Adenosine Receptor Antagonists and Retinal Neovascularization in Vivo

Abstract

purpose. The role of adenosine receptor (AdoR) antagonists in human retinal endothelial cell function in vitro has previously been determined. In this study, efficacy of AdoR antagonist administration in reducing retinal neovascularization was examined in a mouse pup model of oxygen-induced retinopathy.

methods. A previously described model of oxygen-induced retinal neovascularization in newborn mouse pups was used to examine the effect of various AdoR antagonists on neovascularization. The nonselective AdoR antagonist xanthine amine congener (XAC), the A_2A-selective antagonist ZM241385, the A_2B-selective antagonists 3-N-propylxanthine (enprofylline) and 3-isobutyl-8-pyrrolidinoxanthine (IPDX), and the A₁-selective antagonist cyclopentyl-1,3-dipropylxanthine (CPX) were used. After the hyperoxia exposure the animals received daily intraperitoneal injections of pharmacologically relevant doses of AdoR antagonists for 5 days. Control animals received vehicle (0.1% dimethyl sulfoxide [DMSO]) alone. The animals were then killed and perfused with fluorescein-dextran. Wholemounts of retinas from one eye were prepared and examined, whereas the retinas of the contralateral eye were embedded, sectioned, and stained for counting neovascular nuclei extending beyond the internal limiting membrane into the vitreous.

results. Angiography of wholemount retinas showed reduction of neovascular tufts in animals treated with selective A_2B AdoR antagonists. Quantification of the extraretinal neovascular nuclei showed that only animals treated with XAC, enprofylline, or IPDX showed a significant reduction in retinal neovascularization. By contrast, neither CPX nor ZM241385 had an effect on neovascularization.

conclusions. The A_2B-selective AdoR antagonists inhibited oxygen-induced retinal neovascularization in vivo and may provide a basis for developing pharmacologic therapies for the treatment of proliferative retinopathies.

Vascular eye diseases, such as retinopathy of prematurity (ROP) and proliferative diabetic retinopathy (PDR), are characterized by the abnormal growth of blood vessels across the retina. Although ROP and PDR differ in many respects, it is thought that neovascular growth arises in both diseases as a result of ischemic injury to retinal blood vessels. ¹ Premature infants are exposed to hyperoxia during postnatal management to compensate for pulmonary insufficiency. On their return to normal air, representing a hypoxic state relative to the hyperoxic environment in which they were kept for some time, they often develop abnormal retinal vasculature. ROP usually regresses but may lead to severe visual impairment, retinal detachment, and blindness. ² Diabetes leads to a prothrombotic hematologic state with increased levels of coagulant proteins, hyper aggregable platelets, and poorly deformable red blood cells, all of which can lead to ischemic injury. In addition, leukostasis and anatomic abnormalities such as capillary dilation and acellularity result in further areas of capillary nonperfusion and retinal ischemia, which in turn may lead to neovascularization. Although glycemic control has been implicated in prevention of progression of diabetic retinopathy, ³ the only current available therapy for PDR is photocoagulation of neovascular areas. ⁴  

Tissue hypoxia and ischemia initiate a series of events that lead to the development of collateral blood vessels, ⁵ in a process referred to as compensatory angiogenesis. ⁶ Potential mediators of compensatory angiogenesis include vascular endothelial growth factor (VEGF), ^{7 8} basic fibroblast growth factor (FGF-2), ⁹ insulin-like growth factor (IGF-I), ¹⁰ hepatocyte growth factor-scatter factor (HGF/SF), ¹¹ platelet-derived growth factor (PDGF), ¹² and nucleosides such as adenosine. ¹³  

Adenosine is released in increased amounts in response to an ischemic insult. ¹⁴ In the retinal vasculature adenosine causes vasodilation, contributing to improved oxygenation of tissues ¹⁵ and has protective effects on neuronal cells. ¹⁶ Moreover, adenosine stimulates the proliferation of a variety of cell types including capillary endothelial cells ^{13 17 18} and promotes angiogenesis. ^{19 20} High levels of adenosine are associated with areas of vasculogenesis in the normal neonatal dog retina as well as sites of angiogenesis in a canine model of oxygen-induced retinopathy. ^{21 22}  

