1.

Introduction

Sodium iodate (${\mathrm{NaIO}}_3$) is a drug commonly used to induce outer retina degeneration in animal models, specifically in rodents. It is known for its toxic effect on the retinal pigment epithelium (PRE) cells¹ that leads to retinal pigmented epithelium (RPE) cell necrosis, followed by photoreceptor (PR) cell apoptosis.^2,3 The effect of this drug on animal retinas has been studied using different methods such as histology and immunohistochemistry,^1–9 auto-fluorescence microscopy,¹ vitreous fluorophotometry,⁴ fundus photography, and fluorescent angiography.¹⁰ Although histology and immunohistochemistry provide accurate and subcellular level resolution cross-sectional images of the retina, these methods are invasive (require animal euthanasia at each time point of a longitudinal study) and are prone to artifacts associated with the sample preparation. Fundus photography and fluorescent angiography allow for non- or minimally invasive in vivo imaging of the retinal surface and blood perfusion; however, they cannot provide high-resolution depth-resolved images of the layered retinal structure to evaluate the drug-induced morphological changes on individual retinal cell types.

Furthermore, a number of studies have been conducted in rodents to investigate the short-term effect of ${\mathrm{NaIO}}_3$ on the retina morphology within the first 24 h of drug administration. These studies utilized research methods such as histology, immunohistochemistry, electron microscopy^2,5,9 and vitreous fluorophotometry.⁴ Defragmentation of RPE cell nuclei at $\sim 2\mathrm{h}$, disorganization of the rod outer segment discs⁹ at $\sim 6\mathrm{h}$, swelling of the PRE cytoplasmic organelles at $\sim 6\mathrm{h}$,⁵ decreased number of RPE cell nuclei at $\sim 12\mathrm{h}$,² and breakdown of the blood-retina barrier at $\sim 24\mathrm{h}$ post drug administration⁴ were reported. Another study reported reduction in the adhesion between the RPE and the Bruch’s membrane within the first 10 min, and between the RPE and the neuro-sensory retina about 2 h later.¹¹

Although ultrahigh resolution optical coherence tomography (UHR-OCT) has been used in the past to image in vivo healthy and diseased rodent retinas,^12–17 until recently, this technology has not been used to image ${\mathrm{NaIO}}_3$ treated rodent retinas. Recently, our research group demonstrated the effectiveness of high-speed UHR-OCT for noninvasive, in vivo three dimensional imaging of the morphological changes induced by ${\mathrm{NaIO}}_3$ in the rat retina.¹⁸ In that study, rats were imaged on days 1, 3, and 7 and at months 1, 2, and 3 post drug injection, to evaluate the long term, acute effect of the drug toxicity.

In this study, we report for the first time, ${\mathrm{NaIO}}_3$ induced changes in the retina morphology and optical properties, observed in vivo with UHR-OCT within the first 12 h of the drug administration. We show that the drug toxicity results in a temporary fluid accumulation at the OS/RPE boundary, commencing within the first 1 to 2 h post drug administration, that causes tearing of the OS from the RPE and precedes the apoptotic and necrotic processes in the photoreceptor and RPE cells previously reported in the literature.

2.

Methods

3.

