P2X₇ Receptor Activation Mediates Retinal Ganglion Cell Death in a Human Retina Model of Ischemic Neurodegeneration

Abstract

Purpose.: There is evidence implicating ischemia and excitotoxicity in the pathogenesis of glaucoma. ATP-mediated excitotoxicity via activation of the P2X₇ receptor (P2X₇R) has been proposed to play a role in retinal ganglion cell (RGC) degeneration in this disease. The aim of this research was to determine whether stimulation of the P2X₇R mediated ischemia-induced RGC death in the human retina.

Methods.: Human organotypic retinal cultures were exposed to the P2X₇R agonist 2′,3′-O-(4-benzoylbenzoyl)-ATP (BzATP) and simulated ischemia (oxygen/glucose deprivation) in the presence or absence of the P2X₇R antagonist, Brilliant Blue G (BBG). Neuronal death in the RGC layer was quantified by neuronal nuclei (NeuN)-positive cell counts and quantitative real-time PCR for THY-1 mRNA. The P2X₇R was localized by immunohistochemistry and P2X7R mRNA profiling using a cryosectioning technique.

Results.: P2X₇R stimulation by BzATP (100 μM) induced loss of RGC markers in human organotypic retinal cultures (HORCs), which was inhibited by BBG (1 μM). Simulated ischemia led to loss of RGCs that was also inhibited by BBG, indicating that ischemia-induced RGC degeneration was mediated by the P2X₇R. The P2X₇R was immunolocalized to the outer and inner plexiform layers of the human retina, and P2X7R mRNA expression was confirmed in the inner retina and ganglion cell layer.

Conclusions.: These studies demonstrated that stimulation of the P2X₇R can mediate RGC death and that this mechanism plays a role in ischemia-induced neurodegeneration in the human retina.

Introduction

There is currently considerable interest relating to the P2X₇ receptor (P2X₇R) in mediation of neurodegeneration, with mounting evidence indicating a role in chronic diseases including Alzheimer's disease, Parkinson's disease, and Huntingdon's disease, as well as acute conditions including stroke and spinal cord injury. ^1–3 As a chronic neurodegenerative condition, glaucoma is characterized by loss of retinal ganglion cells (RGCs) resulting in progressive optic neuropathy and consequent visual field loss. Reduction in blood flow to the optic nerve head and the consequent ischemia has been suggested as a mechanism of RGC death in glaucoma. ^4–6 Recent studies have also provided evidence that the P2X₇R may have a role in glaucomatous RGC death. ^7–11  

The P2X₇R belongs to the family of purine (P2) receptors mediating signaling via extracellular ATP. ¹² As with other members of the P2X family, the P2X₇R is a ligand-gated ion channel, that when activated results in sodium and calcium influx. ^12,13 Unlike other members of the P2X family, the P2X₇R requires relatively high concentrations of ATP for activation, which are usually only achieved following significant tissue damage, inflammation, or mechanical stress. Sustained activation can lead to formation of a pore that allows nonselective permeability of molecules up to 900 Da. ¹² The increase in permeability, in turn, can mediate cell death. ^14,15 It has been well documented that the P2X₇R mediates release of cytokines from cells of monocytic lineage including microglia, ^16–18 whereas its function in neurotransmission is less well defined than for other P2X receptors. ¹⁹ P2X receptors have been identified throughout the mammalian retina, ²⁰ and P2X₇R immunolocalization has been reported in the inner and outer plexiform layers, ^21,22 on RGCs, ^23–25 and Müller glial cells. ²⁶ P2X₇R localization suggests an active role in visual processing within the retina. However, in a similar manner to glutamate-induced excitotoxicity, elevated extracellular ATP, acting by stimulation of P2X₇Rs, has been found to be toxic to RGCs, both in vitro and in vivo, ^7,8,10 providing evidence for a potential role of the P2X₇R in mediation of glaucomatous RGC death. 

We have developed a human organotypic retinal culture (HORC) system to investigate RGC death, ²⁷ which was used in the present study to assess a potential role for the P2X₇R in loss of RGCs within the human retina. We demonstrated for the first time that simulated ischemia (oxygen/glucose deprivation [OGD]) resulted in loss of RGCs via a mechanism that involved P2X₇R activation. 