Adenosine modulates a variety of cellular functions by interacting with specific adenosine receptors (AdoR) on the cell surface. Four subtypes of AdoRs, termed A₁, A_2A, A_2B, and A₃, have been identified. ²³ A number of antagonists have been developed with differing selectivity for the AdoR subtypes. To date, the absence of a potent, selective A_2B AdoR antagonist has hampered the characterization of the cellular functions modulated by the activation of this receptor subtype. Most studies examining the role of the A_2B AdoR relied on excluding any effect mediated through the other receptor subtypes, using selective antagonists for those receptor subtypes. Recently, a novel, potent selective A_2B AdoR antagonist, 3-isobutyl-8-pyrrolidinoxanthine (IPDX) has been designed and synthesized ²⁴ and was made available to us. 

We have previously reported that cultured human retinal endothelial cells (HRECs) express the A_2B AdoR subtype and that this receptor mediates the mitogenic effects of adenosine. ^{17 18} In the present study, we determine the potential efficacy of administering various AdoR antagonists, especially A_2B-selective antagonists, to reduce the degree of retinal neovascularization observed in the mouse pup model of oxygen-induced retinopathy initially described by Smith et al. ²⁵ In this model, 7-day-old mouse pups are exposed to 75% oxygen for 5 days, after which they are returned to normal air and kept for an additional 5 days. Animals thus treated undergo a hypoxic insult, resulting in retinal vascular diseases, such as vascular engorgement and tortuosity, increased peripheral perfusion, and neovascular tufts. This model thus presents a simple method for testing the in vivo efficacy of systemically administered antiangiogenic agents. 

Materials and Methods

Animals

All animals were treated in accordance with The Guiding Principles in the Care and Use of Animals and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. C57BL/6J mice were obtained from Jackson Laboratory (Bar Harbor, ME). Breeding-age animals were housed in our institutional facilities. Females were examined daily for signs of pregnancy and placed in individual cages when their status was confirmed. Timed-pregnancy animals were occasionally purchased from this vendor, as well. Animals were killed by an overdose of ketamine-xylazine mixture (injections of 10 mg/ml ketamine HCl, 2 mg/ml xylazine in 0.9% NaCl) administered subcutaneously at a dose of 5 μl/g body weight. 

Reagents

Xanthine amine congener (XAC), 8-cyclopentyl-1,3-dipropylxanthine (CPX), 3-N-propylxanthine (enprofylline), and fluorescein isothiocyanate (FITC)-dextran 2,000,000 were purchased from Sigma-Aldrich (St. Louis, MO). ZM241385 and IPDX were provided by two of the authors (LB and IB, respectively). 

Hyperoxia Treatment and Drug Administration

Mouse pups, along with their nursing mothers, were placed in 75% oxygen beginning at postnatal day (P)7 and maintained at this oxygen concentration for a period of 5 days (P12). They were then returned to normal air and maintained for another 5 days (P17). From P12 to P17, the pups received daily intraperitoneal injections of the test agent. Each antagonist was administered at doses of 0.3 μg/kg, 3.0 μg/kg, and 30.0 μg/kg of body weight. Mice from at least three litters were used to test each condition to control for variability in litter size, degree of maternal attention, and nutrition. Vehicle control mice were injected with 0.1% dimethyl sulfoxide (DMSO) in isotonic saline used to dissolve the adenosine antagonists. Normoxia control mice of age identical with the hyperoxia group were maintained at room air but otherwise treated identically. Table 1 itemizes the number of mice used for each condition. At P17 the pups were anesthetized (as described earlier) and perfused through cardiac puncture with 4% formaldehyde in 0.1 M sodium phosphate (pH 7.4) containing 5 mg/ml FITC-dextran. 