Results

Figure 1(a) shows a representative cross-sectional UHR-OCT image of the healthy rat retina acquired at baseline, where all retinal layers are clearly visualized. Since this project focused on examining the early effect of ${\mathrm{NaIO}}_3$ toxicity on the RPE and photoreceptor IS and OS, it was important to have a good understanding of the healthy rat retina morphology and specifically the outer retina, as visualized in the UHR-OCT images. Figure 1(b) shows a magnified view of the rat outer retina [area marked with the red rectangle in Fig. 1(a)], while Fig. 1(c) is an anatomically correct schematic of a photoreceptor and an RPE cell, similar to those found in retinal anatomy books. As observed in the UHR-OCT tomogram [Fig. 1(b)], the photoreceptor IS and OS layers consist of alternating low and high reflective bands, which starting from the top and following the analysis presented in a recent publication from Spaide and Curcio,²⁰ can be interpreted as: a very narrow and highly reflective external limiting membrane (ELM), composed from synapses of the Mueller cells; a low reflective band corresponding to the myoid, the upper section of the photoreceptors inner segment (IS1), which is filled with transparent cytoplasm and cellular organelles such as microfilaments, Golgi apparatus and ribosomes that are less than 1 μm in size, thus resulting in weak optical reflectivity; a highly reflective band corresponding to the ellipsoid, the lower part of the photoreceptors inner segment (IS2), which is filled with densely packed mitochondria with significantly higher refractive index ($\sim 1.4$)²¹ as compared to the surrounding cytoplasm ($\sim 1.35$)^22,23; a low reflective outer segment (OS1) consisting of double lipid membrane dicks of fairly low refractive index; a highly reflective band (OS2), corresponding to the OS/RPE junction, filled with interspaced photoreceptor OS and highly reflective microvilli of the RPE cells, which contain large quantities of melanin granules of high optical refractive index ($\sim 1.67$)²⁴; the very thin and highly reflective line of the RPE layer, which contains large quantities of melanin. The detailed analysis of the rat outer retina morphology as visualized in the UHR-OCT images presented in Fig. 1 was used as a reference point to examine any structural changes in the photoreceptor and RPE layers caused by the ${\mathrm{NaIO}}_3$ toxicity at different time points of the longitudinal study. UHR-OCT tomograms of the same appearance and quality were acquired from all rats at baseline and at all time points in the control animals.

Fig. 1

A representative UHR-OCT tomogram acquired from a healthy rat retina with all retinal layers clearly visible and labeled (a) Magnified view of the photoreceptor and RPE layers (b) allows for direct correlation of the alternating low and highly reflective photoreceptor IS and OS bands to anatomical features of the rod photoreceptor cell in the schematic representation (c).

Figure 2 shows representative UHR-OCT images from the rat retina acquired at all time points of the longitudinal study from one of the ${\mathrm{NaIO}}_3$ injected rats. The typical layered structure of the normal healthy rat retina imaged at baseline is shown in Fig. 2(a). A magnified view ($3\times$) of a section of the outer retina, marked with the red box in Fig. 2(a) is shown in Fig. 2(b), where the bright and dark bands of the PR layer are labeled following the method described in Fig. 1. Figure 2(c) shows a cross-sectional image of the same location in the retina, acquired 1 h post ${\mathrm{NaIO}}_3$ injection. The previously clear band corresponding to the OS1 in Fig. 2(b), is now more optically scattering, and a new very low reflectivity layer (LRL) appears at the OS/RPE boundary. In the enlarged view [Fig. 2(d)], the RPE appears structurally intact, while the dark, band above it, most likely corresponds to loose ends of the photoreceptor OS that have been detached from the RPE layer. The end of the OS layer, located immediately above the LRL appears darker as compared with the baseline image, which suggest rearrangement of the loose photoreceptor OS, that can scatter light more, as compared to the case of the healthy rat retina, where the OS ends are neatly arranged in between the RPE microvilli and attached to the RPE.

Fig. 2

Representative UHR-OCT cross-sectional images acquired from the same region of the central retina, showing progressive changes in the outer retina due to the toxic effect of the injected drug: baseline (a) and (b) 1 h (c) and (d), 3 h (e) and (f), 6 h (g) and (h), 8 h (i) and (j), 10 h (k) and (l), 12 h (m) and (n). This rat showed the smallest reaction to the drug toxicity among all rats that were not hydrated with PBS.

