Intraretinal Oxygen Distribution in the Monkey Retina and the Response to Systemic Hyperoxia

Abstract

purpose. To measure the intraretinal oxygen distribution and consumption in the fovea, the parafovea, and the inferior retina in the monkey eye and determine the influence of graded systemic hyperoxia.

methods. Oxygen sensitive microelectrodes were used to measure the oxygen tension as a function of depth through the retina in anesthetized monkeys (n = 8) under normoxic and hyperoxic conditions. Oxygen consumption rates in the avascular regions of retina were determined by fitting the oxygen profiles to established oxygen consumption models.

results. Under normoxic conditions, in the foveal area, the intraretinal oxygen distribution reflected the absence of retinal capillaries and the predominantly choroidal origin of retinal oxygenation. A similar shape of oxygen distribution was seen in the parafoveal retina with the addition of local perturbations in the inner retina attributed to the presence of retinal capillaries. In the inferior retina the same general shape was found. Oxygen consumption in the outer retina was higher in the parafoveal region, and the minimum oxygen tension was lower. During hyperoxia, choroidal oxygen levels in all areas increased dramatically, but the increase in oxygen tension in the inner retina was much less. The avascular nature of the foveal area allowed oxygen consumption analysis of both the inner and outer retina and showed that inner retinal oxygen consumption increased significantly during hyperoxic ventilation to a level equivalent to that of the outer retina.

conclusions. In the outer retina of the monkey the Po ₂ minimum is lower, and the oxygen consumption rate is higher in the parafoveal region. During systemic hyperoxia, outer retinal oxygen consumption is unaffected, but in the foveal area, total oxygen consumption increases. This regulation of oxygen consumption in the monkey retina is comparable to that reported in lower mammals and may represent an important mechanism in retinal homeostasis.

Measurements of intraretinal oxygen distribution can provide detailed information about oxygen delivery and consumption within the living retina under normal or pathologic conditions. Unfortunately, such techniques cannot be applied to the human eye, so we are confined to the use of animal models. To date, the animal model closest to humans in which the intraretinal oxygen distribution has been measured is the macaque monkey. ¹ That study provided information about the oxygen distribution in the parafoveal area in two animals and included measurements in the foveal area in one of the animals. The present study attempts to expand this work in monkeys, using a considerably larger sample size and also incorporating a study of the effects of induced systemic hyperoxia. Intraretinal oxygen distributions under normal and hyperoxic conditions were also compared in three areas of retina: the fovea, the parafovea, and the inferior retina. 

In other species, such as the rat, ^{2 3} pig, ⁴ and cat, ⁵ there is now ample evidence that oxygen regulatory mechanisms exist in the retina which act to reduce the influence of systemic hyperoxia on the inner retinal layers. This study seeks to determine whether such effects are also present in the monkey retina. If present, then it is reasonable to expect that similar mechanisms could be present in the human retina, the purpose of which remains to be elucidated. 

Methods

General

The experimental techniques were similar to those reported in our earlier publications in a range of other species. ^{6 7 8 9} Eight macaque monkeys were used in this study (body weight, 3–8 kg). Sedation was induced by an intramuscular injection of ketamine (25 mg/kg body weight, intramuscularly; Ketalar; Pfizer, New York, NY), followed by an intravenous infusion of alphaxalone-alphadone (6–9 mg/kg body weight; Saffan; Schering-Plough Animal Health Ltd., Baulkham Hills, NSW, Australia) at a rate sufficient to maintain anesthesia throughout the experiment. The femoral artery was cannulated for systemic blood pressure monitoring and intermittent sampling of blood for blood gas analysis (Rapidlab 248; Bayer Corp., Medfield, MA). The animals were initially ventilated with 20% oxygen at 35 breaths per minute with a tidal volume sufficient to produce blood gas levels within the normal range. Stepwise hyperoxia was induced by increasing the oxygen percentage in the ventilation gas with a commercially available oxygen mixer (OxyCycler, model SH10; BioSpherix, Redfield, NY). Arterial blood pressure was monitored continuously throughout the experiment and recorded on a multichannel chart recorder (model LR8100; Yogogawa, Tokyo, Japan). Experiments usually lasted 6 hours, after which the monkey was killed with an anesthetic overdose before neural tissue was harvested for unrelated studies. All procedures conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Intraretinal Oxygen Profiles

