Hyperoxia Improves Oxygen Consumption in the Detached Feline Retina

Abstract

purpose. To investigate the effects of hyperoxia on retinal oxygenation and oxygen consumption in the detached feline retina.

methods. Retinal detachment was created in nine intact anesthetized cats by injecting 0.25% sodium hyaluronate in balanced salt solution into the subretinal space. Oxygen microelectrodes were used to collect spatial profiles of retinal Po ₂ in both the attached and detached retina. A diffusion model was fitted to quantify photoreceptor oxygen consumption (Q _av).

results. In the detached retina, the Po ₂ at the border between the retina and the fluid layer under the retina decreased; hyperoxia increased it to a level that was not significantly different from the control (attached retina, air breathing). Detachment did not change the Po ₂ at the border between the avascular and vascularized retina; hyperoxia significantly increased the level. Oxygen consumption decreased to 47% ± 18% of the control value in the detached retina during normoxia; hyperoxia increased Q_av to 68% ± 17% of control. Hyperoxia increased the average inner retinal Po ₂ (P _IR) in the detached retina to a level higher than that during normoxia. Detachment did not change P _IR during normoxia.

conclusions. Hyperoxia has been shown to improve photoreceptor survival in the detached retina. The present work suggests that hyperoxia is protective because it allowed increased photoreceptor oxygen consumption. Whereas normal Po ₂s were maintained at the inner and outer border of the avascular region during hyperoxia, Q_av was not restored to normal, suggesting that other factors are involved in photoreceptor dysfunction during detachment in addition to insufficient oxygen delivery.

Retinal detachment (RD) can cause significant vision loss. Morphologic changes and cell death occur in the detached retina, ^{1 2 3 4 5 6} but these changes can be alleviated with supplemental O₂ at the onset of RD in detached feline and ground squirrel retinas. ^{7 8 9 10} Mathematical simulations predict a severe O₂ deficit in the photoreceptors in even a small RD because the photoreceptors are farther from the choroid, which supplies most of their nutrients. ¹¹ Simulations also predict that supplemental O₂ inspiration will restore photoreceptor oxygen consumption. ¹¹ Increasing the availability of oxygen may prevent the deconstruction of the cytoskeleton of the photoreceptors and their ultimate degeneration. 

Hyperoxia may have a role in the clinical management of RD. Retinal reattachment surgery is frequently needed once a detachment is detected, but patients often must wait for a period before reattachment surgery. We suspect that there are some morphologic changes during the period when patients are waiting for surgery, due to hypoxia in the photoreceptors. Based on the protective effects of hyperoxia that have already been demonstrated in animals, treating RD with supplemental O₂ may be a safe, relatively easy way to improve visual recovery during and after reattachment surgery. 

Although the beneficial effects of hyperoxia have been demonstrated histologically, and a theoretical explanation has been proposed, there are no oxygen data available to validate the theory or assess the extent of protection that hyperoxia might afford. In addition, there is no information on the function of the retinal circulation during retinal detachment. In order to address these issues, this study investigated retinal oxygenation and oxygen consumption of the photoreceptors in the detached feline retina during inspiration of air and 100% O₂. 

Methods

Animal Preparation

All experimental procedures were in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The cat was initially given 0.4 mg/kg butorphanol (intramuscularly) at the animal facility. Anesthesia was induced with an intravenous injection of 5% pentothal sodium (22 mg/kg) followed by additional pentothal as needed during surgery. Intramuscular ketamine (25 mg/kg) was used when the cat was difficult to handle. Urethane (20%, 200 mg/kg loading dose followed by 20–40 mg · kg⁻¹· h⁻¹) was used to maintain long-term anesthesia. The cat was paralyzed by an intravenous injection of 2 mL 1% pancuronium bromide (Pavulon; Organon International, Roseland, NJ) and was artificially ventilated. Body temperature was monitored by a rectal probe and was maintained at 39°C by a feedback-controlled heating pad. Arterial blood samples were analyzed with a blood gas analyzer (860; Bayer, Leverkeusen, Germany) to monitor Pao ₂, Paco ₂, pH_a, and glucose levels. The cat was ventilated with air most of the time and was switched to 100% O₂ when necessary. At the end of the experiment, pentobarbital sodium solution was injected intravenously to euthanize the cat when the eye was saved for histology; otherwise, saturated KCl was used. 