Three hours after the ${\mathrm{NaIO}}_3$ injection, the new LRL is no longer visible [Fig. 2(e)]. This is confirmed by the magnified view of the outer retina in Fig. 2(f), where the new LRL is missing, while the photoreceptor OS layer appears almost homogeneous and highly reflective and can no longer be separated in to two distinct bands of low (OS1) and high (OS2) reflectivity, as in the healthy retina. The higher scattering is most likely induced by re-arrangement of the OS, disruption of the membrane disks or both. Six hours post ${\mathrm{NaIO}}_3$ injection [Fig 2(g) and 2(h)], the reflectivity of the photoreceptor OS layer is somewhat reduced as compared to the previous time point and faint appearance of two bands (OS1 and OS2) is observed, which suggests partial re-arrangement of the OS. This process continues gradually at later time points as observed in Fig. 2(1), 2(g), 2(k), and 2(l), acquired at 8 and 10 h post ${\mathrm{NaIO}}_3$ injection, respectively. Twelve hours after ${\mathrm{NaIO}}_3$ injection [Fig. 2(m) and 2(n)], the appearance of the alternating low and highly reflective bands in the IS and OS layers resembles closely the one of the healthy retina [Fig. 2(a) and 2(b)]; however, the overall image contrast of the photoreceptor layer is lower as compared to the baseline image.

The timing of the appearance and disappearance of the LRL and its maximum thickness varied significantly for all rats in the study, as the toxicity of the ${\mathrm{NaIO}}_3$ drug causes somewhat different systemic reaction in each animal. The results presented in Fig. 2 show an extreme case of the smallest LRL thickness, as well as the earliest and most rapid appearance and disappearance of LRL, that were observed in all animals of this study.

Results presented in Fig. 3 show another extreme case, where the LRL thickness was the largest and the layer disappeared sometime between 10 and 12 h post drug injection. In this case, the LRL appeared as early as 1 h post ${\mathrm{NaIO}}_3$ injection [Fig. 3(b)], gradually increased in thickness over the next few hours, reaching a maximum at $\sim 6$ to 8 h post drug injection [Fig. 3(e) and 3(f)] and completely disappeared by 12 h [Fig. 3(h)].

Fig. 3

Representative UHR-OCT cross-sectional images acquired at the same central location of a rat retina depicting the largest reaction to the drug toxicity in the case of no hydration with PBS: baseline (a) and 1 h (b), 2 h (c), 3 h (d), 6 h (e), 8 h (f), 10 h (g), 12 h (h) post ${\mathrm{NaIO}}_3$ injection.

Representative results from the histological and immohistochemical analysis are summarized in Fig. 4. The baseline H&E histological cross-section [Fig. 4(a)] shows the typical normal appearance of the healthy rat retina with clearly defined retinal layers, while the baseline immunohistochemistry image [Fig. 4(f)] shows that the labeling IgG protein remains as expected within the retinal blood vessels, signifying an intact blood-retina barrier (black arrow). One hour post ${\mathrm{NaIO}}_3$ administration, swelling of the photoreceptor OS is observed in the histological cross-section [Fig. 4(b)]. The toxic photoreceptor damage is even more pronounced in the histology image at 3 h, at which time point, the photoreceptors appear large, friable, and pale due to hydropic change and impaired uptake of the H&E dyes [Fig. 4(c)]. The most drastic morphological change was observed at the sixth-hour time point, when the fluid accumulation causes complete detachment of the photoreceptor OS from the underlying RPE [Fig. 4(d)]. Substantial staining of IgG was detected outside the lumen of inner retinal blood vessels in the immunohistochemistry image acquired at 6 h post ${\mathrm{NaIO}}_3$ administration [Fig. 4(g)], confirming protein leakage as a consequence of a breakdown of the blood-retinal barrier. The H&E histological cross-section corresponding to 12 h post drug administration [Fig. 4(e)] appears similar to the one acquired at baseline [Fig. 4(a)], except for the swollen and broken RPE cells.

Fig. 4

Representative H&E (a) through (e) and IgG (f) and (g) stained histology images acquired at various time points of the study. Baseline (a) and (f), 1 h (b), 3 h (c), 6 h (d) and 12 h (e) post sodium iodate injection.

