The Effect of Retinal Ganglion Cell Injury on Light-Induced Photoreceptor Degeneration

Abstract

purpose. To determine the effect of optic nerve transection (ONT) and excitotoxic retinal ganglion cell (RGC) injury on light-induced photoreceptor degeneration.

methods. Age- and sex-matched rats underwent unilateral ONT or received intravitreal injections of N-methyl-d-aspartate (NMDA). The fellow eye received sham treatment, and 7 or 21 days later each eye was subjected to an intense photic injury. Maximum a- and b-wave amplitudes of the flash electroretinogram (ERG) were measured at baseline, after the RGC insult, and 5 days after the photic injury. Semiquantitative reverse transcription-polymerase chain reaction analysis and immunoblot analysis were used to assess rod opsin mRNA and rhodopsin kinase protein levels and to measure defined trophic factors 7 or 21 days after ONT or injection of NMDA. Structural changes after the insults were determined histologically and immunohistochemically.

results. ONT caused time-dependent reductions in the mean a- and b-wave amplitudes. Seven days after intravitreal NMDA the b-wave amplitude was reduced, but the a-wave was unaffected. ONT and NMDA injection attenuated the light-induced reductions in the a- and b-wave. Rod opsin mRNA levels and rhodopsin kinase protein levels were also significantly greater in the axotomized and NMDA-treated eyes compared with the sham-treated fellow eyes after the photic injury. Structural protection in the RGC-injured eyes was also evident histologically. Fibroblast growth factor (FGF)-2, ciliary neurotrophic factor (CNTF), and glial fibrillary acidic protein (GFAP) were significantly upregulated after ONT and NMDA.

conclusions. ONT and intravitreal injection of NMDA protect against subsequent photic injury. This protection may relate to the activation of retinal glial cells and the possible action of trophic factors such as FGF-2 and CNTF.

In recent years, various insults (mechanical injury, bright light, and ischemia) ^{1 2 3} have been shown to protect the retina against subsequent light-induced photoreceptor degeneration (LIPD). A particularly curious form of protective insult was reported by Bush and Williams, ⁴ who found that optic nerve transection (ONT) affords histologic protection to the photoreceptors against light-induced injury. An unusual aspect of the ONT-induced photoreceptor protection is that it specifically affects the retinal ganglion cells (RGCs), yet subsequently protects the outer retina. This phenomenon implies the existence of retrograde communication systems within the retina, possibly involving Müller cells and FGF-2, ^{5 6} but does not exclude the possibility that the effect is specific to ONT. A nonspecific effect would suggest that similar responses might be occurring in a wide range of optic neuropathies. We hypothesized that the protective effect of ONT may be a generalizable effect and that other forms of inner retinal injury such as excitotoxic injury may also protect against LIPD. Furthermore, although FGF-2 has been implicated as the agent responsible for the ONT effect, other trophic factors may be involved. For example, intraocular injection of brain-derived neurotrophic factor (BDNF), neurotrophin (NT)-3, ciliary neurotrophic factor (CNTF), and glial-derived neurotrophic factor (GDNF) rescue photoreceptors in animal models of retinal degeneration. ^{7 8 9 10} In addition, the temporal relationship between ONT and the degree of protection against LIPD has not been reported. 

Hence, the aims of the present study were (1) to confirm that ONT protects against LIPD and to examine the time-dependent nature of this effect, (2) to correlate the degree of protection with changes in retinal trophic factor levels, (3) to determine the effect of NMDA-induced retinal injury on LIPD, and (4) to compare changes in retinal trophic factor levels after NMDA-induced excitotoxicity with changes after ONT. 

Methods

Treatment of Animals

Procedures used in this study conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Home Office in the United Kingdom. Adult female Wistar rats (200–250 g) housed in a 12-hour light–dark cycle were used for all experiments. Food and water were provided ad libitum. Anesthesia was achieved with a combination of intramuscular Hypnorm 0.4 mL/kg (fentanyl citrate 0.315 mg/mL and fluanisone 10 mg/mL; Janssen Pharmaceutica, Beerse, Belgium) and diazepam 0.4 mL/kg. Animals were killed by an overdose of intraperitoneal pentobarbitone. 