Materials and Methods

Human Organotypic Retinal Cultures

Donor human eyes were obtained within 24 hours post mortem from the East Anglian Eye Bank with full ethical approval by the Norfolk Research ethics committee and under the tenets of the Declaration of Helsinki. Eyes with known ocular disease were excluded. Human organotypic retinal explant cultures were produced as described by Niyadurupola et al. ²⁷ Briefly, the human retina was dissected and 5 × 4 mm diameter punch explants, taken from the paramacular region, were cultured RGC side up in serum-free Dulbecco's Modified Eagle Medium (DMEM)/HamF12 (Invitrogen, Paisley, UK) in 35-mm cell culture dishes. One HORC was fixed/frozen immediately as a t = 0 h sample, where t indicates time. The remaining four HORCs were then incubated under four different experimental conditions, giving one replicate per retina. All media were supplemented with gentamicin (50 μg/ml; Sigma-Aldrich, Poole, UK). HORCs were incubated at 35°C in a humidified atmosphere of 95% air/5% CO₂. 

In experiments where the P2X₇R agonist 2′,3′-O-(4-benzoylbenzoyl)-ATP (BzATP; Sigma-Aldrich) was used, it was incubated with HORCs in serum-free DMEM/HamF12 for 24 hours. Control HORCs were also incubated for 24 hours. The P2X₇R antagonist Brilliant Blue G (BBG; Sigma-Aldrich), where used, was incubated with HORCs in serum-free DMEM/HamF12 for 1 hour prior to addition of BzATP or 1 hour prior to OGD. 

Oxygen/Glucose Deprivation (Simulated Ischemia)

HORCs were placed in serum-free DMEM without glucose and pyruvate, previously warmed and bubbled with 95% N₂/5% CO₂ for 10 minutes to remove oxygen. HORCs were placed in a modular incubator container (Billups-Rothenburg, Del March, CA) and gassed at 2 psi for 10 minutes, with 95% N₂/5% CO₂ before sealing the chamber. HORCs were incubated for another 50 minutes (total ischemic period of 60 minutes) then switched to serum-free DMEM/HamF12 and incubated for 24 hours. Control HORCs were placed for 60 minutes into DMEM with glucose for the initial incubation period. 

Quantitative Real-Time Polymerase Chain Reaction (QRT-PCR)

Human Organotypic Retinal Cultures.

Total RNA was extracted from HORCs using the RNeasy Mini Kit column-based method (Qiagen, Croyden, UK). A 15 minutes DNaseI incubation step was included to eliminate contamination from genomic DNA. Total RNA, measured using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE), was reverse transcribed to synthesize cDNA, as previously described. ²⁷ The ABI Prism 7700 Sequence Detection System (Applied Biosystems, Warrington, UK) was used for QRT-PCR. cDNA was mixed with Human THY-1 probe (Hs00174816_m1; Applied Biosystems), or Human P2X7R probe (Hs0017521_m1; Applied Biosystems), and TaqMan PCR Master Mix (Applied Biosystems). PCR amplification was carried out at 50°C for 2 minutes, 95°C for 10 minutes, then 40 cycles of 15 seconds at 95°C, and 1 minute at 60°C. 

The geNorm programme (available in the public domain, http://medgen.ugent.be/∼jvdesomp/genorm) identified that the most stable housekeeping genes for these experiments with HORCs were cytochrome c-1 (CYC1) and topoisomerase DNA I (TOP1). The geometric mean of C_T values of CYC1 and TOP1 were used, as described previously, ²⁸ to normalize the data for THY-1 in the HORC experiments. Note that there is a loss of THY-1 signal over the 24 hour culture period. ²⁷  

Retinal Sections.