Retinal Detachment

Retinal detachments were created in the area centralis under dim light by injecting 0.25% sodium hyaluronate in balanced salt solution into the subretinal space when the cat was breathing 100% oxygen. The injection was performed with a subretinal injector (Surmodics, Eden Prairie, MN). The diameter of RD was between 3 and 5 mm, with an injection of no more than 0.1 mL solution. Based on the geometry, we calculated that a detachment of 5 mm has a height of 500 μm in the center. Figure 1shows a typical RD. 

Intraretinal Oxygen Profiles

Most experimental methods were similar to those reported earlier. ^{12 13} The eye was stabilized by attaching the conjunctiva to an eye ring that is part of the microelectrode manipulator. Double-barreled oxygen microelectrodes ¹⁴ were inserted into the eye 6 mm behind the limbus through a guide needle. These electrodes were used to collect both Po ₂ as a function of distance across the retina (Po ₂ profiles) during electrode withdrawal and the electroretinograms (ERGs) during penetration. Oxygen profiles were collected in both the attached and detached parts of the retina during dark adaptation. In the attached retina, the microelectrode penetrated all the way to the choroid, signaled by the transepithelial potential (TEP) when the electrode crossed the retinal pigment epithelium (RPE). In the detached retina, the ERGs were consistently absent. The microelectrode did not always reach the choroid to prevent unnecessary damage to the retina caused by the increased diameter of the microelectrode along the penetration depth. As a result, most Po ₂ profiles collected in the detached region did not start from the choroid. In general, the microelectrode penetrated as far as 600 μm and was withdrawn even if the transepithelial potential had not been observed. 

Data Analysis

Outer Retina.

For the attached retina, a one-dimensional, three-layer diffusion model ^{12 15} was fitted to the part of the oxygen profile in the avascular region to quantify photoreceptor oxygen consumption (Q_av). The model assumes that only the second layer (layer 2), corresponding to the inner segment, consumes oxygen. This layer is the location of all the photoreceptor mitochondria, with the exception of those in the synapses. Diffusion is assumed to be the only mechanism for oxygen delivery to the photoreceptors from the choroid and from the retinal circulation, so the outer retina can be treated as a slab of tissue through which oxygen diffuses. The model equations for Po ₂ as a function of distance (x) from the choroid are the solutions to Fick’s law at steady state. 

Fick’s Law:  

\[\frac{d^{2}P}{dx^{2}}{=}\ \frac{Q}{Dk}\]

Solving Fick’s law for each layer  

\[\mathbf{P}_{1}(\mathbf{x}){=}{\alpha}_{1}x{+}{\beta}_{1}\ 0{\leq}\mathbf{x}{\leq}\mathbf{L}_{1}\]

 

\[\mathbf{P}_{2}(\mathbf{x}){=}\ \frac{\mathbf{Q}_{2}}{2\mathbf{Dk}}\mathbf{x}^{2}{+}{\alpha}_{2}\mathbf{x}{+}{\beta}2\ \mathbf{L}_{1}{\leq}\mathbf{x}{\leq}\mathbf{L}_{2}\]

 

\[\mathbf{P}_{3}(\mathbf{x}){=}{\alpha}_{3}\mathbf{x}{+}{\beta}_{3}\mathbf{L}_{2}{\leq}\mathbf{x}{\leq}\mathbf{L}\]

where P _i (x) is the partial pressure of Po ₂ in layer i; L ₁ is the position of the inner segment (IS)-outer segment (OS) boundary; L ₂ is the position of the IS-outer nuclear layer (ONL) boundary; L is the position of the inner–outer retinal boundary (the inner end of the avascular retina); Q ₂ is the oxygen consumption of the IS layer, between L ₁ and L ₂; D is the diffusion coefficient of oxygen in the retina, k is the solubility of oxygen, and α _i and β _i are constants in layer i (i = 1, 2, or 3) derived from the boundary conditions on P, as well as continuity of flux and Po ₂ at the borders between layers. 