Optic Nerve Transection

A superior conjunctival approach to the optic nerve was chosen because it minimized operating time (<5 minutes per eye) and globe manipulation. Anesthetized rats underwent a lateral canthotomy, the superior lid was retracted with a 5-0 silk traction suture, and a superior fornix-based conjunctival flap was created. The superior muscle complex was divided (episcleral vortex veins were not transected) and the optic nerve exposed by blunt dissection. The posterior ciliary artery was identified on the ventromedial aspect of the nerve sheath and the optic nerve sheath was incised longitudinally, exposing the optic nerve. The nerve was then transected 3 mm from the globe, and the fundus was checked to ensure that retinal blood flow had not been interrupted. Sham surgery was identical and continued for the same duration, except that the optic nerve was not transected. 

Excitotoxic RGC Injury

Anesthetized rats received a 5-μL intravitreal injection of 4 mM NMDA (Tocris Neuramin, Essex, UK) in one eye and a 5-μL injection of sterile isotonic saline in the fellow eye. 

Photic Injury

Rats were placed separately in cages and exposed to evenly distributed, bright light. The floor of each cage was illuminated by approximately 2000 lux. The temperature inside each cage was maintained at 24°C, and the animals had free access to food and water. The rats were placed under these conditions at the same time of day for each experiment (1700 hours) and returned to their normal housing after 48 hours of constant exposure to light. 

Electroretinography

Maximum amplitude responses from dark-adapted rats were recorded as follows: Rats were dark adapted for at least 2 hours and prepared under dim red illumination. After anesthesia, the pupils were dilated with cyclopentolate (1%), the eyelids were separated by placing a cotton thread loosely around the equator of the globe, and rats were placed on a stereotaxic frame facing the stimulus at a distance of 50 cm. A reference electrode was placed through the tongue, a grounding electrode was attached to the scruff of the neck, and a platinum electrode was placed in contact with the central cornea. The cornea was intermittently irrigated with a physiologic saline solution (BSS; Alcon, Fort Worth, TX) to maintain the baseline recording and to prevent keratopathy from exposure. The dim red illumination was switched off during recordings, and the rats were kept warm during and after the procedure. 

Ten responses to a flash (10 μs, 0.1 Hz) from a photic stimulator (PS33-plus; Grass Telefactor, Inc., West Warwick, RI) set at maximum brightness (setting 16; approximately 17.5 cd-s/m²) were amplified (gain set at 300), filtered (100-Hz low-pass filter, DC high-pass filter, and 50-Hz filter, notch activated) and averaged (model 1902 Signal Conditioner/1401 Laboratory Interface; CED, Cambridge, UK) The b-wave amplitude was measured from the trough of the a-wave to the peak of the b-wave, and the a-wave was measured as the difference in amplitude between the recording at onset and the trough of the negative deflection. Baseline recordings were taken 1 to 4 days before ONT or injection of NMDA, 1 day before the photic injury, and 5 days after the completion of the photic injury. 

Assessment of Retinal mRNA Levels

The retinal levels of glyceraldehyde-3-phoshate dehydrogenase (GAPDH), rod opsin, FGF-2, glial fibrillary acidic protein (GFAP), CNTF, BDNF, nerve growth factor (NGF), GDNF, and Thy-1 mRNA were determined with a semiquantitative reverse transcription-polymerase chain reaction technique (RT-PCR), as described previously. ¹¹ Briefly, total RNA was isolated, and first-strand cDNA synthesis performed on 2 μg DNase-treated RNA. The individual cDNA species were amplified in a 10 μL reaction, containing the 2 μL cDNA aliquot, PCR buffer (10 mM Tris-HCl [pH 8.3] and 50 mM KCl), 4 mM MgCl₂, 200 μM of each dNTP, 4 ng/μL of both the sense and antisense primers and Taq polymerase (2.5 U). Reactions were initiated by incubating at 94°C for 10 minutes and PCR (94°C, 15 seconds; 52°C, 55°C or 56°C, 30 seconds; and 72°C, 30 seconds; performed for a suitable number of cycles followed by a final extension at 72°C for 3 minutes). Prior experiments had determined the linear phase of amplification for each set of primers, and the PCR product from each set of primers was sequenced to ensure its validity. Interexperimental variations were avoided by performing all amplifications in a single run. The oligonucleotides primer pairs and their annealing temperatures are shown in Table 1 . The PCR products of all primer pairs yielded single bands corresponding to the expected molecular weights, which are also shown in Table 1 . PCR reaction products were separated on 1.5% agarose gels, with using ethidium bromide used for visualization. The relative abundance of each PCR product was determined by digital analysis of gel photographs on computer (Labworks software; Ultra-violet Products, Cambridge, UK). For semiquantitative analysis, the ratio of the densitometric readings between the preconditioned and sham-treated eyes was calculated and compared with an internal standard mRNA ratio (GAPDH). 

