Expert Opinion on Drug Discovery最新文献

Introduction: Human norovirus is the leading global cause of viral acute gastroenteritis, with an estimated ~685 million cases and ~200,000 deaths annually. No licensed antivirals or vaccines are currently available. Despite historical limitations in robust in vitro models, structure and ligand-based computational approaches - supported by protease and polymerase crystal structures - have identified multiple chemotypes as potential antivirals.

Areas covered: This review provides an overview of all studies reported to date, indexed in public databases, in which computer-aided drug discovery (CADD) techniques have been employed. The authors report the computational methodologies used, the chemical structures of the identified compounds, and, if available, their biological activities. Where in silico results lack experimental validation, the authors highlight limitations and propose minimal validation assays.

Expert opinion: The absence of 3D structures for most viral proteins has limited the identification of novel chemotypes through CADD approaches. Furthermore, the lack of biological validation after in silico studies may slow down progress in this field, as researchers might focus on compounds that seem promising only at the computational level. Emerging systems such as human intestinal enteroids, together with AI/ML augmented CADD, can accelerate optimization and triage of non-nucleoside and covalent protease inhibitors.

Introduction: Despite remarkable advances in computational methods, pre-clinical drug discovery continues to grapple with rising timelines and costs. Software, data, and automation are more powerful than ever, and increasingly these technologies are being embraced. However, there is still work to be done to translate this potential into meaningful reductions in cost and time.

Areas covered: This perspective discusses the growth in drug discovery capability, exploring modern data infrastructure including cloud-native platforms, active learning, and laboratory automation. It covers emerging technologies such as LLM-based orchestration and emulation. Implementation examples illustrate successes and challenges.

Expert opinion: AI presents an opportunity to envisage a new approach to drug discovery, but cultural and technological changes are required. The exponential growth in computational drug discovery tools requires solutions that enable researchers to access scalable and robust capabilities more easily. Data generation is usually the slowest and most expensive part of the design cycle; we advocate for a rigorous application of statistical methods focussing on learning efficiency from data over absolute predictive accuracy of models. Automation also plays a critical role in enabling rapid, high-quality data generation. Focussing on modular interoperable automated units with more attractive economics will drive much wider adoption.

Introduction: Aromatase inhibitors (AIs) are a group of drugs used to cease the production of estrogens that have been used to treat mainly post-menopausal women with both early and advanced estrogen receptor positive (ER+) breast cancer for more than 40 years. They are the first-line therapy for hormone-dependent cancers either alone or in combination with other therapies. This review aims to demonstrate the relevance of research on steroidal AIs.

Areas covered: This work focuses on steroidal AIs and aims to provide a comprehensive explanation of the most potent ones (IC₅₀ < 10 µM), the respective structure-activity relationships (SAR), and the particular biological mechanisms of action, published during the last 20 years, from 2005 until June 2025. The literature search was made on the Web of Science database, using the topic 'steroidal aromatase inhibitors.'

Expert opinion: Potent hit steroidal AIs have been discovered. However, further studies are needed, including tests on more complex models, namely using 3D cell lines and organoids as well as animal studies, with the aim of progressing on the drug discovery process, ultimately to discover new potent steroidal AIs with less side effects, especially those related to bone and cardiovascular health and capable to overcome endocrine resistance.

Introduction: Fluorinated amino acids (FAAs) are at a focal point of two key current strategic areas within drug discovery being both important for the design/development of new small molecule drugs as well as having the potential to be exploited in the rapidly expanding area of peptide-based therapeutics. Their exquisite modularity, synthetic versatility, and extensive commercial availability coupled with the robust understanding of the roles that fluorine can play, and which improves their potency and physicochemical properties of drug candidates, have enabled FAAs to be widely used in numerous drug discovery programs.

Areas covered: This review provides an overview of the use of fluorinated amino acids in small drug discovery focusing initially on their impact across a diverse range of therapeutic areas before evaluating methods to access them synthetically. Furthermore, in-depth analyses of programs focusing on the design of thrombin, γ-secretase, and FXII inhibitors demonstrate how the strategic introduction of a specific fluorinated amino acid within a molecule can significantly favorably modulate the efficacy and/or the physicochemical properties of the lead through enhancing the electronic/steric interactions with the desired biological target.

Expert opinion: While significant advances have been made enabling access to a broad range of fluorinated amino acids and their derivatives, there are several active ongoing research areas in this space. These most notably include developing methods for the synthesis of more-constrained fluorinated bicyclic amino acid derivatives potentially as bioisosteric replacements of aromatic moieties to increase the 3D-dimensionality of a compound while retaining both conformational rigidity and defined orientation of the functional vectors.

Introduction: Radiation-exposed victims are often subjected to additional traumas such as wounds, burns, hemorrhages, or infections, commonly referred to as radiation combined injury (RCI). Though significant advances have been made over the last three decades toward the development of effective drugs for RCI, no specific agent for the syndrome has yet been approved by the Food and Drug Administration (FDA).

Areas covered: This article covers the development of drugs to treat RCI with critical evaluation of both small (mouse, rats) and large (canines and swine) animal models. These models of RCI have been analyzed for strengths and weaknesses relative to drug development. The major categories of medicinals of interest include new classes of a) anti-radiation agents (prophylactics/mitigators/therapeutics), b) tissue-reparative recombinants (growth factors/cytokines), c) blood products (artificial blood cells, stem cells), and d) new generation(s) of broad-spectrum antibiotics (ciprofloxacin). This review is based on the PubMed search of literature covering the period up to October 2025.