UHR-OCT and histological images of the ${\mathrm{NaIO}}_3$ induced damage to the rat outer retina were very similar in terms of morphological changes such as the retinal detachment, the presence of the new LRL and the altered retinal layers thickness and reflectivity, in all imaged animals from both study groups. However, we observed significant difference in the timing of the appearance and disappearance, as well as the maximum thickness of the LRL between animals of the same group and across the two study groups. This difference was much more pronounced for the animals in Group #1, which have not received supplementary subcutaneous PBS injections, intended to alleviate the systemic effects of the ${\mathrm{NaIO}}_3$ toxicity.

Volumetric OCT images acquired from all animals in each of the two groups were used for statistical analysis of the retinal layer thickness and optical reflectivity. Quantitative thickness and reflectivity measurements were carried out by segmenting individual retinal layers and the choroid at baseline as well as 3, 8, and 12 h. These time points were selected based on less inter-animal variability. Figure 5 shows a representative UHR-OCT tomogram from a healthy rat retina (baseline measurement) with all retinal layers clearly visible [Fig. 5(a)], along with the same image after segmentation with labels indicating the regions used for quantitative statistical analysis [Fig. 5(b)].

Fig. 5

A representative tomogram of a healthy retina with all layers labeled (a) and segmented (b) Brackets show the regions used for statistical analysis of the data; IPL, ONL, PR Layer, choroid, total retina.

Figures 6 and 7 summarize the statistical results for the measure retinal layer thickness and reflectivity. The control group showed no significant difference at any time point compared to the baseline. Therefore, the average (gray solid line) and the standard deviation (gray dotted line) of the values at different time-points acquired from each layer is embedded in the graphs showing data from ${\mathrm{NaIO}}_3$-injected rats for convenient comparison. Figure 6 shows the variation in the retinal and choroidal thickness as a function of time. The total retinal thickness [Fig. 6(a)] showed significant increase ($p=0.0000017$) at 3 h after drug injection, which returned to normal value at 8 and 12 h. The choroid [Fig. 6(b)], the IPL [Fig. 6(c)], and the ONL [Fig. 6(d)] showed no statistically significant thickness changes. The thickness of the PR layer (from ELM to RPE) increased significantly ($p=0.000015$) by 3 h before reducing back to the normal value at later time points as the LRL disappeared [Fig. 6(e)].

Fig. 6

Statistical analysis of changes in layer thickness at different imaging time points in three sodium iodate treated rats receiving extra 1 mL of subcutaneous PBS injection at 1 and 3 h post drug administration. Graphs represent total retina (a), choroid (b), IPL (c), ONL (d), and PR layer (e) thickness measured at baseline and 3, 8, and 12 h. Mean (gray solid line) and standard deviation (gray dotted line) of the control values acquired from three control rats are presented in each graph.

Fig. 7

Statistical analysis of changes in layer reflectivity (intensity) at different imaging time points in three sodium iodate treated rats receiving extra 1 mL of subcutaneous PBS injection at 1 and 3 h post drug administration. Graphs represent total retina (a), choroid (b), IPL (c), ONL (d), and PR layer (e) intensities measured at baseline and 3, 8, and 12 h. Mean (gray solid line) and standard deviation (gray dotted line) of the control values acquired from three control rats are presented in each graph.

Figure 7 shows reflectivity (intensity) changes in the tissue over time within the first 12 h of ${\mathrm{NaIO}}_3$ injection. The intensity of each layer as well as the total retina and the choroid was normalized to that of the corresponding IPL to compensate for image contrast variability during the different imaging sessions due to slight variation in the system performance, light coupling in the animal eye, and possible differences in the image quality in different animals. This normalization was based on the assumption that the IPL remained unaffected by the ${\mathrm{NaIO}}_3$ induced outer retina degeneration. As reported in previous studies^3,7,9 and confirmed in our study, the IPL did not change significantly over time [Figs. 6(c) and 7(c)]. The reflectivity values are presented in percentage by setting the baseline value as 100% for convenient comparison. The total retina reflectivity decreased significantly at 8 h ($p=0.00238$) and remained low ($p=0.00017$) until 12 h [Fig. 7(a)]. The choroid [Fig. 7(b)], the IPL [Fig. 7(c)], and the ONL [Fig. 7(d)] did not show any significant changes in the reflectivity. The photoreceptor (PR) layer however, showed the highest degree of reflectivity changes. The intensity value dropped significantly at 8 h ($p=0.00299$; Fig. 7(d) and remained low until 12 h [$p=0.00016$; Fig. 7(e)].