Histopathology and Immunohistochemistry

Anesthetized rats were transcardially perfused with 50 mL of 10 mM phosphate-buffered saline, followed by 4% paraformaldehyde. The eyes were enucleated and immersion fixed for 1 hour in 4% paraformaldehyde, transferred to 10% neutral-buffered formalin overnight, and processed for routine paraffin-embedded sectioning on an automated tissue processor (Shandon Pathcentre, Thermo Shandon Inc., Pittsburgh, PA). Eyes were embedded sagittally and 5-μm serial sections were cut with a rotary microtome (Microm HM 330, McBain Instruments, Chatsworth, CA) and stained with hematoxylin and eosin. 

For immunohistochemistry, the eyes of saline-perfused rats were enucleated and immersion fixed in 4% paraformaldehyde for 20 minutes, hemisectioned, fixed for a further 20 minutes, and placed in 30% sucrose and 0.1 M sodium phosphate at 4°C overnight. Frozen sections (10 μm) were cut on a cryostat (Bright Instruments Ltd., Huntingdon, UK) and suitable hemiglobe sections within approximately 1 mm of the optic nerve head were recovered onto gelatin-coated slides. Air-dried slides were incubated at 4°C overnight with the primary antibody (rat anti-ChAT, 1:10; Roche Diagnostics, East Sussex, UK) and at room temperature for 45 minutes with the secondary antibody conjugated to FITC (anti-rat IgG-FITC conjugate, 1:100; Sigma-Aldrich, St. Louis, MO). 

Immunoblot Analysis

Proteins were extracted simultaneously with RNA from retina samples by using a standard procedure (Tri Reagent; Sigma-Aldrich). Protein samples were solubilized in buffer (20 mM Tris-HCl [pH 7.4], containing 2 mM EDTA, 0.5 mM EGTA, 1% SDS, 0.1 mM phenylmethylsulfonyl fluoride, 50 μg/mL aprotinin, 50 μg/mL leupeptin, and 50 μg/mL pepstatin A). An equal volume of sample buffer (62.5 mM Tris-HCl [pH 7.4], containing 4% SDS, 10% glycerol, 10% β-mercaptoethanol, and 0.005% bromophenol blue) was added, and the samples were boiled for 3 minutes. Electrophoresis of samples was performed using 10% polyacrylamide gels containing 0.1% SDS. Proteins were then blotted onto nitrocellulose. The blots were incubated with anti-actin (1:1000; Chemicon, Chandlers Ford, UK), anti-rhodopsin kinase (1:1000; Cambridge Bioscience, Cambridge, UK), anti-FGF-2 (1:200; Santa Cruz Biotechnology, Santa Cruz, CA), or anti-NF-L (1:1000; Chemicon) for 3 hours at room temperature, and appropriate secondary antibodies conjugated to horseradish peroxidase were subsequently used. Nitrocellulose blots were developed with a 0.016% (wt/vol) solution of 3-amino-9-ethylcarbazole in 50 mM sodium acetate (pH 5.0) containing 0.05% (vol/vol) Tween-20 and 0.03% (vol/vol) H₂O₂. 

Experimental Groups

The effect of ONT and NMDA-induced excitotoxicity on LIPD was investigated in three experimental groups. The first group (ONT-7 day group) was subjected to photic injury 7 days after unilateral ONT, the second group (ONT-21 day group) was subjected to photic injury 21 days after unilateral ONT (fellow eyes received a sham ONT), and the third group received a unilateral intravitreal injection of NMDA (fellow eyes received saline) 7 days before the photic injury (NMDA group). 

Statistical Analysis

For comparison of independent samples and normalizing for slight day-to-day variation in the ERG, the ratio of the a- and b-wave amplitudes between both eyes of each rat (one eye treated and the fellow sham-treated) were used as the unit of statistical analysis. Similarly, the paired-eye ratio of the RT-PCR and immunoblot densitometric readings was used as the unit of statistical analysis. A Student’s t-test or one-way ANOVA was used to compare means between two or more independent groups, a Tukey honest significant difference (HSD) test was used for post hoc comparisons, and a Bonferroni correction applied for comparisons at multiple time points within an individual experiment. A repeated-measures ANOVA was used to compare serial changes within a group. All statistical determinations were performed on computer (SPSS for Windows, ver.10; SPSS Science, Inc., Chicago, IL), and all data are expressed as the mean ± SEM, with P < 0.05 considered statistically significant. 

