Decreased Visual Function after Patchy Loss of Retinal Pigment Epithelium Induced by Low-Dose Sodium Iodate

Abstract

purpose. To correlate damage to the retinal pigment epithelium (RPE) with decreased visual function after the systemic administration of sodium iodate (NaIO₃).

methods. Damage was produced in mice by injection of 15, 25, or 35 mg/kg NaIO₃. Visual function was assessed with the cued water maze (WM) behavioral test and the optokinetic reflex (OKR) measurement at different times after injection. Autofluorescence in whole eye flatmounts was quantified, and hematoxylin and eosin staining of paraffin sections was performed to assess changes in the outer retina.

results. After 15 mg/kg NaIO₃, cued WM test results were normal, whereas OKR measurements were significantly decreased at all times. Focal RPE loss began on day 21, but no significant damage to the outer nuclear layer was observed. After 25 mg/kg NaIO₃, the cued WM test was transitionally reduced and the OKR measurement again decreased at all times. Large areas of RPE loss occurred on day 14 with a reduced outer nuclear layer on the same day. With 35 mg/kg NaIO₃, the cued WM test was reduced beginning on day 14 with complete obliteration of the OKR beginning on day 3, large areas of RPE loss on the same day, and a reduced outer nuclear layer on day 7.

conclusions. Stable, patchy RPE loss was observed with a low concentration of NaIO₃. The OKR measurement showed changes in visual function earlier than the cued WM test and before histologic findings were observed.

Alterations in the RPE monolayer are part of physiological aging and of pathophysiological processes. Several characteristics of a normal, aged pigment epithelium are a decrease in RPE density, a clinically observed decrease in the pigmented appearance of the RPE cells, and the accumulation of lipofuscin within RPE cells. ^{1 2} RPE dysfunction and atrophy can be found in several retinal disorders, including hereditary disorders such as retinitis pigmentosa ³ and Stargardt’s disease ⁴ and age-related macular degeneration (AMD). In AMD, the initial morphologic changes are associated with lipofuscin accumulation and with the formation of drusen and other deposits on Bruch’s membrane. Subsequently, RPE cell loss occurs presumably through apoptosis associated with the loss of cell attachment. ^{5 6} Loss of RPE, observed as areas of bare Bruch’s membrane, is also found after submacular surgery for choroidal neovascularization performed in patients with exudative or “wet” AMD. ⁷ Furthermore, damage to the RPE has been observed after uncomplicated photodynamic therapy, an established treatment for choroidal neovascularization. ⁸ Therefore, the selective loss of the RPE monolayer in the rodent after the intravenous injection of sodium iodate (NaIO₃) may be a useful model for studying RPE cell regeneration on a normal BM. ⁹  

We previously reported a decrease in electrophysiological function (the electroretinogram) and visual function (the cued water maze [WM] test), after the intravenous injection of NaIO₃ in the rodent. These changes were observed at various times after injection and were dependent on the concentration used. ^{10 11}  

In this study, we compared a decrease in visual functional behavior in pigmented C57BL/6 mice after systemic treatment with low doses of NaIO₃ using either the cued WM test or the optokinetic reflex (OKR) and correlated these changes with structural alterations in the neural retina. We observed that after damage to the RPE with low doses of NaIO₃, the OKR was more sensitive than the cued WM test in detecting a decreased behavioral response and that the functional decrease correlated with morphologic changes in the neural retina. 

Methods

Animals

Four- to 6-week-old male C57BL/6 mice (Charles River WIGA, Sulzfeld, Germany), treated according to the regulations in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, received single intravenous injections of sterile 0.5% or 1% NaIO₃ in saline (15, 25, or 35 mg/ kg; Sigma-Aldrich, Buchs SG, Switzerland). Control animals received 100 μL 0.9% NaCl. Visual function was assessed in parallel in the cued WM and by OKR measurement 3, 7, 14, 21, and 28 days and 3 months postinjection (PI) with five animals per group and time. Additional animals were euthanatized at each time point for assessment of morphologic changes in retinal tissue. 