The oxygen consumption in layer 2 (Q ₂), choroidal Po ₂ (Pc), the Po ₂ at the boundary between inner and outer retina (PL), and the positions of the borders between layers 1 and 2 (L ₁) and layers 2 and 3 (L ₂), are obtained by least-squares fitting of the model to Po ₂ data. Although Pc and PL are fitted parameters, they are very tightly constrained by the measured values, and in practice, it is Q ₂, L ₁, and L ₂ that are obtained from fits. The average oxygen consumption of the outer retina (Q _av) was then calculated: Q _av = Q ₂ (L ₂ − L ₁)/L. Q _av is reported rather than Q ₂ because this product has a smaller coefficient of variation and is less sensitive to distortion in the profiles. ¹⁵  

For the detached retina, there was a fluid layer between the choroid and the photoreceptors. Rather than use a different model with that layer incorporated, we excluded the fluid layer from the model fit, so that only the part of the profile within the outer retina itself was fitted. In this way, the model remained equally valid in the detached retina. Fits in the detached retina required a new way to determine the position of the outer surface of the retina itself. Figure 2shows how these boundaries were determined from an intraretinal oxygen profile in the detached retina. The profile through the OS (layer 1) and the ONL (layer 3) is expected to consist of straight lines, because these two layers have no oxygen consumption, ¹⁵ whereas the profile through layer 2 has a quadratic shape because of its oxygen consumption. Consequently, straight lines could be fitted by eye to determine the locations of layers 1 and 3. The outer border of layer 1 (X= 0), which is the outer edge of the retina, was determined by a slope change at this point. This occurred because the diffusion coefficient in the retina was approximately 70% of that in the fluid under the retina. ¹⁶ The other boundary, X= L, was the point at the proximal end of layer 3. A deviation from the line at the proximal end of this layer indicates a region of supply or consumption, which is expected in the inner retina. The Po ₂s at the boundaries of the avascular layer, called Po ₂ (X= 0) and PL, and Q _av were obtained from the model fit for both the attached and detached retina. 

Inner Retina.

The diffusion model cannot be applied to the inner retina, because oxygen gradients in the presence of a capillary bed are three-dimensional. A three-dimensional model, along with three-dimensional oxygen data would be necessary, and it is feasible to collect only one-dimensional data. Instead, the inner retina was characterized by average inner retinal Po ₂, which was calculated for each oxygen profile in both attached and detached retina. 

Statistics

Parameters derived from all profiles obtained under a given condition in each cat were averaged, and the results are reported as the mean ± SD across cats. Values normalized to the control condition are shown in the figures so that trends can be seen, but these were not used for statistical testing. ANOVA was applied first to test for significant differences among the three experimental conditions (attached retina during air breathing [control, air] RD, O₂; and RD, air). Paired t-tests were used to test for significant differences between any two conditions thereafter. The difference was considered significant if P < 0.05. 

Results

Po ₂ Profiles

Typical oxygen profiles in the attached and the detached retina during normoxia and hyperoxia are shown in Figure 3 . The O₂ profiles in the attached retina had a characteristic shape (Fig. 3A) . The Po ₂ was highest in the choroid and started to decrease immediately once the electrode left the choroid, and it reached a minimum close to 0 mm Hg in the IS. In addition to the Po ₂ gradient from the choroid to the IS, there was a Po ₂ gradient from the inner retina to the IS as well. These two Po ₂ gradients indicate that both choroidal and retinal circulations supply oxygen to the photoreceptors. 