In each experiment the head of the animals was immobilized in a stereotaxic frame, and the eye was stabilized by suturing to a fixed eye ring at the limbus. Oxygen-sensing microelectrodes were manufactured in our laboratory with techniques developed by Whalen et al. ¹⁰ The electrodes were calibrated before and after the experiment. ⁷ The oxygen-sensitive microelectrode (1-μm tip size) entered the eye through a small hole at the pars plana and was visualized inside the eye by an operating microscope via a plano-concave contact lens. The electrode was oriented through a custom built stereotaxic apparatus. ¹¹ The small size of the electrode tip coupled with high acceleration piezoelectric translation of the electrode through the retina produces highly reproducible measurements of intraretinal oxygen distribution. Under microscope observation, the electrode tip was placed just anterior to the chosen area of retina in a region free of major retinal vessels in either the inferior retina (1.5–2 disc diameters inferior to the disc), the parafovea, or the avascular area of the fovea. The electrode was then stepped through the retina (10 μm steps), under computer control, until a peak oxygen level within the choroid was reached. The measurement was then repeated during stepwise withdrawal of the electrode. The withdrawal profiles were used for data analysis as they tended to be less influenced by artifacts associated with mechanical stress on the electrode tip during penetration. ⁷ Typically, three profiles from each area were used for data analysis for air-breathing conditions, and a single set of profiles was used from each location for each animal under hyperoxic conditions. The order in which each location was measured was randomized. All measurements were performed under photopic conditions. 

Measurements were repeated after stepwise increments (20%) in oxygen percentage in the ventilation gas, thus generating data for 20%, 40%, 60%, 80%, and 100% oxygen ventilation. At least 10 minutes was allowed for intraretinal oxygen levels to stabilize after each increase in systemic oxygenation, and each profile measurement took approximately three minutes. The nonperpendicular nature of the penetration means that distances are expressed as track length through the retina, rather than as absolute retinal depth. 

Oxygen Consumption Analysis

The portion of the intraretinal oxygen profiles deeper than Bruch’s membrane was discarded and an established three layer model of retinal oxygen consumption ¹² was fitted to the oxygen profiles in the outermost 150 μm of each profile. This analysis extracts a value for outer retinal oxygen consumption and generates a best-fit curve for each profile. For the oxygen profiles in the foveal area a more extensive analysis was used. This method is appropriate because the absence of retinal capillaries in this area means that the full thickness of the retina can be included in the consumption analysis. ^{9 13 14} To allow a comparison between oxygen consumption in the inner and outer regions of the foveal area of retina, the oxygen consumption model was expanded to include five layers as previously used in the avascular region of the rabbit retina. ^{9 13 14} Three layers were confined to the outer 150-μm track length and the additional layers were confined to the remaining retina. Once again, the analysis extracted the best-fit curve for each profile, along with levels of inner and outer retinal oxygen consumption. 

Statistics

All statistical testing was performed on computer (SigmaStat; SSPS Scientific, Chicago, IL). Two-way ANOVA with an acceptance level of P < 0.05 was used to determine any significant differences in oxygen tension before and after alterations of inspired oxygen percentage. The significance of any change in oxygen consumption with inspired oxygen percentage was tested using one-way ANOVA, with a significance acceptance level of P < 0.05 for the F value. When appropriate, Student’s t-test was used. All data are expressed as the mean ± SE, and all error bars on graphs are also SE. 