Cued Water Maze Test

Before NaIO₃ treatment, mice underwent a pretraining acclimation session during which they were allowed to swim in the testing pool with no platform present. After this explanatory session, mice were trained twice on the visible platform (cued) task in a WM commonly used to assess sensorimotor deficits in rodents. ¹² The WM consisted of a circular, galvanized steel pool (diameter, 1.8 m; height, 0.7 m). A white platform (diameter, 20 cm) was placed inside, and the tank was filled with water (23°C) that was made opaque with white nontoxic emulsion paint (Bahag AG, Mannheim, Germany), which rendered the submerged portion of the platform invisible. The location of the platform was identified with a visual stimulus consisting of a plastic cylinder (height, 5 cm; radius, 10 cm) that extended approximately 1.5 cm above the surface. Four points along the perimeter of the maze, arbitrarily designated as north, south, east, and west, served as the release points for the mice. Swim tracks were recorded with a high-resolution camera (CCD-IRIA; Sony) placed above the pool and connected to a computer system. Test sessions consisted of four trials, with the platform moved to a new location in the pool after each trial. Once a mouse located the platform on a given trial, it was allowed to remain there for 15 seconds before it was removed from the tank. If a mouse did not locate the platform within 60 seconds, it was manually guided to it and also allowed to rest for 15 seconds. Swim latency, path length, and swim speed were calculated using the recorded data (EthoVison; Noldus Information Technology, Wageningen, The Netherlands). To ensure that any behavioral changes could be attributed to alterations in visual performance, the physical condition of each animal was monitored PI. 

Optokinetic Reflex Measurement

Measurement of the OKR was performed after recording the WM performance (OptoMotry system; CerebralMechanics, Lethbridge, AB, Canada). A virtual cylinder comprising a vertical sine wave grating was projected in three-dimensional (3-D) coordinate space on computer monitors arranged in a square around a testing arena. ¹³ The testing arena was bordered by a mirror bottom with a platform positioned 13 cm above the floor, a mirrored lid on the top of the apparatus, and four 17” LCD computer monitors (model 1703FP; Dell, Phoenix, AZ) attached to each outside wall so that all monitors projected into the arena. A video camera (DCR-HC26; Sony) was positioned directly above the platform. A computer program was used to project on the monitors a virtual cylinder in 3-D coordinated space. Visual stimuli were drawn on the walls of the cylinder, the image of which was extended by the floor and ceiling mirrors. Therefore, from the perspective of the platform, each monitor appeared as a window on a surrounding 3-D world. The software also controlled the speed of rotation and geometry of the cylinder and the spatial frequency of the stimuli and enabled live video feedback of the testing arena. 

Mice standing unrestrained on the centered platform tracked the grating with reflexive head movements. As the mouse moved about the platform, the experimenter followed the mouse’s head with a crosshair superimposed on the video image. The x-y positional coordinates of the crosshair were used to center the rotation of the cylinder at the mouse’s viewing position, thereby maintaining the virtual walls of the cylinder at a constant distance from the animal. When a grating perceptible to the mouse was projected on the cylinder wall and the cylinder was rotated (12°/s), the mouse would usually stop moving its body and begin to track the grating with reflexive head movements in concert with the rotation. An experimenter assessed whether the animals tracked the cylinder by monitoring in the video window the image of the cylinder, the animal, and the crosshair simultaneously. If the mouse’s head tracked cylinder rotation, which was evident as movement against the stationary arms of the crosshair, it was judged that the animal could see the grating. If, during the course of testing, an animal slipped or jumped off the platform, it was simply returned to the platform and testing was resumed. Whenever possible, experimenters were masked to the treatment and to the animal’s previously recorded thresholds. All animals were habituated before the outset of testing by being placed on the platform for a few minutes at a time. 

For measurement of visual acuity, the animal was assessed for tracking behavior for a few seconds; this procedure was repeated until unambiguous tracking was observed. The short testing reduced the possibility of adapting to the stimulus and established that each animal was capable of tracking when a changed stimulus was present, and initiating the testing with a low spatial grating enabled each mouse’s optomotor response to be typified. Using a staircase procedure, spatial frequency of the grating was randomly increased until the animal no longer responded. The process of changing the spatial frequency of the test grating was repeated a few times until the highest spatial frequency the mouse could track was identified as the threshold. 

