Background and Aims: Extrachromosomal circular DNAs (eccDNAs) are prevalent in cancer genomes and emerge as a class of crucial yet less characterized oncogenic drivers. However, the structure, composition, genome-wide frequency, and contribution of eccDNAs in hepatocellular carcinoma (HCC), one of the most fatal and prevalent cancers, remain unexplored. In this study, we provide a comprehensive characterization of eccDNAs in human HCC and demonstrate an oncogenic role of microRNA-17-92-containing eccDNAs in tumor progression. Approach and Results: Using the Circle-sequencing method, we identify and characterize more than 230,000 eccDNAs from four paired samples of HCC tumor and adjacent non-tumor liver tissues. EccDNAs are highly enriched in HCC tumors, preferentially originate from certain chromosomal hotspots, and are correlated with differential gene expression. Particularly, a series of eccDNAs carrying the microRNA-17-92 cluster are validated by outward PCR and sanger sequencing. Quantitative PCR analyses reveal that microRNA-17-92-containing eccDNAs, along with the expression of their corresponding microRNAs, are elevated in HCC tumors and associated with poor outcomes and the age of HCC patients. More intriguingly, exogenous expression of artificial DNA circles harboring the miR-17-92 cluster, which are synthesized by the ligase-assisted minicircle accumulation method, can significantly accelerate HCC cell proliferation and migration. Conclusions: These findings delineate the genome-wide eccDNAs profiling of HCC and highlight the functional significance of microRNA-containing eccDNAs in tumorigenesis, providing insight into HCC pathogenesis and cancer therapy, as well as eccDNA and microRNA biology. Export
Alterations to N6-methyladenosine modifications can promote malignant progression by modulating gene expression through regulation of transcript metabolism. Quantifying the causal impact of m6A dysregulation at the population level could help to guide personalized therapeutic interventions. Here, we developed a causal framework that enables precise estimation of the driving effects (DE) of m6A dysregulation on tumor survival, establishing DE-based enhanced prioritization rules for anti-m6A therapies. The global average DE of m6A dysregulation across 9,647 tumors resulted in a decrease in overall survival of 180.1 days considering a 5-year survival endpoint, suggesting potential overall benefits of treatment. Profiling of tumors most susceptible to m6A dysregulation and application of a modifier-mining tool revealed an effect modification in the PI3K/AKT/mTOR pathway. Specifically, the benefits of m6A targeted treatment varied depending on baseline mTOR levels, which was validated in vitro and in vivo. Overall, focusing on m6A, this study established a paradigm leveraging DE over conventional risk to optimize personalized therapy.