Single Nucleus Transcriptome and Chromatin Accessibility Landscapes of Human Pituitaries

Abstract

The pituitary gland regulates key physiological functions, including growth, sexual maturation, reproduction, and lactation. Here, we present a paired single-nuclei (sn) transcriptome and chromatin accessibility characterization of six post-mortem human pituitaries. These samples were from juvenile, adult, and elderly male and female subjects. Well-correlated snRNAseq and snATACseq datasets facilitated robust identification of the major pituitary cell types in each sample. Using latent variable pathway analysis, we uncovered previously unreported coordinated gene expression modules and chromatin accessibility programs for each major cell type as well as an age-specific program across all the endocrine cell types. These largely appear to be congruent between human and mouse datasets. Given the importance of murine models in the study of human pituitary disorders and pituitary physiology, we next sought to compare expression profiles of pituitary cell types in mouse vs. human. Murine and human cell types were well correlated, exemplified by coordinated gene expression programs, especially for undifferentiated stem cells (SCs). In both species, we identified clusters corresponding to naive and committing SCs. All human SC clusters expressed the established SC markers SOX2 and SOX9, as well as genes involved in SC regulatory pathways (WWTR1, YAP1 andPITX2). Additional markers previously reported in murine pituitary SCs were also found in human SC, including WIF1, LGR5, FOS, CDH1, EGFR, LGR4, and WLS. Remarkably, in human, the main naive SC cluster was roughly divided into a high-JUN and a low-JUN expressing subgroup, whereas Jun expression was less pronounced in the murine SC cluster. In both species, committing SC clusters expressed the endocrine markers for POU1F1, TSHB, or POMC, while SCs committing to an intermediate lobe/melanotrope cell identity were distinguishable based on PAX7 expression. In addition, in the human datasets we identify a population of cells as originating from the pars tuberalis. We offer a range of markers that can be utilized for in vivo validation of these cells. Overall, the characterization of the murine and human pituitary SCs strongly suggests the co-existence of subpopulations with different lineage commitments in addition to a single uncommitted SC population. This sn atlas of the human pituitary is a valuable resource that will be made web-accessible.