RPE Autofluorescence

For evaluation of morphologic changes in the outer retina, eyes were enucleated and immersion fixed in 4% paraformaldehyde (Sigma-Aldrich) in phosphate-buffered saline (PBS) overnight and processed for either retinal cross-sections or whole eye flatmounts. In the latter, connective and muscle tissue were removed, and an incision was made at the ora serrata, followed by a circumpolar cut. The anterior part of the eye, as well as the lens and sensory retina, were discarded, and the eyecup was washed in PBS and flattened onto a glass slide with three centered cuts. The continuity of RPE fluorescence in these flatmounts was studied on a confocal microscope (SP 2; Leica Microsystems, Heerbrugg SG, Switzerland) with an argon laser (wavelength, 488 nm), and digital images were acquired to compare control and NaIO₃-treated specimens. Finally, the area without autofluorescence was quantified with the use of imaging software (MetaMorph; Molecular Devices, Downingtown, PA). 

Histology

For cross-sections eyes were dehydrated and embedded in paraffin, and 7-μm transverse sections were cut on an ultramicrotome (RM2245; Leica Microsystems) and stained with hematoxylin and eosin (Sigma). Outer nuclear layer (ONL) thickness was measured (Image Manager 50; Leica Microsystems), and the number of nuclei in the ONL was counted at six different locations (400, 1000, and 1600 μm inferior and superior from the edge of the optic nerve head) in each eye. Locations sampled were all at the level of the optic nerve head. 

Statistical Analysis

For each time point and dosage, the mean ± SD (functional tests) or SEM (morphometric data) of these measures were computed, and differences were assessed with one-way ANOVA. Differences were considered statistically significant with P < 0.05. 

Results

Cued Water Maze

Visual function, as measured by the cued WM test, was significantly decreased with 35 mg/kg NaIO₃ beginning on day 14. In other words, increased swim latency (Fig. 1A)and path length (Fig. 1B)were observed. In contrast, the cued WM test was only transiently diminished with 25 mg/kg NaIO₃ and unchanged with 15 mg/kg NaIO₃. These observations at all three doses did not change over the 3-month observation period (data not shown). Mean body weight and swim speed of these animals were not affected over time, indicating no physical impairment from NaIO₃. 

OKR Measurement

Similar measurements of visual acuity were made using the reflexive head movement (optomotor response) to changes in the spatial frequency of the grating. In contrast to the results with the cued WM test, the OKR was significantly reduced with 15 mg/kg NaIO₃ beginning on day 7 and remained reduced throughout the observation period. With 25 mg/kg NaIO₃, there was no response on day 3, and responses thereafter were markedly reduced. With 35 mg/kg NaIO₃, no OKR response was detectable over the entire 3 months (Fig. 2) . 

Autofluorescence

Autofluorescence was used to measure the extent of RPE damage after the injection of NaIO_3. An increase in the area of RPE cell loss was observed (Fig. 3)beginning on day 7. Over time this effect progressed, with the maximum increase measured on day 28 (15 mg/ kg, 263%; 25 mg/kg, 384%; 35 mg/ kg, 395%; compared with control, 100%). However, at 3 months there was a significant reduction in the autofluorescence-free area compared with day 28 at all concentrations. Figure 4shows representative photographs taken after treatment with different NaIO₃ concentrations at several time points. 

Measurement of Outer Nuclear Layer

To investigate the effect of NaIO₃ on the outer neural retina, histologic sections were stained with hematoxylin and eoxin, after which the thickness of the ONL and the number of ONL nuclei were counted. Thereby, a significant decrease in the thickness (Fig. 5A)and in the number of rows of photoreceptor nuclei (Fig. 5B)was observed. The lowest concentration of NaIO₃, 15 mg/ kg, showed the least damage, and the number of photoreceptor nuclei was reduced only at 3 months. Increasing the concentration of NaIO₃ led to the augmentation of degeneration in the outer neurosensory retina. With 25 mg/kg NaIO₃, ONL thickness was reduced beginning on day 14; with 35 mg/kg NaIO₃, it was reduced beginning on day 3. These changes were noted at all times, and, in contrast to RPE autofluorescence, no recovery was observed. Figure 6shows representative photographs with different NaIO₃ concentrations at several time points. 

