Berberine (BBR) has seen growing application in ophthalmology, yet the ocular toxicity of BBR and its metabolites remains poorly understood. This study aimed to evaluate the ocular toxicity of BBR and its major metabolite M1 and unravel their underlying mechanisms. Ocular toxicity was evaluated in human corneal epithelial cells and wild-type AB zebrafish. Mechanistic studies utilized fluorescence imaging, biochemical quantitative assays, and qPCR analyses in AB zebrafish and transgenic mitochondrial fluorescent zebrafish (strain Tg(Xla.Eef1a1:mlsEGFP)). Both BBR and M1 induced significant ocular toxicity across models, with BBR showing higher toxicity than M1. Mechanistic analyses revealed their toxicity stemmed from photoreceptor cell damage and Sirtuin 3 (SIRT3) inhibition, triggering a cascade of pathological events: mitochondrial dysfunction, oxidative stress, autophagic dysfunction, apoptosis, and inflammation. This study provides a reference for individualized risk assessment and clinical management of BBR-based therapies and paves the way for developing BBR derivatives with reduced ocular toxicity.

Hand, foot, and mouth disease (HFMD) caused by enterovirus A71 (EV-A71) infection, currently lacks specific preventive and therapeutic interventions. Here, we demonstrated that Pien Tze Huang (PZH) could dose-dependently inhibit EV-A71 replication at the cellular level, resulting in significant reductions in EV-A71 virus protein 1 (VP1) expression and viral yields in Vero and human rhabdomyosarcoma cells. More importantly, we confirmed that PZH could protect mice from EV-A71 infection for the first time, with Ribavirin serving as a positive control. PZH treatment reduced EV-A71 VP1 protein expression, viral yields in infected muscles, and improved muscle pathology. Additionally, we conducted a preliminary mechanism study using quantitative proteomics. The results suggested that the suppression of the PI3K/AKT/mTOR and NF-κB signaling pathways may contribute to the anti-EV-A71 activity of PZH. These findings provide strong evidence supporting the potential therapeutic application of PZH for EV-A71 infection management.

Levornidazole, a nitroimidazole compound, has been linked to hepatotoxic adverse effects in clinical settings. However, the hepatotoxicity of levornidazole and its impurities has not been fully elucidated. This study aimed to predict and evaluate the potential hepatotoxicity of levornidazole, and elucidate the underlying mechanisms of action. Computational models based on support vector machines (SVM) and artificial neural networks (ANN) predicted that levornidazole, ornidazole, and impurity II exhibited hepatotoxic effects. The hepatotoxicity of levornidazole and impurity II was confirmed using a zebrafish toxicity study, with impurity II demonstrating hepatotoxicity at lower doses. Molecular structure analysis revealed that the electronegativity of the side-chain groups and the molecular polarity structure were correlated with the degree of hepatotoxicity. The toxic response was primarily associated with specific structural domains of the molecule, including the 2-methyl-5-nitro-1H-imiddaster-1-yl structure and the substituent groups of 1-chloro and 2(S)-2-methyloxirane. Transcriptome sequencing analysis indicated that levornidazole and impurity II affect multiple metabolic processes in the liver, including glucose, lipid, protein, hormone, and drug metabolism. These findings highlight the potential hepatotoxic risks associated with levomeprazole and its impurities, emphasizing the importance of further investigation and regulatory attention to ensure patient safety.

Drug-induced ocular toxicity is difficult to predict and evaluate, particularly for prostaglandin F2α (PGF2α) analogs used in the management of glaucoma. Traditional global predictive models, which are built with large and various datasets (n = 6187), provide systematic false-negative results for groups of structurally uniform compounds. To address this limitation, we developed special local classification models for PGF2α analogs. A structurally consistent local training dataset (n = 350) was assembled using Murcko scaffold filtering and Tanimoto similarity selection (≥0.5). Binary classification models were built using three different molecular fingerprints in combination with machine learning and two deep learning methods. Although two-dimensional molecular representations do not encode stereochemistry, computational predictions still apply to the shared two-dimensional scaffold of these compounds. These improved models were used to predict the ocular toxicity of latanoprost and its related impurities. The local models (n = 350) provided accurate toxicity predictions for latanoprost, latanoprost acid, 15(S)-latanoprost, and the trans-5,6-latanoprost isomer, whereas all 16 global computational models provided false-negative results. The experimental assessment performed with primary rabbit corneal epithelial cells (pRCECs) and a human corneal epithelial cell line (HCE-T) confirmed the cytotoxic effects, which were in agreement with the predictions made by the models. Among the tested compounds, 15(S)-latanoprost exhibited the highest cytotoxicity (IC50 = 86.22 μM, 95% CI: 82.90–89.56 μM), followed by trans-5,6-latanoprost (IC50 = 106.70 μM, 95% CI: 103.4–110.0 μM) and latanoprost (IC50 = 112.60 μM, 95% CI: 106.7–118.6 μM). At the standard therapeutic dosage (0.005%), no significant toxic response was observed. Virtual molecular docking was employed to explore the mechanism, and all analogs docked favorably into the quinone-binding channel of mitochondrial complex I and the catalytic cleft of SIRT3. These results show how a localized modeling approach is better able to capture structure–toxicity correlations among chemically similar compounds and highlight the critical need for rigorous impurity management in latanoprost products, particularly regarding 15(S)-latanoprost.