Optimizing the sodium iodate model: effects of dose, gender, and age

Abstract

Sodium iodate (NaIO₃) is a commonly used model for age-related macular degeneration (AMD), but its rapid and severe induction of retinal pigment epithelial (RPE) and photoreceptor degeneration can lead to the premature dismissal of potentially effective therapeutics. Additionally, little is known about how sex and age affect the retinal response to NaIO₃. This study aims to establish a less severe yet reproducible regimen by testing low doses of NaIO₃ while considering age- and sex-related effects, enabling a broader range of therapeutic evaluations. In this study, young (3–5 months) and old (18–24 months) male and female C57Bl/6J mice were given an intraperitoneal (IP) injection of 15, 20, or 25 mg/kg NaIO₃. Damage assessment one week post-injection included in vivo imaging, histological examination, and qRT-PCR analysis. The results revealed that young mice showed no damage at 15 mg/kg IP NaIO₃, with varying degrees of damage observed at 20 mg/kg. At 25 mg/kg, most young mice displayed widespread retinal damage, with females exhibiting less retinal thinning than males. In contrast, older mice at 20 and 25 mg/kg displayed a more patchy degeneration pattern, outer retinal undulations, and greater variability in degeneration than the young mice. The most effective model for minimizing damage while maintaining consistency utilizes young female mice injected with 25 mg/kg NaIO₃. The observed sex- and age-related differences underscore the importance of considering these variables in research, aligning with the National Institutes of Health’s guidance. While the model does not fully replicate the complexity of AMD, these findings enhance its utility as a valuable tool for testing RPE/photoreceptor protective or replacement therapies.

1. Introduction

Age-related macular degeneration (AMD) is a highly prevalent disease that affects the high-acuity vision produced by the macula (Thomas et al., 2021). The disease affects more than 12% of people above the age of 40 in the United States (Rein et al., 2022) with symptoms ranging from mild distortion of central vision to complete central vision loss. Although AMD arises from many factors, a common feature is retinal pigment epithelial (RPE) dysfunction and death (Somasundaran et al., 2020). Despite great research efforts, therapeutic options for AMD remain limited, particularly for patients with the early-stage dry AMD who are seeking to limit its progression into more advanced forms of the disease such as geographic atrophy and wet AMD. Currently these early-stage patients are treated with the antioxidant AREDS vitamins, which only limit the progression of the disease by about 25% (Age-Related Eye Disease Study Research Group, 2001). Because of this, there is a great need for novel therapeutics that halt the progression of AMD.

Sodium iodate (NaIO₃) is a small molecule that has been used to test potential AMD therapeutics. Its effects on the retina were first discovered over 80 years ago (Sorsby, 1941) and over the last few decades has become widely used because of its ability to selectively damage RPE cells and photoreceptors. It damages the retina through oxidative stress, which is a key component that drives the progression of AMD. Damage is induced through a single dose, commonly given at 40 mg/kg. This is below what has been reported to damage other organs such as the liver and kidney (50–75 mg/kg IV) (Yang et al., 2014) and well below the LD₅₀ (108 mg/kg IV or 119 mg/kg IP) (Fiume, 1995), although research focusing on NaIO₃’s effects outside of the retina is limited. It is primarily administered intravenously (IV) or intraperitoneally (IP) (Chowers et al., 2017), and some researchers have injected it intravitreally (Cho et al., 2016) or subretinally (Bhutto et al., 2018; Monés et al., 2016). It should be noted when comparing doses across studies that IV injections cause more damage than IP injections at the same concentration.

The popularity of using NaIO₃ to model AMD arose because of its ability to easily and quickly induce retinal oxidative stress, with RPE and photoreceptor death occurring within a week of injection. Although the speed of the model is practical, NaIO₃’s rapid severe onset does not align with the chronic nature of AMD. This poses a challenge, as therapies effective in slowing AMD’s gradual progression might not be able to counter the swift and severe damage caused by NaIO₃. To reduce the chance that researchers discard useful therapies while still keeping the practicality of a fast model, it is important to explore methods for decreasing NaIO₃-induced damage without introducing increased variability among mice.