The shape of the oxygen profiles in the detached retina did not remain completely the same as those in the attached retina, but they had some characteristics in common (Figs. 3B 3C) . The highest Po ₂ was still at the choroid and the retinal circulation was clearly present in the inner retina. However, the Po ₂ decreased more gradually from the choroid to the minimum Po ₂ in the detached retina. In the fluid layer, the Po ₂ did not decrease linearly with the distance from the choroid in most cases, and there was always a slope change at the border between the fluid layer and the outer edge of the retina. These features were probably caused by (1) a contribution of convection to oxygen transport as the electrode moved through the fluid layer ¹⁷ and (2) a larger oxygen diffusion coefficient in the fluid layer than in the retina. ¹⁶  

In the oxygen profiles shown in Figure 3 , the microelectrode reached the choroid. This was possible only in cases when the detachment height was not too high. Under this condition, detachment height could be calculated. It was simply the distance between the choroid and the outer edge of the retina. The detachment height was approximately 210 and 250 μm, respectively, for the profiles in Figures 3B and 3C . 

Po ₂ at the Outer Edge of the Retina

For each cat, the average value of each parameter during air breathing in the attached retina was taken to be 1.0, and other conditions were normalized to this. Normalized Po ₂ (X= 0) in each cat and the average normalized Po ₂ (X= 0) of nine cats are shown in Figure 4 . The Po ₂ (X= 0) decreased to 31% ± 16% of the control value under normoxia in the detached retina (P ≪ 0.001). Hyperoxia increased the Po ₂ (X= 0) in the detached retina so that it was not significantly different from the control value (111% ± 52%, P = 0.40) and was significantly higher than the value in the detached retina during air breathing (P = 0.003). In all animals, hyperoxia maintained Po ₂ (X= 0) better than air breathing in RD. 

Po ₂ at the Inner Border of the Avascular Layer

The Po ₂ at the inner border of the avascular layer (PL) provides information about the supply of oxygen from the retinal circulation to the photoreceptors. Normalized PL in each cat and the average normalized PL of nine cats are shown in Figure 5 . Retinal detachment did not make PL during air breathing significantly different from the control value (272% ± 320%, P = 0.13). The average was higher because, in one cat, PL was very high. In this cat, the electrode was probably close to a blood vessel. If this cat was excluded, the average was 172% ± 119%. Hyperoxia increased PL in the detached retina to 410% ± 290% of the control (P = 0.007); this value was also significantly higher than the value in the RD, air condition (P = 0.045). 

Photoreceptor Oxygen Consumption

Normalized photoreceptor oxygen consumption (Q _av) for each cat and the average normalized Q _av are shown in Figure 6(n = 9). Normalized Q _av significantly decreased to 47% ± 18% of the control value during the RD,air condition (P = 0.0001). Hyperoxia increased Q _av to 68% ± 17% of the control value. This was still significantly lower than the control value (P = 0.003) but was higher than the value during air breathing (P = 0.002). In some animals, hyperoxia restored Q _av to almost normal, whereas in other cases it maintained Q _av at approximately 40% of the control level. In all but one animal, hyperoxia maintained Q _av better than air; in this animal, normoxia and hyperoxia were equally effective. 

The fraction of oxygen delivered by the retinal circulation to the photoreceptors (F _r) was calculated based on the parameters derived from each fit. ¹² F _r was significantly higher under RD,air (49.34% ± 27.4%) than under RD,O₂ (23.83% ± 25.2%) (P = 0.001), and it was also significantly higher than that in the attached retina under normoxia (5.67% ± 5.44%, P = 0.002). There was no significant difference between the control and RD,O₂ (P = 0.089), but in six of the nine animals, F _r was higher during RD,O₂ than the control. Thus, the retinal circulation frequently made a larger contribution to photoreceptor oxygenation in the detached retina than in the attached retina. 

