 1.29423E−05, which designate compounds that pass through the human BBB) can pass through the spheroid cells of the BBB. Our validation of in vitro experiments indicated that the in silico prediction of small-molecule permeation in the BBB model is accurate. Transformer-based models like MegaMolBART, leveraging the SMILES representations of molecules, show great promise for applications in new drug discovery. These models have the potential to accelerate the development of novel targeted treatments for disorders of the central nervous system.

Toehold region triggered CRISPR/Cas12a trans-cleavage for detection of uracil-DNA glycosylase activity.

DNA glycosylases are a group of enzymes that play a crucial role in the DNA repair process by recognizing and removing damaged or incorrect bases from DNA molecules, which maintains the integrity of the genetic information. The abnormal expression of uracil-DNA glycosylase (UDG), one of significant DNA glycosylases in the base-excision repair pathway, is linked to numerous diseases. Here, we proposed a simple UDG activity detection method based on toehold region triggered CRISPR/Cas12a trans-cleavage. The toehold region on hairpin DNA probe (HP) produced by UDG could induce the trans-cleavage of ssDNA with fluorophore and quencher, generating an obvious fluorescence signal. This protospacer adjacent motif (PAM)-free approach achieves remarkable sensitivity and specificity in detecting UDG, with a detection limit as low as 0.000368 UmL-1. Moreover, this method is able to screen inhibitors and measure UDG in complex biological samples. These advantages render it highly promising for applications in clinical diagnosis and drug discovery.

Single-Step 3D Bioprinting of Alginate-Collagen Type I Hydrogel Fiber Rings to Promote Angiogenic Network Formation.

In the advent of tissue engineering and regenerative medicine, the demand for innovative approaches to biofabricate complex vascular structures is increasing. We describe a single-step 3D bioprinting method leveraging Aspect Biosystems RX1 technology, which integrates the crosslinking step at a flow-focusing junction, to biofabricate immortalized adult rat brain endothelial cell (SV-ARBEC)-encapsulated alginate-collagen type I hydrogel rings. This single-step biofabrication process involves the strategic layer-by-layer assembly of hydrogel rings, encapsulating SV-ARBECs in a spatially controlled manner while optimizing access to media and nutrients. The spatial arrangement of the SV-ARBECs within the rings promotes spontaneous angiogenic network formation and the constrained deposition of cells within the hydrogel matrix facilitates tissue-like organized vascular-like network development. This approach provides a platform that can be adapted to many different endothelial cell types and leveraged to better understand the mechanisms driving angiogenesis and vascular-network formation in 3D bioprinted constructs supporting the development of more complex tissue and disease models for advancing drug discovery, tissue engineering, and regenerative medicine applications.

Deep representation learning of chemical-induced transcriptional profile for phenotype-based drug discovery

Artificial intelligence transforms drug discovery, with phenotype-based approaches emerging as a promising alternative to target-based methods, overcoming limitations like lack of well-defined targets. While chemical-induced transcriptional profiles offer a comprehensive view of drug mechanisms, inherent noise often obscures the true signal, hindering their potential for meaningful insights. Here, we highlight the development of TranSiGen, a deep generative model employing self-supervised representation learning. TranSiGen analyzes basal cell gene expression and molecular structures to reconstruct chemical-induced transcriptional profiles with high accuracy. By capturing both cellular and compound information, TranSiGen-derived representations demonstrate efficacy in diverse downstream tasks like ligand-based virtual screening, drug response prediction, and phenotype-based drug repurposing. Notably, in vitro validation of TranSiGen’s application in pancreatic cancer drug discovery highlights its potential for identifying effective compounds. We envisage that integrating TranSiGen into the drug discovery and mechanism research holds significant promise for advancing biomedicine.

