Oxygen Distribution in the Mouse Retina

Abstract

purpose. To make the first measurements of intraretinal oxygen distribution in the mouse, an animal model of increasing importance in ophthalmic research.

methods. Oxygen-sensitive microelectrodes were used to measure oxygen tension as a function of depth through the retina and choroid in anesthetized mice (n = 8). All measurements were performed under light-adapted conditions, with the animals spontaneously inspiring room air. The oxygen distribution in the avascular portion of the outer retina was analyzed by an established three-layer mathematical model that determines outer retinal oxygen consumption.

results. The intraretinal oxygen distribution in the inner retina in individual profiles was often characterized by sharp peaks associated with elements of the retinal microvasculature, but, in the outer retina, the oxygen distribution was much more predictable and reflected the high oxygen uptake of the photoreceptors. Average choroidal oxygen tension was 42.0 ± 1.2 mm Hg (mean ± SE), and Po ₂ at the surface of the retina was 21.7 ± 0.8 mm Hg. The average minimum oxygen tension in the outer retina was 4.2 ± 0.5 mm Hg. Average outer retinal oxygen consumption in the mouse was 193.3 ± 10.6 nL O₂/min per square centimeter, which is similar to that previously reported in the outer retina in vascularized areas of retina in the rat and monkey using similar techniques.

conclusions. The intraretinal oxygen distribution in the mouse is qualitatively and quantitatively similar to that in other species with vascularized retinas. The rate of oxygen consumption in the outer retina is also similar. These baseline data can now be used in studies employing mouse models of retinal disease.

Mouse models of retinal disease are increasing in popularity, and this trend seems set to continue, given the rapid expansion of gene-targeted “knockout” and over expression studies that are readily applicable to mice. The mouse model of oxygen-induced retinopathy is also widely used in pharmacological studies relevant to the treatment of vascular proliferation. However, at present, there is no information in the literature about the normal distribution of oxygen tension across the mouse retina. Such measurements have been performed in many other species, and the heterogeneous distribution of oxygen supply and consumption in the retina has been well established. ^{1 2 3 4 5 6 7} These studies required the use of microelectrode-based measurements, in which a small-tipped electrode penetrates the retina and measures oxygen level as a function of retinal depth. The manipulation of electrodes within the eye becomes more difficult in smaller animals because the eyes are smaller and the relative size of the lens increases, leaving a very small volume of vitreous in which the electrode may be manipulated. The smallest mammal in which the intraretinal oxygen distribution has been published is the rat, which has a vitreous volume of approximately 55 μL. ⁴ This article describes the first measurements of intraretinal oxygen distribution in the mouse, a species with an eye volume approximately one eighth that of the rat ⁸ but that has had rapidly increasing application in ophthalmic research. 

Methods

The experimental techniques were similar to those reported in our earlier work in rats, ⁴ guinea pigs, ⁶ and rabbits. ⁹ Eight Arc(S)(Swiss) albino mice (Animal Resources Centre, Perth, Western Australia, Australia) with an average body weight of 35.3 ± 1.3 g were included in the study. Mice were anesthetized with ketamine 80 mg/kg and xylazine 4 mg/kg given intramuscularly and top-up doses given as required. The animal was then placed in a robotic stereotaxic apparatus, ¹⁰ and the eye was stabilized by suturing to a fixed eye ring at the limbus using 10-0 sutures at five points around the eye. The upper jaw was stabilized by a bite bar, and the head was fixed in position by ear bars. The mice were intubated but allowed to breathe room air under spontaneous ventilation conditions. Body temperature was monitored and maintained with a homeothermic blanket and feed-back controller (Harvard Apparatus, Holliston, MA). 

A small hole at the pars plana allowed entry of an oxygen-sensitive microelectrode. The electrodes were manufactured in our laboratory using techniques based on those described by Whalen et al. ¹¹ The electrode was visualized inside the eye via a plano concave contact lens and operating microscope (OPMI; Carl Zeiss Meditec, Jena, Germany). The electrode was positioned so that its tip was placed close to the surface of the inferior retina in a region free of major retinal vessels. Figure 1schematically illustrates the electrode in position inside the mouse eye. Only the tapered portion of the electrode enters the eye, thus minimizing vitreous displacement. All electrode movements during intraretinal penetrations were under computer control, with the oxygen level being recorded at 10-μm intervals through the retina. Measurements were made during penetration and subsequent withdrawal from the retina. Withdrawal profiles were used for analysis, because they were less affected by artifacts associated with mechanical stress on the electrode tip during tissue penetration. Four intraretinal profiles from each animal were used for subsequent analysis. The nonperpendicular nature of the penetration means that distances are expressed as track length through the retina, rather than as absolute retinal depth. For the quantitative 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 ² was fitted to the oxygen profiles in the outermost 150 μm of each profile. ¹² This analysis extracts a value for outer retinal oxygen consumption and generates a best-fit curve for each profile. All measurements were performed in light-adapted conditions. All average values are shown as the mean ± SE, and all error bars on graphs are also standard errors. Supplemental anesthesia was administered if slight eye movements were observed in monitoring the position of the microelectrode relative to the fundus. Typically, the first profile measurements began within an hour of anesthesia induction, and the experiment lasted for a further 2 hours, during which one or two top-up doses of anesthetic were required. At the conclusion of the experiment, the animal was killed with an anesthetic overdose. All procedures conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Results

