PBMC circRNA Profiles Distinguish Atypical From Idiopathic Parkinsonism and Track Clinical Severity Across Syndromes

Abstract

Early separation of atypical Parkinsonian syndromes (APS) from idiopathic Parkinson's disease (iPD) remains challenging. Given that many brain-biased circular RNAs (circRNAs) are stable and expressed in blood cells, peripheral blood mononuclear cell (PBMC) circRNAs were profiled in 140 individuals (iPD 43, MSA 35, PSP 19, healthy controls 43) using RT-qPCR across 46 a priori candidates. Group differences were analyzed using the Kruskal–Wallis test with FDR correction; clinical associations were examined using Bonferroni-corrected Spearman correlations with a bootstrap stability assessment; and discrimination was assessed using logistic regression with nested repeated cross-validation, calibration analysis, and ROC analysis. Six circRNAs differed across groups after FDR adjustment (SLC8A1_circ_0000994, MAPK9_circ_0001566, MGA_circ_0000591, AGTPBP1_circ_0007162, SFMBT2_circ_0000211, HIPK3_circ_0000284), with median levels lowest in iPD and highest in PSP. Internally cross-validated multi-marker panels achieved AUCs of 0.88 for PSP versus controls, 0.78 for PSP versus iPD, and 0.67 for MSA versus iPD. Several circRNAs, including non-differential ones, showed bootstrap-stable associations with cognitive, mood, working-memory, and praxis measures. ENCORI- and CircInteractome-based analyses placed the differentially expressed circRNAs within a shared RBP interaction landscape and identified modest but significant miRNA-target overlap. Findings indicate that PBMC circRNA patterns aid APS-iPD differentiation and align with clinical impairment within this cohort; external validation, longitudinal assessment, and mechanistic studies are needed to establish diagnostic performance and biological relevance.