Spatial single cell analysis of tumor microenvironment remodeling pattern in primary central nervous system lymphoma

To determine the overall tumor microenvironment (TME), characteristics, and transition mechanisms in primary central nervous system lymphoma (PCNSL), we performed spatial transcriptomics and matched the corresponding single-cell sequencing data of PCNSL patients. We found that tumor cells may achieve a “TME remodeling pattern” through an “immune pressure-sensing model”, in which they could choose to reshape the TME into a barrier environment or a cold environment according to the immune pressure. A key FKBP5+ tumor subgroup was found to be responsible for pushing tumors into the barrier environment, which provides a possible way to evaluate the stage of PCNSL. The specific mechanism of the TME remodeling pattern and the key molecules of the immune pressure-sensing model were identified through the spatial communication analysis. Finally, we discovered the spatial and temporal distributions and variation characteristics of immune checkpoint molecules and CAR-T target molecules in immunotherapy. These data clarified the TME remodeling pattern of PCNSL, provided a reference for its immunotherapy, and provided suggestions for the TME remodeling mechanism of other cancers. Overall design: Here, 14964 single-cell transcriptomes of PCNSL were combined with spatial transcriptome analyses of hot, IME, IMS, and cold TMEs. The TMEs were characterized. In brief, the main cell types in the TME of PCNSL patients were identified by single cell sequencing, and were mapped to the corresponding spatial locations in 4 TMEs. Through integrative analysis, we defined and annotated tumor cell subpopulations according to their spatial distribution and functional characteristics and found that various tumor cells work together to reshape the TME into a barrier or cold environment through an "immune pressure-sensing model", thus realizing the "TME remodeling pattern". A key FKBP5+ tumor subgroup was found to be closely associated with TME remodeling, which provides a possible method for assessing brain tumor staging. The spatial communication mode between tumor cells and immune cells in each TME was also identified, thus refining the specific mechanism of the TME remodeling pattern and identifying the key molecules of the immune pressure-sensing model. Corresponding treatment suggestions were proposed according to the spatial characteristics of each TME in current PCNSL immunotherapy. These results reveal the spatial heterogeneity of PCNSL, highlight the localization and status of cell types and potential intercellular signals in the TME, and provide resources for further research on TMEs and updating immunotherapy methods.