Discussion

We observed selective loss of the RPE monolayer in the mouse after the intravenous injection of NaIO₃, and a direct correlation was seen between decreased visual function and anatomic cell loss. Furthermore, the extent of RPE damage was dependent on the concentration of NaIO₃ injected and the time elapsed after injection. The threshold between significant and nonsignificant morphologic changes appeared to be around 35 mg/kg NaIO₃, a concentration at which we did not find significant changes in earlier studies ¹⁰ but could detect this time. 

NaIO₃ is known to specifically target the RPE with secondary degeneration of the photoreceptors and choriocapillaris. ¹⁴ The events triggered by NaIO₃ result in immediate changes, specifically in the destruction of the basal membrane of the RPE and loss of the c-wave in the electroretinogram. ¹⁵ Within 24 hours breakdown of the blood-retinal barrier occurs; subsequently, the neurosensory retina and choriocapillaris are affected. ^{16 17} Cell ultrastructure and TUNEL labeling studies indicate that death of the RPE cells results from necrosis and that of the photoreceptors results from apoptosis. ¹⁸  

The selective effect on RPE has been attributed to several mechanisms. NaIO₃ inhibits various enzymes (e.g., acid phosphatase) in RPE cells and destroys the zonula occludens, which is the anatomic basis for the blood-retinal barrier. In addition, it damages anionic sites on both sides of the basal membrane of RPE cells ¹⁹ and has immediate access to the pigment epithelial layer through the choriocapillaris. ²⁰ Baich and Ziegler ²¹ described an effect on melanin, a major component of the RPE, by increasing the ability of melanin to convert glycine to glyoxylate. This chemical reaction was suggested as a partial explanation of the specificity of iodate toxicity toward RPE. 

With a low concentration of NaIO₃ (15 mg/kg), we were able to mimic the patchy loss of the RPE seen in various RPE dystrophies and in geographic atrophy. However, we do not mean to imply that this is a model of AMD but rather a model of RPE loss with subsequent damage to the outer neurosensory retina and choriocapillaris. A major difference with AMD is the normal Bruch’s membrane in this model. Nevertheless, the NaIO₃ model of RPE may represent a useful experimental tool for a number of reasons. First, NaIO₃ can be administered by a single intravenous injection; second, its effect is measurable after short PI times; third, the patchy loss of RPE cells is dose dependent, and one can modulate the extent of the RPE damage by variation in the dose. 

In previous studies, we reported a decrease in electrophysiological function (i.e., the electroretinogram) and the cued WM test using a higher dose of NaIO₃ (50 mg/kg). ¹⁰ However, this dose resulted in widespread destruction of the RPE that was devastating to the overlying retina. Thus, we wanted to investigate the effect of lower doses of NaIO₃ and to identify a visual behavioral test that would correlate with patchy loss of the RPE. 

We used two different visual behavior tests in this study, the cued WM test and the OKR. Our results showed that with 35 mg/kg NaIO₃, the cued WM test is significantly reduced on day 14 but the OKR is obliterated almost immediately after injection. With lower concentrations of NaIO₃ (15 mg/ kg), visual function loss could not be detected by the cued WM test but was readily measurable and detectable by the OKR. Furthermore, by providing thresholds for visual acuity, the OKR measurement gives the ability to quantify changes in visual function. Other investigators have also reported that a virtual optomotor system provides a precise method for quantifying mouse vision without the limitations of other behavioral tests. ¹³  

Highly reliable values with low intersubject differences can be obtained in a matter of minutes with the OKR. Thresholds of visual acuity measured in our control group (mean over all time points, 0.379 cyc/deg) were stable and comparable with published data. ¹³ The data from our treatment groups also had low intersubject differences and were easily distinguished from controls. Thus, the OKR is a versatile tool that can be used to detect even small changes in visual function. 

