Betacellulin induces increased retinal vascular permeability in mice

Abstract

Background: Diabetic maculopathy, the leading cause of vision loss in patients with type 2 diabetes, is characterized by hyper-permeability of retinal blood vessels with subsequent formation of macular edema and hard exudates. The degree of hyperglycemia and duration of diabetes have been suggested to be good predictors of retinal complications. Intervention studies have determined that while intensive treatment of diabetes reduced the development of proliferative diabetic retinopathy it was associated with a two to three-fold increased risk of severe hypoglycemia. Thus we hypothesized the need to identify downstream glycemic targets, which induce retinal vascular permeability that could be targeted therapeutically without the additional risks associated with intensive treatment of the hyperglycemia. Betacellulin is a 32 kD member of the epidermal growth factor family with mitogenic properties for the retinal pigment epithelial cells. This led us to hypothesize a role for betacellulin in the retinal vascular complications associated with diabetes.

Methods and findings: In this study, using a mouse model of diabetes, we demonstrate that diabetic mice have accentuated retinal vascular permeability with a concomitant increased expression of a cleaved soluble form of betacellulin (s-Btc) in the retina. Intravitreal injection of soluble betacellulin induced retinal vascular permeability in normoglycemic and hyperglycemic mice. Western blot analysis of retinas from patients with diabetic retinopathy showed an increase in the active soluble form of betacellulin. In addition, an increase in the levels of A disintegrin and metalloproteinase (ADAM)-10 which plays a role in the cleavage of betacellulin was seen in the retinas of diabetic mice and humans.

Conclusions: These results suggest that excessive amounts of betacellulin in the retina may contribute to the pathogenesis of diabetic macular edema.

Figure 1. Diabetic mice show augmented retinal vascular permeability and increased expression of cleaved betacellulin in the retina.

Leakage of Evans Blue dye from mouse retinal vessels in representative a) normoglycemic and b) hyperglycemic mice was quantitated in whole mount retinas one hr after perfusion with the dye(n = 8). (c). Representative Western blot analysis of betacellulin in pancreas (d) and retinas (e) from control and streptozotocin-induced diabetic mice. (f) densitometric quantitation of the ratio of sBTC to ProBTC in mouse retina from control (n = 12) and diabetic (n = 12) mouse eyes. ***p = 0.008 (ANOVA).

Figure 2. Intravitreal injection of betacellulin induces increased retinal haemorrhage and augmented retinal vascular permeability in normal and diabetic retinas.

Dose response of the effect of intravitreal injection of betacellulin (50–300 ng), VEGF (200 ng) and PBS in normoglycemic mice on a) retinal haemorrhage as seen in a bright field image of dissected posterior pole of mouse eyes. Arrowheads indicate areas of hemorrhage. For BTC-300, inset is magnification of boxed area b) Evans Blue dye leakage in flat mounted retinas. Leakage of Evans blue dye from retinal vessels of normoglycemic (c,d) and hyperglycemic (e,f) mice injected intravitreally with PBS(c,e) or 200 ng recombinant mouse betacellulin (d,f). Results are quantitated (g). * p<0.009.

Figure 3. Increased expression of betacellulin in retinas of humans with diabetes.

Western blot analysis of (a) BTC (b) GAPDH in retina tissue from human donors with no diabetes, diabetes or diabetes with diabetic retinopathy. (c) Densitometric quantitation of the ration of sBTC to full length BTC using GAPDH as a reference. ** p = 0.01 * p = 0.03.

Figure 4. Increased expression of ADAM-10 in diabetic retinas.

Western blot analysis of ADAM-10 in retina tissue from (a) control and STZ induced diabetic mice (b) control and intravitreal VEGF injected mice and (c) human retinas from patients with no diabetes, diabetes or diabetes with diabetic retinopathy(d) Densitometric quantitation of mature ADAM-10 (mADAM-10) relative to actin on the same blot. ** p≤0.03.