Qualitative Assessment of Retinal Neovascularization

The retina from one eye from each P17 pup was dissected and flatmounted as described by D’Amato et al. ²⁶ The flatmounted retinas were examined by fluorescence microscopy and photographed. At least three retinas from each treatment group were examined qualitatively using this methodology (Table 1) . 

Quantitative Assessment of Retinal Neovascularization

The contralateral eye from each P17 pup was embedded in paraffin for serial sectioning as described by Smith et al. ²⁷ Sections were stained with hematoxylin and eosin to visualize cell nuclei. Individuals masked to the identity of treatment counted all cell nuclei above the internal limiting membrane for 10 sections from each eye. Cross sections that included the optic nerve were not sampled because normal vessels emanating from the optic nerve, although distinguishable from neovascularization extending into the vitreous, fulfilled the counting criteria and would have increased the error. Vascular cell nuclei were considered to be associated with new vessels if they were found on the vitreous side of the internal limiting membrane. Pericytes were not morphologically identifiable in the neovascular tufts. It is possible that pericytes or pericyte precursors were included in our cell counts. 

Statistical Analysis

Data were analyzed by one-way analysis of variance (ANOVA) followed by Bonferroni t-test, with either normoxia- and vehicle-treated or hyperoxia- and vehicle-treated used as the determinant as appropriate. Results are reported as mean ± SE. P < 0.05 was deemed significant. 

Results

The pattern of vascular development and neovascularization observed in the FITC-dextran–perfused retinas were similar to those reported in previous studies using this model. ^{25 26} The retinas of animals exposed to hyperoxia showed increased perfusion at the periphery, dilation and tortuosity of radial vessels, and persistent absence of central perfusion. Figure 1 depicts the characteristic results seen in vehicle-treated, hyperoxia-exposed retina (Fig. 1A) and contrasts those to drug-treated, hyperoxia-exposed retina (Figs. 1B 1C 1D 1E 1F) . Notable was the absence of (or marked decrease in) neovascular tufts observed in the groups receiving the nonselective AdoR antagonist XAC (Fig. 1B) and the A_2B-selective antagonists enprofylline and IPDX, (Figs. 1C 1D , respectively). In contrast, the groups receiving the A₁-selective antagonist CPX or the A_2A-selective antagonist ZM241385 (Figs. 1E 1F , respectively) did not appear to have less neovascularization than did the vehicle-treated group. 

Quantitative assessment of cell nuclei above the internal limiting membrane agreed with the qualitative assessment. Exposure to hyperoxia followed by injection with vehicle alone showed an increased number of neovascular nuclei (P < 0.001) compared with normoxia-treated controls injected with vehicle. However, vehicle alone did not affect the extent of neovascularization compared to uninjected hyperoxia-treated animals (Fig. 2) . The nonselective AdoR antagonist XAC, as well as the A_2B-selective antagonists enprofylline and IPDX reduced the neovascular response (P < 0.001). In contrast, neither the A₁-selective antagonist CPX nor the A_2A-selective antagonist ZM241385 affected the neovascular response (Fig. 2) . The data for all antagonists represents the 30 μg/kg body weight dosage. None of the antagonists demonstrated a consistent or significant effect at the lower doses in hyperoxia-treated animals, nor did they adversely affect the apparent health of the normoxia mice at any dose (data not shown). 

Discussion

The cellular and molecular responses to ischemic injury that lead to neovascularization are complex and remain to be fully elucidated. Adenosine is a critical mediator of blood flow changes in response to ischemia. It is a significant component of the retina’s compensatory hyperemic response to ischemia, hypoxia, and hypoglycemia. ²⁸ In most cell types and organ systems, adenosine activates A₁ AdoRs to decrease work (decrease O₂ demand), whereas A₂ AdoRs increase O₂ supply. ¹⁴ Thus, adenosine, by increasing O₂ supply (activation of A₂ AdoR) and by decreasing O₂ demand (activation of A₁ AdoR), is an ideal candidate to rectify imbalances between O₂ supply and demand. 

