Journal of Neuro-Ophthalmology, March 2001, Volume 21, Issue 1

Functional and Morphologic Comparison of Two Methods of Produce Transient Retinal ischemia in the Rat

OBJECTIVES: Much of our knowledge of the pathophysiology of retinal ischemic injury is from a multitude of studies that use in vitro or in vivo animal models of retinal ischemia followed by reperfusion. The objective of this study was to compare histopathologic and electrophysiologic (electroretinography) parameters using two different models of transient retinal ischemia: high intraocular pressure (HIOP) and suture ligation of the optic nerve (SL). METHODS: Transient retinal ischemia was induced using the HIOP model or the SL model in the Sprague-Dawley rat for either 30 or 60 minutes. Histopathologic outcome was determined at 1 and 7 days after ischemia. In addition, electroretinography (ERG) was performed at 2 hours, I day, 3 days, and 7 days after ischemia. RESULTS: At 1 and 7 days after 30 minutes of ischemia, there were no significant histopathologic abnormalities in the retina with either model, except for a slight decrease of the cell count in the ganglion cell layer (GCL) with the SL method. After 60 minutes of ischemia, there was significant thinning of the inner retina. There was a significant early dropout of cells at 1 day in the inner nuclear layer (INL) in the HIOP method compared to the SL method where the dropout was delayed and gradually progressive. Dropout of cells in the GCL was early (I day) and gradually progressive in both models but more severe in HIOP than SL. There was a significant decrease in the ERG b-wave amplitudes as early as 2 hours after both 30 and 60 minutes of ischemia compared to preischemic baselines. CONCLUSIONS: The degree of retinal injury after transient retinal ischemia was more severe at 1 day after reperfusion in the HIOP method compared to the SL method but was similar at 7 days in both models. Furthermore, our data suggests that functional assessment of ischemic damage by electroretinography may be a more sensitive parameter than conventional histopathologic quantification. The timing of either measurement relative to the ischemic stimulus is critical because histologic measurements performed too early after ischemia may underestimate the degree of injury.

Journal of'A] euro- Ophthalmology 21( 1): 62- 68, 2001. © 2001 Lippincott Williams & Wilkins, Inc., Philadelphia Functional and Morphologic Comparison of Two Methods to Produce Transient Retinal Ischemia in the Rat Daniel M. Rosenbaum, MD, Pearl S. Rosenbaum, MD, Manjeet Singh, MD, Gaurav Gupta, MD, Himanshu Gupta, MD, Bing Li, MD, and Steven Roth, MD Objectives: Much of our knowledge of the pathophysiology of retinal ischemic injury is from a multitude of studies that use in vitro or in vivo animal models of retinal ischemia followed by reperfusion. The objective of this study was to compare histo­pathologic and electrophysiologic ( electroretinography) param­eters using two different models of transient retinal ischemia: high intraocular pressure ( HIOP) and suture ligation of the optic nerve ( SL). Methods: Transient retinal ischemia was induced using the HIOP model or the SL model in the Sprague- Dawley rat for either 30 or 60 minutes. Histopathologic outcome was deter­mined at 1 and 7 days after ischemia. In addition, electroretin­ography ( ERG) was performed at 2 hours, 1 day, 3 days, and 7 days after ischemia. Results: At 1 and 7 days after 30 minutes of ischemia, there were no significant histopathologic abnormalities in the retina with either model, except for a slight decrease of the cell count in the ganglion cell layer ( GCL) with the SL method. After 60 minutes of ischemia, there was significant thinning of the inner retina. There was a significant early dropout of cells at 1 day in the inner nuclear layer ( INL) in the HIOP method compared to the SL method where the dropout was delayed and gradually progressive. Dropout of cells in the GCL was early ( 1 day) and gradually progressive in both models but more severe in HIOP than SL. There was a significant decrease in the ERG b- wave amplitudes as early as 2 hours after both 30 and 60 minutes of ischemia compared to preischemic baselines. Conclusions: The degree of retinal injury after transient retinal ischemia was more severe at 1 day after reperfusion in the HIOP method compared to the SL method but was similar at 7 days in both models. Furthermore, our