Pancreatic Islet Production of Vascular Endothelial Growth Factor-A Is Essential for Islet Vascularization, Revascularization, and Function

Development of islet vasculature and establishment of islet blood flow occur concomitantly with islet formation. A–C: Colocalization of VEGF-A (red) and insulin (Ins; green). D: Colocalization of VEGFR2 (R2; red) and insulin (Ins; green). E–G: Colocalization of Ang-1 (red) and insulin (Ins; green). H: Colocalization of Tie2 (red) and glucagon (Glu; green). I–N: Blood flow in developing pancreas. Embryo at e16.5 was infused with 5 μl endothelium-binding tomato lectin (TL). I: Photograph of dissected digestive organs at e16.5. Black solid line shows boundaries of the pancreas (D, duodenum; P, pancreas; S, spleen; St, stomach). J: Micrograph of 10-μm DAPI-counterstained (blue) section from the pancreatic area marked by blue dotted rectangle (I). Lectin (TL; green) is detected in large blood vessels with visible presence of erythrocytes (I) and also in microvascular structures marked by two white dotted rectangles (arrows point to the panels showing enlargement of these two areas). Lectin⁺ microvascular structures are mostly found to be associated with endocrine cells. K–M: Pancreatic section with lectin-labeled vasculature (TL; green) (K) was costained subsequently with antibodies to VEGFR2 (R2; red) (L) and cocktail of antibodies to three islet hormones (insulin, glucagon, somatostatin, Endo; blue) (M). M: Overlay of lectin label with staining for VEGFR2 (L) and endocrine cells (Endo; blue). Coalescing endocrine cells (blue) are adjacent to the vessels with blood flow (vascular structures double positive for lectin [green] and VEGFR2 [red]). N: Colocalization of insulin (Ins; green), glucagon (Glu; red), and somatostatin (Som; blue) reveals distribution of individual endocrine cells in the developing islets (the same endocrine clusters as in M). O: Model of pancreatic islet vascularization. Bar in A–N represents 50 μm.

Reduced VEGF-A production by β-cells results in abnormal islet vasculature. A: VEGF-A production by islets from VEGF^fl/fl (white), Rip-Cre;VEGF^fl/wt (gray), and Rip-Cre;VEGF^fl/fl (black) mice. *P < 0.05, **P < 0.001 compared with VEGF^fl/fl mice; †P < 0.05 compared with Rip-Cre;VEGF^fl/wt mice. B: Colocalization for insulin (green), glucagon (red), and somatostatin (blue) reveals normal islet morphology in mice with β-cell–reduced VEGF-A expression. C: Phase contrast images of islets in the pancreas (top panels) and corresponding micrographs of pancreatic vasculature labeled with fluorescein isothiocyanate–conjugated tomato lectin (bottom panels; dotted yellow line marks islet boundaries). D: 3D reconstructed optical sections through VEGF^fl/fl (top panel) and Rip-Cre;VEGF^fl/fl (bottom panel) pancreas (30-μm thick histological sections) demonstrate reduced number of vessels and reduced vessel size in VEGF-A–deficient islet (bottom panel). PECAM-1 (red) is colocalized with somatostatin (blue); dotted white line in both panels marks islet boundaries. Vessel density in Rip-Cre;VEGF^fl/fl islet is higher around perimeter where VEGF-A production is maintained by non–β-cells. E: Vessel density in islets declines progressively with VEGF-A reduction. VEGF^fl/fl (white), Rip-Cre;VEGF^fl/wt (gray), and Rip-Cre;VEGF^fl/fl (black) mice. **P < 0.001 compared with VEGF^fl/fl mice; †P < 0.001 compared with Rip-Cre;VEGF^fl/wt mice. Vessel density in exocrine tissue was normal across all three genotypes; 597 ± 22, 576 ± 19, and 584 ± 24 count/mm² in VEGF^fl/fl, Rip-Cre;VEGF^fl/wt, and Rip-Cre;VEGF^fl/fl mice, respectively. F: Area per vessel indicates similar reduction of vessel size and/or branching in Rip-Cre;VEGF^fl/wt (gray) and Rip-Cre;VEGF^fl/fl (black) islets compared with VEGF^fl/fl (white) mice. **P < 0.001 compared with VEGF^fl/fl mice. Rip-Cre;VEGF^fl/wt and Rip-Cre;VEGF^fl/fl mice were not statistically different. Area per vessel in exocrine tissue was normal across all three genotypes; 50 ± 3, 47 ± 2, and 48 ± 2 μm² in VEGF^fl/fl, Rip-Cre;VEGF^fl/wt, and Rip-Cre;VEGF^fl/fl mice, respectively. Bar in B–D represents 50 μm.

