Identifying novel tumor-related antigens and immune phenotypes for developing mRNA vaccines in lung adenocarcinoma.

The mRNA vaccines have been a novel strategy of immunotherapies for multiple cancers. Although several types of mRNA vaccines have been investigated and validated in some studies, their efficacy among patients with lung adenocarcinoma (LUAD) remains largely unknown. The number of tumor-associated antigens is not enough and no study focuses on stratifying the subgroup of LUAD patients suitable for vaccination. Based on the expression profiles of immune-related genes, consensus clustering was performed to identify the most appropriate phenotype for vaccination. The immune landscape of LUAD was shown via the graph learning-based dimensionality reduction analysis. We screened for five mutated and upregulated LUAD-related antigens (CCNB1, KIAA0101, PBK, OIP5 and PLEK2) that were highly correlated with immune infiltrating cells and unfavorable clinical outcomes. And three distinct immune phenotypes were identified in the TCGA and GSE72094 cohorts. Group S1 was an immunological "hot" cluster and related to a better prognosis, whereas Group S2&S3 was an immunological "cold" cluster and associated with a poorer prognosis. At last, the results revealed heterogeneity of LUAD patients in the immune landscape. We identified five potential cancer-related antigens for mRNA vaccines, and Group S2&S3 were the most suitable phenotypes for vaccination.