Results

The Effect of ONT and NMDA on the ERG

The changes in the a- and b-wave amplitudes 3, 10, and 21 days after ONT are shown in Figure 1 . The mean a-wave amplitude in the axotomized eyes showed a small but significant reduction at day 3 and 10 (P = 0.018 and 0.01, respectively) but had returned to baseline by 21 days (Fig. 1A) . In contrast, the b-wave amplitude in the axotomized eyes was unchanged from baseline at day 3, but was significantly reduced at days 10 and 21 (P = 0.009 and 0.006, respectively; Fig. 1B ). 

Six days after the intravitreal injection of NMDA, the mean b-wave amplitude in the NMDA-injected eyes (as a percentage of the saline-injected fellow eye) was significantly smaller than the ratio at baseline (P < 0.001, n = 9; Fig. 2B ) but the mean a-wave amplitude was unchanged (Fig. 2A) . 

Protective Effect of ONT and NMDA-Induced Excitotoxicity

Electroretinographic Data.

Figure 2 shows the protective effect that ONT and NMDA-induced excitotoxicity had on the a- and b-wave amplitudes of all three groups. Although, 6 days after ONT the mean a-wave amplitude of the axotomized eyes in the ONT-7 day group was significantly smaller than the ratio at baseline (P < 0.035, n = 12), 5 days after the photic injury, the amplitude was significantly greater than baseline (P < 0.001; Fig. 2A ). Similarly, the mean b-wave amplitude of the axotomized eyes in the ONT-7 day group was significantly greater after the photic injury than the amplitude at baseline (P < 0.001, n = 12; Fig. 2B ). In the ONT-21 day group, the mean a-wave amplitude of the axotomized eyes was also significantly greater after the photic injury than the amplitude at baseline (P < 0.001, n = 8; Fig. 2A ); and, although, 20 days after ONT, the mean b-wave amplitude of the axotomized eyes was significantly smaller than the amplitude at baseline (P = 0.007, n = 8), 5 days after the photic injury, the amplitude was significantly greater than baseline (P < 0.001, n = 8; Fig. 2B ). Hence, ONT provided functional protection against LIPD, and this protection was evident at both 7 and 21 days after the axotomy. 

In the NMDA group 5 days after the photic injury, both the a- and b-wave amplitudes were significantly greater than baseline (P < 0.001, n = 8; Fig. 2 ), indicating that the NMDA-injected eyes also displayed relative functional protection. 

There was no statistically significant difference between the ONT-21 day and the NMDA groups in the mean a-wave (P = 0.38) or b-wave (P = 0.75) amplitudes after photic injury. However, the mean a-wave amplitudes in both the ONT-21 day and the NMDA groups were significantly greater after the photic injury than the mean a-wave amplitudes in the ONT-7 day group (P = 0.01 and 0.028, respectively). Similarly, the mean b-wave amplitudes in these groups were significantly greater than the mean b-wave amplitude in the ONT-7 day group (P < 0.001). Representative ERG recordings from the ONT-21 day group and the NMDA group are shown in Figure 3 . 

RT-PCR Data.

Figure 4 shows the mean GAPDH, rod opsin, and Thy-1 mRNA levels in all three groups 5 days after the photic injury. In all three experimental groups, the level of GAPDH mRNA (the internal control gene) did not differ significantly (P = 0.87) between eyes, as indicated by the nearly 100% ratios in all three groups (Fig. 4) . In the ONT-7 day group, the level of rod opsin mRNA was significantly greater than the level of GAPDH (P = 0.006, n = 8), whereas the Thy-1 mRNA level was significantly smaller than the GAPDH level (P < 0.001, n = 8), indicating relative protection of the photoreceptor cell bodies (the location of the rod opsin mRNA) in the axotomized eyes, and a significant damage to RGCs. Similarly, in the ONT-21 day group, the rod opsin mRNA level was significantly greater than the GAPDH level (P = 0.002, n = 6), and the Thy-1 mRNA level was significantly smaller (P < 00.001, n = 6). The NMDA group also displayed a similar pattern. The rod opsin mRNA level was significantly greater than the GAPDH level (P = 0.003, n = 6), and the Thy-1 mRNA level was significantly smaller (P < 0.001, n = 6). 

Immunoblot Analysis.

Figure 5 shows the relative preservation of rhodopsin kinase (an outer segment protein) 5 days after photic insult in both the axotomized eyes (P = 0.035, n = 4) and the NMDA-injected eyes (P = 0.011, n = 4–6). In both the ONT-21 day group and the NMDA group the level of actin (the internal control protein) did not differ significantly (P = 0.87) between eyes, as indicated by the nearly 100% ratios in all three groups (Fig. 5) . There was a dramatic loss of neurofilament light (NFL, an axonal marker) by 21 days after ONT (P = 0.004) and a smaller, but significant loss of NFL 7 days after NMDA injection (Fig. 5 ; P = 0.023). 