Using low-dose NaIO₃ increases the potential for therapeutics to demonstrate retinal protection. Several studies in recent years have demonstrated ineffectiveness of their therapeutic at higher doses of NaIO₃ but significant protection at lower doses (Ildefonso et al., 2016; Yang et al., 2023). In this context, “low-dose” can be defined as less than 30 mg/kg for IP injections and 20 mg/kg for IV injections, thresholds surpassed in most therapeutic investigations. While some studies have provided insight into the damage induced by low-dose NaIO₃ (Machalińska et al., 2013; Wolk et al., 2020; Yang et al., 2021; Zhang et al., 2021), including studies that have looked at a range of doses (Franco et al., 2009; Wang et al., 2014; Zhou et al., 2014), concerns have arisen regarding the uniformity in the extent of damage (Chowers et al., 2017), an important factor in the context of therapeutic investigations. Consequently, further research is needed to determine the lowest dose that reliably causes consistent damage.

Sex has been a frequently overlooked aspect in NaIO₃ studies, with the majority of studies focusing on males or not reporting sex at all (Yang et al., 2021). A recent study focusing on the RPE response to low-dose NaIO₃ found that female mice had a larger area of damaged RPE than males given the same dose (Yang et al., 2021). In contrast, another study briefly mentioned that male mouse retinas thinned more than female retinas following a 50 mg/kg IP NaIO₃ injection (Schnabolk et al., 2020). Understanding how the male and female retinas differ in response to NaIO₃ is not only important to optimize protocols for inducing minimal consistent damage, but also aligns with the guidance from the National Institutes of Health (NIH) regarding the inclusion of both sexes in research (National Institutes of Health, 2015). Furthermore, further investigation may reveal how sex influences oxidative stress responses, impacting our understanding of retinal degeneration modifiers and aiding the development of AMD therapeutics.

Although age stands as a pivotal factor in AMD, it surprisingly remains an understudied variable in the NaIO₃ model. A partial insight into the effect of age in this model was offered by a study that compared 3- and 15-month-old mice in a low-dose (10 mg/kg IV) NaIO₃ study (Upadhyay et al., 2020). Though the dose was insufficient to induce degeneration, the researchers observed a reduced antioxidant response capacity in older mice, hinting that higher doses could lead to more extensive retinal degeneration in older mice compared to their younger counterparts. Another study also suggested that a diminished ability to increase the antioxidant response when NaIO3 was administered lead to greater oxidative stress in the RPE (Sachdeva et al., 2014). Because oxidative stress and age both contribute to AMD, investigating the potential influence of age within the NaIO₃ model becomes an important pursuit. The purpose of this study was threefold: (1) To provide a strong foundation for researchers interested in lower doses of NaIO₃ for therapeutic testing. Because consistency is important for these studies and typically decreases with the dose, a high number of mice were used to clearly define what effect the doses have on the mouse retina. IP injections were used because they provide an easy, consistent way of administering NaIO₃. (2) To describe the variations in NaIO₃ response with respect to sex. This was not only to establish a methodology for inducing minimal and consistent damage but also to contribute valuable data to an area with limited and contradictory information (Schnabolk et al., 2020; Yang et al., 2021). (3) To provide novel insights on the effect of age in the NaIO₃ model, with an additional examination of the interaction between sex and aging.

