Modeling cross-talk of RNA modification enzymes reveals tumor microenvironment-associated clinical significance and immunotherapy prediction in hepatobiliary malignancy

RNA modification includes four main types, N6-methyladenosine, N1-methyladenosine, alternative polyadenylation (APA), and adenosine-to-inosine (A-to-I) RNA editing, involving 41 enzymes that serve as "writers", "readers" and "erasers". By collecting RNA modifying enzyme information in 1759 hepatobiliary malignancy (HBM) samples from 11 datasets, an RNA modification HBM Score (RH_score) was established based on unsupervised cluster analysis of RNA modification-associated differentially expressed genes (DEGs). We identified the imbalanced expression of 41 RNA modification enzymes in HBM, which was scientifically categorized into two groups: RH_Score high and RH_Score low. A high RH_Score was associated with a worse prognosis and more immature immune cells in the tumor microenvironment (TME), while a low RH_Score was associated with a better prognosis and more mature immune cells in the TME. Further analysis using single-cell databases showed that the high RH_Score was immune exhaustion in the TME. RH_Score was involved in transcriptional regulation and post-transcriptional events in HBM. Additionally, resistant and sensitive drugs were selected based on RNA modification, and anti-PD-L1 therapy responded better with low RH_Score. In conclusion, our study comprehensively analyzes RNA modification in HBM, which induces TME changes and transcriptional and posttranscriptional events, implying potential guiding significance in prognosis prediction and treatment options.