The average intraretinal oxygen level as a function of penetration track length is shown in Figure 2 . The Po ₂ at the retinal surface was 21.7 ± 0.8 mm Hg and the Po ₂ at the peak in the choroid was 42.0 ± 1.2 mm Hg. The minimum Po ₂ in the outer retina was 4.2 ± 0.5 mm Hg, and in the inner retina it was 5.0 ± 0.5 mm Hg. Although, in individual profiles, the size and the position of the “peaks” in the inner retina were quite variable, there was still a detectable “bump” in the average Po ₂ distribution in the inner retina. The outermost 150 μm of track length in the outer retina of each profile was fitted to the mathematical model of outer retinal oxygen consumption, and the average of all the best fit curves is shown in Figure 3 . The average fitted Po ₂ at Bruch’s membrane was 42.2 ±1.4 mm Hg, and the minimum Po ₂ in the outer retina was 4.26 ± 0.6 mm Hg. The average oxygen consumption rate calculated for the outer retina was 193.3 ± 10.6 nL O₂/min per square centimeter. 

Discussion

Both insufficient and excessive oxygenation of the retina can be damaging. For example, retinal hypoxia is thought to be a key factor in the sight-threatening consequences of retinal ischemia, and retinal hyperoxia is thought to be a major factor in retinopathy of prematurity. A knowledge of the intraretinal oxygen distribution provides valuable information about supply and consumption of oxygen in the mammalian retina. Previous measurements of intraretinal oxygen distribution in monkeys, ^{5 12} pigs, ¹³ cats, ^{1 2} rabbits, ⁷ guinea pigs, ⁶ and rats ⁴ have yielded a wealth of information about retinal oxygen metabolism in health and disease. It is clear that retinal oxygen supply is closely matched to tissue demands and that only minor perturbation in oxygen supply or demand can lead to tissue hypoxia. Each species has been shown to have distinct patterns of oxygen supply and consumption related to their particular retinal structure, metabolic requirements, and the presence or absence of retinal vasculature in the region studied. The mouse is now becoming an increasingly useful animal in ophthalmic research, mainly due to the availability of new mouse models of retinal disease. The small size of the mouse eye has previously discouraged the use of microelectrode-based technologies in the eye. The only data available for oxygenation status of the mouse eye is restricted to preretinal measurements of hyperoxia induced oxygen changes using noninvasive nuclear magnetic resonance (NMR) techniques ¹⁴ or a mixture of retinal and choroidal Po ₂ in two-dimensional maps generated by fluorescence lifetime measurements. ¹⁵ Neither of these techniques allows the determination of the oxygen status or consumption rates of specific retinal layers. Having previously adapted oxygen-sensitive microelectrode technology for use in the eyes of young rats, ^{16 17} we thought that a further refinement of such technology would make measurements of intraretinal oxygen distribution in the mouse eye feasible. The main difficulty to overcome was the very limited vitreous space in which the electrode can be manipulated. Using an estimate that the mouse eye is one eighth of the volume of a rat eye, and a vitreous volume in the rat of 55 μL, an estimate for the vitreous volume in the mouse is ∼7 μL. Very careful attention to the initial placement of the electrode in the eye was therefore required, a process that was greatly aided by the precision of our robotic electrode orientation system. ¹⁰ We were not able to monitor intraocular pressure (IOP) during our measurements. The mouse eye is too small to cannulate the anterior chamber with the plano concave contact lens and eye ring in position, and corneal applanation techniques are precluded by the contact lens and the stabilizing eye ring. It may be expected that the penetration hole at the pars plana leads to a small transient drop in IOP as vitreous gel “bulges” out of the entry hole. The fact that the vitreous “bulges” rather than leaks means that there is scope for normalization of IOP through the normal control mechanisms. When the electrode is placed into the eye, a small gap remains around the electrode taper, even when the electrode is fully inserted. The diameter of the electrode shaft 3 mm back from the 1-μm tip is approximately 100 μm, so during a typical retinal penetration with a track length of ∼400 μm the volume of additional vitreous displaced by the electrode is approximately 3.1 nL. Compared with the eye volume of ∼18.8 μL, this can be considered negligible. We are therefore confident that movement of the electrode within the eye during a retinal penetration does not result in increases in IOP large enough to affect retinal and choroidal perfusion significantly. In the present study in mice, we were also unable to monitor arterial blood gases during the oxygen measurements. Cannulation of the carotid artery would create unwanted surgical trauma, and the low blood volume of the mouse (2–3 mL) ¹⁸ precludes repetitive blood sampling that we have used in larger animals. ¹⁹ Consequently, we cannot claim that the animals were under normal physiological conditions. However, there is nothing in the nature of the oxygen profiles obtained that is suggestive of poor physiological conditions. Studies in larger animals have indicated that intraretinal hypoxia is an early indicator of suppressed blood pressure, ⁴ low blood gases, ²⁰ or raised IOP. ^{21 22} The absence of intraretinal hypoxia in our study suggests that oxygen supply and consumption were not significantly affected by abnormal blood gases, low blood pressure, or raised IOP. 

