Tracing the molecular route to progression in miRNA biogenesis-defective thyroid lesions [Spatial_transcriptomics]

Germline and somatic alterations in DICER1 and DGCR8 microprocessors confer risk of developing benign and malignant thyroid lesions, yet the molecular events driving malignant transformation remain unclear. We trace the molecular trajectories from benignity to malignancy in DICER1/DGCR8-mutated thyroid lesions using multi-omic profiling on 27 DICER1/DGCR8-mutated samples. Our findings reveal a progressive, specific, and linear accumulation of genetic alterations, which when combined with enhanced downregulation of miRNAs distinguished DICER1/DGCR8-malignant lesions from their benign counterparts. Compensatory hypomethylation of miRNA encoding genes characterised DICER1/DGCR8-benign lesions, but as the tumors progressed to malignancy, methylation was partly reimposed, reversing the attempts to activate miRNA-encoded genes and further compromising miRNA production. Transcriptomic analyses revealed mutation-specific effects on the microenvironment, whereby DICER1 mutations activated canonical thyroid cancer progression pathways, whereas altered DGCR8 associated with immune-related alterations. This work unveils specific molecular events underlying malignant progression of miRNA-biogenesis related thyroid tumors and identifies potential biomarkers and disease etiology mechanisms. Overall design: We explored the spatial transcriptonal profiles of 1 sporadic DICER1-mutated case and 1 germline DGCR8-mutated case, capturing different histological components in each case. For the DICER1-case, this included a tumour core from a DICER1-mutated PDTC, the tumor's capsule (DICER1-wildtype), a non-neoplastic nodular area (DICER1-wildtype benign thyroid lesion) and a norma lthyroid area (DICER1-wildtype). For the DGCR8-case, this included a tumour core from a DGCR8-mutated micropapillary thyroid carcinoma, a normal thyroid area adjacent to the tumour (DGCR8-mutated, 3 non-neoplastic nodular areas (DGCR8-mutated benign thyroid lesions) and a normal thyroid area (DGCR8-mutated). Each histological component was further segmented into PanCK+ (representing the epithelial population of the histological component) and VIM+ areas (representing the stroma-enriched population of the histological component). Differential expression analyses were performed between the different histological components at the region of interest level (PanCK+/VIM+), the epithelial level (PanCK+/VIM-) and stromal level (PanCK-/VIM+).