An Integrative RNA Spliceosomic Landscape of Pancreatic Neuroendocrine Tumors Identifies Clinically Relevant Molecular Subgroups

Alterations in alternative splicing are emerging as a novel cancer hallmark, offering new insights into tumor biology. However, integrative analyses of splicing are still scarce, particularly in rare cancers like pancreatic neuroendocrine tumors (PanNETs), whose striking heterogeneity complicates patient diagnosis and treatment. Here, we provide the first comprehensive characterization of the RNA splicing landscape in PanNETs through integrative analysis of RNA-seq data from 174 tumor samples. We identified three robust spliceosomic groups (SPN1, SPN2, SPN3) each associated with unique clinical and molecular characteristics. SPN1 displayed intermediate clinical behavior alongside enhanced mTOR signaling; SPN2 was characterized by a less secretory phenotype, enrichment in alpha-cell markers and somatostatin receptors, increased metastasis, and frequent mutations in MEN1 and DAXX/ATRX genes; in contrast, SPN3 was composed mainly by low grade tumors with beta-cell marker expression and the lowest mutational rate, yet it also contained all the highly proliferative neoplasms. Moreover, each group had a specific alternative splicing events signature, revealing an unprecedented discovery: the association between the expression profile of the splicing machinery and its product, the splicing variants. We provide a detailed characterization of the molecular and functional consequences of the splice variants defining each of the spliceosomic groups. These findings underscore the previously unrecognized yet significant impact of RNA splicing on PanNET heterogeneity and suggest that detailed splicing profiles could serve as valuable tools for identifying novel biomarkers and therapeutic targets. Thus, beyond providing crucial insights into PanNET molecular biology, our study offers a foundation for future studies exploring personalized therapeutic strategies based on splicing features.