Discriminating epigenetic landscapes: multi-omics characterization of benign thyroid nodules versus papillary thyroid carcinomas.

BACKGROUND: Discriminating the epigenetic landscapes of coincidental benign thyroid nodules (particularly follicular adenoma subtypes) from papillary thyroid carcinoma (PTC) remains a critical unresolved challenge, impeding mechanistic insights into their divergent pathogenic trajectories. METHODS: To address this knowledge gap, we performed integrative multi-omics profiling of histologically paired benign thyroid nodules and PTC lesions from the same patients, synergizing chromatin accessibility mapping (ATAC-seq), whole-exome sequencing, transcriptomics, and ATAC-seq-derived extrachromosomal circular DNA (eccDNA) detection. RESULTS: Three pivotal mechanisms emerged from our cross-omics analyses to delineate the benign-malignant dichotomy. First, chromatin architecture interrogation revealed spatially colocalized PTC-specific accessible regions with somatic mutation hotspots, suggesting coordinated interplay between epigenetic remodeling and genomic instability in malignant transformation. Second, we uncovered ARHGEF28 and ARHGEF24 as novel potential benign-specific master regulators, where TEAD4-binding motif enrichment in benign-hyperaccessible chromatin drives their coordinated overexpression, forming a self-reinforcing regulatory loop unique to benign thyroid nodules. Third, eccDNA-centric profiling delineated a different regulatory paradigm: benign thyroid noduless exhibited preferential enrichment of T-cell signaling related elements on eccDNA scaffolds, whereas PTCs eccDNA were enriched in the DNA replication signaling pathways. This multidimensional atlas not only maps lineage-specific regulatory topologies of thyroid neoplasms but also establishes the ARHGEF28/24-TEAD4 axis as potential association with benign lineage. CONCLUSIONS: By elucidating chromatin-based thresholds of malignant progression, our findings provide a molecular framework for differential diagnosis and mechanistic dissection of transformation checkpoints.