4.

Discussion

Considering the very low optical reflectivity and dynamics (appearance and disappearance on a time scale of only a few hours) of the LRL, and the clean separation between the photoreceptor OS and the RPE observed in the histological images acquired at 3 and 6 h post injection of ${\mathrm{NaIO}}_3$, most likely the LRL corresponds to fluid accumulation in the subretinal space. This fluid accumulation may result from the early breakdown of the blood retina barrier and loss of RPE adhesion and active outward transport, as an early response to the ${\mathrm{NaIO}}_3$ toxicity. This fluid accumulation is followed by changes in the spatial arrangement of the photoreceptor OS, as seen in the histological cross-sections [Fig. 4], as well as changes in the optical reflectivity of the IS and OS layers (Figs. 6 and 7). The spatial rearrangement observed in our study correlates well with disorganization in the photoreceptor OS discs, observed previously with electron microscopy.⁹ We speculate that detachment of the photoreceptor OS from the RPE, and impaired RPE function disrupt the outer segment disc phagocytosis and could lead to disc accumulation of cellular debris, that could explain the local increase of optical scattering. Our statistical analysis shows that the changes in the total retina and PR layer thickness and reflectivity correspond well with the dynamics of the LRL and alterations observed in the PR layer. The PR layer thickness increased by $\sim 68\%$ ($\sim 28\mu \text{m}$) 3 h after drug injection [Fig. 6(e)] due to formation of the LRL leading to subsequent $\sim 20\%$ ($\sim 33\mu \text{m}$) increase in the total retina thickness [Fig. 6(a)]. As the fluid in the LRL subsides later, the thickness of both PR and total retina decrease and do not show any significant difference from the baseline at 8 and 12 h. Reflectivity changes in the total retina also correspond well with the changes observed in the PR layer. The PR layer intensity [Fig. 7(e)] decreased by $\sim 48\%$ at 8 h post injection due to clearing of the highly reflective material previously observed in the PR space as well as reduced reflectivity of the dark IS2 and OS2 bands. The intensity was decreased further at the last time point (12 h) post ${\mathrm{NaIO}}_3$ injection by 65% compared with the baseline.

Results from this study also suggest that the early effect of ${\mathrm{NaIO}}_3$ toxicity on the rat retina varies strongly from animal to animal depending on the animal physiological response to the toxicity of the ${\mathrm{NaIO}}_3$ drug. The strongest variation in the observed morphological changes in the retina was recorded from Group #1, where the animals were not hydrated with additional subcutaneous PBS injections. These animals also showed severe adverse side effects to the drug toxicity such as blood in the urine. The second group of animals that were hydrated with subcutaneous PBS injections showed less inter-animal variability in the retinal morphological changes and had considerably less adverse side effects to the drug.

5.

Conclusions

We have observed and quantified in vivo for the first time, the early effect of ${\mathrm{NaIO}}_3$ on the morphology of the rat outer retina using high speed UHR-OCT. Comparison with histology and immunohistochemistry suggests that changes in the structure and optical reflectivity of the ${\mathrm{NaIO}}_3$ treated retina correspond to cell swelling, photoreceptor OS rearrangement, and fluid accumulation at the boundary between the OS and the RPE. Our quantitative statistical analysis results suggest that the photoreceptor layer thickness and reflectivity could serve as markers for noninvasive monitoring of outer retinal degeneration. Results from this study could potentially improve our understanding of the mechanisms of outer retina degeneration induced by ${\mathrm{NaIO}}_3$ and the interplay between the early changes in the photoreceptor morphology and disrupted RPE metabolic and blood-retina barrier function.