Hepatocellular carcinoma (HCC) remains a global health challenge with a high recurrence rate, highlighting the need to decipher its underlying molecular mechanisms. While circular RNAs (circRNAs) are known to regulate gene expression, the landscape of competitive splicing-the dynamic balance between back-splicing and linear splicing-during HCC initiation and recurrence remains largely unexplored. Leveraging rRNA-zero RNA-seq data from GSE169289 (comprising 34 primary, recurrent, and non-recurrent HCC samples with paired controls), this study utilized an upgraded SUVA pipeline to systematically profile circRNA-associated alternative splicing events. By integrating four distinct algorithms, we identified 9,652 high-confidence circRNAs. Our analysis revealed extensive differential splicing events distinguishing tumor from normal tissues, and notably, recurrent from non-recurrent tumors, with host genes enriched in chromatin remodeling and ubiquitin-dependent pathways. Specifically, we observed a conserved attenuation in the competitive splicing of the LARP1B circRNA in tumor tissues, which correlated with host gene downregulation and poor prognosis. In the context of recurrence, circ-KLHL8 exhibited specific splicing suppression; in silico modeling suggests a potential mechanism where circ-KLHL8 loss relieves the sponging of miRNAs, thereby suppressing the tumor suppressor GLYCTK. Furthermore, RBP-circRNA co-regulatory network analysis implicated HNRNPU and CPSF4 as potential upstream regulators of initiation-related splicing, and DKC1 as a modulator of recurrence-associated events. Collectively, this study provides the first systematic characterization of the competitive splicing landscape in HCC, proposing a novel theoretical framework and candidate targets for diagnostic and therapeutic interrogation.
Jiang et al. (Thu,) studied this question.