Transcriptomic deconvolution of human endometrial cell types

The human endometrium is a highly heterogeneous tissue comprising multiple cell-types and cellular states that change in cycle-dependent manner. This tissue heterogeneity is a major drawback when interpreting bulk (whole biopsy) transcriptomic data. To address this hurdle, RNA-seq analysis was performed on four isolated stromal cell populations [SUSD2+ perivascular stromal cells (PVCs), endometrial stromal cells (EnSCs), SUSD2+ clonal cells (MSCs), and clonal transit amplifying cells (TAs)], glandular epithelial cells, uNK cells, and matched whole tissue. Transcriptomic data were obtained for 3 whole endometrial biopsies and 6 matched, purified cell populations. Three independent midluteal biopsies were used to isolate, culture and sequence different endometrial cell types and stromal subsets. Following enzymatic digestion, epithelial cells (EpC) were separated from the stromal and immune cell fraction. Primary EpC were subjected to gland organoid formation and cultured for 10 days. Stromal cells were then subjected to magnetic activated cell sorting (MACS) using the W5C5 antibody that recognises SUSD2, perivascular cell marker. PVC and EnSC were propagated as standard cultures as well as subjected to colony forming unit (CFU) assays. After overnight incubation, the supernatant of the stromal cell fraction was subjected to MACS to isolate uNK cells using a PE-conjugated anti-human CD56 monoclonal antibody. All samples were frozen and then subjected to RNA extraction and RNA library preparation. Each library was sequenced on two separate flow cells.