The low dosage of NaIO₃ used in our study led to secondary degeneration of the ONL in the retina, as described earlier. ^{10 18} In contrast to higher dosages, 15 mg/kg NaIO₃ not only caused less damage to the RPE and neurosensory retina but the damage to visual function remained stable over the 3-month period of observation. Interestingly, the area of RPE autofluorescence in the whole eye flatmounts was found to have increased at 3 months, suggesting possible recovery or regeneration of the RPE, though visual function remained decreased presumably because of earlier permanent damage to the outer neurosensory retina. 

In conclusion, 15 mg/kg NaIO₃ injected intravenously in the mouse produced patchy RPE loss that remained stable over 3 months. At higher concentrations of NaIO₃, widespread damage to the RPE was observed with more severe secondary damage to the ONL of the retina. The OKR measurement was more sensitive than the cued WM test to evaluate visual behavior in this murine model. Thus, the NaIO₃ model of RPE damage in the mouse provides a unique system in which the onset and severity of RPE damage, as well as its functional visual consequences, can be studied noninvasively. This experimental model, in conjunction with OKR-based threshold measurements, can be useful to test the effectiveness of cell replacement therapy because it allows the correlation of morphologic, electrophysiological, and visual behavioral testing. Future experiments could include assessment of neuroprotective traits of a specific compound or neurotrophic factor and regeneration experiments with stem cells. Furthermore, the OKR and autofluorescence measurement could be additional means of assessing RPE morphology and function. 

²  These authors contributed equally to the work presented here and should therefore be regarded as equivalent authors.

 Supported in part by grants from the Swiss National Science Foundation (VE: 310000-119894) and the Berne University Research Foundation (VE).

 Submitted for publication September 17, 2008; revised February 3, 2009; accepted June 8, 2009.

 Disclosure: L.M. Franco, None; R. Zulliger, None; U.E.K. Wolf-Schnurrbusch, None; Y. Katagiri, None; H.J. Kaplan, None; S. Wolf, None; V. Enzmann, 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.

 

Swim latency (A) and path length (B) were diminished after NaIO₃ intravenous injection compared with controls. Significant differences appeared at day 14 after injection (n = 5; mean ± SD; *P < 0.05) using 35 mg/kg NaIO₃ and were detectable over the time period investigated. Injection of 25 mg/kg NaIO₃ led to transient increases in latency and path length between days 7 and 21, whereas 15 mg/kg NaIO₃ did not significantly alter WM performance.

Figure 2.

 

OKR measurement after NaIO₃ treatment over time. Visual acuity was significantly reduced at all time points. Using 15 and 25 mg/kg an effect was measurable beginning on day 7 and day 3, respectively. Improvement was detectable with the latter concentration between days 3 and 14. Injection of 35 mg/kg NaIO₃ suppressed the OKR response completely.

Figure 3.

 

Comparison of whole flatmount RPE autofluorescence after NaIO₃ treatment. The graph depicts the mean ± SEM area without autofluorescence (i.e., RPE) measured using imaging software. Comparison was made to control (solid line) ± SEM (dashed line). *P < 0.05 (control). **P < 0.05 (day 28 PI).

Figure 4.

 

Autofluorescence in whole eye flatmount preparations of control (A) and NaIO₃-treated mice (B–F). Representative photographs of several times PI with different doses are shown: (B) 25 mg/kg, day 7 PI; (C) 35 mg/kg, month 3 PI; (D) 15 mg/kg, day 28 PI; (E) 25 mg/kg, day 28 PI; (F) 35 mg/kg, day 28 PI. Patchy loss of RPE was detected by the decrease in autofluorescence (i.e., increase in black areas). Original magnification, ×1000.

Figure 5.

 

Comparison of retinal layer thickness (μm; mean ± SEM) and number of rows of nuclei in the ONL after treatment with different concentrations of NaIO₃. Comparison was made with control (solid line) ± SEM (dashed line). *P < 0.05.

Figure 6.

 

Histologic sections of the eye stained with hematoxylin and eosin after NaIO₃ treatment compared with control. Representative photographs of several times PI with different doses are shown (15 mg/kg, day 28 PI; 25 mg/kg, day 28 PI; 35 mg/kg, month 3 PI). No alteration in the inner retina was observed with any dose of NaIO₃. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; Segm, photoreceptor segments.

The authors thank Monika Kilchenmann and Franziska Flückiger for their excellent technical assistance.