Description

InDrops – Single cell RNA sequencing for characterizing tumor microenvironments

Ambrose Carr^1,2, Vaidotas Kiseliovas^1,3, Juozas Nainys^2,3, Elham Azizi¹, Andrew Cornish^1,2, Sandhya Prabhakaran¹, George Plitas¹, Kasia Konopacki¹, Alexander Rudensky¹, Linas Mazutis^1,3, Dana Pe`er^1*

^*Coresponding author

¹Computational and Systems Biology Program, Sloan Kettering Institute, New York, USA; ²Department of Biological Sciences, Columbia University, New York, USA; ³ Institute of Biotechnology, Vilnius University, Vilnius, Lithuania

The tumor-immune microenvironment varies considerably between and within patients, with a heterogeneous array of cell subtypes likely necessary for effective elimination of tumor cells. A more precise delineation of these subtypes and their interplay is crucial to determining mechanisms of elimination of tumor cells and improving treatments. Towards understanding and harnessing the heterogeneity of tumor-infiltrating immune cells, we aimed to characterize immune cell populations in breast cancer in an unbiased genome-wide approach. We used InDrops – droplet based single cell RNA sequencing, to collect a dataset of over 50,000 single immune cell transcriptomes from 8 patients. The CD45+ cells were collected from different tissues including tumor, peripheral blood, lymph node and normal tissue from prophylactic mastectomies. For normalization, clustering, and imputing dropouts we used the Biscuit algorithm, obtaining 82 cell subpopulations revealing significant diversity of immune cells across tissues and patients. We observed that while T-cells share surface markers between blood and tissue resident cells, their gene expression is dramatically different at the global level resulting in 32 T cell clusters, where T cells in tumor span a range of activation, with the least activated T cells overlapping with those in normal tissue. Cytotoxic and regulatory T cell clusters are more abundant in tumor tissue. T cells in blood however are phenotypically different and exhibit naive characteristics. We also observed more activated tumor-associated macrophages, monocytes and pDCs in tumor tissue, and less activated monocytic cells in normal tissue.