Average Inner Retinal Po ₂

Figure 7shows the normalized average inner retinal Po ₂ (P _IR) from eight cats. Data from one cat were excluded because the absolute P _IR from that cat in the control condition was more than 2 standard deviations greater than the average of absolute P _IR from all nine cats. P _IR was 166% ± 110% of the control value during detachment under normoxia; this was not a significant increase (P = 0.19). Hyperoxia increased P _IR to 240% ± 143% of the control value during detachment; this value was substantially higher than both the control value (P = 0.008) and the RD,air value (P = 0.043). In all eight cases, P _IR was higher than the control value during detachment under hyperoxia. 

Internal Correlations

Although hyperoxia was beneficial, its effect was variable. In particular, Figure 6shows that in some cases Q _av could be maintained at almost normal levels by hyperoxia, whereas in other cases it could only be kept at 40% to 50% of normal. We wanted to know whether this variability in Q _av may depend on variation in other parameters, and so we examined correlations between Q _av and Po ₂ (X= 0; Fig. 8A ), and Q _av and PL (Fig. 8B)with linear regression. The correlation between PL and P _IR (Fig. 8C)was examined as well, to determine the extent to which the Po ₂ at one boundary of the inner retina (PL) was representative of that in the entire inner retina (P _IR). As observed previously, Q _av was strongly dependent on Po ₂ (X= 0; P = 0.002, R ² = 0.77) in the attached retina under normoxia. ^{12 18} The strong dependency of Q _av on Po ₂ (X= 0) disappeared during detachment under both normoxia (P = 0.15, R ² = 0.28) and hyperoxia (P = 0.29, R ² = 0.16). Q _av was not dependent on PL under normoxia in the attached retina (P = 0.97, R ² = 0.0002), which is consistent with the small fraction of oxygen that the photoreceptors usually derive from the retinal circulation. However, the dependency of Q _av on PL was almost significant (P = 0.064, R ² = 0.41) under hyperoxia during detachment. These correlations reflect a shift in the importance of the two circulations. PL was strongly dependent on P _IR (P < 0.05, R ² = 0.74, R ² = 0.87, R ² = 0.70) in all three cases. It is interesting that they are well correlated since PL is just one point in the average P _{I R}. 

Discussion

Effects of Hyperoxia on Retinal Oxygenation and Oxygen Consumption

At this time, there are only two ways to obtain information on the rate of photoreceptor oxidative metabolism in vivo. One is by using the Fick principle—that is, multiplying the difference between arterial and venous oxygen saturations by blood flow rate. This would have to be done for both the retinal and choroidal circulations, because photoreceptors derive oxygen from both. In addition to the difficulty of obtaining absolute measurements of retinal and choroidal blood flow and venous saturation, the Fick principle has two further limitations that prevent its use under the conditions we were interested in. First, it would have been impossible to assess how much of the consumption value obtained from the retinal circulation was used to supply the photoreceptors, and second, it cannot be used to obtain values for different retinal regions (e.g., attached and detached). The second method for assessing oxidative metabolism, which we used in the present study, is to record oxygen profiles and fit a mathematical model to those data. The three-layer model we use was selected from several candidate models, ¹⁵ provides outstanding fits to the data, ^{12 15} is consistent with the known structure of photoreceptors, and fits data for all conditions and animals in which it has been used. ^{12 18 19 20 21} As a result, the consumption values extracted from the model provide the best measure of in vivo photoreceptor oxidative metabolism that is available. 

By combining the measurements and modeling, we have demonstrated that hyperoxia improves oxygen consumption in the detached retina. Once the retina was detached, the Po ₂ at the outer edge of the retina, Po ₂ (X= 0), decreased substantially, making it impossible for photoreceptors to obtain enough O₂ to maintain normal Q _av. This is what happens in systemic hypoxia as well. ¹² Hyperoxia elevated Po ₂ (X= 0) to normal and increased PL well above normal, but this restored Q _av only part way to the control value (68% ± 17%). The failure to restore Q _av even though the oxygen supply at the boundaries of the avascular layer is normal or above normal suggests that there are other factors in addition to hypoxia that have detrimental effects on the photoreceptors in the detached retina. These additional factors could include acidosis or reduced supply of other diffusible substances. 