Following dissection from the globe, a 4-mm diameter sample was taken from the macular region of each retina. The sample was mounted carefully on filter paper and placed on a freshly cut surface of frozen optimal cutting temperature (OCT) medium (Sakura Finetek, Zoeterwoude, Netherlands). Further OCT medium was used to cover the sample, which was frozen prior to cryosectioning (20-μm sections) in the plane of the retinal nuclear layers using a Hacker Bright OTF 5000 cryostat (Bright Instruments, Huntingdon, UK). Individual sections were maintained at −80°C until total RNA was extracted using the RNeasy Micro Kit (Qiagen). Following RNA quantification (NanoDrop Technologies) samples were converted to cDNA followed by QRT-PCR, as described above. Recoverin, calbindin, choline acetyltransferase, and Thy-1 were used as markers for photoreceptors, horizontal cells, amacrine cells, and RGCs, respectively, using the following probe/primers: RCVRN: Hs00610056-m1; CALB: Hs00191821-m1; CHAT: Hs00758143-m1; THY-1: Hs00174816_m1; P2X7R: Hs0017521_m1; RBFOX3: Hs00876928-m1; and POU4F1: Hs00366711-m1 (Applied Biosystems). Expression in samples from different donors was aligned by superimposing the expression profile for RCVRN. 

Immunohistochemistry

Neuronal Nuclei (NeuN).

HORCs were fixed in 4% paraformaldehyde for 24 hours and then cryopreserved in 30% sucrose in PBS for a further 24 hours. HORCs were then placed into cryostat block moulds filled with OCT medium (Sakura Finetek) and frozen on dry ice. Retinal slices (13 μm) were cut using a Hacker Bright OTF 5040 cryostat (Bright Instruments) and collected on 3′aminopropyl-triethoxyl silane (TESPA; Sigma-Aldrich)-coated glass slides. Immunohistochemistry for NeuN on HORC sections has been previously described. ²⁷ The primary monoclonal antibody for NeuN (Millipore, Watford, UK) was used at 1:200 and was incubated at 4°C overnight. The secondary antibody was Alexa Fluor 488–conjugated (Invitrogen) used at 1:1000, and incubation was for 2 hours at room temperature. Samples were protected from light from this point onwards. Nuclei were counterstained with 4′,6-diamidino-2-phenyindole dilactate (DAPI; Invitrogen) (1:100) for 10 minutes at room temperature prior to mounting in hydromount (National Diagnostics, Hull, UK) immunofluorescence mounting medium. A wide-field Zeiss Axiovert 200M fluorescent microscope with a 100W mercury arc lamp was used for immunofluorescent imaging (Carl Zeiss, Welwyn Garden City, UK). Zeiss Axiovision 4.7 software was used and images were captured with a cooled monochrome CCD camera (Zeiss AxioCam; Carl Zeiss). 

The number of NeuN-labeled RGC layer cells was used as a measure of RGC death. Note that there is no major staining of NeuN in the inner nuclear layer (INL) suggesting that NeuN is not labeling amacrine cells to the same extent as that seen in RGCs. Six nonconsecutive immunohistochemical slices were prepared for each HORC and the numbers of NeuN-stained RGC layer neurones were counted (in a masked fashion) in three, 200-μm sections of each slice. The mean for the six slices was taken as one biological replicate. Note that there is no loss of NeuN staining over the 24 hour culture period. ²⁷  

P2X₇ Receptor.

Macular samples were fixed with 4% formaldehyde in PBS (24 hours; 4°C), then dehydrated through graded ethanol (30%, 50%, 70%, 90%, and 100% EtOH: water; 30 minutes each; 4°C), followed by 50:50 xylene: EtOH, then 100% xylene for 30 minutes at room temperature (RT). The sample was embedded in wax (60°C; 60 minutes). The sample was then sectioned (5 μm) and mounted. Prior to immunohistochemistry, the slide was immersed in 100% xylene for 5 minutes, then graded EtOH (100%, 90%, and 70%) for 1 minute, then washed prior to permeabilization in PBS containing 5% goat serum and 0.2% Triton X-100 (90 minutes; RT). Following washing in PBS, the sample was incubated with primary antibody (P2X₇R rabbit polyclonal antibody raised against the extracellular domain; diluted 1:200; Sigma-Aldrich) overnight at 4°C. The sample was further washed, and then incubated with secondary antibody (1:1000; Alexa 568-conjugated) for 2 hours at room temperature. Nuclei were stained with DAPI (1:100) for 10 minutes at room temperature, washed, and then mounted in hydromount immunofluorescence mounting medium (National Diagnostics, Hull, UK). 

Statistical Analysis

Significance was evaluated using SPSS 16.0 software (IBM United Kingdom Ltd., Portsmouth, UK) using one-way ANOVA followed by Tukey's post hoc multiple comparison test. A P value less than 0.05 was considered statistically significant. 