MDB-68. DISCOVERING DRUG SENSITIVITIES IN MEDULLOBLASTOMA USING PATIENT- AND PDX-DERIVED TUMOROIDS

Abstract BACKGROUND Medulloblastoma (MB) is one of the most common malignant pediatric brain tumor types and is subdivided into four major types: WNT, SHH, Group 3 and Group 4. Treatment, including resection, cranio-spinal radiotherapy, and chemotherapy, often leads to undesirable long-term side effects. MB prognosis depends on type, subtype, and genetic alterations, reflecting the high level of heterogeneity observed in these tumors and indicating that specific patient populations with particularly poor prognosis require alternative treatment approaches. Therefore, preclinical models are essential for recapitulating MB heterogeneity and developing subtype-specific treatments. METHODS Long-term in vitro MB tumoroid models were developed from patient and PDX tumors and molecularly characterized using DNA methylation and transcriptomic analyses. High-throughput single drug screening (225 compounds) and combination screening were performed to identify (subtype-specific) drug sensitivities. RESULTS Fifteen tumoroid lines were established from TP53-mutated SHH (n=6), Group 3 (G3) MYC-amplified (n=4) and non-amplified (n=5) patient (n=6) and PDX (n=9) tumors. DNA methylation analyses confirmed the maintenance of molecular identity in these tumoroids after long-term culture (&amp;gt;5 passages). Transcriptome analyses reveal a more proliferative, progenitor-like phenotype in these tumoroids compared to patient/PDX tumors. Hight-throughput drug screens identified both common and type-specific drug sensitivities within MB. Notably, drugs targeting CDK7, CDK12, CHECK1, BCL2L, and HDAC demonstrated efficacy in both G3 and TP53-mutated SHH-MB, while TOP1/2 and MDM2 inhibitors were specific for G3-MB. Currently, these drugs are being further investigated to select the most effective synergistic drug combinations for in vivo validation in PDX models. CONCLUSIONS Thus far, we have established already fifteen MB tumoroid models representing SHH-MB and G3-MB and identified several drug targets, including cell cycle kinases and HDACs. Tumoroids closely resemble patient tumors, based on methylome and transcriptome, enhancing their relevance for drug discovery applications in a high-throughput fashion and for prioritizing new treatment strategies for MB patients in clinical trials.

Regiodivergent and Enantioselective Synthesis of Cyclic Sulfones via Ligand-Controlled Nickel-Catalyzed Hydroalkylation.

Cyclic sulfones have demonstrated important applications in drug discovery. However, the catalytic and enantioselective synthesis of chiral cyclic sulfones remains challenging. Herein, we develop nickel-catalyzed regiodivergent and enantioselective hydroalkylation of sulfolenes to streamline the synthesis of chiral alkyl cyclic sulfones. The method has broad scope and high functional group tolerance. The regioselectivity can be controlled by ligands only. A neutral PYROX ligand favors C3-alkylation whereas an anionic BOX ligand favors C2-alkylation. This control is kinetic in origin as the C2-bound Ni intermediates are always thermodynamically more stable. Reactivity study of a wide range of relevant Ni intermediates reveal a NiI/NiIII catalytic cycle with a NiII-H species as the resting state. The regio- and enantio-determining step is the insertion of this NiII-H species into 2-sulfolene. This work provides an efficient catalytic method for the synthesis of an important class of organic compounds and enhances the mechanistic understanding of Ni-catalyzed stereoselective hydroalkylation.

The Goldilocks paradigm: comparing classical machine learning, large language models, and few-shot learning for drug discovery applications

Recent advances in machine learning (ML) have led to newer model architectures including transformers (large language models, LLMs) showing state of the art results in text generation and image analysis as well as few-shot learning (FSLC) models which offer predictive power with extremely small datasets. These new architectures may offer promise, yet the ‘no-free lunch’ theorem suggests that no single model algorithm can outperform at all possible tasks. Here, we explore the capabilities of classical (SVR), FSLC, and transformer models (MolBART) over a range of dataset tasks and show a ‘goldilocks zone’ for each model type, in which dataset size and feature distribution (i.e. dataset “diversity”) determines the optimal algorithm strategy. When datasets are small ( < 50 molecules), FSLC tend to outperform both classical ML and transformers. When datasets are small-to-medium sized (50-240 molecules) and diverse, transformers outperform both classical models and few-shot learning. Finally, when datasets are of larger and of sufficient size, classical models then perform the best, suggesting that the optimal model to choose likely depends on the dataset available, its size and diversity. These findings may help to answer the perennial question of which ML algorithm is to be used when faced with a new dataset.