data suggests that func­tional assessment of ischemic damage by electroretinography may be a more sensitive parameter than conventional histo­pathologic quantification. The timing of either measurement relative to the ischemic stimulus is critical because histologic measurements performed too early after ischemia may under­estimate the degree of injury. Manuscript received August 18, 2000; accepted December 15, 2000. From the Departments of Neurology ( DMR, MS, GG, HG), Neuro-science ( DMR, MS, GG, HG), Ophthalmology and Visual Sciences ( DMR, PSR), and Pathology ( PSR), Albert Einstein College of Medi­cine, Bronx, New York; and the Department of Anesthesia and Critical Care ( BL, SR), University of Chicago, Chicago, Illinois. Address correspondence and reprint requests to Daniel M. Rosenbaum, MD, Albert Einstein College of Medicine, 1410 Pelham Parkway S., K341, Bronx, NY 10461; e- mail: drosenba@ aecom. yu. edu. Key Words: Ischemia- Retina- Rat- Suture ligation- High intraocular pressure. Retinal ischemia may occur clinically in various set­tings, such as after acute retinal arterial occlusion, ca­rotid artery disease, or other ocular disorders in which ischemia may play a pathogenetic role ( e. g., diabetes mellitus, hypertension, or glaucoma) ( 1). Much of our knowledge of the pathophysiology of retinal ischemic injury is from a multitude of studies that use in vitro or in vivo animal models of retinal ischemia followed by reperfusion. In vitro models allow the opportunity to study separately hypoxia and substrate deprivation, the two major components of ischemia ( 2,3). However, in­terpretation of the results of these experiments is com­plicated by the contribution of toxic metabolic products and the inability to remove these products. In vivo mod­els cannot separate the influence of hypoxia and sub­strate deprivation, and results may be strongly influenced by changes in blood flow ( 4,5). Nonetheless, in vivo models are of great significance and widely studied, be­cause they are believed to be more representative of clinical disease. Experimental studies have used such species as mon­key ( 6), cat ( 7), rabbit ( 8), and rat ( 9- 11). The use of less costly animals in the smallest number possible, while providing reproducible results relevant to human disease, is a goal of these studies. There are various practical and scientific advantages in the use of rats for these types of studies, and, as a result, many investigations of retinal ischemia involving this species have been reported ( 9,11- 14). The rat retina is well vascularized ( 15), unlike that of other small animals such as rabbits ( 16), and is analogous to the human retina. In larger more costly animals such as cats, the functional recovery, especially in the early hours after ischemia, seems more robust ( 17). This observation is a disadvantage for designing experi­ments that seek to measure the degree of functional im­provement after ischemia. Moreover, long- term follow-up after ischemia is more difficult and costly in such animal species. Despite the large number of earlier stud­ies, the preferred method for inducing experimental isch­emia in the rat retina in vivo is not known. At least four 62 RETINAL ISCHEMIA MODELS IN RATS 63 different approaches have been described: 1) increase intraocular pressure ( IOP) above systolic arterial blood pressure ( 9,10,18) ( high intraocular pressure [ HIOP] method), 2) ligation of the central retinal artery ( 9,11, 19), including the optic nerve, using an occluding suture ( SL), 3) a photothrombotic method ( 20), and 4) tempo­rary carotid artery occlusion ( 21). The photothrombotic method ( 20) is relatively simple, but it suffers from several substantial disadvantages: reti­nal damage is variable, perhaps because of different de­grees of light exposure throughout the rat retina ( 20); the retina is inevitably detached ( 20); the injury itself re­sembles a complete infarction, producing widespread retinal necrosis ( 22); and because capillary thrombosis ( the degree of which is not yet known) occurs ( 23,24), the model is that of permanent rather than reversible ischemic insult, effectively precluding the study of post­ishemic reperfusion events. Temporary occlusion of the carotid arteries ( 21) is simple to apply and is reversible, but its drawbacks include the need for extensive surgical exposure ( invasiveness), effects of such ischemia on or­gans other than the retina ( i. e., brain), and relative retinal sparing even in the presence of prolonged carotid occlu­sion. Because of these limitations, the photothrombotic and temporary carotid artery occlusion methods are less preferred in experimental studies of retinal ischemia. In contrast, the SL and HIOP methods are reversible, simple to apply, and require little specialized equipment or surgical manipulation. Because these two methods re­sult in transient reversible ischemia, it seems more ap­propriate to extrapolate results from studies using these techniques to humans, because they more closely parallel the usual clinical situation. The major limitation in in­terpreting results using these techniques is that it is clini­cally unusual to find complete cessation of retinal flow in patients. Although SL and HIOP methods produce ischemia by similar mechanisms, there has not been a previous con­trolled comparison of these two methods in the rat. Moreover, most previous studies have examined either electrophysiologic or histologic measurements of dam­age in isolation, and little data are available demonstrat­ing electrical recovery over an extended period of time after ischemia. The purpose of this study was to deter­mine the relative efficacy of the HIOP and the SL meth­ods in studies of transient retinal ischemia and subse­quent recovery in order to determine which of these two methods would be preferred in in vivo studies of retinal ischemia. Specifically, this study compares the histo­pathologic and electrophysiologic ( electroretinography [ ERG]) parameters using these two methods in the rat. Additionally, this study attempts to correlate retinal elec-trophysiology to histopathology subsequent to transient retinal ischemia. METHODS Animals and anesthesia Procedures used in this investigation conformed to the Association for Research in Vision and Ophthalmology ( ARVO) Resolution on the Use of Animals in Research and were approved by our animal care committee. We studied Sprague- Dawley rats, 150 to 200 gms, purchased from Taconic ( Germantown, NY), maintained in a 12- hour on/ 12- hour off light- dark cycle. Animals were fasted overnight before surgery. The rats were anesthe­tized with an intraperitoneal mixture of ketamine 30 to 40 mg/ kg and intramuscular xylazine 2.5 mg/ kg injec­tion. Adequacy of anesthesia was tested by tail clamping with a hemostat, and supplemental intramuscular doses of ketamine and xylazine were administered as needed. The rat was positioned prone during all of the measure­ments. Body temperature was maintained throughout all experiments at 37° C with a heating pad. Induction of ischemia High intraocular pressure method. This procedure has been described in detail previously ( 10). In brief, the anterior chamber OD was cannulated using a 27- gauge, ^- inch needle attached by a three- way stopcock to an infusion of sterile 0.9% saline and a manometer. Under direct vision, the needle tip was placed within the ante­rior chamber, and the corneal puncture site was sealed with cyanoacrylate cement. IOP was raised to 150 mm Hg in order to exceed systemic arterial blood pressure. After completion of the target period of ischemia, the needle was withdrawn and the IOP normalized. Suture ligation method. We have described this pro­cedure in recent publications ( 11,25- 27). A sterile 2- 0 suture was placed around the retrobulbar optic nerve and blood vessels OD, and the suture was pulled through a short length of polyethylene tubing ( PE- 200). By push­ing the tubing toward the eye while clamping the suture to maximal tightness, we were able to produce complete ocular ischemia for the target period of time. Complete loss of the electroretinogram b- wave as well as obliteration of retinal vessels by fundoscopic exami­nation served as evidence of retinal ischemia in both experimental models. The contralateral eye of each ani­mal served as a nonischemic control. One drop of gen-tamicin ophthalmic solution was applied topically to the ischemic eye before and after the eye was ren­dered ischemic. Electroretinography The procedures used were those we have previously reported ( 5,11,26). In brief, rats were dark- adapted over­night, their pupils dilated with tropicamide 0.5% and Cyclomydril ( Alcon Laboratories, Inc., Ft. Worth, TX). A platinum electroencephalogram ( EEG) electrode was placed on the topically anesthetized cornea, a reference electrode was placed on the ipsilateral mastoid, and a