Our results have shown that the intraretinal oxygen environment in the mouse is similar to that in vascularized retinas, such as the cat, pig, rat, and monkey. The familiar minimum oxygen tension in the outer retina is evident along with steep oxygen gradients to a peak oxygen tension within the choroid. Based on the choroidal Po ₂ of 42 mm Hg and a preretinal Po ₂ of 21.7 mm Hg in the present mouse studies, it seems unlikely that anesthesia and spontaneous ventilation result in significant systemic hypoxia. In larger rodents such as the rat, studies using direct blood gas sampling and forced mechanical ventilation showed average choroidal Po ₂ to be 42.3 mm Hg and preretinal Po ₂ to be 18.6 mm Hg under normal systemic conditions. ¹⁹ The minimum Po ₂ in the outer retina is also comparable, at 4.5 mm Hg in the rat and 4.2 mm Hg in the mouse. We believe that blood pressure and heart rate were also close to normal, because in more recent experiments with the same anesthetic protocol, we have been able to use a commercial heart rate–blood pressure monitor (RTBP; Harvard Apparatus) to confirm that blood pressure and heart rate are in the normal range for mice. 

In the outer retina, the absence of intrinsic vasculature means that the oxygen distribution can be analyzed to determine the rate of outer retinal oxygen consumption. The resultant estimate of outer retinal oxygen in consumption in the inferior retina of the mouse (193.3 ± 10.6 nL O₂/min per square centimeter) is similar to that reported by this group for equivalent locations in the rat (148 ± 11 nL O₂/min per square centimeter) and the monkey (216.9 ± 12.3 nL O₂/min per square centimeter). ^{12 23} Direct comparison with studies from other groups should take into account applied correction factors for electrode penetration angle ⁵ and/or tissue stretch during withdrawal of the electrode. ² It must also be remembered that, in many species, outer retinal oxygen consumption is considerably higher in dark-adapted conditions. ^{2 5}  

The inner retinal oxygen consumption in vascularized retinas can generally not be determined by analysis of the intraretinal oxygen distribution because of the influence of the retinal vasculature. In the rat, this problem was partially overcome by the highly layered nature of the retinal capillary beds. ^{23 24} However, in the present studies in the mouse, the unpredictable nature of inner retinal troughs and peaks in the oxygen profile indicates a much more random distribution of capillaries or other vascular structures influencing the oxygen profile. Consequently, quantitative analysis of inner retinal oxygen consumption was not performed. 

The present study provides the first normative data for intraretinal oxygen distribution and consumption in the mouse. These data provide the baseline for studies in which intraretinal oxygen distribution in the mouse can be manipulated by environmental, genetic, and experimental models of retinal disease. 

 Supported by the National Health and Medical Research Council of Australia: Australian Research Council under the ARC Centres of Excellence Program.

 Submitted for publication August 23, 2005; revised September 29 and November 7, 2005; accepted January 5, 2006.

 Disclosure: D.-Y. Yu, None; S.J. Cringle, 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 scale drawing of the vertical plane of the mouse eye, as described by Remtulla and Hallet, ⁸ shown with a schematic oxygen-sensitive microelectrode entering the eye at the pars plana and in position in front of the inferior retina. The electrode tip is 1 μm in diameter, and the diameter of the electrode at the entry point into the eye is approximately 100 μm. The eye is fixed in position by suturing to an eye ring at the limbus. A plano concave contact lens on the cornea allows the electrode to be visualized inside the eye with the aid of an operating microscope. Scale bar, 1 mm.

Figure 2.

 

Average oxygen tension as a function of electrode track length through the retina of anesthetized mice (32 measurements in eight mice). Average Po ₂ at the retinal surface was 21.7 ± 0.8 mm Hg, at the inner retinal minimum, 5.0 ± 0.5 mm Hg; and at the outer retinal minimum, 4.2 ± 0.5 mm Hg before rising steeply to reach a peak of 42.0 ± 1.2 mm Hg in the choroid. Error bars, ±SE.

Figure 3.

 

The average of the best fit curves obtained when fitting the data from the outermost 150-μm track length to the mathematical model of outer retinal oxygen consumption. The average fitted value for the outer retinal peak was 42.2 ± 1.4 mm Hg and the fitted minimum Po ₂ in the outer retina was 4.26 ± 0.6 mm Hg. The calculated value for outer retinal oxygen consumption was 193.3 ± 10.6 nL O₂/min per square centimeter. Error bars, ±SE.

The authors thank Dean Darcey, Judi Granger, Paula Yu, Er-Ning Su, and Megan Dallas for expert technical assistance.