Results

Stimulation of the P2X₇R Caused Death of Human Retinal Ganglion Cells

In order to determine whether stimulation of the P2X₇R resulted in loss of RGCs in human retina, HORCs were incubated with the P2X₇R agonist BzATP for 24 hours, following which levels of RGC markers were assessed. ²⁷ BzATP (100 μM) caused a significant reduction in THY-1 mRNA of approximately 40% compared with control HORCs at 24 hours (P < 0.05, n = 5) (Fig. 1A). Because a loss of THY-1 could be due to a change in expression levels, it was important to also look directly at RGC number, by assessment of NeuN-labeled cells in the RGC layer. Similarly, 100 μM BzATP caused a significant reduction of approximately 55% in the numbers of NeuN-labeled cells of HORCs at 24 hours compared with control HORCs (P < 0.05, n = 4) (Figs. 1B, 1D, 1F). The P2X₇R agonist BzATP, therefore, reduced levels of RGC markers in the human retina indicating a loss of RGCs as a result of BzATP exposure. 

Figure 1

 

The P2X₇ agonist BzATP caused loss of RGCs in HORCs that was inhibited by the P2X₇ antagonist BBG. HORCs were incubated for 24 hours in the presence or absence of 100 μM BzATP, in the presence and absence of 1 μM BBG, or in 1 μM BBG alone. (A) The level of THY-1 mRNA assessed by QRT-PCR as a measure of RGC loss. THY-1 expression was normalized to the housekeeping genes TOP1 and CYC1, then expressed as a percentage of control (mean ± SEM; n = 5; one-way ANOVA with Tukey's post hoc test, *P < 0.05 versus control, † P < 0.05 vs. 100 μM BzATP). (B–E) Immunohistochemistry of (B) control, (C) 1 μM BBG-, (D) 100 μM BzATP-, and (E) 1 μM BBG + 100 μM BzATP-treated HORCs at 24 hours labeled with antibody to the RGC marker NeuN (green) and the nuclear stain DAPI (blue). (F) Corresponding NeuN-labeled neuronal cell counts expressed as a percentage of control (mean ± SEM; n = 4; one-way ANOVA with Tukey's post hoc test, *P < 0.05 versus control, † P < 0.05 vs. 100 μM BzATP).

To confirm whether the BzATP-induced loss of RGC markers was due to activation of the P2X₇R, BBG, a noncompetitive antagonist of the P2X₇R, ²⁹ was utilized. BBG has been found to be an effective and selective inhibitor of the P2X₇R at concentrations of up to 100 nM in rats. ²⁹ Inhibition of the human P2X₇R, however, has been found to require 10-fold higher concentrations than the rat P2X₇R, ²⁹ so a concentration of 1 μM was used in the HORC experiments. BBG (1 μM) gave almost full protection from BzATP-induced loss of THY-1 mRNA at 24 hours (P < 0.05 versus BzATP group, n = 5) (Fig. 1A). Similarly, BBG significantly reduced the BzATP-induced loss of NeuN-labeled neuronal cells from the RGC layer of HORCs at 24 hours (P < 0.05 versus BzATP group, n = 4) (Figs. 1B–F). 

Ischemia Caused Death of Human RGCs by Stimulation of the P2X₇R

Ischemia has been suggested to play a role in glaucomatous RGC death. We have shown previously that 60 minutes of OGD (simulated ischemia) caused the death of RGCs in HORCs. ²⁷ In the present study, we showed that a 41% reduction in THY-1 mRNA from HORCs subjected to 60 minutes of OGD was significantly inhibited by 1 μM BBG at 24 hours (P < 0.05, n = 4) (Fig. 2A). A similar pattern was observed in the numbers of NeuN-labeled neuronal cells in the RGC layer of HORCs, although statistical significance was not achieved (P > 0.05, n = 5) (Figs. 2B–F). 

