Effect of P2X7 Receptor Activation on the Retinal Blood Velocity of Diabetic Rabbits

Objective  To test the effect of activation of P2X₇ receptors on retinal blood velocity in diabetic rabbits.

Methods  Immunohistochemical analysis was performed on healthy and diabetic rabbit eyes using P2X₇ receptor antibodies. Diabetes was induced using alloxan. Retinal blood velocity was measured with the laser speckle circulation analyzer. Visual function was assessed using electroretinography.

Results  Cells in inner retinal layers were positive for anti-P2X₇ receptor antibodies in healthy rabbits. The distribution of positive cells extended to outer layers and some small vessels stained in diabetic rabbits. When assayed 24 hours after an intravitreal injection of 150 nmol of benzoylbenzoyl adenosine triphosphate (BzATP), a P2X₇ agonist, the retinal blood velocity in healthy rabbits was reduced by approximately 30%; this reduction continued for at least 4 weeks. Only in diabetic rabbits did an injection of 50 nmol of BzATP reduce retinal blood velocity by approximately 30% and the amplitudes of electroretinography a waves, b waves, and oscillatory potentials for at least 4 weeks.

Conclusions  Soon after the onset of alloxan-induced diabetes, retinal blood velocity and function become more vulnerable to reduction initiated through P2X₇ receptors.

Clinical Relevance  Our findings support the hypothesis that the retinal circulation disorder accelerated by activation of P2X₇ receptors may be involved in the early changes of diabetic retinopathy.

The pathophysiologic mechanism of diabetic retinopathy is not yet fully elucidated; however, the apoptotic death of microvascular pericytes and endothelial cells is one of the hallmarks of diabetic retinopathy.¹ The loss of pericytes seems to play an important role in the progression of diabetic retinopathy.² Many studies have examined retinal blood flow in diabetes.^3,4 As the disease progresses before vasoproliferation, retinal blood flow was found to be above normal.^5-7 However, some researchers have reported reduced retinal blood flow in patients with diabetes mellitus without diabetic retinopathy or with background diabetic retinopathy.^8,9 These changes in retinal blood flow might be related to the dysfunction or loss of microvascular pericytes.

Our previous immunocytochemical and electrophysiologic studies have shown that microvessels of the retina express functional purinergic P2X₇ receptors.¹⁰ We also demonstrated that retinal microvasculatures became more vulnerable to cell death initiated through P2X₇ receptors within 2 weeks after the onset of streptozotocin-induced diabetes.¹¹ We reported that there were potent mechanisms that minimize purinergic vasotoxicity under physiologic conditions.¹² Extracellular adenosine triphosphate (ATP) probably serves as one of the vasoactive molecules in the retinal microvasculature according to the previously mentioned experiment¹⁰ using pericyte-containing capillaries of the retina. Additional researchers reported the similar effect of P2X₇ receptor activation in other tissues.¹³ However, changes induced by P2X₇ receptor stimulation in the retinal blood flow at the capillary level have not yet been recorded. In addition, it is questionable whether those changes are enhanced in diabetic animals. We studied whether activation of P2X₇ receptors alters retinal blood velocity in healthy and diabetic rabbits using the laser speckle circulation analyzer that was developed in Japan.^14,15

Adenosine triphosphate, the P2X₇ agonist benzoylbenzoyl-ATP (BzATP), the P2X₇ antagonist oxidized ATP, and alloxan monohydrate were purchased from Sigma Chemical Co (St Louis, Mo).

Male albino rabbits that weighed 2.5 to 3.4 kg were purchased from Shimizu Laboratory Supplies (Kyoto, Japan). They were housed in an air-conditioned room at a mean ± SD temperature of 22°C ± 1°C and a mean ± SD level of humidity of 66% ± 3% humidity with a 24-hour light-dark cycle. They were handled in accordance with the Annual Meeting of the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

Two healthy rabbits and 3 rabbits made diabetic by alloxan for approximately 6 weeks were killed by injecting a lethal dose of pentobarbital sodium, and the eyes were removed. The excised eyes were fixed in 4% paraformaldehyde for approximately 24 hours, dehydrated, and embedded in paraffin. Continuous 5-μm sections were then prepared using a microtome. After deparaffinization, sections were treated with a 3% hydrogen peroxide solution for 5 minutes and incubated with primary antibodies (anti–human P2X₇ IgG at a 1:20 ratio for 60 minutes [Santa Cruz Biotechnology Ltd, Santa Cruz, Calif]) and secondary antibodies (biotin-conjugated goat anti–rabbit IgG, at a 1:200 ratio for 30 minutes [Vector Laboratories, Burlingame, Calif] and the avidin-biotin-peroxidase complex for 30 minutes [Vector Laboratories]). The resulting sections were stained with diaminobenzidine, and nuclear staining was performed using 3% methyl green.