Results

The average data for oxygen tension as a function of penetration depth through the foveal region of the monkey retina is shown in Figure 1 . Innermost retinal Po ₂ was 5.8 ± 0.9 mm Hg, which decreased on penetration into the retina, to reach a minimum of 2.3 ± 0.5 mm Hg in the outer retina, before rising steeply to a peak of 65.2 ± 2.8 mm Hg in the choroid. The small oxygen gradient in the inner retina indicates that there is some oxygen contribution to the outer retina from the vitreous side, but the large gradient in the outermost retina indicates that the dominant source of oxygen supply to this region of the retina is from the choroid. The lack of any local variation in oxygen gradient in the inner retina reflects the lack of retinal capillaries in this region. 

The average data for oxygen tension as a function of penetration depth through the parafoveal region of retina is shown in Figure 2 . In the parafoveal area, the oxygen contribution from the retinal capillary beds was evident by the present of significant bumps in the oxygen profile in the inner retinal layers of individual profiles. These bumps tended to be smoothed out by the averaging process across animals, but they are still evident in the averaged data. Average inner retinal surface Po ₂ was 5.6 ± 0.8 mm Hg, which was not significantly different from equivalent locations in the foveal area (P = 0.547), whereas the minimum Po ₂ in the outer retina was 1.3 ± 0.2 mm Hg, which was significantly different from equivalent locations in the foveal area (P < 0.05). The Po ₂ peak in the choroid (64.0 ± 2.1 mm Hg) was not significantly different from that in the fovea (P = 0.993). 

The average data for oxygen tension as a function of penetration depth through the inferior retina is shown in Figure 3 . In the inferior retina, the oxygen contribution from the retinal capillary beds was also evident. The average inner retinal surface Po ₂ of 13.7 ± 1.4 mm Hg and the minimum Po ₂ of 5.7 ± 0.9 mm Hg in the outer retina were both significantly higher than the corresponding locations in either the foveal (both P < 0.001) or parafoveal regions (both P < 0.001). The peak choroidal Po ₂ of 60.4 ± 1.4 mm Hg in the inferior retina was not significantly different from that in the foveal and parafoveal areas (P = 0.391 and P = 0.434, respectively). 

The average of all the best-fit curves for the outermost 150-μm track length in each of the three regions is shown in Figure 4 . The average Po ₂ values at Bruch’s membrane were not significantly different (P = 0.266), but in the region more than 60-μm track length from Bruch’s membrane the oxygen level in the inferior retina was higher than in the foveal or parafoveal areas. 

The average oxygen consumption rates for the outer retina in each location are shown in Figure 5 . No correction for angle of penetration was attempted, as the exact angle is unknown, but it is likely to be similar in each location because the angular adjustment necessary to move between the three locations is relatively small. The outer retinal oxygen consumption was higher in the parafoveal area than in the fovea (P = 0.001) or inferior retina (P = 0.002). There was no significant difference between outer retinal oxygen consumption in the fovea and inferior retina (P = 0.82). 

The average systemic arterial blood gas levels at different levels of ventilatory oxygen percentage are shown in Figure 6 . The P_a o ₂ of 79.0 ± 4.0 mm Hg, P_a co ₂ of 36.6 ± 1.9 mm Hg, and pH 7.4 ± 0.01 (n = 8) for 20% oxygen ventilation are in the normal range of values expected. The stepwise increases in inspired oxygen percentage produced stepwise increases in P_a o ₂, as expected, whereas pH and P_a co ₂ were not significantly affected (P = 0.965 and P = 0.884). 

The average response of intraretinal oxygen levels in the foveal area after stepwise systemic hyperoxia is shown in Figure 7 . Each increase in the percentage of oxygen in the ventilation mixture created a significant rise in choroidal oxygen tension, but a much smaller increase in inner retinal oxygen levels. Choroidal Po ₂ increased from 61.3 ± 9.9 mm Hg (n = 8) at 20% oxygen ventilation to reach 357.6 ± 23.2 mm Hg at 100% oxygen ventilation. The corresponding values at the surface of the inner retina were 6.0 ± 1.5 and 40.7 ± 14.0 mm Hg for 20% and 100% oxygen ventilation, respectively. At higher levels of systemic hyperoxia the shape of the intraretinal oxygen distribution indicates that there is a net gradient of oxygen from the retina to the vitreous. This reversal of oxygen gradient at the retinal surface means that this region becomes a net source of vitreal oxygen. 

