환경 미세먼지의 피부-면역-뇌 축 교란: 다중오믹스와 AI 기반 접근

Environmental Fine Dust and the Skin–Immune–Brain Axis: Mechanistic Insights and AI-Driven Multi-Omics Approaches

Background: Consisting of sulfates, nitrates, metals, volatile organic compounds, and polycyclic aromatic hydrocarbons, particulate matter (PM) is a major component of air pollution. Chronic exposure to PM has traditionally been associated with respiratory and cardiovascular disorders; however, increasing evidence suggests that its detrimental effects extend to the skin, immune system, and central nervous system (CNS). Objectives: This review aims to summarize current knowledge on the pathophysiological mechanisms by which PM affects the skin–immune–brain axis and to explore the potential of AI-driven predictive models for early diagnosis and therapeutic development. Methods: We integrated evidence from epidemiological, clinical, and experimental studies that investigated PM-induced alterations in skin barrier function, immune activation, and neuroinflammation. In addition, we examined recent advances in multi-omics analyses and artificial intelligence (AI) applications for biomarker discovery and predictive modeling. Results: PM exposure disrupts skin barrier integrity, induces oxidative stress, and promotes pro-inflammatory cytokine release, thereby exacerbating inflammatory skin diseases such as atopic dermatitis and psoriasis. PM also triggers Th2 immune polarization; increases IgE, histamine, and chemokine signaling; and promotes systemic inflammatory responses. Systemic immune activation and cytokine signaling subsequently affect the CNS, leading to microglial activation, blood–brain barrier disruption, and neuroinflammation, which are linked to depression, anxiety, Alzheimer’s disease, and Parkinson’s disease. Recent AI-driven multi-omics studies such as Graph Neural Networks (GNN) and variational autoencoder–based integration have identified key SIB axis biomarkers including AhR–CYP1A1, IL-6, IL-1β, SOD2, and NLRP3, improving prediction of PM related disease risk. Conclusions: PM acts as a systemic environmental risk factor that drives interconnected pathological processes across the SIB axis. Integrating multi-omics with AI-based prediction models may facilitate early detection and personalized interventions for PM related skin, immune, and neuropsychiatric disorders.