Hydrogen Sulfide Is a Novel Protector of the Retinal Glycocalyx and Endothelial Permeability Barrier

Abstract

Significantly reduced levels of the anti-inflammatory gaseous transmitter hydrogen sulfide (H₂S) are observed in diabetic patients and correlate with microvascular dysfunction. H₂S may protect the microvasculature by preventing loss of the endothelial glycocalyx. We tested the hypothesis that H₂S could prevent or treat retinal microvascular endothelial dysfunction in diabetes. Bovine retinal endothelial cells (BRECs) were exposed to normal (NG, 5.5 mmol/L) or high glucose (HG, 25 mmol/L) ± the slow-release H₂S donor NaGYY4137 in vitro. Glycocalyx coverage (stained with WGA-FITC) and calcein-labeled monocyte adherence were measured. In vivo, fundus fluorescein angiography (FFA) was performed in normal and streptozotocin-induced (STZ) diabetic rats. Animals received intraocular injection of NaGYY4137 (1 μM) or the mitochondrial-targeted H₂S donor AP39 (100 nM) simultaneously with STZ (prevention) or on day 6 after STZ (treatment), and the ratio of interstitial to vascular fluorescence was used to estimate apparent permeability. NaGYY4137 prevented HG-induced loss of BREC glycocalyx, increased monocyte binding to BRECs (p ≤ 0.001), and increased overall glycocalyx coverage (p ≤ 0.001). In rats, the STZ-induced increase in apparent retinal vascular permeability (p ≤ 0.01) was significantly prevented by pre-treatment with NaGYY4137 and AP39 (p < 0.05) and stabilized by their post-STZ administration. NaGYY4137 also reduced the number of acellular capillaries (collagen IV + /IB4-) in the diabetic retina in both groups (p ≤ 0.05). We conclude that NaGYY4137 and AP39 protected the retinal glycocalyx and endothelial permeability barrier from diabetes-associated loss of integrity and reduced the progression of diabetic retinopathy (DR). Hydrogen sulfide donors that target the glycocalyx may therefore be a therapeutic candidate for DR.

Keywords: diabetes; glycocalyx; hydrogen sulfide; inflammation; mitochondria; retinal permeability; slow-release hydrogen sulfide donors.

FIGURE 1

Retinal permeability analysis model. (A) Fundus fluorescein angiography was performed in non- and diabetic Norway Brown rats treated ± H₂S donors, on days 0 and 7. The angiograms shown are representative of a non-diabetic animal. (B) The mean intensity of NaF in the retinal tissue (Adamis and Berman, 2008) and a main retinal vessel (Ahmad et al., 2016) were measured every 200 frames for 3 min. (C) Apparent permeability (Permeability.Surface Area product, PA) was calculated from the rate of change of extravascular fluorescence (= ΔI_f/Δt) per unit concentration difference (ΔC) between blood and tissue. P = ΔI_f/Δt/(ΔC x A). ΔI_f/Δt ∼ slope. ΔC = difference between intra- and extravascular intensity.

FIGURE 2

Administration of the slow-release H₂S donor NaGYY4137 completely reversed the loss of retinal EC glycocalyx integrity induced by hyperglycemia. Retinal ECs were incubated with normal (5.5 mmol/L, NG) or high glucose (25 mmol/L, HG) ± GYY4137 (500 μmol/L) for 24 h before assessment of glycocalyx integrity with fluorescently labeled lectin (WGA-FITC). Data are mean ± SD, presented as fluorescence intensity/μg protein. *p ≤ 0.01 vs. control, **p ≤ 0.01 vs. HG, n = 12–30. One-way ANOVA.

FIGURE 3

Administration of the slow-release H₂S donor NaGYY4137 completely attenuated the increased leukocyte adhesion to retinal ECs induced by hyperglycemia. Retinal ECs were incubated with normal (5.5 mmol/L, NG) or high glucose (25 mmol/L, HG) ± NaGYY4137 (500 μmol/L) for 24 h before examination of leukocyte adhesion under flow (1 dyne/cm²). (A) Number of adherent U937 cells. Data are mean ± SD. *p ≤ 0.001 vs. control, **p ≤ 0.001 vs. HG, n = 4. One-way ANOVA. (B) Representative bright field, phase contrast (upper panel) and corresponding fluorescence images (lower panel) of adherent U937 cells stained with calcein acetoxymethyl ester adhering to a BREC monolayer. Scale bar, 75 μm; original magnification, ×100.

FIGURE 4

Chemically induced diabetes significantly increased retinal vascular permeability in Norway Brown rats compared to non-diabetic controls. (A) Fundus fluorescein angiography measurements from a single rat carried out on day 0 (before STZ) and day 7 (1 week of STZ treatment). (B) Permeability was measured from control (blood glucose 7.99 ± 0.21 mmol/L), and (C) diabetic (blood glucose 29.75 ± 1.63 mmol/L) Norway Brown rats on days 0, 7, 14, 21, and 28. Apparent permeability was calculated as described in the methodology text and depicted in Figure 1. Data including mean ± SEM. *p ≤ 0.001 vs. day 0, **p ≤ 0.0001 vs. diabetic day 0. One-way ANOVA.

FIGURE 5

Ocular administration of the slow-release H₂S donors reduced retinal permeability in chemically induced diabetic rats. Type I diabetes was successfully induced in rats following a single dose of STZ and animals were maintained for 1 week without insulin supplementation. Retinal vascular leak was measured on day 0 and day 7 in non- and diabetic animals following treatment. No observed increases in retinal permeability were measured on day 7 in non-diabetic animals treated with saline (A,C). In diabetic animals treated with saline a significant (*p ≤ 0.05) increase in retinal permeability was measured on day 7 vs. day 0 (B,C). NaGYY4137 and AP39 significantly (*p < 0.05, paired t-test) reduced retinal permeability in non- and diabetic rats (D,G,F) on day 7 when given as a single intraocular preventative dose, on day 0. In addition, NaGYY4137 and AP39 reduced retinal permeability in non-diabetic rats (E,F,H,I) when administered therapeutically (day 6). In diabetic rats, NaGYY4137 and AP39 stabilized retinal permeability on day 7 in rats with pre-existing diabetes (E,F,H,I).

FIGURE 6

NaGYY4137 protected against diabetes-induced increase in acellular capillaries. Norway Brown rats were given a single dose of streptozotocin (50 mg/kg) to induce Type I diabetes. Rats were treated with an intraocular injection of either saline or 1 μmol/L H₂S donor NaGYY4137 at day 0 (prevention) or day 6 (treatment). (A) Rats were enucleated at day 7 and whole retinae were mounted and stained with Hoescht, IB4, and collagen IV. The tissue was imaged using 20 × lens. The number of capillaries positively stained for collagen IV but lacking IB4 (white arrows) were counted in three areas of the retina. (B) Acellular capillaries are expressed per mm². Error bars represent standard errors of the mean. Statistical analysis was performed using a one-way analysis of variance corrected for false discovery rate. *p ≤ 0.05.