Abstract
RNA m6 methyladenosine (m6A) modifications impact tumor biology and immune processes, particularly in hepatocellular malignant tumors. Using a consensus clustering algorithm on 371 hepatocellular carcinoma (HCC) samples, we identified three m6A-modified subtypes and correlated them with positive tumor microenvironment (TME) markers for distinct immune phenotypes. Stratifying patients based on m6A scores revealed a low presentation group with better immune penetration, lower tumor mutation load, and increased expression of immune checkpoint markers like CTLA-4 and PD-1, suggesting enhanced responsiveness to immunization therapy. A machine-learning model of 23 m6A genes was constructed. Single-cell analysis revealed a surprising enrichment of IGFBP3 in astrocytes, prompting the exploration of associated signaling pathways. Experimental verification shows that IGFBP3 is significantly enhanced in normal tissues, while immunohistochemical analysis shows that its expression is lower in tumor tissues, indicating its protective effect in HCC and a good prognosis. Importantly, high IGFBP3 expression is associated with better outcomes in patients receiving immunotherapy. Moreover, cytotoxic T lymphocyte (CTL) experiments have confirmed that high expression of IGFBP3 is associated with stronger T cell-killing ability. In summary, the comprehensive evaluation of m6A modification, immune characteristics, and single-cell analysis in this study not only revealed the TME of HCC but also made significant contributions to the progress of personalized HCC immunotherapy targeting IGFBP3. This study provides a solid theoretical foundation for clinical translation and emphasizes its potential impact on developing effective treatment strategies.