Aging is a multifaceted biological phenomenon driven by cellular damage accumulated through chronic oxidative stress. While traditional antioxidant systems mitigate this damage, their stoichiometric, scavenging-based reduction mechanisms present inherent shortcomings, including potential pro-oxidant effects and disruption of essential physiological redox signaling. To address these limitations, this paper introduces a cysteine-centered, dual-mechanism framework that integrates traditional thiol networks with the advanced super-sulfide axis to manage sulfur-dependent redox control. This model operates via two interdependent pathways: first, catalytic networks, including the glutathione and thioredoxin systems, which actively remove peroxides, repair oxidized proteins, support detoxification, and buffer thiol redox potential; second, catenated sulfur species—such as hydropersulfides, polysulfides, and protein persulfides—which trap radicals, transfer sulfur, and drive reversible protein persulfidation. Protein persulfidation serves as the primary molecular mechanism, regulating core aging processes by tuning protein functions, stimulating NRF2-driven antioxidant responses, preserving mitochondrial integrity, and preventing ferroptosis. Crucially, the advanced super-sulfide axis operates on a strict hormetic curve, where excessive concentrations disrupt cellular respiration and induce toxicity, highlighting the necessity of maintaining a tight physiological window. Preclinical in vitro and animal data demonstrate that super-sulfides offer an efficient, specific, and sustained protective profile compared to the transient effects of classical antioxidants. From a public health perspective, dietary interventions utilizing local, affordable sources of antioxidants and super-sulfide precursors represent a culturally practical strategy for promoting healthy aging. In conclusion, focusing on the sulfur metabolism super-sulfide axis represents a paradigm shift from simple radical scavenging to the intelligent regulation of cellular defense mechanisms. However, further population-specific studies, clinical trials, and the development of targeted biomarkers are mandatory to validate these preclinical findings and exploit this therapeutic potential for the global aging population.
Background Obesity affects one in eight people worldwide and its metabolic consequences drive neuroinflammation and cognitive decline. Although GLP-1 receptor agonists are widely prescribed for obesity and type 2 diabetes, recent phase 3 studies onneurodegenerative disease have not reproduced their preclinical promise. Here, we examined whether delayed exendin-4 treatment could improve peripheral metabolic stress and preserve neuronal architecture. Methods C57BL/6 J mice receiving a high-fat high-cholesterol diet for 25 weeks were administered exendin-4 during the last 5weeks. Results Exendin-4 was linked to metabolic and redox-inflammatory changes without proportionate weight loss, reinforced hepatic antioxidant transcription and a trend toward reduced collagen deposition. It induced an adipose thermogenic program, alongside a systemic shift toward IGF-1 signaling. Cortical transcriptomic profiling revealed remodeling of G protein-coupled receptor signaling and glycosphingolipid and ceramide catabolic programs. In vitro, conditioned media from high-fat high-cholesterol mimicking microglia disrupted growth-cone F-actin structure and suppressed actin regulation, and exendin-4 attenuated inflammatory factors and was associated with preserved neurite complexity in both Neuro-2A cells and primary cortical neurons. In vivo, Map2 and synaptic protein expression improved and along with spatial working memory. The cytoskeletal signature was replicated in leptin-deficient Ob/Ob mice. Conclusion These findings identify neuronal cytoskeletal integrity as a key structural correlate of systemic metabolic improvement. This supports central synaptic preservation with peripheral recovery after established disease without proportionate weight loss and suggests metabolic stratification as a potential enrichment strategy for neuroprotection trials. Limitations Mouse and cell models do not establish clinical efficacy in humans.
Background Existing network pharmacology workflows for traditional Chinese medicine (TCM) remain fragmented across databases and software tools, with limited support for systematic target prioritization. This study aimed to develop TCMDATA, which is an integrated R package for reproducible, target-oriented TCM network pharmacology analysis. Methods TCMDATA is an integrated R package providing a scriptable, target-oriented workflow for TCM network pharmacology. It integrates herb–compound–target associations from multiple public resources, comprising 1808 herbs, 2068 target-linked compounds, and 15,889 human protein targets after data cleaning and gene-symbol standardization. For target prioritization, TCMDATA natively implements 19 PPI centrality metrics, four community detection algorithms, and a configurable network perturbation framework for assessing target importance. A multi-algorithm machine learning module supports consensus candidate selection, while an AI agent module enables task routing and customizable skills for automated analysis and interpretation. Results In a case study of Ganoderma lucidum in sepsis, TCMDATA prioritized four candidate targets, including MAPK14, CASP1, ARG1, and PPARG. External disease databases, literature evidence, and molecular docking further supported the relevance of these targets to sepsis, illustrating the utility of TCMDATA as a reproducible analytical resource for TCM network pharmacology. Conclusion TCMDATA provides an integrated and reproducible R framework for target-oriented TCM network pharmacology, supporting transparent data integration, network analysis, and machine learning-based candidate prioritization. Interpretation of its outputs should consider the coverage of available public databases, and further experimental validation will help establish the mechanistic relevance of prioritized targets. TCMDATA is now freely available at https://github.com/YuLab-SMU/TCMDATA. Public transcriptomic datasets used in the secondary analyses are available from NCBI GEO under accession numbers GSE142025, GSE157103, GSE154918, and GSE185263.