ground electrode was placed on the lower dorsum. The average response to 3 to 4 white- light flashes generated at a distance of 15 cm from the rat's eyes was recorded. The ERG data were analyzed as previously described ( 5,11,26). Light microscopy The animals were anesthetized and the eyes enucle­ated at the chosen survival time points and then fixed in Trump fixative, consisting of 11.6 g of sodium J Neuro- Ophthalmol, Vol. 21, No. 1, 2001 64 D. M. ROSENBAUM ET AL. monophosphate and 2.7 g of sodium hydroxide dissolved in 500 mL of distilled water, 100 mL of 37% formalde­hyde, and 20 mL of glutaraldehyde, with further addition of distilled water to a volume of 1000 mL. Enucleated globes were then sectioned in the vertical meridian and the inferior portion of the eyeball ( retina, choroid, and sclera) embedded in epoxy resin. One micron- thick sec­tions were stained with 1% toluidine blue. The retinal histoarchitecture was evaluated as previously described by light microscopy ( 10,11). Measurements of the thick­ness of the retinal layers were performed as follows: 1) outer limiting membrane ( OLM) to inner limiting mem­brane ( ILM), 2) outer nuclear layer ( ONL), 3) outer plexiform layer ( OPL), 4) inner nuclear layer ( INL), and 5) IPL to ILM. The mean value for these measurements taken in four adjacent areas of the inferior retina within 1 mm of the optic nerve was calculated. Additionally, manual cell counts of the INL and ganglion cell layer ( GCL) were performed over a length of 200 microns in the inferior peripapillary region of the retina. These mea­surements were performed in the same area of retina in all the eyes in order to prevent any effect on the results because of possible regional anatomic variation. Studies Each experimental group ( HIOP or SL model) was divided so that the retina of each set of animals was rendered ischemic for 30 or 60 minutes. Repeat ERG examination was performed 2 hours, 1 day, 3 days, and 7 days after the end of 30 or 60 minutes of ischemia. These ischemic times were chosen on the basis of earlier studies that reported severe histologic damage after 60 minutes of iscliemia in the rat and lesser degrees of dam­age after shorter periods of ischemia ( 9,14). Eyes were harvested for histologic examination at either 1 or 7 days after ischemia. Statistics Electroretinography b- wave amplitudes were normal­ized to baseline values and expressed as a percent of the baseline. To account for variation in the ERG amplitudes ( e. g., day- to- day variation within a subject), values ob­tained for follow- up examinations after ischemia ended were corrected by dividing the normalized ischemic value by the normalized control value ( control ERG am­plitude at a given time point divided by the baseline control) ( 11). Repeated measures of analysis of variance ( ANOVA) were used to examine the changes in wave amplitude over time compared to baseline. Unpaired t tests were used to compare results between groups at matched follow- up time points after ischemia. Histologic data was examined using paired t tests to compare control to ischemic retina of paired eyes with­in groups, on either day 1 or day 7 after ischemia, and unpaired t tests to compare ischemic results be­tween groups. RESULTS Histopathology Ischemia sustained for 30 minutes using HIOP or SL resulted in no significant changes compared with the control in retinal thickness 1 or 7 days later. There was, however, a slight decrease in the GCL count at 7 days in the SL model. Ischemia sustained for 60 minutes using either HIOP ( Figs. 1A, IB, and 2) or SL ( Figs. 1C, ID, and 2) resulted in the typical histopathologic features expected subsequent to acute retinal ischemia ( 9,10,14). One day after ischemia, a decrease in the ML cell count was evident in the eye subjected to HIOP only, and a significant decrease in cell counts in the GCL in both models was noted ( Fig. 1C, D). Light microscopy at 7 days after ischemia revealed similar histopathologic fea­tures in HIOP and SL models of retinal ischemia. There was a significant decrease in overall retinal thickness, with marked thinning of the inner retinal layers and ex­tensive disorganization of the inner retinal histoarchitec­ture ( Fig. 1A, C). The number of cells in the ML and GCL of the ischemic retinas were reduced in both mod­els ( Fig. IB, D). Of note, there was no further decline in the INL cell count after one day in the HIOP model as compared to the