Stereoselective Synthesis of Azobenzene-Based Glycomacrocycles

AbstractCarbohydrate-based macrocyclic compounds are of particular interest because of their multifunctionality, their unique structural and physicochemical properties as well as their potential applications in chemistry, biology, and drug discovery. Introducing a molecular photoswitch into the skeleton of glycomacrocycles makes possible the reversible control of properties of the resulting photoswitchable glycomacrocycles by light illumination. Therefore, development of stereoselective synthesis of this class of glycomacrocycles is of great interest. Two new azobenzene-based glycomacrocycles have been synthesized through an intramolecular glycosylation approach. Excellent 1,2-cis stereoselectivity has been achieved for the mannosylation.

QupKake: Integrating Machine Learning and Quantum Chemistry for Micro-pKa Predictions.

Accurate prediction of micro-pKa values is crucial for understanding and modulating the acidity and basicity of organic molecules, with applications in drug discovery, materials science, and environmental chemistry. This work introduces QupKake, a novel method that combines graph neural network models with semiempirical quantum mechanical (QM) features to achieve exceptional accuracy and generalization in micro-pKa prediction. QupKake outperforms state-of-the-art models on a variety of benchmark data sets, with root-mean-square errors between 0.5 and 0.8 pKa units on five external test sets. Feature importance analysis reveals the crucial role of QM features in both the reaction site enumeration and micro-pKa prediction models. QupKake represents a significant advancement in micro-pKa prediction, offering a powerful tool for various applications in chemistry and beyond.

Benchmarking compound activity prediction for real-world drug discovery applications

Identifying active compounds for target proteins is fundamental in early drug discovery. Recently, data-driven computational methods have demonstrated promising potential in predicting compound activities. However, there lacks a well-designed benchmark to comprehensively evaluate these methods from a practical perspective. To fill this gap, we propose a Compound Activity benchmark for Real-world Applications (CARA). Through carefully distinguishing assay types, designing train-test splitting schemes and selecting evaluation metrics, CARA can consider the biased distribution of current real-world compound activity data and avoid overestimation of model performances. We observed that although current models can make successful predictions for certain proportions of assays, their performances varied across different assays. In addition, evaluation of several few-shot training strategies demonstrated different performances related to task types. Overall, we provide a high-quality dataset for developing and evaluating compound activity prediction models, and the analyses in this work may inspire better applications of data-driven models in drug discovery.

Editorial on the Special Issue “Natural Compounds Applications in Drug Discovery and Development”

Nature is an amazing source of natural bioactive compounds derived from numerous species of plants, marine bacteria, and fungi [...]

Optimizing Nodal, Wnt and BMP signaling pathways for robust and efficient differentiation of human induced pluripotent stem cells to intermediate mesoderm cells.

Several differentiation protocols have enabled the generation of intermediate mesoderm (IM)-derived cells from human pluripotent stem cells (hPSC). However, the substantial variability between existing protocols for generating IM cells compromises their efficiency, reproducibility, and overall success, potentially hindering the utility of urogenital system organoids. Here, we examined the role of high levels of Nodal signaling and BMP activity, as well as WNT signaling in the specification of IM cells derived from a UCSD167i-99-1 human induced pluripotent stem cells (hiPSC) line. We demonstrate that precise modulation of WNT and BMP signaling significantly enhances IM differentiation efficiency. Treatment of hPSC with 3μM CHIR99021 induced TBXT+/MIXL1+ mesoderm progenitor (MP) cells after 48h of differentiation. Further treatment with a combination of 3μM CHIR99021 and 4ng/mL BMP4 resulted in the generation of OSR1+/GATA3+/PAX2+ IM cells within a subsequent 48h period. Molecular characterization of differentiated cells was confirmed through immunofluorescence staining and RT-qPCR. Hence, this study establishes a consistent and reproducible protocol for differentiating hiPSC into IM cells that faithfully recapitulates the molecular signatures of IM development. This protocol holds promise for improving the success of protocols designed to generate urogenital system organoids in vitro, with potential applications in regenerative medicine, drug discovery, and disease modeling.

