Endocrine Abstracts

Background: Molecular mechanisms underlying the pathogenesis of adrenocortical adenomas (ACAs) and autonomous cortisol secretion remains frequently unexplained despite previous comprehensive genomics studies.

Aim: To gain novel insights into molecular pathogenesis of adrenocortical tumours by investigating transcriptome profiles of ACAs at single-nuclei resolution (snRNA-Seq), using adult human normal adrenal glands (NAGs) as reference.

Methods: We isolated single nuclei from 6 NAGs and 12 ACAs, including 7 cortisol-producing adenomas (CPAs) and 5 endocrine inactive adenomas (EIAs) with different genetic background. snRNA-Seq was performed using inDrop^TM technology. Data analysis, integration and exploration was performed using Seurat R package. Pathway enrichment analysis was performed using pathfindR withKEGG pathways as reference. Transcriptome profile from ACA was integrated with NAG using anchor pairs between the two datasets. Identification of tumour- and mutation-specific cell subpopulations (i.e. clusters) was done by differential gene expression analysis.

Results: Within the NAGs, we identified different satellite clusters of immune, myeloid and vascular origin, in addition to main clusters representing the three adrenal cortex zones, medulla and capsule. We also identified two subpopulations potentially representing adrenocortical and adrenomedullary progenitor cells, located within and underneath the capsule. Comparative analysis of the transcriptional profiles of NAGs and ACAs revealed the presence of six ACA-specific clusters, namely four “tumour-specific” (TC1-4), “tumour microenvironment” (TME), and one cluster overexpressing genes of cholesterol pathway (Chol-upreg). Specifically, the TC1 was mostly found in 2 CTNNB1-mutated samples (one EIA and one CPA), where a significant overexpression of AFF3, FTO and ISM1, as well as genes of spliceosome (fold enrichment, FE=7.6), ECM-receptor interaction (FE=7.3) and Hippo signaling (FE=2.8) pathways were observed. The TC2 cluster was more abundant in EIA and characterized by overexpression of genes like MMP26, SP100 and EIF4H, as well as genes of NOD-like receptor (FE=3.6) and IL-17 signaling pathway (FE=3.5) associated with tumour promotion. The Chol_upreg cluster was largely represented in CPAs and characterized by a very high expression of genes associated to steroid biosynthesis (FE=34.2) and cortisol synthesis and secretion (FE=9.5), including HMGCS1, SQLE and FDX1. The remaining clusters (TME, TC3-4) were quite homogeneously distributed in all ACAs, independently from mutational status (2 PRKACA-, 1 GNAS-, 6 CTNNB1-, 3 no driver-mutation) and cortisol secretion.

Conclusion: Our human adult NAG single-cell atlas represents a unique source for investigations of adrenal diseases and allowed us to investigate the molecular heterogeneity of ACAs at single-cell level, showing the presence of specific cell populations associated with cortisol secretion and genetic background.