Decrypting cellular engagement and recruitment from spatially resolved transcriptomics

The recruitment of the various types of cells into the tissue microenvironment and how these cells engage with other cells in the tissue sites play critical biological roles. However, it is difficult to study these processes on a genome-wide scale using traditional low-throughput experiments. Spatially resolved transcriptomics (SRT), especially high definition SRT, have offered opportunities in deciphering the determinants of cellular localization at unprecedented scale and resolution. But it is challenging to digest the abundance of transcriptomics and positional information of the numerous cells in SRT data. To address this gap, we developed a fully interpretable multi-instance deep learner, spacer. We deployed spacer to a panel of 17 high definition and 20 low definition SRT datasets, for studying how stromal and immune cells were recruited into tumors and heart during myocarditis. We coupled spacer with orthogonal immuno-peptidomics, spatial T cell receptor (TCR) sequencing, and single cell sequencing experiments. We discovered genes that encode more immunogenic peptides and that are involved in developmental pathways are more potent in recruiting T cells to local tumor sites. We also found, however, that tumor cells tend to down-regulate such genes to avoid T cell recognition. On the other hand, expression of mucins in the tumor cells was found to repel T cell localization. For the engaging cell side, we uncovered a tumor-engaging gene signature for T cells, validated by spatial-TCR-seq data. Spacer also revealed that CD4 ⁺ T cells, though fewer in numbers, are more responsive than CD8 ⁺ T cells in the heart during myocarditis. Collectively, this study establishes a spatially resolved paradigm for studying cellular localization mechanisms in situ and paves the way for the construction of a comprehensive cellular recruiting-engaging interaction atlas in solid tissues.