delayed decrease in the INL cell count that was seen 7 days after ischemia in the SL model. There was slight thinning of the OPL 7 days after isch­emia in the SL model only. The histoarchitecture of the ONL showed disruption of the orderly, vertically ori­ented, columnar arrangement of cells and cytologic ir­regularities ( Fig. 2). Electrophysiology Both techniques produced reliable and reversible reti­nal ischemia. In the early postischemic period ( first 2 hours of recovery) the b- wave recorded similarly in both models regardless of the duration of the ischemic insult. No significant changes in wave amplitudes were found over time compared to baseline in the nonischemic eyes of any of the groups. There was complete absence of ERG activity during ischemia and varying degrees of reappearance of the waveforms during recovery, depend­ing upon the duration of ischemia. The b- wave amplitude increased throughout the recovery period, after 30 minutes of ischemia, to a final value of 74.9 ± 21.9% ( p < 0.0095 vs baseline) at 7 days after ischemia for HIOP and 73.9 ± 7.1% for SL ( Table 1). In the 60- minute ischemic group, recovery of the b-wave plateaued by 120 minutes after ischemia ended, and the final value of 8.5 ± 2.1% ( p < 0.00009 vs base­line) at 7 days after ischemia for HIOP and 11.7 ± 1.8% for SL did not differ significantly from values recorded at earlier recovery periods. The b- wave amplitude fol­lowing 60 minutes of ischemia declined at 1 day, at a time when the histology only demonstrated mild patho­logic changes and remained depressed for up to 7 days ( Table 1). DISCUSSION The mechanisms of cell death caused by ischemia and reperfusion in the retina are not yet fully understood. The time course of postischemic neuronal damage in the retina is believed to be similar to that observed in other regions of the central nervous system. It appears that a / Neuro- Ophthalmol, Vol. 21, No. 1, 2001 RETINAL ISCHEMIA MODELS IN RATS 65 60 min HlOP 200 180 160 E 140 ^. CO 1 20 CO LU 100 Z o 80 200 180 160 ^ 140 CO LU 100 Z ^ 80 o X 60 40 20 0 n =] CONTROL Y///////// A \ DAY ^ ^ ^ 7 DAYS i t n^ g E o 60 min SL ^ J 100 O O 60 LU 40 o 20 0 i 1 CONTROL Y///////// A \ DAY mmm 7 DAYS _ n ooZ> Q_ o 60 min HlOP T I = I CONTROL V//////// A \ DAY fei^&& aa 7 DAYS n INL GCL 60 min SL m i 1 CONTROL V///////// A \ DAY mmm 7 DAYS R INL GCL FIG. 1. A, C: Measurements ( mean ± standard error of mean [ SEM]) of the thickness of retinal layers of nonischemic eyes ( control), and eyes 1 day and 7 days after 60 minutes ischemia show a significant decrease in overall retinal thickness ( OLM- ILM) and in the inner retinal layers ( INL, IPL- ILM) in both the HlOP ( A) and SL ( C) models. B, D: Manual cell counts of the INL and GCL in control eyes and 1 day and 7 days after 60 minutes ischemia. The INL cell count is significantly reduced at 1 day in the HlOP model ( B); a delayed decrease in the INL cell count is noted at day 7 in the SL model ( D). The GCL cell count is significantly reduced at 1 day and 7 days in both the HlOP ( B) and SL ( D) models ( n = 5 per group). * p < 0.05. f p < 0.01. " maturation" phenomenon is present, whereby ischemic damage, at least according to histologic criteria, becomes more evident with increasing recovery times following ischemia. The initial ischemic insult results in cellular perturbations that continue to progress despite, or per­haps because of, reperfusion of the ischemic tissue ( 28). It is commonly accepted that ischemia and reperfusion lead to the generation of oxygen free radicals and excit­atory amino acids, leading to cellular damage primarily from massive influx of Ca+ 2. Ca+ 2 influx results in the activation of enzymes such as lipases, proteases, endo-nucleases, nitric oxide synthase, and damage to cell membranes and DNA ( 4). Activation of these enzymes could lead to further increases in the production of dam­aging oxygen free radicals ( 29- 32). Delayed cell death in the retina could also be the result of ischemia- triggered programmed cell death, or apoptosis; however, the mechanisms of this latter phenomenon remain to be de­termined ( 10,33- 35). The primary implication of the evolving injury after ischemia is that the extent of postischemic injury may be underestimated if examination is performed too early in the recovery period. In the present study, we have shown that the functional ( ERG) and conventional histologic changes after ischemia progress