Histopathology and Immunocytochemistry.

The axotomized eyes showed a marked loss of RGCs at both 7 and 21 days that was associated with slight thinning of the inner plexiform layer (IPL), but, after the photic injury, the axotomized eyes exhibited better preservation of the outer nuclear layer (ONL; Figs. 6A 6B 6C 6D ) than did the sham-axotomized eyes. The NMDA-injected eyes also exhibited loss of RGCs and thinning of the IPL, but showed better preservation of the ONL after the photic injury than the saline-injected eyes (Figs. 6E 6F ). A comparison of the cell body counts in the ONL is shown in Table 2 . 

The Effect of ONT and NMDA on ChAT Immunostaining

The characteristic ChAT staining 21 days after ONT was similar to that in the sham-axotomized eyes, indicating that cholinergic amacrine cells, in the inner nuclear layer (INL) and the ganglion cell layer (GCL; displaced amacrine cells) were relatively unaffected by ONT (Figs. 7A 7B) . Similarly, ChAT staining was unaffected in the saline-injected eyes (Fig. 7C) , but was abolished in the NMDA-injected eyes (Fig. 7D) , indicating some degree of injury to subganglionic retinal layers. 

The Effect of ONT and NMDA on Trophic Factor Levels

RT-PCR Data.

Table 3 shows the increase in mRNA levels in the axotomized and NMDA-injured eyes compared with the sham-treated eyes. Of the factors measured, GFAP and FGF-2 displayed the greatest increase after both axotomy and NMDA injection. FGF-2 mRNA levels were significantly greater 21 days after axotomy than at 7 days after (P = 0.006). BDNF displayed a small but significant reduction 21 days after axotomy compared with 7 days after (P = 0.032). CNTF levels were also elevated after ONT and NMDA-induced injury (P < 0.001). NGF levels were only slightly elevated by ONT and unaffected by NMDA injection. Retinal GDNF levels were not significantly altered by the ganglion cell injuries. 

Immunoblot Data.

Figure 8 shows the relative increase in FGF-2 protein after ONT and NMDA-induced injury compared with the internal control protein actin. Both 7 and 21 days after ONT, the expression of FGF-2 was significantly increased compared with the expression of actin (P = 0.032 and 0.015, respectively, n = 4). Similarly, the expression of retinal FGF-2 was also increased 7 days after NMDA injection, compared with the expression of actin (P = 0.018, n = 4). 

Discussion

Reports of the effect of ONT on the flash ERG are conflicting. ONT has been reported to have no effect on the flash ERG in rats and cats, ^{12 13} and the few recordings of the flash ERG in humans after traumatic or surgical ONT proximal to the entry of the central retinal artery were reported as normal or very mildly affected. ^{14 15} In contrast, Gargini et al. ¹⁶ have reported that the flash ERG is significantly altered 3 to 4 weeks after ONT in rats. They showed b-wave changes over a wide range of stimulus intensities and more subtle a-wave changes at high intensities. They attributed the changes to FGF-2–induced suppression of retinal function, an explanation supported by the finding that an intravitreal injection of FGF-2 also suppressed the b-wave. Vaegan et al. ¹⁷ have recently reported that the flash ERG in cats is mildly affected by ONT, and evidence has accrued indicating that at least one optic neuropathy (glaucoma) affects the flash ERG. ^{18 19 20} A unified interpretation of these observations is complicated by the varied and dynamic effects that ONT has on the retina: ONT not only disconnects the retina from the remainder of the brain, triggering retrograde death of RGCs and causing possible transsynaptic degeneration, but also acts as a form of conditioning, albeit extreme, that protects photoreceptors from light-induced injury. Furthermore, all these effects appear to be species dependent. 