Expert opinion: Several animal models are currently being developed to study RCI and drugs for its treatment. These animal models are important for regulatory approval of RCI drugs designed to enhance survival outcomes.

Introduction: Traditional drug discovery is hampered by high costs, long timelines, and low success rates due to inefficient screening and inadequate model systems. The convergence of artificial intelligence (AI) and functional protein micropatterning offers a novel paradigm to address these limitations by accelerating candidate identification and enhancing physiological relevance.

Areas covered: The purpose of this critical perspective is to provide the reader with the author's expert opinion on the convergence of AI and micropatterning, synthesizing current evidence and discussing future opportunities and limitations. The Web-of-Science and PubMed databases were used to collate information. Within this article, coverage is given to the recent advances in combining machine learning and deep learning for efficient virtual screening, molecular design, and structural prediction with state-of-the-art protein micropatterning techniques that generate standardized, biomimetic assay platforms.

Expert opinion: The integration of automated imaging and AI-driven data analysis enables high-throughput, information-rich experimental workflows. Persistent challenges include explainable AI requirements, data quality, and evolving regulatory frameworks supporting non-animal models. These AI-enabled functional micropatterning platforms offer significant benefits for drug discovery. Over the next decade, advances in explainable AI and workflow automation will be essential for widespread adoption, regulatory acceptance, and the realization of closed-loop systems (AI-driven experimental iteration) that reshape pharmaceutical research by enabling close collaboration between scientists and intelligent technologies.

Introduction: Fluoroquinolones (FQs) are a class of antibiotics effective against both Gram-positive and Gram-negative bacteria owing to their ability to target DNA gyrase and topoisomerase IV. The growth of bacterial resistance to antimicrobials leads to greater attention to comparative evaluation of biological activities of pharmaceutical preparations. In this regard, a quantitative structure-activity relationship (QSAR) analysis is used for predicting FQ activity. QSAR serves as an effective tool for predicting antibacterial, antiviral, anti-cancer, genotoxic activity, etc. Moreover, the QSAR approach provides possibilities to estimate FQ contribution to the environment and ecosystems.

Areas covered: This review summarizes more than 100 publications on QSAR and machine learning studies of various medicinal, chemical, and biological characteristics of FQs.

Expert opinion: The authors expect the appearance of novel FQs with improved bactericidal activity. This will be done with in silico assistance of computer-aided drug design and artificial intelligence techniques in drug development. Future QSAR models will be effectively applied to three aspects of FQ activity: 1) detection of FQs with the use of immunoassays; 2) high photo- and biodegradability of antibiotics in the environment; 3) physicochemical activity, in particular photochemical production of singlet oxygen and free radical species.

Background: Skin aging is linked to the overactivity of matrix metalloproteinases (MMPs) and elastase, making their inhibition a promising approach for antiaging. This study aimed to discover novel antiaging peptides from Chlorella proteins using high-throughput virtual screening.

Methods: Batch molecular docking protocol with a custom Python script for 3D peptide structure modeling and AutoDock Vina was applied to predict inhibitory peptides on MMPs and elastase from 1,965 peptides theoretically resistant to gastrointestinal digestion. The top candidates were synthesized for activity assay, and MD simulation illustrated the binding mechanism of potent peptides.

Results: Seventeen peptides with a binding energy < -7.0 kcal/mol showed IC₅₀ ≤ 150 μM. Peptide DGSY acted high potency against MMP-1 (IC₅₀ = 32.6 μM), and HDISHW inhibited MMP-9 and elastase at the lowest IC₅₀ (20.1, 16.5 μM). GAASF inhibited all three enzymes (IC₅₀ = 54.0, 41.9, 62.5 μM). MD simulations confirmed the stability of these peptide-protein complexes, which coincided with the in vitro activity well.

Conclusion: The virtual strategy efficiently identified multifunctional antiaging peptides and could accelerate the discovery of bioactive peptides for cosmetic and therapeutic use. Additionally, its efficiency makes it useful for building high-quality training sets in deep learning models for bioactive structure discovery.

Introduction: The recent surge in cholera outbreaks worldwide, partly driven by climate change, highlights its potential as a significant public health threat. The absence of definitive treatments underscores the urgent need for developing effective targeted therapeutics. The cholera holotoxin comprises a catalytically active A1 subunit, which mediates ADP-ribosylation to induce secretory diarrhea, and a pentameric B subunit responsible for toxin-host cell attachment via GM1 receptors. A1 activation requires binding to human ADP-ribosylation factor 6 (ARF6). Although the inhibition of B-pentamer - GM1 binding has been extensively investigated, several structural and pharmacokinetic challenges remain.

Areas covered: This article is based on a keyword-based literature survey across relevant research repositories, covering studies published up to 2025. It summarizes the structure- and ligand-based molecular modeling approaches employed for identifying inhibitors targeting toxin-host binding, including GM1 mimetics, glycomimetics, and natural compounds. Alternative avenues of toxin inhibition, including occlusion of the B-pentamer pore, A1 catalytic site, and the A1-ARF6 interface to disrupt toxin assembly, ADP-ribosylation, and A1 activation, respectively, are also discussed.

Expert opinion: Targeting the B-pentamer pore, A1 active site, or A1-ARF6 interface holds significant therapeutic potential against cholera-induced dehydration and hypovolaemic shock. These underexplored yet promising druggable targets warrant further investigation for developing effective, targeted therapies.