over time and that ERG disturbance predate morphologic changes. In earlier studies, it has been shown that ganglion cell loss in the rat retina after ischemia increases with time ( 14). How­ever, most studies examining retinal function ( via mea­surement of the ERG) have confined measurements to the early hours after ischemia ( 36- 39). Therefore, studies of the effects of various interventions on the outcome after a period of ischemia may yield deceptive informa­tion. In addition, the examination of either histologic or J Neuro- Ophthalmol, Vol 21, No. 1, 2001 66 D. M. ROSENBAUM ET AL. ILM GCL ILM i GCL IPLI INL ONL OLM D ^ r# W « E * t* lft*- F ^ K'T'^ FIG. 2. Representative photomicrographs showing the histo­logic appearance of the nonischemic ( control) and ischemic reti­nas 1 day and 7 days after 60 minutes ischemia. The overall retinal thickness is decreased with marked thinning of the inner retina in the ischemic retinas. Toluidine blue; original magnifica­tion, x 80. A: control; B: 1 day, HIOP; C: 7 days, HIOP; D: control; E: 1 day, SL; F: 7 days, SL. functional changes alone after ischemia is not an optimal approach. Functional outcome is the most relevant mea­surement to use in designing future clinical translation of research findings ( 40), whereas conventional histologic measurement is useful for localizing the site( s) of injury in studies of the basic mechanisms of retinal ischemia. TABLE 1. Electroretinography b- wave (% baseline) Time ( reperfusion) 30 minutes of ischemia 2 hrs 1 day 3 days 7 days 60 minutes of ischemia 2 hrs 1 day 3 days 7 days HIOP ( n = 10) 59.6 ± 17.7 64.5 ± 23.7 66.8 ± 23.7 74.9 ± 21.9 16.1 ± 2.6 13.6 ± 3.9 11.8 ± 3.9 8.5 ± 2.1 SL ( n = 11) 54.1 ± 3.7 57.9 ± 4.1 50.9 ± 4.5 73.9 ± 7.1 16.7 ± 2.6 9.6 ± 2.1 4.7 ± 1.7 11.7 ± 1.8 HIOP, high intraocular pressure; SL, suture ligation. The present study provides new data showing simulta­neously the functional and histologic progression of changes after retinal ischemia in the rat. The duration of ischemia has a significant impact on postischemic recovery, and the results were similar using either model. After 30 minutes of ischemia with SL or HIOP, there was progressive b- wave recovery to nearly 75% of preischemic baseline 7 days later. However, 30 minutes of ischemia with HIOP or SL produced no sig­nificant retinal structural changes. These results could indicate that conventional light microscopy is not sensi­tive enough to detect changes in retinal structure. The findings are similar to those following 45 minutes of bilateral carotid artery occlusion in the rat ( 41) and are consistent with the lack of significant change in the num­ber of retinal ganglion cells after 30 minutes of ischemia by increased IOP in the rat retina ( 14). Thus, 30 minutes of ischemia using either HIOP or SL may be a useful duration to examine the mechanisms responsible for re­covery of retinal function after ischemia. Both retinal structure and function were significantly altered after 60 minutes of ischemia. These functional alterations were evident immediately after ischemia and became increasingly evident over time. That functional alterations show a greater magnitude in comparison to conventional histologic changes (> 90% decrement in function compared to a 45% decrease in inner retinal thickness) suggests that ERG may be a more sensitive means of assessing the extent of retinal damage in these models. Earlier reports using the HIOP method showed a greater susceptibility of the inner retina to damage com­pared to the outer retina ( 9,15,39,42). Our present study demonstrates that whereas the HIOP and SL methods of retinal ischemia result in inner retinal damage with rela­tive sparing of the outer retina, the outer retina is more adversely affected in the SL model. The apparent selec­tive vulnerability of the inner retina to ischemia as com­pared to the outer retina cannot be attributed to a pressure phenomenon, because even in the SL method, the inner retina showed greater histopathologic alteration than did the outer retina. Nor can this selective inner retinal vul­nerability be attributed to alteration in retinal blood flow alone because diminution of the retinal and choroidal blood flows have been demonstrated to occur to a similar degree by HIOP in the cat and by SL in the rat ( 43). Other mechanisms, such as variation in glycogen stores or greater sensitivity to excitotoxic and apoptotic damage ( 10,35), may be