Activation of the Wnt pathway plays a pathogenic role in diabetic retinopathy in humans and animal models

Abstract

Although Wnt signaling is known to mediate multiple biological and pathological processes, its association with diabetic retinopathy (DR) has not been established. Here we show that retinal levels and nuclear translocation of beta-catenin, a key effector in the canonical Wnt pathway, were increased in humans with DR and in three DR models. Retinal levels of low-density lipoprotein receptor-related proteins 5 and 6, coreceptors of Wnts, were also elevated in the DR models. The high glucose-induced activation of beta-catenin was attenuated by aminoguanidine, suggesting that oxidative stress is a direct cause for the Wnt pathway activation in diabetes. Indeed, Dickkopf homolog 1, a specific inhibitor of the Wnt pathway, ameliorated retinal inflammation, vascular leakage, and retinal neovascularization in the DR models. Dickkopf homolog 1 also blocked the generation of reactive oxygen species induced by high glucose, suggesting that Wnt signaling contributes to the oxidative stress in diabetes. These observations indicate that the Wnt pathway plays a pathogenic role in DR and represents a novel therapeutic target.

Figure 1

Activated Wnt signaling in the human retina with DR. Retinal sections from five non-DM donors and six diabetic donors with NPDR were immunostained with an antibody for β-catenin. The signal was developed with the diaminobenzidine method (brown color). Representative retinal images from two non-DM (A and B) and two NPDR donors (C and D) showed more intensive signals of β-catenin in the inner retina and increased β-catenin staining in the nuclei of the retinal cells from the DM-NPDR donors, compared with that from the non-DM subjects. Scale bar = 20 μm. E: β-catenin signal was quantified by using morphometric analysis software and expressed as arbitrary units (mean ± SD). **P < 0.01.

Figure 2

Increased β-catenin levels in the retinas of Akita mice, STZ-induced diabetic rats, and OIR rats. The retinas from Akita mice at 16 weeks of age, STZ-DM rats at 16 weeks following the STZ injection, OIR rats at the age of P16, and age-matched nondiabetic or normoxic controls were used for Western blot (A–C) and immunohistochemistry (D–Q) analyses of β-catenin. A–C: The same amount (50 μg) of retinal proteins from each animal was blotted with an antibody specific for β-catenin. The membranes were stripped and reblotted with an antibody for β-actin. Each lane represents an individual animal. D–Q: Representative retinal sections from Akita mice (G–I) and their nondiabetic littermates (D–F), STZ-DM rats (M–O) and non-DM rats (J–L), OIR rats (Q), and age-matched normal rats maintained under constant normoxia (P) were immunostained with an antibody for β-catenin. F, I, L, and O: The nucleus was counterstained with 4’, 6-diamidino-2-phenylindole (DAPI) (colored red) and merged with β-catenin signal. Red arrows (in I and O) indicate the nuclei showing green or orange color as a result of increased β-catenin signal in the nuclei of diabetic retinas, while the white arrows (in F and L) point to nuclei (red color) in nondiabetic retinas. Scale bar = 20 μm.

Figure 3

Up-regulated expression of LRP5/6 in the retinas of STZ-diabetic and OIR rats. A and B: The same amount of retinal proteins (100 μg) from three STZ-induced diabetic rats 16 weeks after the onset of diabetes and age-matched nondiabetic rats (A), and four OIR rats and normal rats at age of P16 (B) was used for Western blot analysis using an antibody specific for LRP5/6 (Santa Cruz Biotechnology). The same membranes were stripped and reblotted with an antibody for β-actin. C–F: Retinal sections from STZ-diabetic rats (D) and non-DM controls (C), and those from OIR rats (F) and their normoxic controls (E) were immunostained with the antibody against LRP5/6 (green). The nuclei were counterstained with DAPI (red). Original magnification, ×400.

Figure 4

Induction of Wnt signaling by hypoxia and oxidative stress. A: RCEC were exposed to 2% oxygen and normoxia for 14 hours. Levels of total β-catenin were determined by Western blot analysis using the same amount (50 μg) of total proteins from each sample and normalized to β-actin levels. Note that the blot represents two independent experiments. B–D: RCEC were treated with low glucose (LG; 5 mmol/L glucose and 25 mmol/L mannitol, B), high glucose (HG, 30 mmol/L glucose, C), and high glucose plus 10 μmol/L aminoguanidine (HG+AG, D) for 24 hours. The subcellular distribution of β-catenin was revealed by immunocytochemistry by using the antibody for β-catenin. E: The same amount of isolated nuclear proteins (50 μg) from each of the above groups was blotted with an antibody for β-catenin and normalized to TATA box-binding protein (TBP) levels.

Figure 5

DKK1 ameliorates retinal inflammation, vascular leakage, and NV, and inhibits ROS generation. A: Various doses of purified DKK1 were injected into the vitreous of the right eye of STZ-diabetic rats at 16 weeks following the onset of diabetes, and the same amounts of BSA were injected into the contralateral eyes for controls. Soluble ICAM-1 concentrations in the retina were measured by enzyme-linked immunosorbent assay, normalized by total protein concentrations, and expressed as ng per mg of proteins (means ± SD, n = 3). B: Purified DKK1 was injected into the vitreous of the right eye (1.2 μg/eye) of STZ-diabetic rats at 16 weeks following the onset of diabetes, and the same amounts of BSA were injected into the contralateral eyes for controls. Retinal vascular leakage was measured 48 hours after the injection by using Evans blue as a tracer, normalized by total protein concentrations, and expressed as μg of Evans blue per mg of retinal proteins (means ± SD, n = 4). C and D: At the age of P14, the OIR rats received an intravitreal injection of DKK1 (1 μg/eye) into the right eye and the same amount of BSA into the contralateral eyes. The retinas were harvested at P16, and the same amount of retinal proteins (20 μg) was loaded for Western blot analysis by using antibodies specific for COX2 (C) and VEGF (D), and normalized by β-actin levels. E: OIR rats at P14 received an intravitreal injection of DKK1 at doses as indicated. Retinal vascular leakage was measured at P16 by using Evans blue as a tracer, normalized by total protein concentrations, and expressed as μg of Evans blue per mg of retinal proteins (means ± SD, n = 3). *P < 0.05. F–J: OIR rats received an intravitreal injection of 2 μg/eye DKK1, and BSA into the contralateral eyes at age of P14. At P18, retinal vasculature was visualized by fluorescein angiography on the whole-mounted retina from the eyes injected with BSA (F and G) and injected with DKK1 (H and I). Original magnification: ×12.5 (F and H); ×100 (G and I). J: Preretinal vascular cells were counted on cross ocular sections from the eyes injected with BSA and DKK1 (means ± SD, n = 5).

Figure 6

The Wnt pathway contributes to the oxidative stress and HIF-1 activation. A–D: DKK1 inhibits HIF-α activation: Primary RCEC were exposed to 5 mmol/L glucose and 25 mmol/L mannitol (A), 30 mmol/L glucose (B), TNF-α (C), and 30 mmol/L glucose with 1 μg/ml DKK1 (D) for 4 hours. HIF-1α nuclear translocation was determined by using immunocytochemistry with an anti-HIF-1α antibody. Scale bar = 50 μm. E: RCEC were exposed to low glucose (5 mmol/L glucose plus 25 mmol/L mannitol) or high glucose (30 mmol/L glucose), or 1 μg/ml TNF-α in the absence or presence of various concentrations of DKK1 (6.25 to 100 nmol/L). Aminoguanidine (AG; 10 μmol/L) was used as a positive control. Intracellular ROS generation was measured and expressed as fluorescent unit per well (means ± SD, n = 3).