GLP-1 receptor signaling increases PCSK1 and β cell features in human α cells

Abstract

Glucagon-like peptide-1 (GLP-1) is an incretin hormone that potentiates glucose-stimulated insulin secretion. GLP-1 is classically produced by gut L cells; however, under certain circumstances α cells can express the prohormone convertase required for proglucagon processing to GLP-1, prohormone convertase 1/3 (PC1/3), and can produce GLP-1. However, the mechanisms through which this occurs are poorly defined. Understanding the mechanisms by which α cell PC1/3 expression can be activated may reveal new targets for diabetes treatment. Here, we demonstrate that the GLP-1 receptor (GLP-1R) agonist, liraglutide, increased α cell GLP-1 expression in a β cell GLP-1R-dependent manner. We demonstrate that this effect of liraglutide was translationally relevant in human islets through application of a new scRNA-seq technology, DART-Seq. We found that the effect of liraglutide to increase α cell PC1/3 mRNA expression occurred in a subcluster of α cells and was associated with increased expression of other β cell-like genes, which we confirmed by IHC. Finally, we found that the effect of liraglutide to increase bihormonal insulin+ glucagon+ cells was mediated by the β cell GLP-1R in mice. Together, our data validate a high-sensitivity method for scRNA-seq in human islets and identify a potentially novel GLP-1-mediated pathway regulating human α cell function.

Keywords: Diabetes; Endocrinology; Islet cells; Metabolism.

Figure 1. β Cell GLP-1R signaling increases islet GLP-1 production and PC1/3 expression.

(A) Schematic of the mouse study design. (B) Representative images of mouse pancreas sections immunostained for GLP-1 (red), glucagon (green), and DAPI in WT and KO mice treated with saline (CTRL) or liraglutide (LIRA). (C) Average GLP-1 staining per islet. (D) Representative images of pancreas sections immunostained for glucagon (red), PC1/3 (green), and DAPI. Examples of glucagon and PC1/3 colocalization are indicated by white arrows. (E) PC1/3 colocalization with glucagon. (F) Representative images of pancreas sections immunostained for glucagon (green) and DAPI. (G) Percentage of islets with centrally located α cells. Data are presented as mean ± SEM. n = 6 per group. **P < 0.01, ***P < 0.001 CTRL WT vs. LIRA WT, ^###P < 0.001 LIRA WT vs. LIRA KO by 2-factor ANOVA. ⁺P < 0.05 CTRL WT vs. CTRL KO by 2-tailed Student’s t test. Scale bar: 20 μm. See also Supplemental Figures 1 and 6.

Figure 2. DART-Seq improves detection of low-abundance transcripts in human islets.

(A) Schematic of DART-Seq modification of the DROP-Seq nano bead. (B) Box plot showing PCSK1 reads per cell detected by DROP-Seq and DART-Seq. (C) Expression level of INS (red), GCG (green), and PCSK1 (purple) projected on the UMAP generated from human islet cells using DROP-Seq and DART-Seq, respectively. n = 1 per group. See also Supplemental Figure 2 and Supplemental Table 1.

Figure 3. DART-Seq assessment of human islets treated with saline or a GLP-1R agonist.

(A) Expression level of INS (red), GCG (green), and SST (blue) projected on the UMAP generated from human islets treated with saline (CTRL) or liraglutide (LIRA). (B) Percentage of α, β, and δ cells in saline- and liraglutide-treated islets from the 3 individual donors (D1–D3) and in the combined data set (CD). (C) UMAP projection of all endocrine cells with respective subclusters in saline- and liraglutide-treated islets. (D) Dot plots showing the expression level of key identity genes in α cell (right) and β cell (left) subclusters. (E) Pseudotime trajectory of α and β cell subclusters showing relative expression of key identity genes, ARX, DNMT1, MAFA, and PDX1. (F) Percentage of α and β cell subclusters in saline- and liraglutide-treated islets. Percentage calculated using combined data set. n = 3 per group. See also Supplemental Figure 3 and Supplemental Table 2.

Figure 4. A GLP-1R agonist increases PCSK1 expression and markers of β cell fate in a subcluster of α cells.

(A) Expression levels of PCSK1, IAPP, and MAFA projected on the UMAP generated from human islet cells treated with saline (CTRL) or liraglutide (LIRA). (B) Volcano plot showing genes that are differentially regulated in the α-1 subcluster upon liraglutide treatment. (C) GO term analysis showing the biological processes enriched in the α-1 subcluster. (D) Different representation of data presented in Figure 3C. Left: Saline- (CTRL, red) and liraglutide-treated (blue) islet cells projected on the UMAP and colored by treatment. Right: Subclusters of α and β cells in the liraglutide sample with the arrow indicating proposed direction of change. (E) Volcano plot showing genes that are differentially regulated in the β-1 subcluster upon liraglutide treatment. n = 3 per group. See also Supplemental Figure 4.

Figure 5. A GLP-1R agonist increases active GLP-1 production and expression and bihormonal insulin⁺ glucagon⁺ cells in human islets.

(A) Average GLP-1 staining per islet. (B) Representative images of human islets at baseline and after 24 hours of saline or liraglutide treatment. Islets were immunostained for GLP-1 (red), glucagon (green), and DAPI. (C) Insulin and glucagon colocalization in the islet. (D) Islets immunostained for insulin (red), glucagon (green), and DAPI. Examples of bihormonal insulin⁺ glucagon⁺ cells are indicated by white arrows. (E) Active GLP-1 concentrations measured in lysate of human islets following 24 hours of saline or liraglutide treatment. Scale bar: 20 μm. n = 4–8 per group. *P < 0.05 by Student’s t test. See also Supplemental Figure 5.

Figure 6. β Cell GLP-1R signaling increases bihormonal insulin⁺ glucagon⁺ islet cells in mice.

(A) Representative images of pancreas sections immunostained for insulin (red), glucagon (green), and DAPI in β cell GLP-1R WT and KO mice treated with saline (CTRL) or liraglutide (LIRA). Examples of bihormonal insulin⁺ glucagon⁺ cells indicated by white arrows. (B) Percentage of bihormonal insulin⁺ glucagon⁺ cells per islet. (C) Percentage of centrally located bihormonal insulin⁺ glucagon⁺ cells per islet. Data are presented as mean ± SEM. n = 6 per group. **P < 0.01, ***P < 0.001 CTRL WT vs. LIRA WT, ^##P < 0.01, ^###P < 0.001 LIRA WT vs. LIRA KO by 2-factor ANOVA. Scale bar: 20 μm.