The average response of intraretinal oxygen levels in the parafoveal area and inferior retina after stepwise systemic hyperoxia are shown in Figures 8 and 9 , respectively. Each increase in the percentage of oxygen in the ventilation mixture induced a significant increase in choroidal oxygen tension reaching 378.1 ± 56.5 mm Hg in the parafovea and 411.8 ± 44.3 in the inferior retina at 100% oxygen ventilation. These choroidal values were not significantly different than those in the foveal region (P = 0.718 and P = 0.880, respectively). The oxygen levels in the innermost retina in the parafovea during 100% oxygen ventilation (32.0 ± 9.2 mmHg) were not significantly different from that in the foveal area (P = 0.152). The oxygen level of the innermost retina in the inferior retina during 100% oxygen ventilation (20.3 ± 6.8 mm Hg) was lower than in the foveal or parafoveal areas (both P < 0.05). The oxygen gradients at the retinal surface are very flat during systemic hyperoxia in the parafoveal and inferior retina, indicating very little oxygen flux to the vitreous in these regions under the hyperoxic condition. 

In the subset of oxygen profiles measured under different levels of inspired oxygen, the three-layer oxygen consumption analysis revealed that outer retinal oxygen consumption did not change significantly with inspired oxygen level in any of the three retinal locations tested. In the special case of the completely avascular foveal area, the five-layer oxygen consumption model for the full thickness of the foveal profiles was applied and the average fitted curves for each oxygen ventilation condition are shown in Figure 10 . The vertical dotted line indicates the arbitrary boundary (150 μm) between the outermost retina and the remaining innermost retina. The oxygen consumption rates of the inner and outer regions of the fovea as a function of inspired oxygen percentage are shown in Figure 11 . Under normoxic conditions, the oxygen consumption of the inner region of the foveal area is only a small proportion of the total retinal oxygen consumption. However, the oxygen consumption of the innermost retina increased as a function of inspired oxygen percentage, becoming significantly greater at 60% oxygen and above (P < 0.001). With 80% and 100% oxygen ventilation, inner retinal oxygen consumption was not significantly different from outer retinal oxygen consumption (P = 0.371 and P = 0.473, respectively). The total oxygen consumption of the foveal area approximately doubles during 100% oxygen ventilation. 

Discussion

The presence of a humanlike macula and foveal specialization in the monkey retina makes this species particularly attractive for the study of the metabolic requirements of these specialized areas of retina. The only published report of the intraretinal oxygen distribution in the fovea of a monkey pointed to unusual oxygen requirements of the retina in this region. ¹ Retinal oxygen consumption was essentially confined to the outer retina, with the inner retinal oxygen consumption being described as very low. ¹ This remarkable suggestion that the inner retina in a region of high visual acuity could have a low oxygen consumption was recently supported by equivalent findings in the high acuity visual streak of the rabbit retina. ¹³ In the rabbit, this was explained by the possibility of high levels of anaerobic metabolism in the inner retina, consistent with earlier findings of high levels of glycogen in the inner retina. ¹⁵  

In the present study, the large number of measurements made in the avascular region of the fovea probably contains some close encounters with the foveal pit. The exact location of the electrode track with respect to the foveal pit cannot be assessed funduscopically, so we lumped all foveal zone measurements together to produce an average measurement. We are confident that we were able to position the electrode in the avascular foveal zone, even though the surrounding capillary net was not visible funduscopically. The avascular zone in monkeys has been described as almost circular, with a diameter of 666 μm. ¹⁶ The approximate center of the avascular zone is not difficult to estimate, given the pattern of surrounding feeder vessels that are readily visible. We estimate that the angle at which the electrode penetrates the retina is ∼40° from the perpendicular, so in traversing a track length of ∼350 μm, the sideways displacement would be approximately 225 μm, which is still within the foveal area when the electrode reaches the outermost retina. 