Mice with reduced production of VEGF-A by β-cells exhibit impaired glucose tolerance. Fasted male (A) and female (B) VEGF^fl/fl (○), Rip-Cre;VEGF^fl/wt (□), and Rip-Cre;VEGF^fl/fl (▪) mice at 16 weeks of age (n = 10−12 per genotype in both male and female groups of mice) received glucose by intraperitoneal injection (2 g/kg of body wt). *P < 0.05 when Rip-Cre;VEGF^fl/wt and Rip-Cre;VEGF^fl/fl mice compared with VEGF^fl/fl mice. C: Plasma insulin levels (males and females combined) normalized for blood glucose concentration, and 15 min intraperitoneal glucose tolerance test are reduced in Rip-Cre;VEGF^fl/wt (gray, n = 10) and Rip-Cre;VEGF^fl/fl (black, n = 18) mice compared with VEGF^fl/fl mice (white, n = 23). **P < 0.01 compared with VEGF^fl/fl mice. Rip-Cre;VEGF^fl/wt and Rip-Cre;VEGF^fl/fl mice were not statistically different. D: Pancreatic insulin content (males and females combined) is similar in VEGF^fl/fl (white, n = 6), Rip-Cre;VEGF^fl/wt (gray, n = 6), and Rip-Cre;VEGF^fl/fl (black, n = 6) mice.

Reduced VEGF-A expression in islets impairs insulin output into vascular system. A: Insulin secretory response of isolated islets in cell perifusion system. Insulin secretion of islets isolated from VEGF^fl/fl (○), Rip-Cre;VEGF^fl/wt (□), and Rip-Cre;VEGF^fl/fl (•) mice was analyzed in response to glucose and isobutylmethylxanthine (IBMX). The integrated response to 16.7 mmol/l glucose was 80.4 ± 24.4 ng insulin in VEGF^fl/fl mice vs. 80.9 ± 29.5 ng insulin in Rip-Cre;VEGF^fl/wt mice and 72.9 ± 9.6 ng insulin in Rip-Cre;VEGF^fl/fl mice (n = 3; P = 0.963). The integrated response to 45 μmol/l isobutylmethylxanthine in the presence of 16.7 mmol/l glucose was 418 ± 51 ng insulin in VEGF^fl/fl mice vs. 443 ± 62 ng insulin in Rip-Cre;VEGF^fl/wt mice and 383 ± 32 ng insulin in Rip-Cre;VEGF^fl/fl mice (n = 3; P = 0.705). B: Insulin secretory response of the in situ perfused pancreas. Insulin secretion from the pancreas of Rip-Cre;VEGF^fl/fl mice (•) and their wild-type littermates (VEGF^fl/fl; ○) was analyzed in response to glucose, isobutylmethylxanthine (IBMX), and arginine. The integrated response to 16.7 mmol/l glucose was 120 ± 17 ng insulin in VEGF^fl/fl mice vs. 75.1 ± 15.2 ng insulin in Rip-Cre;VEGF^fl/fl mice (n = 5; P = 0.043). The integrated response to 45 μmol/l isobutylmethylxanthine in the presence of 16.7 mmol/l glucose was 136 ± 27 ng insulin in VEGF^fl/fl mice vs. 74.0 ± 12.9 ng insulin in Rip-Cre;VEGF^fl/fl mice (n = 5; P = 0.0356). The integrated response to 20 mmol/l arginine in the presence of 16.7 mmol/l glucose was 877 ± 151 ng insulin in VEGF^fl/fl mice vs. 469 ± 63 ng insulin in Rip-Cre;VEGF^fl/fl mice (n = 5; P = 0.015). C: Ultrastructural changes to intra-islet capillaries in islets with diminished VEGF-A levels. Transmission electron microscopy images show intra-islet endothelial cells with adjacent islet cells of wild-type (left panel), Rip-Cre;VEGF^fl/wt (middle panel), and Rip-Cre;VEGF^fl/fl (right panel) mice. Arrows point to fenestrations and caveolae. Images are representative, taken from either group. BM, basement membrane; C, caveolae; F, fenestration; L, capillary lumen; SG, secretory granule; 40,000× magnification. Bar in C represents 500 nm.