In some cats, hyperoxia maintained Q _av better than in others. Disorganized photoreceptors were occasionally encountered in the detached region in histologic slides from some cats ¹⁷ although they remained intact in most cases. The disorganization of the photoreceptors would be likely to cause a decrease in Q _av, even during hyperoxia, and the protective effects of hyperoxia are expected to be less pronounced under those conditions. 

Average inner retinal Po ₂ (P _IR) in the detached retina was not significantly different from the control value, indicating that inner retinal oxygenation was preserved, and the retinal circulation was functional during detachment. We are not aware of any previous work on this subject. Hyperoxia increased P _IR to 240% ± 143% of the control value. This increase in P _IR was important in elevating Q _av during detachment and would be even more important in larger detachments because the choroid is too far away to supply oxygen. 

In the attached retina, a strong dependency of Q _av on Po ₂ (X= 0) was observed as reported previously. ^{12 18} This dependence occurs because usually the photoreceptor oxygen consumption is limited by the amount of oxygen that can be obtained from the choroid, which depends on the Po ₂ gradient between the choroid and inner segments. The strong dependency of Q _av on Po ₂ (X= 0) disappeared in the detached retina under both normoxic and hyperoxic conditions, suggesting that Po ₂ (X= 0) does not play as important a role in maintaining oxygen consumption in the detached retina as it does in the attached retina. However, the fraction of oxygen supplied by the retinal circulation was considerably higher during detachment, and the dependency of Q _av on PL was nearly as significant during hyperoxia. These observations indicate that oxygen supply from the retinal circulation plays a more important role than Po ₂ (X= 0) when the retina is detached. This unusual phenomenon can be appreciated in the following way. In the attached retina, choroidal Po ₂ is much higher than PL, and retinal circulation supplies only a small portion of oxygen to the photoreceptors. ¹² However, in the detached region, PL may not be substantially lower than Po ₂ (X= 0) due to a substantial decrease in Po ₂ (X= 0). Under that condition, photoreceptors rely more on the retinal circulation. In larger detachments, the dependency of Q _av on PL is expected to be stronger. This is because Po ₂ (X= 0) would decrease tremendously, whereas PL depends on the retinal circulation, which seems to be functioning normally. 

Clinical Implications

The oxygenation data in this study are in good agreement with, and provide a possible mechanism for, the protective effects of hyperoxia on retinal structure during retinal detachments that have been observed in the previous studies in cats and ground squirrels. ^{7 8} These oxygenation and histologic data indicate that hyperoxia has a potential for treating retinal detachment during the time before a detachment is surgically repaired. Because oxygen cannot be stored in tissue, hyperoxia would probably have to be maintained continuously to achieve a benefit. Clinically, 70% O₂ would be more suitable than 100% O₂, to reduce lung toxicity. This level is known to be effective, as it has been used in histologic studies. ¹⁰ Hyperoxia itself can be toxic to rat photoreceptors, ²² but in the histologic studies of hyperoxia in detachment, there were no apparent detrimental effects of 6 days exposure to 70% O₂ to the photoreceptors in the attached part of the retina. ¹⁰  

In the present study, most detachments were created in the area centralis. The oxygenation of the area centralis in the cat is similar to that of the parafovea in the monkey, ¹⁹ and the oxygen data presented herein do not provide complete information on cases in which there is a detachment of the macula. A further limitation is that the measurements reported were obtained in the detached retina and do not prove that hyperoxia administered during detachment would contribute to recovery of metabolism or function after the retina is reattached. For that reason, an important next step is to assess oxygen consumption after reattachment and compare the conditions in which the animal breathes air or oxygen during the time of the detachment. 