The oxygen consumption of the outer retina was shown to be significantly greater in the parafoveal area of retina than in the fovea or inferior retina. This is consistent with earlier work in the monkey fovea and parafovea in which statistical analysis was not performed, but it was suggested that light-adapted parafoveal oxygen consumption could be higher than in the fovea. ¹ Our finding that oxygen consumption of the inner retinal tissue in the foveal area is small under air-breathing conditions (Fig. 11)is also consistent with earlier work, in which investigators that did not quantify the oxygen consumption of this region but described it as very low. ¹  

A major finding of the present study is that oxygen levels within the inner retina of the monkey show a relative immunity to the effects of systemic hyperoxia. The muted oxygen response in the innermost retina was not due to inadequate time for the oxygen response to develop. We routinely monitored the preretinal oxygen response after each change in the percentage of oxygen in the ventilation gas and waited for the new oxygen equilibrium to stabilize. This was usually no more than 10 minutes, which is consistent with recent studies of vasoactive response times to systemic hyperoxia in humans. ¹⁷ Muted intraretinal oxygen changes in the face of extreme systemic hyperoxia has been reported in several other mammals, such as the pig, ⁴ cat, ¹⁸ rat, ¹⁹ and guinea pig. ²⁰ However, this phenomenon has not been measured in the fovea or posterior pole of the primate eye. Landers et al. ²¹ measured preretinal oxygen changes in monkey eyes 6 to 9 months after photocoagulation therapy in treated and untreated areas of nasal retina. They reported a preretinal oxygen tension of 8.6 ± 4.5 mm Hg over untreated retina during air breathing, which increased to 86.2 ± 41 mm Hg during oxygen ventilation. However, the scatter in their measurements was large, and the example data presented shows a preretinal level of ∼7 mm Hg during air breathing and between 20 and 40 mm Hg during 100% oxygen ventilation. These values are very similar to those that we report in the foveal, parafoveal, and inferior areas of retina in our monkeys. An earlier study in monkey eyes ²² in which only relative measurements of preretinal Po ₂ were reported also suggests that 100% oxygen ventilation leads to less than a doubling of the air breathing preretinal Po ₂. 

In the present study the completely avascular nature of the primate fovea allows the full retinal thickness to be analyzed using a multilayer model of inner and outer retinal oxygen consumption. ^{9 13 14} This analysis has shown that increased oxygen consumption in the inner portion of the foveal retina during systemic hyperoxia is responsible for the muted oxygen response, with the oxygen consumption of the outer retina remaining unchanged. This is consistent with earlier findings in the vascularized area of retina in the rat during systemic hyperoxia, for which more complex models were needed to accommodate the oxygen input from the retinal circulation. ² Whatever mechanisms are involved in this regulation of inner retinal oxygen exposure may be important to retinal function. Although systemic hyperoxia is not an environmental concern in the natural world, the retina may well be equipped to deal with the local variation of oxygen level due to changes in supply or demand. For example, in the developing retina the inner retina may well be subject to hyperoxic oxygen levels before the development of the photoreceptors. Both too little, and too much oxygen can be damaging to the retina. Retinal hypoxia is thought to play a major role in ischemic retinal diseases. Retinal hyperoxia is damaging to the developing human retina, as evidenced by the sight-threatening consequences of oxygen supplementation in premature infants. The ability of the retina to cope with variations in oxygen environment may be an important component of retinal homeostasis. The primate retina must also cope with a marked degree of spatial variation in photoreceptor density and type, ²³ which very likely requires individual photoreceptors to have different oxygen demands based on the local oxygen availability and the needs of its neighbors. 