Figure 2

 

The P2X₇ antagonist BBG protected against simulated ischemia-induced loss of RGCs in HORCs. Simulated ischemia was induced in HORCs by 1 hour of OGD (simulated ischemia), followed by incubation for 23 hours in control medium either in the presence and absence of 1 μM BBG or in 1 μM BBG alone. (A) Upper panel: BBG inhibited the loss of THY-1 mRNA by simulated ischemia (mean ± SEM; n = 4; one-way ANOVA with Tukey's post hoc test, *P < 0.05 versus control, † P < 0.05 versus OGD). THY-1 expression was normalized to the housekeeping genes TOP1 and CYC1, then expressed as a percentage of control. (B–E) Immunohistochemistry of (B) control, (C) 1 μM BBG-, (D) OGD-, and (E) 1 μM BBG + OGD-treated HORCs at 24 hours labeled with antibody to the RGC marker NeuN (green) and the nuclear stain DAPI (blue). (F) Lower panel: Corresponding NeuN-labeled neuronal cell counts expressed as a percentage of control (mean ± SEM; n = 5; one-way ANOVA with Tukey's post hoc test, *P < 0.05 versus control).

Expression Profiles of the P2X₇R in HORCs

In order to determine the pattern of expression of the P2X₇R in the human retina, P2X₇R immunoreactivity and P2X₇R mRNA profiles were assessed. Immunolocalization of the P2X₇R in transverse retinal sections (Fig. 3) was carried out in macular samples due to the higher density of RGCs in the macula. The results indicated that P2X₇R expression was located primarily in the outer plexiform layer (OPL). Less intense labeling was apparent in the inner plexiform layer (IPL), with little evidence of labeling in the somatic region of RGC cells. The antibody used for these studies was raised against the extracellular domain of the P2X₇R and, therefore, should have been able to detect all reported functionally expressed P2X₇R splice variants. ^30,31 However, the selectivity of P2X₇R antibodies has been the subject of much debate ^31,32 making it essential that other methods were used to investigate expression. For this reason a further novel approach was taken to investigate the pattern of expression of P2X₇R mRNA (P2X₇R). Using a cryosectioning technique, serial horizontal sections of the retina were taken in order to provide samples in the plane of the retinal layers. Consecutive sections, each 20 μm, were taken from the outer to the inner retina. In order to validate the technique, mRNA levels of known markers in each of the sections were assessed (Fig. 4). Peak mRNA levels for the photoreceptor marker recovering (RCVRN) and the RGC marker Thy-1 (THY-1) occurred in the outer and inner retina, respectively, while the horizontal and amacrine cell markers (calbindin [CALB] and choline acetyltransferase [CHAT]) showed peaks in the intermediate sections. The expression profiles of the markers corresponded well with localization of the cell bodies of the specific neurons in the retina. Further, RGC markers were also used (NeuN [RBFOX3] and Brn3a [POU4F1]) and their pattern of expression coincided with that seen for THY-1 (data not shown). P2X7R expression was low in the outer retina, with no coincidence with the photoreceptor marker RCVRN. The initial peak of P2X7R expression corresponded with CALB expression. The P2X7R expression level then remained elevated in all the following sections, including the RGC layer. This pattern of expression was also apparent in paramacular sections (see Supplementary Material and Supplementary Fig. S1). 

Figure 3

 

P2X₇ receptor immunolocalization in the human macula. Immunohistochemistry of HORCs labeled with DAPI (blue), NeuN (green), P2X7R (red). P2X₇ receptor immunoreactivity was predominant in the OPL, but there was also labeling in the IPL.

Figure 4

 

Expression of the P2X₇ receptor in human macula. Serial sections (20 μm) of the human macula were analyzed for P2X₇ receptor mRNA by QRT-PCR and compared with the retinal neuron markers recoverin (RCVRN; photoreceptors), calbindin (CALB; horizontal cells), Choline acetyltransferase (CHAT; amacrine cells), and THY-1 (RGCs). Data were expressed relative to the sample with maximal expression level (mean ± SEM; n = 6). Retinal neuron markers showed the expected distribution, validating the technique and indicating that P2X7R expression was found in the inner layers of the retina corresponding with CALB, CHAT, and THY-1, but not with RCVRN, expression.