involved. The results of this study show remarkable similarity between the two commonly used experimental tech­niques, histopathologically as well as electrophysiologi-cally. An earlier report comparing similar models in the cat demonstrated that HIOP led to greater injury than SL ( 44). There were, however, several important differences between that study and ours, including the use of cats in place of rats, a greater increase in IOP ( to 160 mmHg), short- term follow- up of the ERG ( only until 390 minutes postischemia), a greater degree of surgical manipu- / Neuro- Ophthalmol, Vol. 21, No. 1, 2001 RETINAL ISCHEMIA MODELS IN RATS 67 lation to produce ischemia with SL, and no histologic measurements. In conclusion, both SL and HIOP are suitable experi­mental models for the study of retinal ischemia in the rat, because ischemia is easily produced, reversible, and quantifiable by histologic or functional criteria. The combination of functional ( ERG) and histologic quanti­fication of ischemic damage is important in the evalua­tion of ischemic injury, because functional measure­ments may reveal injury at a time when retinal morphol­ogy appears relatively normal. Conventional histologic measurements performed too early after ischemia may underestimate the degree of injury. Thus, the timing of either measurement relative to the ischemic stimulus is critical in designing studies of retinal ischemia. Acknowledgments: This work was supported by National Institutes of Health Grants EY11253 ( DMR), EY10343 ( SR), and Research to Prevent Blindness, New York, NY ( PSR). We thank Ms. Antoinette Barnecott for her expert assistance in the preparation of this manuscript. REFERENCES 1. Ffythce TJ. A rationalization of treatment of central retinal artery occlusion. Trans Ophthal Soc UK. 1974; 94: 468- 79. 2. Zeevalk GD, Nicklas WJ. Chemically induced hypoglycemia and anoxia: relationship to glutamate receptor- mediated toxicity in retina. J Pharmacol Exp Ther 1990; 253: 1285- 92. 3. Zeevalk GD, Nicklas WJ. Nitric oxide in retina: relation to excit­atory amino acids and excitotoxicity. Exp Eye Res 1994; 58: 343- 50. 4. Goldstein IM, Ostwald P, Roth S. Nitric oxide: a review of its role in retinal function and disease. Vision Res 1996; 36: 2979- 94. 5. Ostwald P, Park SS, Toledano AY, et al. Adenosine receptor blockade and nitric oxide synthase inhibition: impact upon blood flow and the electroretinogram following retinal ischemia. Vision Res 1997; 37: 3453- 61. 6. Hayreh SS, Weingeist TA. Experimental occlusion of the central artery of the retina. IV: retinal tolerance time to acute ischaemia. Br J Ophthalmol 1980; 64: 818- 25. 7. Roth S, Pietrzyk Z. Blood flow after retinal ischemia in cats. Invest Ophthalmol Vis Sci 1994; 35: 3209- 17. 8. Yoon YH, Marmor MF. Dextromethorphan protects retina against ischemic injury in vivo. Arch Ophthalmol 1989; 107: 409- 11. 9. Hughes WF. Quantitation of ischemic damage in the rat retina. Exp Eye Res 1991; 53: 573- 82. 10. Rosenbaum DM, Rosenbaum PS, Gupta A, et al. Retinal ischemia leads to apoptosis which is ameliorated by aurintricarboxylic acid. Vision Res 1997; 37: 3443- 51. 11. Roth S, Li B, Rosenbaum PS, et al. Preconditioning provides com­plete protection against retinal ischemic injury in rats. Invest Oph­thalmol Vis Sci 1998; 39: 777- 85. 12. Hangai M, Yoshimura N, Hiroi K, et al. Inducible nitric oxide synthase in retinal ischemia- reperfusion injury. Exp Eye Res 1996; 63: 501- 9. 13. Hayashi A, Koroma BM, Imai K, et al. Increase of protein tyrosine phosphorylation in rat retina after ischemia- reperfusion injury. In­vest Ophthalmol Vis Sci 1996; 37: 2146- 56. 14. Selles- Navarro I, Villegras- Perez MP, Salvador- Silva M, et al. Retinal ganglion cell death after different transient periods of pres­sure- induced ischemia and survival intervals. Invest Ophthalmol Vis Sci 1996; 37: 2002- 14. 15. Buchi ER, Suivaizidis I, Fu J. Pressure- induced retinal ischemia in rats: an experimental model for quantitative study. Ophthalmo-logica 1991; 203: 138- 47. 16. Aim A, Bill A. Ocular circulation. In: Moses A, Hart C, eds. Adler's Physiology of the Eye. St Louis: C. V. Mosby, 1987: 183- 203. 17. Gehlbach P, Purple RL. Enhancement of retinal recovery by con­jugated deferoxamine after ischemia- reperfusion. Invest Ophthal­mol Vis Sci 1994; 35: 669- 76. 18. Larsen AK, Osborne NN. Involvement of adenosine in retinal isch­emia: studies in the rat. Invest Ophthalmol Vis Sci 1996; 37: 2603- 11. 