Integrative machine learning and single-nucleus transcriptomics reveal MAPK-centered neuronal responses underlying...

Hexafluoropropylene oxide-dimer acid (GenX or HFPO-DA) is a novel per- and polyfluoroalkyl substance developed as a replacement for legacy compounds, yet its potential neurotoxic effects remain poorly understood. In this study, we combined bibliometric profiling, in silico target prediction, in vivo behavioral assessments, single-nucleus RNA sequencing (snRNA-seq), network biology, machine learning, molecular docking, and in vitro validation to explore the potential relevance of GenX in Alzheimer's disease (AD). Bibliometric analysis revealed increasing research attention to GenX-associated health hazards, including emerging concerns regarding brain-related effects. Target prediction identified 301 putative GenX-related genes, which were significantly enriched in AD-related pathways. Behavioral analyses demonstrated that chronic GenX exposure impaired recognition memory and spatial learning in mice. Analysis of human prefrontal cortex snRNA-seq data revealed pronounced transcriptional alterations in AD neurons and identified 200 AD-related neuronal DEGs. Integration with GenX targets yielded 10 overlapping genes, which were further prioritized through protein-protein interaction (PPI) network analysis. Machine learning further identified an 8-gene signature with robust diagnostic performance across training and external validation cohorts. Molecular docking showed favorable binding affinities between GenX and core target proteins. Finally, GenX exposure reduced SH-SY5Y cell viability, activated the RAS-RAF-MEK-ERK cascade, and promoted apoptosis-related alterations, supporting a MAPK-centered neurotoxic mechanism potentially relevant to AD-associated neuronal vulnerability. Collectively, this integrative multi-level analysis provides mechanistic insights into the potential neurotoxic effects of GenX and underscores its possible relevance to neurodegeneration-associated molecular processes in AD.