ERG changes after ONT must be interpreted cautiously, because they may reflect inadvertent ischemic changes. The vascular supply to the rat retina is derived from a single vessel (the posterior ciliary artery), a branch of the ophthalmic artery that travels along the ventromedial aspect of the optic nerve and divides into retinal and uveal branches as it enters the posterior globe. ²¹ Ischemic injury can occur if the globe is retracted for extended periods, because traction interferes with the retinal blood supply, as evidenced by whitening of the fundus during this maneuver. In addition, the posterior ciliary artery may be transected during the division of the optic nerve; however, if this occurs, complete ocular ischemia occurs, and the ERG is flat. Intracranial transection of the optic nerve may exclude the possibility of disruption of the retinal blood supply, but may not affect the retina in the same way as an intraorbital axotomy. ²²  

Given that there is no evidence to suggest that RGCs contribute to the flash ERG, the serial changes in the ERG observed after ONT and NMDA injection could be explained either by a direct or indirect insult to the anatomic substrates that are responsible for the ERG in the middle and outer retina. ONT per se directly affects only the GCL, but inadvertent disruption of retinal blood supply during ONT may affect deeper layers. However, immunostaining with ChAT, a marker for cholinergic amacrine cells that forms distinct laminae in the IPL, ²³ was normal after ONT, as previously described in neonatal rats. ²⁴ ChAT-containing retinal neurons are very sensitive to ischemic injury ²³ ; hence, the normal staining after ONT is further evidence that the ONT did not significantly affect the blood supply. Therefore, our findings suggest that ONT causes subtle indirect effects to the middle and outer retina, possibly mediated by Müller cell activation, indicated by elevated GFAP mRNA levels after ONT—findings that are consistent but less marked than those reported by Gargini et al. ¹⁶  

In 1991, Bush and Williams ⁴ found that ONT protects photoreceptors against light-induced injury. To our knowledge, the present report is the first confirmation of this intriguing effect. Our histologic results were very similar to those previously reported ⁴ and the ERG, RT-PCR, and immunoblot data all provided evidence in support of the protective effect of ONT against LIPD. The relatively greater loss of rhodopsin kinase protein compared with rod opsin mRNA may be explained by the greater light-induced loss of rod outer segments compared with the loss of cell bodies in the ONL. It is interesting to note that this protection of the axotomized eye occurs despite a behavioral tendency for the rats to shield the sighted eye during intense light exposure, a phenomenon also reported by Bush and Williams. ⁴  

Kostyk et al. ⁵ showed that FGF-2 immunoreactivity was increased in the ONL of the retina 3 to 4 weeks after ONT and related the known neurotrophic properties of FGF-2 to the photoreceptor protection. Several insults have been shown to be protective against light-induced injury and to upregulate FGF-2. ^{1 2 3 25} This fact, together with the neuroprotective properties of exogenous FGF-2, ¹ has led to the concept that FGF-2 may be an important endogenous survival factor in the rat retina, particularly for photoreceptors. ^{26 27} We showed a significant upregulation of FGF-2 mRNA and protein after ONT, an increase that was greater 21 days after ONT than at 7 days, corresponding with the greater degree of photoreceptor preservation that we observed at this later time point. There was, however, a discrepancy between the level of FGF-2 protein and mRNA expression after ONT and NMDA injection. Other investigators, ¹⁶ using in situ hybridization to measure FGF-2 mRNA, have reported a large discrepancy between the expression of FGF-2 protein and mRNA after ONT, but it is possible that the discrepancy in the present study reflects the use of semiquantitative rather than quantitative techniques. However, in general, the findings support the concept that endogenous FGF-2 is an important photoreceptor protectant. 

LaVail et al. ²⁶ have demonstrated that a variety of substances delivered intravitreally protect mouse photoreceptors against injury, and many of these survival factors are synthesized endogenously. Increased CNTF expression by Müller cells has been reported to occur after ONT. ²⁸ However, given that photoreceptors lack receptors for CNTF and BDNF, ^{29 30} but that Müller cells have receptors for most of the molecules involved in photoreceptor rescue and become activated after retinal injury, it has been suggested that Müller cells play an integral role in endogenous neuroprotection. ^{6 26 31 32 33 34} Harada et al. ^{33 34} have provided evidence for a macroglial–neuronal and microglial–macroglial communication system involved in endogenous neuroprotection after LIPD. Of note, they hav shown that p75^NTR ligand (probably NGF) reduces FGF-2 production in Müller cells, ³⁴ which is consistent with the lack of upregulation of NGF, as found in the present study. 

Stone et al. ²⁷ have hypothesized that the reason FGF-2 expression is not permanently upregulated is that the structural protection comes at a functional cost. They suggest that FGF-2 induces a form of metabolic suppression in the photoreceptors and predicted that the a-wave of the ERG would be suppressed by FGF-2. We observed precisely this effect 3 and 10 days after ONT, but the response had normalized by day 21. Therefore, our results support the findings of Stone et al. and lend weight to their prediction; however, the recovery of the a-wave, despite the increasing expression of FGF-2, implies the presence of other mechanisms. 