 Supported by National Eye Institute Grant R01EY05034 (RAL).

 Submitted for publication July 21, 2006; revised October 5, 2006; accepted January 16, 2007.

 Disclosure: S. Wang, None; R. A . Linsenmeier, 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.

 

A typical fundus picture with a retinal detachment in the cat eye. Arrows: attached and detached parts of the retina.

Figure 2.

 

Determination of the two boundaries of the outer retina from an intraretinal oxygen profile collected in the detached retina. The abscissa is labeled according to the distance in the mathematical model. X= 0 is the border between the fluid layer and the outer edge of the retina; X = L is the boundary between the outer and the inner retina. Data were obtained during electrode withdrawal from the retina, after the retina was penetrated from the vitreous. Lines superimposed on the data show layers 1 and 3 of the diffusion model (OS and ONL respectively), where oxygen consumption is zero, and so the gradient is linear. The avascular retina is about half of the retinal thickness, and therefore X= 2L represents the vitreous–retina border.

Figure 3.

 

Examples of oxygen profiles collected in different parts of the retina under air and 100% O₂ breathing in cat 377. Profiles were collected (A) in the attached retina during air breathing (Control, air); (B) in the detached retina under air breathing (RD, air); and (C) in the detached retina under 100% O₂ breathing (RD,O₂). OR, outer retina (determined as described in the text and Fig. 2 ); IR, inner retina; FL, fluid layer under the retina.

Figure 4.

 

(A) Normalized Po ₂ at the outer edge of the retina, Po ₂ (X= 0), in the attached part of the retina during normoxia (Control, air) and in the detached part of the retina during both normoxia (RD,air) and hyperoxia (RD,O₂) in each cat (n = 9). Each data point is the average of all profiles obtained in one animal under a particular condition. (B) The mean ± SD of normalized Po ₂ at the outer edge of the retina for all cats (n = 9). *Statistical significance between conditions (P < 0.05) was based on paired t-test of the absolute values after ANOVA, not on the mean values shown. Under control conditions, Po ₂ (X= 0) was 43.77 ± 17.01 mm Hg.

Figure 5.

 

(A) Normalized Po ₂ at the inner border of the avascular region (PL) in the attached part of the retina under normoxia and in the detached part of retina under both normoxia and hyperoxia in each cat (n = 9). (B) The average of normalized PL for all the cats (n = 9). The format is the same as in Figure 4 . Under control conditions, PL was 8.88 ± 4.46 mm Hg.

Figure 6.

 

Normalized average oxygen consumption of the outer retina (Q _av) under both air and 100% O₂ breathing in the attached and detached retina for each cat (A) and for all cats (B) n = 9. The format is the same as in Figure 4 . Under control conditions, Q _av was 2.67 ± 1.26 mL O₂ · 100 g⁻¹ · min⁻¹.

Figure 7.

 

Normalized P _IR in different regions of the retina under both air and 100% O₂ breathing in the attached and detached retina for each cat (A) and for all cats (B); n = 8. The format is the same as in Figure 4 . Under control conditions, P _IR was 15.51 ± 4.44 mm Hg.

Figure 8.

 

(A) Correlation between Q _av and Po ₂ (X= 0) under different breathing conditions in the detached and the attached parts of the retina. Each point represents the average value for one cat. The slope of the Control, Air condition was significantly different from zero (P = 0.002). (B) Correlation between Q _av and PL. The function for RD,O₂ approached statistical significance (P = 0.064). (C) Correlation between PL and P_IR. All correlations were significant (P < 0.05).

The authors thank Steven Fisher and Geoff Lewis for useful discussions, and Norbert Wangsa-Wirawan, Gulnur Birol, Ewa Budzynski, Paul Bryar, Christina Enroth–Cugell, Yun Kim, and Karen Tenenhaus for assistance with the experiments.