Discussion

We have shown that P2X₇R activation caused loss of human RGCs. This confirms that research carried out in rodent species ^7–10 is relevant to the human retina. In order to further identify a role for the P2X₇R in the pathogenesis of glaucomatous optic neuropathy we used an established model of simulated ischemia in HORCs. ²⁷ Reduction in blood flow to the optic nerve head and consequent ischemia has been proposed as a mechanism that may contribute to loss of RGCs in glaucoma. ^4–6 Reduced ocular blood flow has been identified in both high- and low-tension glaucoma patients, ^33–36 and recent data shows a close link between decreased retinal blood flow and visual field loss in glaucoma. ³⁷ In addition, vasospastic conditions, such as migraine and Raynaud's disease, have been identified as risk factors for the disease, ^38–41 suggesting that dysfunctional autoregulation in retinal vessels and subsequent ischemia are important in the pathogenesis of glaucoma. Retinal ischemia induced by raised IOP has been shown to cause apoptosis of RGCs in experimental in vivo models, ^42,43 and we have shown previously that simulated ischemia caused death of human RGCs in culture. ²⁷ In the present study, we have shown that loss of human RGCs from HORCs caused by simulated ischemia was prevented by the P2X₇R antagonist BBG. These experiments suggested that ischemia caused death of human RGCs involving activation of the P2X₇R. 

Studies in brain slices suggest that ischemic injury may lead to activation of the P2X₇R, resulting in cell death that is mimicked by application of BzATP. ^44–46 Furthermore, ATP release has been demonstrated in response to ischemia in the brain. ^47,48 Experiments that monitor membrane potential and extracellular ATP demonstrate release of ATP immediately following ischemia-induced depolarization of neurons (anoxic depolarization). ⁴⁹ It is likely that similar release of ATP from retinal neurons would occur in the ischemic retina. Certainly, hypoxic stress to purified rat RGCs has been linked with P2X₇R activation. ⁵⁰ Chemically-induced ischemia has also been shown to increase ATP release from cultured human retinal pigment epithelial cells ⁵¹ indicating a further potential source of ATP within the retina. In addition, oligodendrocytes isolated from rat optic nerve have been shown to release ATP in response to simulated ischemia via activation of pannexin hemichannels. ⁵² Whether equivalent ATP release pathways are effective in retinal glia remains to be determined. 

Inhibiting the P2X₇R with BBG has been found to be effective in inhibiting neurotoxicity in several models in both the retina and brain. It provided neuroprotection in vitro in primary cortical neuron cultures and brain slices as well as in an in vivo model of transient focal cerebral ischemia. ⁴⁶ In the isolated rat optic nerve, it significantly improved survival and recovery of axon function following ischemia ⁵² and in cultured rat retinal neurons, it inhibited hypoxia-induced cell death. ⁵⁰ In addition, increased expression of P2X₇Rs has been observed in experimental cerebral ischemia in in vivo ⁵³ and in vitro ^54,55 rodent models. Taken together with our data, we, therefore, propose that as a result of retinal ischemia, there is a release of ATP, which activates P2X₇Rs, which in turn mediate RGC death. 

Also supporting a role in glaucoma, elevated concentrations of ATP have been found in the vitreous humor of rat eyes subjected to increased IOP in vivo. ⁸ Furthermore, an acute pressure increase in an in vivo model of glaucoma, also led to RGC damage that was inhibited by the addition of apyrase to degrade extracellular ATP. ⁸ Ours is the first data to demonstrate that ischemia-induced RGC degeneration is via P2X₇R activation. It is possible, therefore, that in the in vivo models of glaucoma, the reduction in blood flow occurring as a result of the acute increase in IOP could have a significant role to play in the P2X₇R-mediated loss of RGCs that was observed. 