Organocatalytic skeletal reorganization for enantioselective synthesis of S-stereogenic sulfinamides

The enantioselective synthesis of S-stereogenic sulfinamides has garnered considerable attention due to their structural and physicochemical properties. However, catalytic asymmetric synthesis of sulfinamides still remains daunting challenges, impeding their broad application in drug discovery and development. Here, we present an approach for the synthesis of S-stereogenic sulfinamides through peptide-mimic phosphonium salt-catalyzed asymmetric skeletal reorganization of simple prochiral and/or racemic sulfoximines. This methodology allows for the facile access to a diverse array of substituted sulfinamides with excellent enantioselectivities, accommodating various substituent patterns through desymmetrization or parallel kinetic resolution process. Mechanistic experiments, coupled with density functional theory calculations, clarify a stepwise pathway involving ring-opening and ring-closing processes, with the ring-opening step identified as crucial for achieving stereoselective control. Given the prevalence of S-stereogenic centers in pharmaceuticals, we anticipate that this protocol will enhance the efficient and precise synthesis of relevant chiral molecules and their analogs, thereby contributing to advancements in drug discovery.

Differentiation of Enteric Nervous System Lineages from Human Pluripotent Stem Cells.

The human enteric nervous system, ENS, is a large network of glial and neuronal cell types with remarkable neurotransmitter diversity. The ENS controls bowel motility, enzyme secretion, and nutrient absorption and interacts with the immune system and the gut microbiome. Consequently, developmental and acquired defects of the ENS are responsible for many human diseases and may contribute to symptoms of Parkinson's disease. Limitations in animal model systems and access to primary tissue pose significant experimental challenges in studies of the human ENS. Here, a detailed protocol is presented for effective in vitro derivation of the ENS lineages from human pluripotent stem cells, hPSC, using defined culture conditions. Our protocol begins with directed differentiation of hPSCs to enteric neural crest cells within 15 days and yields diverse subtypes of functional enteric neurons within 30 days. This platform provides a scalable resource for developmental studies, disease modeling, drug discovery, and regenerative applications.

Structural Studies of a Complex of a CAG/CTG Repeat Sequence-Specific Binding Molecule and A-A-Mismatch-Containing DNA.

Triplet repeat diseases are caused by the abnormal elongation of repeated sequences comprising three bases. In particular, the elongation of CAG/CTG repeat sequences is thought to result in conditions such as Huntington's disease and myotonic dystrophy type 1. Although the causes of these diseases are known, fundamental treatments have not been established, and specific drugs are expected to be developed. Pyrrole imidazole polyamide (PIP) is a class of molecules that binds to the minor groove of the DNA duplex in a sequence-specific manner; because of this property, it shows promise in drug discovery applications. Earlier, it was reported that PIP designed to bind CAG/CTG repeat sequences suppresses the genes that cause triplet repeat diseases. In this study, we performed an X-ray crystal structure analysis of a complex of double-stranded DNA containing A-A mismatched base pairs and a cyclic-PIP that binds specifically to CAG/CTG sequences. Furthermore, the validity and characteristics of this structure were analyzed using in silico molecular modeling, ab initio energy calculations, gel electrophoresis, and surface plasmon resonance. With our direct observation using atomic force microscopy and DNA origami, we revealed that the PIP caused structural changes in the DNA strands carrying the expanded CAG/CTG repeat. Overall, our study provides new insight into PIP from a structural perspective.

Synthesis, Antibacterial Effects, and Toxicity of Licochalcone C.

Drug-resistant bacteria constitute a big barrier against current pharmacotherapy. Efforts are urgent to discover antibacterial drugs with novel chemical and biological features. Our work aimed at the synthesis, evaluation of antibacterial effects, and toxicity of licochalcone C (LCC), a naturally occurring chalcone. The synthetic route included six steps, affording a 10% overall yield. LCC showed effects against Gram-positive bacteria (MIC = 6.2-50.0 µg/mL), Mycobacterium species (MIC = 36.2-125 µg/mL), and Helicobacter pylori (MIC = 25 µg/mL). LCC inhibited the biofilm formation of MSSA and MRSA, demonstrating MBIC50 values of 6.25 μg/mL for both strains. The investigations by fluorescence microscopy, using PI and SYTO9 as fluorophores, indicated that LCC was able to disrupt the S. aureus membrane, similarly to nisin. Systemic toxicity assays using Galleria mellonella larvae showed that LCC was not lethal at 100 µg/mL after 80 h treatment. These data suggest new uses for LCC as a compound with potential applications in antibacterial drug discovery and medical device coating.