Reorganization of pancreas circadian transcriptome with aging

Abstract

The evolutionarily conserved circadian system allows organisms to synchronize internal processes with 24-h cycling environmental timing cues, ensuring optimal adaptation. Like other organs, the pancreas function is under circadian control. Recent evidence suggests that aging by itself is associated with altered circadian homeostasis in different tissues which could affect the organ's resiliency to aging-related pathologies. Pancreas pathologies of either endocrine or exocrine components are age-related. Whether pancreas circadian transcriptome output is affected by age is still unknown. To address this, here we profiled the impact of age on the pancreatic transcriptome over a full circadian cycle and elucidated a circadian transcriptome reorganization of pancreas by aging. Our study highlights gain of rhythms in the extrinsic cellular pathways in the aged pancreas and extends a potential role to fibroblast-associated mechanisms.

Keywords: RNA transcriptomics; aging; circadian rhythms; pancreas.

Figure 1

Differential and temporal gene expression signatures in aged pancreas. (A) Pairwise correlation of normalized gene expression (transcripts per million; TPM) for young and old RNA-Seq samples (n=12). Colors indicate the Pearson correlation. (B) Principal Component Analysis (PCA) of temporal gene expression from young and old mice samples (n = 12). (C) Top: Volcano plot showing differential gene expression profile with significantly (p < 0.05, fold change ≥ 1.5) upregulated (n = 133 genes) and downregulated (n = 251) genes in old mice when compared to young. (D) Enriched molecular pathways corresponding to up and downregulated genes. (E) Top: Venn diagram of rhythmic genes in young and old mice. Bottom: Heatmap of z-scored rhythmically expressed genes in young and old mice. Heatmaps are double plotted for time.

Figure 2

Circadian transcriptome of the aged pancreas. (A) Examples of circadian profiles of genes that are rhythmic in both young and old mice (Arntl, Per2, Tef), only in old mice (Col1a1, Cdh11), and only in young mice (Cpsf3). (B) Enriched molecular pathways corresponding to rhythmic genes in young and old mice. (C) Peak time maps of all oscillating genes from young and old mice. Each dot represents the peak time of 10 significantly rhythmic genes. Enriched pathways corresponding to each time point are also shown. (D) Clock correlation distance (CCD) of 12 core clock genes from young and old mice pancreas.

Figure 3

Fibroblasts mediate differential rhythmicity in aged pancreas. (A) Bar plots for (left) fibroblast fraction obtained by imputation and (right) Fibroblast to epithelial ratio obtained by Fluorescence Activated Cell Sorting (FACS) analysis for young and old mice pancreas. (B) Bar plots for fibroblast fraction obtained by imputation for young and old human pancreas from 178 GTeX datasets. (C) Heatmap of significantly rhythmic genes indicative of fibroblast markers in young and old mice. Fibroblast marker Thy1/CD90 is shown in bold. Heatmap represents the p-value of rhythmicity. (D) Top: Comparison of temporal fibroblast fractions obtained by two independent methods: imputation and FACS. Bottom: Similarity between the estimated fibroblast fractions between imputations and FACS methods. (E) Pairwise correlation plot for fibroblast markers and fibroblast responsive genes (FRGs) in young (lower diagonal) and old (upper diagonal) mice. Colors represent the Pearson correlation coefficient. (F) Heatmap of significantly rhythmic FRGs in young and old, time-stamped human datasets. Heatmap represents the p-value of rhythmicity. (G) Heatmap of significantly rhythmic FRGs in young and old mice. Heatmap represents the p-value of rhythmicity.