Although NMDA-induced excitotoxicity has been used as a method of inducing RGC death, ¹¹ it also injures other cells, including a subset of cholinergic amacrine cells in the INL known to express the NMDA receptor. ^{35 36} The effect on these cells was evidenced by the loss of the characteristic ChAT immunostaining in the NMDA-injected eyes. To our knowledge, the effect of NMDA on the ERG has not been reported, but the ischemia-like changes (reduction in the b-wave amplitude with preservation of the a-wave) can probably be explained by NMDA-induced injury to cells within the INL. 

NMDA-induced excitotoxic injury to the inner retina protected against subsequent LIPD in a manner similar to ONT. This indicates that the protective effect is generalizable, at least to some extent, and that inner retinal injury, particularly RGC death initiates a response from the retina that results in a delayed protection of the photoreceptors against photic injury. Furthermore, intravitreal NMDA caused an increase in GFAP mRNA and an increase in retinal FGF-2 mRNA and protein levels, suggesting that macroglial cell activation and the release of FGF-2 may be an essential component of this protective response. 

In conclusion, ONT in adult Wistar rats caused small but significant serial changes in both the a- and b-wave of the flash ERG. Intravitreally injected NMDA caused more marked reduction of the b-wave without affecting the a-wave and clearly affected cholinergic amacrine cells in the INL. Furthermore, ONT and NMDA-induced excitotoxicity produced upregulation of retinal GFAP, FGF-2, CNTF, and NGF, with a temporal response that corresponded to the protection afforded against LIPD, suggesting a causal relationship. These endogenous protective responses from the retina may be responsible in a complex and dynamic manner for both the photoreceptor protection and electroretinographic changes and may play a role in diseases affecting the RGCs. 

 Supported by the European Community (E.C. PRO-AGE-RET Program) QLK6-CT-2001-00385.

 Submitted for publication July 1, 2003; revised September 22, 2003; accepted October 22, 2003.

 Disclosure: R.J. Casson, None; G. Chidlow, None; J.P.M. Wood, None; M. Vidal-Sanz, None; N.N. Osborne, None

 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.

 

Primer Sequences

mRNA	Primer Sequences	Product Size (bp)	Annealing Temperature (°C)	Accession Number
GAPDH	5′-CATCAAGAAGGTGGTGAAGCAGG-3′	206	52	AF106860
	5′-CCACCACCCTGTTGCTGTAGCCA-3′
Rhodopsin	5′-CAGTGTTCATGTGGGATTGACT-3′	365	52	Z46957
	5′-ATGATTGGGTTGTAGATGGAGG-3′
Thy-1	5′-CGCTTTATCAAGGTCCTTACTC-3′	344	52	X03150
	5′-GCGTTTTGAGATATTTGAAGGT-3′
BDNF	5′-AGCTGAGCGTGTGTGACAGTAT-3′	293	55	M61178
	5′-GTCTATCCTTATGAACCGCCAG-3′
GFAP	5′-ATTCCGCGCCTCTCCCTGTCTC-3′	437	55	U03700
	5′-GCTTCATCCGCCTCCTGTCTGT-3′
FGF-2	5′-CGTCAAACTACAGCTCCAAGCAGA-3′	235	55	M22427
	5′-GGATCCGAGTTTATACTGCCCAGT-3′
CNTF	5′-TGGCTAGCAAGGAAGATTCGT-3′	468	56	X17457
	5′-ACGAAGGTCATGGATGGACCT-3′
GDNF	5′-GACTCTAAGATGAAGTTATGG-3′	483	56	L15305
	5′-TTTGTCGTACATTGTCTCGG-3′
NGF	5′-CTGGACTAAACTTCAGCATTC-3′	395	55	M36589
	5′-TGTTGTTAATGTTCACCTCGC-3′

Figure 1.

 

The effect of ONT on the a- and b-wave amplitudes. (A) The a-wave amplitude was significantly reduced 3 and 10 days after ONT compared with sham surgery, but had returned to normal by day 21. (B) The b-wave amplitude was not significantly affected at day 3 but was reduced 10 and 21 days after ONT. Sham surgery did not significantly affect the amplitudes compared with no treatment. Ratios are shown between (▪, n = 8–9) axotomized and untouched fellow eyes, ( , n = 13) sham-axotomized and untreated eyes; and (□, n = 6) untreated right and left eyes. *P < 0.05, **P < 0.01.

Figure 2.

