ICAM-1 on the luminal surface of endothelial cells is induced to a greater extent in mouse retina than in other tissues in diabetes

Abstract

Aims/hypothesis: Induction of intercellular adhesion molecule-1 (ICAM-1) has been implicated in the development of macrovascular and microvascular diseases such as diabetic retinopathy. Lesions of diabetic retinopathy are unique to the retina but the reason for this is unclear, as all tissues are exposed to the same hyperglycaemic insult. We tested whether diabetes induces ICAM-1 on the luminal surface of endothelial cells to a greater extent in the retina than in other tissues and the role of vision itself in that induction.

Methods: Experimental diabetes was induced in C57Bl/6J, P23H opsin mutant and Gnat1^-/- × Gnat2^-/- double knockout mice using streptozotocin. The relative abundance of ICAM-1 on the luminal surface of endothelial cells in retina and other tissues was determined by conjugating anti-ICAM-1 antibodies to fluorescent microspheres (2 μm), injecting them intravenously and allowing them to circulate for 30 min. After transcardial perfusion, quantification of microspheres adherent to the endothelium in tissues throughout the body was carried out by fluorescent microscopy or flow cytometry. Mice injected with lipopolysaccharide (LPS) were used as positive controls. The difference in leucostasis between retinal and non-retinal vasculature was evaluated.

Results: Diabetes significantly increased ICAM-1-mediated adherence of microspheres to retinal microvessels by almost threefold, independent of sex. In contrast, diabetes had a much smaller effect on endothelial ICAM-1 in other tissues, and more tissues showed a significant induction of endothelial ICAM-1 with LPS than with diabetes. The diabetes-induced increase in endothelial ICAM-1 in retinal vasculature was inhibited by blocking phototransduction in photoreceptor cells. Diabetes significantly increased leucostasis in the retina by threefold compared with a non-ocular tissue (cremaster).

Conclusions/interpretation: The diabetes-induced upregulation of ICAM-1 on the luminal surface of the vascular endothelium varies considerably among tissues and is highest in the retina. Induction of ICAM-1 on retinal vascular endothelial cells in diabetes is influenced by vision-related processes in photoreceptor cells. The unique presence of photoreceptors in the retina might contribute to the greater susceptibility of this tissue to vascular disease in diabetes.

Keywords: Adhesion molecule; Diabetes; Diabetic retinopathy; ICAM-1; Leucostasis; Photoreceptors; Phototransduction; Vascular inflammation.

Fig. 1

Quantification of fluorescent microspheres by flow cytometry. Gating was established using microspheres of different sizes for comparison (a). Microspheres were distinguished from cells and microparticles in forward scatter vs side scatter density plots (b) and side scatter vs fluorescence density plots (c). The number of microspheres analysed was obtained from the corresponding microspheres’ histogram gate in (c) as a reason of the fluorescent channel in (d)

Fig. 2

Microsphere concentration and validation. Increasing the number of injected microspheres had no effect on the number of microspheres binding to the vascular endothelium, even after LPS challenge, as shown in the retina (a) and liver (b). (c) Data represent the LPS-induced fold increase in number of adherent microspheres above that in non-diabetic animals in the vasculature of various tissues. Non-diabetic animals showed an overall increment in adherent microspheres with LPS challenge, with a statistically significant difference in the retina, aorta, diaphragm, spleen and lung. Statistical analysis was performed using an unpaired t test with Welch correction for each tissue analysed. Data are mean ± SD; n=6–11 mice. p≤0.05 was considered significant. **^,§§p≤0.01 and ***^{,†††,‡‡‡}p≤0.001 compared with WT non-diabetic mice

Fig. 3

Effect of diabetes on the levels of ICAM-1 on the luminal surface of vascular endothelium throughout the body. Data represent the diabetes-induced fold increase in number of anti-ICAM-1-coated microspheres above that in the same tissue of non-diabetic animals. After 2 months of diabetes, only retina showed a statistically significant induction of ICAM-1 on the vascular surface in male and female mice. Statistical analysis was performed using an unpaired t test with Welch correction for each tissue analysed. Data are mean ± SD; n=11–18 mice. p≤0.05 was considered significant. ***p ≤0.001 compared with WT non-diabetic mice

Fig. 4

Phototransduction inhibition in rods and cones and induction of ICAM-1 in the retinal vascular endothelium. Genetic inhibition of phototransduction (Gnat1^−/− × Gnat2^−/− double knockout mice) or 48h of dark adaptation of WT animals inhibited the diabetes-induced upregulation of ICAM-1 on the luminal surface of the retinal endothelium in male and female mice. Statistical analysis was performed using one-way ANOVA with an uncorrected Fisher’s test. Data are mean ± SD; n=8–16 mice. p≤0.05 is considered significant. ^‡‡p ≤ 0.01 and ***^,†††p ≤ 0.001 as shown. D, diabetic; N, non-diabetic

Fig. 5

Induction of ICAM-1 in the retinal vascular endothelium of P23H opsin mutant mice. The expression of a mutant opsin (P23H) is sufficient to inhibit the diabetes-induced upregulation of ICAM-1 on the luminal surface of the retinal endothelium in male mice. Statistical analysis was performed using one-way ANOVA with an uncorrected Fisher’s test. Data are mean ± SD; n=7–10. p≤0.05 is considered significant. †p≤0.05, **p ≤ 0.01 as shown. D, diabetic; N, non-diabetic

Fig. 6

The number of leucocytes adherent to the vascular endothelium is consistent with the effect of diabetes on ICAM-1 expression. Leucostasis is greater in the retinal vasculature (a, c, e) than in the cremaster vasculature (a, d, f) of animals with diabetes of 2 months’ duration. The induction of ICAM-1 in the retinal vasculature compared with the cremaster vasculature correlates with the number of adherent microspheres (b). (c, e) Representative images of retinal leucostasis (white arrows); (e) is a magnification of (c). (d, f) Representative images of cremaster leucostasis (white arrows); (f) is a magnification of (d). Scale bars are 100μm. Statistical analysis was performed using one-way ANOVA with an uncorrected Fisher’s test. Data are mean ± SD; n=6–7. p≤0.05 is considered significant. *^,§p≤0.05 and ^{††,‡‡}p ≤ 0.01 as shown. D, diabetic; N, non-diabetic