Stage-specific transcriptomic changes in pancreatic α-cells after massive β-cell loss

Abstract

Background: Loss of pancreatic insulin-secreting β-cells due to metabolic or autoimmune damage leads to the development of diabetes. The discovery that α-cells can be efficiently reprogrammed into insulin-secreting cells in mice and humans has opened promising avenues for innovative diabetes therapies. β-cell loss triggers spontaneous reprogramming of only 1-2% of α-cells, limiting the extent of regeneration. Most α-cells are refractory to conversion and their global transcriptomic response to severe β-cell loss as well as the mechanisms opposing their reprogramming into insulin producers are largely unknown. Here, we performed RNA-seq on FAC-sorted α-cells to characterize their global transcriptional responses at different time points after massive β-cell ablation.

Results: Our results show that α-cells undergo stage-specific transcriptional changes 5- and 15-days post-diphtheria toxin (DT)-mediated β-cell ablation. At 5 days, α-cells transiently upregulate various genes associated with interferon signaling and proliferation, including Interferon Induced Protein with Tetratricopeptide Repeats 3 (Ifit3). Subsequently, at 15 days post β-cell ablation, α-cells undergo a transient downregulation of genes from several pathways including Insulin receptor, mTOR and MET signaling.

Conclusions: The results presented here pinpoint novel markers discriminating α-cells at different stages after acute β-cell loss, and highlight additional signaling pathways that are modulated in α-cells in this context.

Keywords: Alpha cell; Beta cell; Conversion; Ifit3; Pancreas; Pancreatic islet; Plasticity; RNA-seq; Regeneration; Transcriptome.

Fig. 1

α-cells undergo stage-selective global transcriptional changes after diphtheria toxin (DT)-induced β-cell loss. A Schematic of the transgenic mice used in this work. B Experimental design. α-cell lineage tracing was performed administering doxycycline for 2 weeks in the drinking water. After 2 additional weeks of DOX clearance, β-cells were ablated by intraperitoneal diphtheria toxin injections on days 1, 3 and 5. α-cells were purified at 5, 15 and 30 days post-DT injection (dpDT). C Hierarchical clustering analysis shows that the transcriptomes from stage-specific α-cell samples present higher correlation with biological replicates than with samples from other stages

Fig. 2

Analysis of the α-cell dynamic transcriptional response to β-cell loss reveals signaling pathways potentially modulating α-cell conversion towards insulin production. Selected Gene Set Enrichment Analysis (GSEA) results are presented (see also Figure S2). Genes with higher expression in 5dpDT α-cells, relative to its timely connected stages (Ctrl and 15dpDT α-cells) are associated with interferon signaling and control of the mitotic process. Genes downregulated in 15dpDT α-cells, relative to its timely connected stages (5dpDT and 30dpDT α-cells) are associated with inhibition of signaling from InsR, mTOR and MET pathways. All results presented are significant, considering a P-value < 0.05 and FDR < 0.25

Fig. 3

Key α-cell transitional gene expression signatures define α-cell stages after acute β-cell loss. A PCA shows that 1522 differentially expressed genes are enough to distinguish between α-cells from different stages after β-cell loss. The bar plot on the left depicts the percent of variance explained by each of the principal components (PC). PC1, accounting itself for 40% of the variance among samples, clearly separates α-cell 15dpDT replicates from the rest of the samples, suggesting that these cells present the most different transcriptomic signature. B Hierarchical clustering of 1522 differentially expressed genes identifies 5 gene clusters associated with specific α-cell signatures that are characteristic of each stage. C Clustered genes’ distribution across the PCA1, PCA2 and PCA3 dimensions. D The gene expression pattern characteristic of each cluster is associated with enrichment at specific α-cell stages. The boxes show the IQR of RNA levels, whiskers extend to 1.5 times the IQR or extreme values and notches indicate 95% confidence intervals of the median. The P-value was calculated with the Wilcoxon rank-sum test. NS: Not significant

Fig. 4

Epigenomic analyses reveal stage-specific marker genes potentially driving the α-cell response to β-cell loss. A Proportion of genes from each cluster associated with transcription factor binding sites, and proportion of cluster genes with their promoter interacting with at least 1 islet enhancer. B Expression profiles for selected cluster 1–5 markers. Codes below each gene name indicate association with at least one transcription factor binding site in human islets (&), promoter physical interaction (pcHi-C) with at least one human islet enhancer (¶) and whether the gene encodes for a transcription factor (#). NH: No human homolog found. * P < 0.05. C Violin plots showing the single-cell expression profiles for selected cluster marker genes in mouse pancreatic islet endocrine cell types (data from [28]). D Integrative map of the IFIT3 locus showing human islet ChIP-seq signal for 5 key pancreatic islet transcription factors, histone modification enrichment profiles associated with active (H3K27ac) promoter (H3K4me3) and enhancer (H3K4me1) regions, islet enhancer hubs, islet regulome regions and arcs representing high-confidence pcHi-C interactions in human islets (data from [18, 26])

Fig. 5

Ifit3 expression is strongly induced in α-cells after acute β-cell loss. A Representative immunofluorescence staining of pancreas sections obtained from Ctrl, 5dpDT and 15dpDT treated mice. B Percentage of cells co-expressing Ifit3 and YFP, expressed relative to the total number of total YFP+ cells in the islets. * P < 0.05. Total number of YFP+ cells (expressed as mean ± SEM) counted per replicate: Control (615 ± 268), 5dpDT (341 ± 184), 15dpDT (206 ± 48) from 3 biological replicates. C Percentage of cells expressing YFP, Ifit3, and co-expressing both markers, expressed relative to the total number of islet cells. Note that YFP+ cells represent the main cell type in 15dpDT islets given that by this timepoint β-cell ablation is almost complete. ** P < 0.01. Total number of islet cells (expressed as mean ± SEM) counted per replicate: Control (2291 ± 613), 5dpDT (1110 ± 455), 15dpDT (404 ± 96) from 3 biological replicates