P08.01 Novel anti-PD1 predictive signature and functional dendritic-cell biomarkers in melanoma identified with systems immunology

Background Despite the remarkable clinical success of anti-PD-1 immunotherapy in melanoma, identifying functional biomarkers to predict response and guide rational treatment combinations remains a major challenge. Current biomarkers used in clinical practice are largely limited to single-parameter indicators, such as PD-1 expression and CD8 + T cell infiltration. However, the existence of non-responsive patients with high CD8+ T cell infiltration underscores the insufficiency of these markers. Emerging evidence suggests that effective dendritic cell (DC)–CD8+ T cell crosstalk within the tumor microenvironment is critical for therapeutic efficacy, yet no predictive signatures currently capture this key interaction.Material and Methods To address this limitation, we developed a systems immunology approach to identify immune signatures from whole-transcriptome data across multiple anti-PD-1–treated metastatic melanoma cohorts. Focusing on DC–CD8 + T cell crosstalk, we used CD74 and CD8A as reference biomarkers to select correlating biomarkers, that were incorporated into a logistic regression framework, resulting in the identification of a 15-gene signature, PRIME. Prediction power of PRIME was evaluated in anti-PD-1 pre-treatment samples from multiple transcriptomic metastatic melanoma datasets. These immune signatures also enabled functional biomarker discovery through unsupervised clustering of gene expression, clinical, and functional data. Potential biomarkers were filtered from antigen processing and presentation -related clusters using spatial transcriptomics and single-cell RNA sequencing datasets. The immunogenic functions of selected biomarkers were evaluated in human monocyte-derived dendritic cells (moDCs) by assessing antigen processing efficacy, expression of costimulatory molecules, and alterations in cytokinome upon biomarker gene expression silencing.Results Our bioinformatic analysis identified 3,359 and 2,902 genes positively and negatively associated with survival of anti-PD-1-treated patients. Of these, 667 genes from the positive group and 129 genes from the negative group showed a significant correlation with DC-CD8+ T cell axis through correlations with CD8A and CD74. Subsequent PRIME model outperformed previously published models in predicting anti-PD-1 response, particularly in tumors with high CD8 + T cell infiltration, reflecting a functional immune state driven by DC–CD8+ T cell interactions. Filtering of potential functional biomarkers highlighted SLAMF7 and TYMP as novel biomarkers associated with DC–CD8+ T cell interactions, based on their expression patterns in dendritic cells and correlation with altered levels of melanoma antigen–specific T cells. Mechanistic studies in moDCs revealed that these biomarkers regulate exogenous antigen processing, expression of costimulatory molecules CD80 and CD40, and secretion of multiple cytokines and chemokines implicated in regulation of DC and T cell functions and migration.Conclusion In conclusion, these results functionally link the identified biomarkers to key immunological processes underlying anti-PD-1 responses. Together, our findings present a new strategy for predicting anti-PD-1 responses in melanoma and uncover new biomarkers with functional roles that may inform future therapeutic development to enhance clinical efficacy.