19. Stefansson E, Wilson CA, Schoen T, et al. Experimental ischemia induces cell mitosis in the adult rat retina. Invest Ophthalmol Vis Sci 1998; 29: 1050- 5. 20. Mosinger JL, Olney JW. Photothrombosis- induced ischemic neu­ronal degeneration in the rat retina. Exp Neurol 1989; 105: 110- 3. 21. Block F, Schwarz M, Sontag K- H. Retinal ischemia induced by occlusion of both common carotid arteries in rats as demonstrated by electroretinography. Neurosci Lett 1992; 144: 124- 6. 22. Mosinger JL, Price MT, Bai HY, et al. Blockade of both NMDA and non- NMDA receptors is required for optimal protection against ischemic neuronal degeneration in the in vivo adult mam­malian retina. Exp Neurol 1991; 113: 10- 7. 23. Moroni F, Lombardi G, Pellegrini- Faussone S, et al. Photochemi-cally induced lesion of the rat retina: a quantitative model for the evaluation of ischemia- induced retinal damage. Vision Res 1993; 33: 1887- 91. 24. Schmidt- Kastner R, Eysel UF. Ischemic damage visualized in flat mounts of rat retina after photochemically induced thrombosis. Brain Res Bull 1994; 34: 487- 91. 25. Roth S, Osinski J, Park S, et al. Measurement of purine nucleoside concentration in the intact rat retina. J Neurosci Methods 1996; 68: 87- 90. 26. Roth S, Park SS, Sikorski CW, et al. Concentrations of adenosine and its metabolites in the rat retina/ choroid during reperfusion after ischemia. Curr Eye Res 1997; 16: 875- 85. 27. Roth S, Rosenbaum PS, Park SS, et al. Ischemia induces signifi­cant changes in purine nucleoside concentration in the retina-choroid in rats. Exp Eye Res 1997; 65: 771- 9. 28. Choi DW. Calcium: still center- stage in hypoxic- ischemic neuro­nal death. Trends Neurosci 1995; 18: 58- 60. 29. Block F, Schwarz M. Effects of antioxidants on ischemic retinal dysfunction. Exp Eye Res 1997; 64: 559- 64. 30. Muller A, Pietri S, Villain M, et al. Free radicals in rabbit retina under ocular hyperpressure and functional consequences. Exp Eye Res 1997; 64: 637- 43. 31. Szabo ME, Droy- Lefaix MT, Doly M. Direct measurement of free radicals in ischemic/ reperfused diabetic rat retina ( review). Clin Neurosci 1997; 4: 240- 5. 32. Zhang H, Agardh CD, Agardh E. Increased catalase levels and hypoxanthine- enhanced nitro- blue tetrazolium staining in rat retina after ischemia followed by recirculation. Curr Eye Res 1995; 14: 47- 54. 33. Nickells RW. Apoptosis of retinal ganglion cells in glaucoma: an update of molecular pathways involved in cell death. Surv Oph­thalmol 1999; 43( suppl 1): S151- 61. 34. Levin LA, Louhab A. Apoptosis of retinal ganglion cells in ante­rior ischemic optic neuropathy. Arch Ophthalmol 1996; 114: 488- 9. 35. Rosenbaum DM, Rosenbaum PS, Gupta H, et al. The role of the p53 protein in the selective vulnerability of the inner retina to transient ischemia. Invest Ophthalmol Vis Sci 1998; 39: 2132- 9. 36. Foulds WS, Johnson NF. Rabbit electroretinogram during recovery from induced ischemia. Trans Ophthal Soc 1974; 94: 383- 93. 37. Brunette JR, Olivier P, Galeano C, et al. Hyper- response and delay in the electroretinogram in acute ischemia. Can J Ophthalmol 1983; 18: 188- 93. 38. Brunette JR, Olivier P, Blondeau P, et al. Electroretinographic recovery from controlled total ischemia. Can J Ophthalmol 1986; 21: 13- 8. 39. Buchi ER. Cell death in the rat retina after pressure- induced isch-aemia- reperfusion insult: electron microscopic study. II. Outer nuclear layer. Jpn J Ophthalmol 1992; 36: 62- 8. J Neuro- Ophthalmol, Vol. 21, No. 1, 2001 68 D. M. ROSENBAUM ET AL. 40. Hunter AJ, Green AR, Cross AJ. Animal models of acute isch-aemic stroke: can they predict clinically successful neuroprotective drugs? Trends Pharm Sci 1995; 16: 123- 8. 41. Barnett NL, Osborne NN. Prolonged bilateral carotid artery occlu­sion induces electrophysiological and immunohistochemical changes to the rat retina without causing histological damage. Exp Eye Res 1995; 61: 83- 90. 42. Buchi ER. Cell death in the rat retina after a pressure- induced ischaemia- reperfusion insult: an electron microscopic study. I. Ganglion cell layer and inner nuclear layer. Exp Eye Res 1992; 55: 605- 13. 43. Lin J, Roth S. Ischemic preconditioning attenuates hypoperfusion after retinal ischemia in rats. Invest Ophthalmol Vis Sci 1999; 40: 2925- 31. 44. Gehlbach PL, Purple RL. A paired comparison of two models of experimental retinal ischemia. Curr Eye Res 1994; 13: 597- 602. / Neuro- Ophthalmol, Vol. 21, No. 1, 2001