Single-cell and spatial transcriptomics analysis of human adrenal aging

Abstract

Objective: The human adrenal cortex comprises three functionally and structurally distinct layers that produce layer-specific steroid hormones. With aging, the human adrenal cortex undergoes functional and structural alteration or "adrenal aging", leading to the unbalanced production of steroid hormones. Given the marked species differences in adrenal biology, the underlying mechanisms of human adrenal aging have not been sufficiently studied. This study was designed to elucidate the mechanisms linking the functional and structural alterations of the human adrenal cortex.

Methods: We conducted single-cell RNA sequencing and spatial transcriptomics analysis of the aged human adrenal cortex.

Results: The data of this study suggest that the layer-specific alterations of multiple signaling pathways underlie the abnormal layered structure and layer-specific changes in steroidogenic cells. We also highlighted that macrophages mediate age-related adrenocortical cell inflammation and senescence.

Conclusions: This study is the first detailed analysis of the aged human adrenal cortex at single-cell resolution and helps to elucidate the mechanism of human adrenal aging, thereby leading to a better understanding of the pathophysiology of age-related disorders associated with adrenal aging.

Keywords: Activator protein 1; Adrenal cortex; Aging; Single-cell RNA sequencing; Spatial transcriptomics; WNT/β-catenin signaling.

Figure 1

Bulk RNA-seq analysis of the young and elderly adrenal cortex. A Violin plots showing gene expression of steroidogenic enzymes. Asterisks indicate significant expression changes. B Bar plot showing the result of gene set enrichment analysis (GSEA) comparing between the young and elderly. Selected gene sets are shown. Gene sets with positive normalized enrichment scores (NES) indicate the positive enrichment of genes in the elderly. C GSEA results for senescence-associated gene sets.

Figure 2

scRNA-seq analysis of the young and elderly adrenal cortex. A UMAP plot of scRNA-seq data colored by cell type annotation. B Dot plot showing the expression of marker genes in the annotated clusters of scRNA-seq data. C Bar plot showing the composition of ZG, ZF, and ZR cells in the young and elderly. D Heatmap showing the expression of steroidogenic enzyme genes in each adrenocortical layer compared between the young and elderly. Genes with positive values indicate upregulation in the elderly. E Dot plot showing the result of gene set enrichment analysis comparing the young and elderly in each adrenocortical layer of scRNA-seq data. Gene sets with significant enrichment are shown. Gene sets with positive normalized enrichment scores (NES) indicated the positive enrichment in the elderly. F TFs ranked by regulon specificity score (RSS), showing the top five TFs with the highest RSS per cell type. The bottom row is a Venn diagram showing the number of overlaps between the young and elderly in the RSS Top 10 TFs.

Figure 3

Effect of perturbation of WNT/β-catenin and AP-1 family members on adrenocortical cell differentiation. A Trajectory analysis inferring the adrenocortical cell differentiation. Adrenocortical cells of the young were dimensionally reduced onto a diffusion map. Left panel shows cell type. Right panel shows the inferred trajectory (gray arrows) and pseudotime of each cell. B Effects of LEF1 knockout and JUN and FOSL2 overexpression on cell differentiation. Black arrows represent shifts in cell identity owing to perturbations. The color of each cell indicates the perturbation score (PS). Negative PS (red) indicates that TF perturbation inhibited differentiation. Positive PS (green) indicates that TF perturbation promoted differentiation.

Figure 4

ST analysis of the young and elderly adrenal cortex. A Split view of the spatial location of ST spots annotated as ZG, ZF, and ZR. B Box plot showing the result of the assessment of continuity in the adrenocortical layers using ST data. Higher assortativity scores indicate that the layers are more continuous. Asterisks indicate significant differences (P value < 0.05) according to the Wilcoxon rank-sum test.

Figure 5

Estimated localization of macrophages in the young and elderly adrenal cortex. A Split view of the spatial distribution of the estimated macrophage proportions in ZG, ZF, and ZR. B Box plot showing the estimated macrophage proportions. Asterisks indicate significant differences (P value < 0.05) according to the Wilcoxon rank-sum test. C Dot plot showing the correlation between the estimated macrophage proportions and the normalized enrichment score of the hallmark inflammatory response and SenMayo gene sets. Dot colors represent the annotated cell types and the blue line represents the linear regression line.

Figure 6

Plasma steroid profiling of the young and elderly. A Scores plot showing the result of partial least squares-discriminant analysis (PLS-DA) between young and elderly. Shaded circles indicate 95% confidence intervals. Dots represent individual samples. B Variable importance in projection (VIP) scores of component 1 of the PLS-DA. Metabolites with a VIP score of 1 or higher are shown. C Volcano plot showing the differences in the concentrations of metabolites between young and elderly. The vertical and horizontal lines indicate the log2 fold-change (log2FC) threshold of 0.25 and the adjusted P-value threshold of 0.05, respectively. Metabolites with positive log2FC indicate the higher concentrations in the elderly than in the young. The top three metabolites with the largest fold changes are highlighted.

Figure 7

Graphical abstract. The human adrenal cortex undergoes the unique functional and structural alteration with aging. Our data suggest that both activation of the AP-1 in ZF and attenuation of WNT/β-catenin signaling in ZG inhibit the centripetal differentiation from ZG through ZF to ZR cells, thereby contributing to the expansion of ZF with reciprocal reduction of ZG and ZR in the elderly. Once, ZF cells cease proliferation and differentiation, might undergo cellular inflammation and senescence, which are phagocytosed by macrophages. However, when such senescent ZF cells exceed the phagocytic capacity of macrophages, they might accumulate and even die, thereby accelerating the process of human adrenal aging.