2. Materials and Methods

2.1. Animals

Two age groups of C57Bl/6 mice were studied. The young mice (3–5 months old) were acquired from Jackson Laboratory (C57Bl/6J; strain number 000664; Bar Harbor, ME) and the old mice (C57BL/6JN; 18–24 months old) were obtained from the National Institute of Aging (NIA; Baltimore, MD). While the mice used in this study weren’t specifically screened for the Rd8 mutation of the Crb1 gene, our laboratory has tested both C57Bl/6J and C57Bl/6JN mice obtained from Jackson Laboratory and NIA, respectively, which confirmed the absence of the Rd8 mutation in these strains. Both sexes were used. Mice were fed a standard laboratory diet ad libitum and were maintained on a 12:12 hours light/dark cycle. The average illuminance, measured from eleven locations one meter above the floor, was 1262 +/− 392 lx (ILT1400 Lux Meter; International Light Technologies, Peabody, MA). All procedures were approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania (Philadelphia, PA) and were performed in accordance with the ARRIVE guidelines and the National Research Council’s Guide for the Care and Use of Laboratory Animals.

A total of 201 mice were used. In the young mouse group there were 131 males (15 mg/kg: 10; 20 mg/kg: 86; 25 mg/kg: 35) and 26 females (25 mg/kg: 26). In the old mouse group there was 24 males (20 mg/kg: 10; 25 mg/kg: 14) and 20 females (20 mg/kg: 7; 25 mg/kg: 13).

2.2. NaIO₃ administration

NaIO₃ (Sigma-Aldrich, St. Louis, MO) was dissolved in 0.9% saline (normal saline; B. Braun, Melsungen, Germany) at a concentration of 1% (w/v) solution. The 1% solution was injected intraperitoneally at 1.5 μl/g, 2 μl/g, or 2.5 μl/g to yield a dose of 15 mg/kg, 20 mg/kg, or 25 mg/kg respectively. A 31 gauge, 8 mm length, 300 μl capacity insulin syringe (BD, Franklin Lakes, NJ) was used. The IP injection was administered on the right side of the mouse’s abdomen, situated between the two right abdominal nipples in females, or in the analogous location for males.

2.3. In vivo imaging

One week after the NaIO₃ was administered, in vivo retinal imaging of both eyes was done using a Spectralis HRA confocal scanning laser ophthalmoscope (cSLO; Heidelberg Engineering, Franklin, MA) and Bioptigen Envisu R2200 ultra-high resolution (UHR) spectral-domain optical coherence tomography system (SD-OCT or OCT; Leica Microsystems, Deerfield, IL).

To prepare mice for in vivo imaging, one drop of a 2:1 mixture of 1% tropicamide and 2.5% phenylephrine (Akorn, Lake Forest, IL) was applied to both eyes several minutes before anesthesia. Anesthesia was induced by IP injection of 93–98 mg/kg ketamine (Dechra Veterinary Products, Overland Park, KS, USA) and 10–11 mg/kg of xylazine (Akorn, Lake Forest, IL). Topical anesthesia with 1% tetracaine (Alcon Laboratories, Geneva, Switzerland) was then applied to both corneas, followed by the application of Refresh Artificial Tears (Allergan, Irvine, CA, USA) and protective eye shields to prevent corneal desiccation and media opacity development (Bell et al., 2014). The mice were then moved to the cSLO imaging platform and imaged before undergoing SD-OCT imaging.

cSLO images were captured using the blue autofluorescence (BAF) channel (486 nm raster scanned laser for excitation and 500–680 nm bandpass range for emission collection) and infrared autofluorescence (IRAF) channel (795 nm raster scanned laser for excitation and 800 nm longpass for emission collection). The images were obtained using the 102° Ultrawide Field (UWF) lens, providing a ~3.4 mm field of view (FOV) of the mouse fundus. The system focus was set on the RPE, and images were collected with the optic nerve centrally located within the image FOV, as described previously (Bell et al., 2015). Twenty-five frames were acquired in high-speed mode (768×768 pixels) from each mouse retina, and real-time co-registration and averaging were performed using the Heidelberg Eye Explorer (HEYEX 1) software’s automatic real-time (ART) processing feature. Averaged images were auto-normalized for optimal contrast.