A major advantage of the organotypic system is that it allows maintenance of regional morphology and cell–cell signaling including synaptic connectivity and neuron-glia interactions. A central question, therefore, in relation to RGC death mediated by the P2X₇R, is whether there is a direct action on the RGCs or whether it occurs via indirect mechanisms. To answer this, it is important to determine whether human RGCs express the P2X₇R. Immunohistochemistry of the human macula showed intense P2X₇R labeling in the OPL with further labeling apparent in the IPL. Our findings support data in rat and marmoset retina where P2X₇R immunoreactivity was seen in both plexiform layers with clustering at synaptic sites in the OPL as well as colocalization with horizontal cell markers. ²¹ Other studies, however, have shown differing patterns of immunoreactivity, with localization seen in the RGC layer in the rat and monkey retina ^23,56 and in both young and aged mice. ²⁴ P2X₇R expression has also been detected in the somatic region of human Müller cells. ²⁶ These are inconsistent with the labeling shown in human macula (Fig. 3), where there was little evidence of specific labeling of RGC cell bodies or of the cells of the INL. Due to these observed inconsistencies, and the documented questions over P2X₇R antibodies, ^31,32 we developed a novel approach in order to examine P2X₇R mRNA expression profiles in the retina. Taking serial sections of the macula, we investigated the distribution of known retinal neuronal markers in order to validate the methodology. It is important to note that using such a technique, the expression will be associated with the cell body of the neurons independent of where the protein is expressed in the neuron. In the outer sections, there was high expression of the photoreceptor marker RCVRN, which subsequently declined to zero. Peak expression of CALB, which is highly expressed in horizontal cells, ⁵⁷ represented the outer aspect of the inner nuclear layer (INL), followed by CHAT, which is highly expressed in starburst amacrine cells, ⁵⁸ thus, enabling the positioning of the inner aspect of the INL. The CHAT signal also showed subsequent peaks consistent with profiles of labeling in the RGC layer. ⁵⁸ Coincident with the initial peak for CHAT was the first peak for the RGC marker THY-1. This initial peak probably represented displaced RGCs resident in the inner aspect of the INL. The initial THY-1 peak was followed by a larger peak showing sustained expression representative of the RGC layer. The RGC markers NeuN (RBFOX3) and Brn3a (POU4F1) showed the same profile. The markers, thus, gave distinct peaks and the expression profile showed excellent correlation with their relative positions within the retinal layers. This novel method can, therefore, be used to establish the expression profile of genes of interest in the human retina. When the P2X₇R was investigated using this technique it was notable that there was no P2X7R expression coincident with the photoreceptor marker RCVRN. However, there was a peak that was coincident with CALB (the horizontal cell marker), this being consistent with immunolocalization in the human retina (Fig. 3), as well as that previously published in other species. ^21,22 A second peak of P2X7R expression, which coincided with the inner aspect of the INL, was also identified. Subsequent to this, P2X7R expression remained high in the inner sections of the retina corresponding to expression in cell bodies within the RGC layer. This pattern of expression coincided precisely with that of the RGC markers. Together, the expression profile and the immunolocalization data suggest that there may be P2X₇ receptors on RGC dendrites, with the P2X₇R mRNA present in the RGC soma in the retinal ganglion cell layer (GCL) and the P2X₇R protein located at the RGC synapses in the IPL. However, more precise immunolocalization looking for colocalization with synaptic proteins would be required to provide unequivocal evidence that the P2X7R is localized at the dendrites of the RGCs. 

Data from the present study have indicated that there is expression of the P2X₇R in RGCs, which would mean that stimulation of this receptor on the RGCs could directly mediate neurotoxicity. This would be consistent with evidence from isolated rat RGCs where P2X₇R stimulation led to calcium overload, caspase activation and subsequent apoptosis. ⁷ The expression at other sites in the human retina also leaves open the possibility of other indirect mechanisms that could contribute to P2X₇R-mediated RGC neurodegeneration. For example, stimulation of P2X₇Rs in the OPL could lead to glutamate release in the IPL leading to N-methyl-D-aspartate receptor-mediated excitotoxicity. Alternatively, stimulation of the P2X₇R could lead to cytotoxic cytokine release. Such indirect mechanisms should be the subject of further investigation. 

Our data demonstrate, for the first time, that RGCs are lost as a result of P2X₇R stimulation under conditions of simulated ischemia. The data suggest that excitotoxicity via the P2X₇R may be an important cause of RGC death in the pathogenesis of glaucoma. As novel P2X₇R antagonists are being trialed for clinical use in the treatment of inflammatory pain and rheumatoid arthritis, the P2X₇R could represent an exciting target for therapeutic intervention to prevent the visual deterioration that occurs in glaucoma. 

Supplementary Materials

pdf S1 

Acknowledgments

The authors thank the staff of the East Anglian Eye Bank and the Humane Research Trust for their support. 

Footnotes

 Supported by grants from the Humane Research Trust and the Norwich Glaucoma Research Fund.

Footnotes

 Disclosure: N. Niyadurupola, None; P. Sidaway, None; N. Ma, None; J.D. Rhodes, None; D.C. Broadway, None; J. Sanderson, None