Substantial evidence supports a role for adenosine in promoting angiogenesis. ^{19 20 29 30 31} Endothelial cells are known to have a very active adenosine metabolism, characterized by a large capacity for uptake and release of the nucleoside. ³² Adenosine can stimulate endothelial cells to alter their pattern of gene expression. ³³ High levels of adenosine are associated with areas of vasculogenesis in the normal neonatal dog retina as well as sites of angiogenesis in a canine model of oxygen-induced retinopathy. ^{21 22} We have previously shown that the stable adenosine analogue 5’-N-ethylcarboxamidoadenosine induces VEGF production in HRECs, induces proliferation and migration, supports endothelial tube formation, and results in increased activation of mitogenic protein kinases. ^{17 18} Furthermore, these effects were inhibited by selective A_2B AdoR antagonists, but not by antagonists selective for other AdoR subtypes. 

The findings presented in this report demonstrate qualitatively and quantitatively a beneficial effect of systemically administered AdoR antagonists on oxygen-induced retinopathy in the mouse. The potent, but nonselective, AdoR antagonist XAC reduced the extent of neovascularization. More important, targeting specific AdoR subtypes was vital to the success of the outcome, because the A_2B-selective antagonists enprofylline and IPDX reduced neovascularization, but the A₁ and A_2A antagonists CPX and ZM241385, respectively, did not. Although the kinetics of systemically administered AdoR antagonists and their penetration into the eye and retina were not explored, improvement of oxygen-induced retinopathy was clearly observed. These findings extend, in an in vivo model, our previous observation that the mitogenic action of adenosine on endothelial cells in vitro is mediated through the A_2B AdoR subtype. Thus, inhibiting A_2B AdoRs could provide a basis for developing pharmacologic therapies designed to prevent or treat the retinal neovascularization characteristic of proliferative retinopathies. 

 Submitted for publication May 16, 2001; revised July 13, 2001; accepted August 1, 2001.

 Commercial relationships policy: P (IB); N (all others).

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Summary of Treatment Groups and Number of Animals Examined in Each Group

	Effector Concentration (μg/kg body weight)
		0	0.3	3.0	30
Normoxia Control
Uninjected	3	—	—	—
Vehicle	5	—	—	—
XAC	—	3	3	5
Enprofylline	—	3	3	4
IPDX	—	3	3	5
CPX	—	3	3	3
Hyperoxia
Uninjected	3	—	—	—
Vehicle	8	—	—	—
XAC	—	3	3	7
Enprofylline	—	3	3	7
IPDX	—	3	3	5
CPX	—	3	3	4

 One eye from each animal was used for qualitative angiography of retinal flatmount. The contralateral eye was sectioned for quantitative assessment of extraretinal neovascular cell nuclei.

Figure 1.

 

Composite photomicrographs depicting wholemounts of retinas from mouse pups exposed to 5 days of hyperoxia followed by 5 days of treatment with adenosine receptor antagonists. Retinas from animals injected with vehicle alone (A, 0.1% DMSO in PBS) appeared identical with uninjected hyperoxia-exposed mice (not shown). Note the loss of central vasculature, the presence of dilated, tortuous blood vessels, neovascular tufts (arrows), and increased peripheral perfusion (arrowhead). Animals injected with the nonselective AdoR antagonist XAC (B) or with the A_2B-selective antagonists enprofylline (C) or IPDX (D), had a dramatic reduction in neovascular tufts of the retinas. In contrast, retinas from animals that received the A₁-selective antagonist CPX (E) or the A_2A-selective antagonist ZM 241385 (F) did not appear different from vehicle control. Original magnification, ×5.

Figure 2.

 

Results of quantification of cell nuclei occurring on the vitreous side of the internal limiting membrane in sections from whole eyes of hyperoxia-treated mouse pups. Quantitative results agree with the qualitative assessment (Fig. 1) of the effects of the various AdoR antagonists. XAC, enprofylline, and IPDX all reduced significantly the degree of oxygen-induced neovascularization, whereas CPX and ZM241385 did not. *P < 0.001 versus vehicle alone.

The authors thank E. Ann Ellis for her invaluable assistance with animal care and histochemical processing.