Retinal blood velocity was measured with the laser speckle circulation analyzer, which permits noninvasive and 2-dimensional measurements of tissue circulation in the retina. The apparatus consists of a fundus camera equipped with an argon laser (wavelength, 488 nm) and an image sensor (100 × 100 pixels). The principle of this method has been described previously.^14,15 The normalized blur (NB) value, a quantitative index of blood velocity in the retinal microcirculation, was recorded 5 times at each time point, and then the mean value was calculated. The image speckles from a retinal field free of visible surface vessels and approximately 1 papillary diameter from the optic nerve head along the medullary rays were recorded to determine the NB value. One eye of each rabbit was measured for blood flow after dilating the pupil with 1 drop of 0.5% tropicamide. During the measurement, the eye to be measured was held open with a Barraque wire speculum with the rabbit under local anesthesia with benoxinate hydrochloride.

The retinal blood velocity was measured at 1 and 3 days and at 2 and 4 weeks after intravitreal injection of 20 μL of ATP (150 and 1500 nmol), the P2X₇ agonist BzATP (50 and 150 nmol), the P2X₇ antagonist oxidized ATP (50 nmol), or physiologic saline solution (Otsuka, Tokyo, Japan), as a control, in nondiabetic rabbits. Five rabbit eyes were used for each dose of ATP, BzATP, oxidized ATP, or control. The relative NB value was determined as the ratio of each absolute value to the previous value. Intravitreal injection was performed through the pars plana using a microsyringe connected to a 30-gauge needle.

Intraocular pressure (IOP) was measured with a calibrated pneumatonometer (model 30 Classic, Medtronic Solan, Jacksonville, Fla) with the rabbit under local anesthesia with benoxinate hydrochloride. The mean blood pressure (MBP) was measured at the front leg by an automatic sphygmomanometer (BP-98E, Softron, Tokyo). A close correspondence between the pressure determined by this sphygmomanometer and that obtained through a pressure transducer cannula placed in the femoral artery had been previously confirmed.¹⁶ Ocular perfusion pressure (OPP) was determined as MBP minus IOP.

To assess the changes of retinal function, electroretinography (ERG) was performed. For recording, a photic stimulator (SLS 4100), a biophysical amplifier (AVM-10), and an averager (DAT-1100) (all from Nihon-Kohden, Tokyo) were used. Before the ERG recordings, the animals were adapted to the dark for 60 minutes. The ERG was performed with a 4.7-cal light stimulus set 20 cm in front of the eye, and recordings were made with a gold ring active electrode on the cornea by averaging 4 responses to the light stimuli at 0.1 Hz. A diffuser was placed before the stimulated eye to ensure a more full-field stimulation, and the mean luminance at the corneal surface was 43.8 lux/s. Bandpass filters were set at 0.5 to 100 Hz for a waves and b waves and at 50 to 300 Hz for oscillatory potentials (OPs). The OPs that appeared first and second were referred to as OP₁ and OP₂, respectively. The amplitudes and the implicit times of the a waves and b waves, OP₁, and OP₂ of the ERGs were measured. Analog data were recorded by a rectilinear pen system and simultaneously stored and digitized using a microcomputer (MacLab 2e; AD Instruments, Castle Hill, Australia). Analysis of the various stored parameters was performed using the microcomputer. We performed ERG before and 2 and 4 weeks after intravitreal injection of 10 μL of BzATP (50 nmol) or physiologic saline solution, as a control, in 5 nondiabetic rabbits.

After an overnight fast, rabbits received an intravenous injection of alloxan (80 mg/kg) diluted in physiologic saline solution. Blood glucose, IOP, MBP, and retinal blood flow were measured at 3 and 6 hours, 3 days, and 1 and 2 weeks after the alloxan application in 7 rabbits. If the blood glucose level was less than 50 mg/dL (<2.78 mmol/L), sufficient glucose (10%) was provided. Two weeks after the indication of diabetes, BzATP (15, 50, or 150 nmol) or physiologic saline solution was intravitreally injected into the eyes of diabetic rabbits. The retinal blood flow and ERG were examined according to the same protocol as the nondiabetic rabbits. Five rabbits were used for each dose of BzATP.