Detailed studies in the rat have suggested that a combination of autoregulation of the retinal oxygen input and an increase in oxygen consumption of the inner and outer plexiform layers are responsible for the muted rise in inner retinal oxygen tension in hyperoxia. ^{2 6} These studies either exploited the highly layered distribution of retinal capillaries in the rat, ² or used occlusion of the retinal circulation to allow extraction of oxygen consumption information from the inner retina (Yu D-Y et al. IOVS 2000;41:ARVO Abstract 97). ⁶ In the guinea pig, a species with a naturally avascular retina, retinal oxygen levels are maintained during systemic hyperoxia by a powerful mechanism regulation choroidal oxygen levels in the face of confirmed systemic hyperoxia. ^{20 24} The absence of such regulatory mechanisms in the avascular region of the rabbit retina, ⁹ coupled with the reported retinotoxic effects of systemic hyperoxia in the rabbit, ²⁵ supports the suggestion that such regulatory mechanisms may be important for healthy retinal function. Given the present findings from the monkey retina, it is reasonable to assume that similar properties may be found in the human retina. The importance of such regulatory mechanisms controlling oxygen levels in the inner retina is difficult to determine. However, such mechanisms would certainly have clinical implications in those diseases in which modulation of the intraretinal oxygen environment is proposed. These include novel ideas for increasing oxygen supply to degenerating photoreceptors or detached retinas, ²⁶ or the more widespread use of laser photocoagulation to reduce photoreceptor oxygen uptake in ischemic retinal diseases such as diabetic retinopathy and occlusive diseases of the retinal circulation. 

 Supported by the National Health and Medical Research Council of Australia and the Australian Research Council Centre of Excellence in Vision Science.

 Submitted for publication June 3, 2005; revised August 18, 2005; accepted October 13, 2005.

 Disclosure: D . -Y. Yu, None; S. J. Cringle, None; E.-N. Su, 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.

 

Figure 1.

 

Average intraretinal oxygen distribution as a function of track distance through the retina in the foveal area of the monkey retina on 20% oxygen ventilation conditions. Error bars, ±SE.

Figure 2.

 

Average intraretinal oxygen distribution as a function of track distance through the retina in the parafoveal area of the monkey retina in 20% oxygen ventilation conditions. Error bars, ±SE.

Figure 3.

 

Intraretinal oxygen distribution as a function of track distance through the retina in the inferior monkey retina in 20% oxygen ventilation conditions. Error bars, ±SE.

Figure 4.

 

Average of the result of fitting the outer 150 μm of the retinal track distance to an established three-layer model of retinal oxygen consumption. There was no significant difference in oxygen tension at Bruch’s membrane in the three regions. In the inferior retina, more than 60 μm track length from Bruch’s membrane, the oxygen level was greater than equivalent locations in the foveal or parafoveal areas.

Figure 5.

 

Average outer retinal oxygen consumption in each of the three regions during 20% oxygen ventilation. The oxygen consumption in the parafoveal region was significantly greater than in the foveal area or in the inferior retina.

Figure 6.

 

Average blood gas data for all monkeys (n = 8), after stepwise increases in inspired oxygen level. Systemic arterial Po ₂ increased linearly with inspired oxygen percentage, whereas system arterial Pco ₂ and pH were not significantly affected.

Figure 7.

 

The effect of stepwise increases in oxygen ventilation on the intraretinal oxygen distribution through the foveal area of the monkey retina.

Figure 8.

 

The effect of stepwise increases in oxygen ventilation on the intraretinal oxygen distribution through the parafoveal area of the monkey retina.

Figure 9.

 

The effects of stepwise increases in oxygen ventilation on the intraretinal oxygen distribution through the inferior area of the monkey retina.

Figure 10.

 

Average of the best-fit curves for oxygen profiles across the full thickness of the foveal area under each ventilation condition.

Figure 11.

 

Average oxygen consumption rates of the inner and outer regions of the foveal area at each oxygen ventilation level. At 60% oxygen and more, the oxygen consumption of the inner region of the fovea was significantly increased. Outer retinal oxygen consumption showed no significant changes with increased oxygen ventilation. At 80% and 100% oxygen ventilation, there was no significant difference between oxygen consumption in the inner and outer regions of the foveal area.

The authors thank Dean Darcey for expert technical assistance, and acknowledge the cooperation of Xinghuai Sun, Wenyi Guo, and Xiaobo Yu and support staff from Fudan University, Shanghai.