 

Relative preservation of the a-wave (A) and b-wave (B) against light-induced photoreceptor degeneration 7 and 21 days after ONT and 7 days after intravitreal injection of NMDA. Recordings after injury were made 6 or 20 days after ONT and 6 days after NMDA; recordings after exposure to light were made 5 days after the completion of the photic injury. Amplitudes are expressed as the mean ratio between the eyes of each rat. One eye receiving either an ONT or an intravitreal NMDA injection, and the fellow eye receiving sham treatment. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 3.

 

Representative ERG recordings. (A) Uninjured control. (B, C, E, F) Recordings made 5 days after the termination of the photic insult: (B) 21-day ONT eye and (C) its fellow sham-treated eye; (E) NMDA-injected eye after photic insult and (F) its fellow vehicle-injected eye. (D) Six days after NMDA injection but before photic insult.

Figure 4.

 

RT-PCR measurements after RGC insults and photic injury. There was relative preservation of rhodopsin mRNA and relative loss of Thy-1 mRNA compared with the housekeeping gene product (GAPDH) after light-induced photoreceptor injury in axotomized eyes and eyes that had previously received an injection of NMDA. Rats underwent a unilateral ONT 7 or 21 days before the photic injury or 7 days before an intravitreal injection of NMDA and were killed 5 days after the photic injury. Mean mRNA levels are expressed as the ratio between the treated eye and the sham-treated eye. **P < 0.01, ***P < 0.001.

Figure 5.

 

(A) Representative immunoblots after RGC insults and photic injury and (B) graphic display of the densitometric analysis. There was relative preservation of RK and relative loss of NFL compared with the housekeeping protein (actin) after light-induced photoreceptor injury in axotomized eyes and eyes that had received an injection of NMDA. Rats underwent a unilateral ONT 21 days before the photic injury or 7 days before an intravitreal injection of NMDA and were killed 5 days after the photic injury. Mean protein levels are expressed as the percentage ratio between the treated eye and the sham-treated eye. *P < 0.05, **P < 0.01. Ax, axotomized eye; Sh, sham-axotomized eye; RK, rhodopsin kinase; NFL, neurofilament light.

Figure 6.

 

Retinal sections demonstrating relative preservation of photoreceptors 5 days after photic injury in eyes that had undergone RGC injury. (A) A sham-axotomized eye that underwent photic insult 7 days later and (B) the axotomized fellow eye showing better preservation of the ONL. (C) A sham-axotomized eye that underwent photic insult 21 days later and (D) axotomized fellow eye showing better preservation of the ONL. (E) Saline-injected eye that underwent photic insult 7 days later, and (F) NMDA-injected fellow eye showing better preservation of the ONL.

Table 2.

 

Effect of Retinal Ganglion Cell Injury Followed by the Photic Insult on the Photoreceptor Cell Bodies

Group	Treatment Group	Control Group
Untouched control	10–12	NA
7-day ONT	6–7	4–5
21-day ONT	6–7	3–4
NMDA	8–9	3–4

 Data are the number of rows of cells in the ONL.

Figure 7.

 

The effect of ONT and NMDA injection on ChAT immunostaining. (A) Twenty-one days after sham axotomy and (B) after axotomy, the characteristic staining pattern is still clearly evident. (C) Saline injection did not affect the staining pattern, but was abolished 7 days after NMDA injection (D). Arrows: amacrine cell bodies.

Table 3.

 

The Effect of ONT and NMDA on Total Retinal mRNA Levels of Various Trophic Factors and GFAP

Treatment	FGF-2	CNTF	BDNF	NGF	GDNF	GFAP
NMDA	141.8 ± 16.8*	79.9 ± 6.9*	2.6 ± 5.8	5.4 ± 20.1	−1.3 ± 13.4	281.9 ± 60.2*
ONT 7 days	67.3 ± 90.0*	48.9 ± 8.7*	−2.39 ± 4.1	24.1 ± 7.9†	−10.7 ± 9.6	92.6 ± 13.2*
ONT 21 days	122.8 ± 21.3*	40.2 ± 5.9*	−8.1 ± 3.5†	23.6 ± 7.1‡	9.9 ± 9.3	76.6 ± 9.8*

 Data are expressed as the mean percentage increase in mKNA in the treated eyes (relative to the control eyes). ± SEM.

 Statistical significance treated eye versus control eye by Student’s paired t-test (*P < 0.001, †P < 0.05, ‡P < 0.01), where n = 8–17 (NMDA), n = 12 (ONT 7 days) and n = 17–23 (ONT 21 days).

Figure 8.

 

Relative upregulation of FGF-2 compared with the house-keeping protein (actin) 7 and 21 days after ONT and 7 days after intravitreal injection of NMDA. *P < 0.05