Statistical analysis was performed using 1-way analysis of variance. If a statistically significant change was detected, further assessment was made with the Dunnett test. Statistical significance was set at P<.05.

The results of immunohistochemical analysis in the retinas of healthy rabbits showed cells positive for an anti-P2X₇ receptor antibody in the inner plexiform and ganglion cell layers (Figure 1B). In the retinas of diabetic rabbits, the distribution of positive cells was extended to the outer nuclear layer. In addition, some small vessels also stained for this antibody (Figure 1C).

Intravitreal injection of 1500 nmol of ATP decreased the retinal blood velocity significantly with a maximum reduction of more than 30% at 2 weeks; 150 nmol of ATP had no significant effect (Figure 2A). Although 150 nmol of the P2X₇ agonist BzATP reduced the blood velocity to almost the same degree as 1500 nmol of ATP, a significant reduction was detected 1 day after the injection and lasted for at least 4 weeks. These changes were inhibited by the P2X₇ antagonist oxidized ATP (Figure 2B).

Blood glucose levels increased in 3 hours, decreased temporarily in 6 hours, and increased to 400 mg/dL (22.2 mmol/L) or more in 3 days. That level was maintained for at least 2 weeks (Table 1). The IOP, MBP, and OPP did not show any significant changes for at least 2 weeks. The retinal blood velocity was reduced at 3 and 6 hours, but it recovered to the initial level within 3 days and then became stable at 2 weeks (Table 1).

A total of 50 nmol of BzATP produced no significant changes in IOP, MBP, or OPP (Table 2). It decreased the retinal blood velocity with a maximum reduction of more than 30% in 4 weeks in diabetic rabbits (Figure 3A), although the same dose had no significant effect in healthy rabbits (Figure 2B). A total of 150 nmol of BzATP increased the blood velocity temporarily in a day but reduced it to the same level as 50 nmol in 4 weeks (Figure 3A). The dose-response curves show the enhancement of BzATP-induced reduction of the retinal blood velocity in diabetic rabbits compared with nondiabetic rabbits (Figure 3B).

As shown in Figure 4, ERG a and b waves did not notably change in the 2 weeks after administration of alloxan; however, 2 weeks after the injection of 50 nmol of BzATP, the amplitudes of the a and b waves and OPs were more reduced than in the controls. A total of 50 nmol of BzATP significantly reduced the amplitudes of a waves, b waves, OP₁, and OP₂ in diabetic rabbits for at least 4 weeks, although it had no significant effect on these parameters in nondiabetic rabbits (Table 3). The implicit times of the a waves, b waves, OP₁, and OP₂ were not significantly changed after BzATP injections, although they have the tendency to increase (data not shown).

To the best of our knowledge, this is the first study that demonstrates that the activation of P2X₇ receptors can cause the persistent reduction of retinal blood velocity and function in vivo. Although a high concentration of the P2X₇ agonist was needed to decrease retinal blood velocity in healthy rabbits, a lower concentration was effective soon after the onset in an alloxan-induced diabetic rabbit model. Therefore, diabetes-induced vulnerability of retinal blood velocity affected by P2X₇ receptor activation may be involved in the mechanism of early changes in diabetic retinopathy.

To measure retinal blood velocity, we used a laser speckle circulation analyzer developed in Japan. Since Tamaki et al¹⁵ have already shown a significant correlation between the relative change in NB and that in the retinal blood flow rate determined using the microsphere technique, we estimated blood flow by measuring blood velocity.

