A specific gene expression program underlies antigen archiving by lymphatic endothelial cells in mammalian lymph nodes.

Lymph node (LN) lymphatic endothelial cells (LEC) actively acquire and archive foreign antigens. Here, we address questions of how LECs achieve durable antigen archiving and whether LECs with high levels of antigen express unique transcriptional programs. We use single cell sequencing in dissociated LN tissue and spatial transcriptomics to quantify antigen levels in LEC subsets and dendritic cell populations at multiple time points after immunization and determine that ceiling and floor LECs archive antigen for the longest duration. We identify, using spatial transcriptomics, antigen positive LEC-dendritic cell interactions. Using a prime-boost strategy we find increased antigen levels within LECs after a second immunization demonstrating that LEC antigen acquisition and archiving capacity can be improved over multiple exposures. Using machine learning we define a unique transcriptional program within archiving LECs that predicts LEC archiving capacity in independent mouse and human data sets. We test this modeling, showing we can predict lower levels of LEC antigen archiving in chikungunya virus-infected mice and demonstrate in vivo the accuracy of our prediction. Collectively, our findings establish unique properties of LECs and a defining transcriptional program for antigen archiving that can predict antigen archiving capacity in different disease states and organisms.