Our results show that vessels and other types of cells in the retina were immunoreactive to P2X₇ receptor antibodies. Other researchers have reported the expression of the P2X₇ receptor in the neurons and Müller glia cells of the retina^17,18; however, they did not direct their attention to the retinal vessels. We confirmed the immunoreactivity of retinal vessels to P2X₇ receptor antibodies, so it is natural that the retinal vasculature responds to P2X₇ agonists. We have already reported that a P2X₇ agonist, BzATP, as well as ATP caused the microvascular lumen to narrow.¹⁰ In the current study, we did not measure the short-term effect of ATP and BzATP on the retinal blood velocity. The BzATP decreased the blood velocity and probably blood flow within 24 hours; therefore, the effect might be partly due to retinal vessel constriction, as shown in the previous study.¹⁰ Since the change lasted 4 weeks, other mechanisms, including apoptosis of the microvascular cells (which we had shown in the previous study),¹¹ might be involved in the circulation reduction. In addition, ATP had no effect on blood velocity at 1 day after injection, suggesting the existence of a different mechanism of ATP and BzATP action. This finding is consistent with the ability of ATP to activate receptors other than P2X₇ receptors. On the other hand, the 50% effective concentration of BzATP to the P2X₇ receptor is much smaller than that of ATP (3:100), meaning that BzATP is the most sensitive P2X₇ agonist.¹⁹ In addition, BzATP is sensitive to the P2X₁ receptor, but the 50% effective concentration of BzATP to the receptor is not smaller than that of ATP (3:1), suggesting that BzATP is not more sensitive to the P2X₁ receptor than ATP.¹⁹ The blood velocity effect of BzATP was also completely inhibited by the P2X₇ antagonist. Taken together, this effect seems to be due to the activation of the P2X₇ receptor.

Some reports have described retinal blood velocity or flow changes in diabetic animal models. In streptozotocin-induced diabetic rats and spontaneous type 2 diabetic rats, reduced retinal blood flow was indicated using retinal angiography.^20,21 In this investigation, we used alloxan to create diabetic rabbits. Alloxan inhibits insulin release from beta cells for 2 to 4 hours, followed by destruction of beta cells and a transient increase of serum insulin in 6 to 12 hours.²² The blood glucose level temporarily increased, followed by a decrease to a fatal level unless glucose was administered to the rabbits. After 3 days the rabbits became hyperglycemic (glucose level >300 mg/dL [>16.6 mmol/L]) and maintained this condition for at least 2 weeks. Retinal blood velocity decreased transiently but returned to the previous level in 3 days and became stable in 2 weeks. The transient reduction of the blood velocity might be due to the decrease in OPP, which was not statistically significant. The BzATP at 50 nmol, which had no significant effect on the blood velocity in healthy rabbits, reduced the blood velocity in diabetic rabbits significantly within 3 days after the intravitreal injection. Since the systemic condition (ie, MBP) and IOP of the animals did not change significantly, the blood velocity change seems to be induced by the response of microvessels to the agonist. Consistent with our previous study¹¹ that used microvascular cells of rats, the retinal blood velocity seemed to react more to a P2X₇ agonist in diabetic rabbits than in healthy ones. The reason that BzATP at 150 nmol increased the retinal blood velocity transiently is not exactly known; however, it might be a result of the inflammation induced by the agonist.²³

The ERG recordings indicated that 50 nmol of BzATP reduced the amplitudes of a and b waves and OPs significantly in diabetic rabbits but not significantly in nondiabetic ones. The ERG b waves and OPs are well-known indicators of inner retinal ischemia in humans and diabetic animals.^24,25 The retinal vascular system in rabbits is extremely underdeveloped, and most parts of the retina are nourished by choroidal circulation. The simultaneous reduction of ERG a and b waves and OPs may indicate that the stimulation of the P2X₇ receptor accelerates functional impairment of the inner and outer retinal segments in diabetic retinas, including photoreceptors and bipolar, amacrine, and Müller cells, probably owing to a reduction in retinal circulation as well as choroidal circulation. Another possibility is that BzATP may directly affect cellular function in the retinas of diabetic animals, since Müller cells express P2X₇ receptors.¹⁸ To fully understand this subject, more research needs to be conducted.

Correspondence: Tetsuya Sugiyama, MD, PhD, Department of Ophthalmology, Osaka Medical College, 2-7 Daigaku-cho, Takatsuki, Osaka 569-8686, Japan (opt017@poh.osaka-med.ac.jp).

Submitted for Publication: August 23, 2005; final revision received March 1, 2006; accepted March 5, 2006.

Financial Disclosure: None reported.

Funding/Support: This study was supported in part by a research grant from the Osaka Eye Bank, Osaka, Japan.

Acknowledgment: We thank Donald G. Puro, MD, PhD, Department of Ophthalmology, University of Michigan (Ann Arbor), for his critical review of our work and Keigo Kawabata, BA, for his technical assistance. We thank San Francisco Edit (www.sfedit.net) for their assistance in editing the manuscript.