Following cSLO imaging, SD-OCT images were obtained using a 45° FOV lens (~1.4 mm B-scan width) with the optic nerve centrally positioned. Orthogonal B-scans (1400 A-scans/2 B-scans×15 frames/B-scan) were collected at 0° and 90° to capture the axial (horizontal) and sagittal (vertical) meridians through the optic nerve. Additionally, peripheral B-scans with the optic nerve off to one side were collected in the temporal, nasal, superior, and inferior regions of both eyes.

After completing the imaging procedures, mice were administered 1.5 mg/kg of atipamezole HCL (Modern Veterinary Therapeutics, Miami, FL, USA) to facilitate anesthesia recovery. To protect the cornea during recovery, the eyes were covered with Puralube Vet Ointment (Dechra Veterinary Products, Overland Park, KS, USA).

2.4. In vivo image analysis

The 15 SD-OCT B-scan image frames from each orthogonal B-scan were co-registered and averaged in ImageJ (NIH, Bethesda, MD) using the OCT Volume Averager plugin (Krebs et al., 2016; Thevenaz et al., 1998) and a custom-made python script. To analyze the peripheral OCT images in ImageJ, the top of the inner nuclear layer (INL) and Bruch’s membrane were outlined and a custom macro was used to measure the average distance between the outlined boundaries every 50 pixels (approximately 109 μm). The top of the INL and Bruch’s membrane were chosen because their lamina remained clear despite the damage from the NaIO₃ distorting other lamina, such as the ones that outline the outer nuclear layer (ONL). The data from the OD and OS eyes were then averaged because of the high inter-eye correlation of NaIO₃-induced observable damage. The values from the temporal and nasal images were plotted together on the horizontal graph and the values from the superior and inferior images were plotted together on the vertical graph.

Damage in the cSLO images was quantified in Photoshop (Adobe, San Jose, CA). Damaged regions were selected and the pixel count of the selected area was compared against the total pixel count of the eye. BAF images were primarily used, but IRAF images were referenced to ensure all damaged regions were included.

2.5. Histology

Immediately after the in vivo imaging (one week after the NaIO₃ injection) mice were euthanized with CO₂ asphyxiation followed by cervical dislocation. The eyes were enucleated while preserving the third eyelid. They were then placed in 2% paraformaldehyde and 2% glutaraldehyde (diluted from 16% paraformaldehyde and 8% glutaraldehyde; Electron Microscopy Sciences, Hatfield, PA) in 1x PBS and stored at 4°C. Later, the eyes were dissected to remove the cornea, iris, and lens. The remaining “eye cup” was then dehydrated by placing it in two washes of 75% EtOH, two washes of 95% EtOH, and two washes of 100% EtOH for 15 minutes each. The eye cups were then put in an infiltration solution (JB-4 Plus Embedding Kit; Polysciences, Warrington, PA) overnight. The next day, the eyes were embedded using a glycol methacrylate-based plastic resin embedding medium (JB-4 Plus Embedding Kit; Polysciences). The eyes were oriented so that sections include the superior and inferior regions.

Using a rotary microtome (Leica RM2165; Leica Biosystems, Wetzlar, Germany), 3 μm sections were collected around the optic nerve. The sections were stained with toluidine blue (Sigma-Aldrich, St. Louis, MO). The toluidine blue was applied for 10 seconds and was then washed off with water. The slides were kept under running water for 2 minutes and then kept in still water for 5 minutes. Once dried, the slides were placed in a wash of 95% ethanol (Decon Labs, King of Prussia, PA) for 1 minute, 95% ethanol for 30 seconds, 100% EtOH for 1 minute, 100% EtOH for 30 seconds, xylenes (Histoprep, ThermoFisher Scientific, Waltham, MA) for 2 minutes, and xylenes for 1 minute. Permount mounting medium (ThermoFisher Scientific, Waltham, MA) was applied immediately before placing a cover slip over the sections.

Initial 20x images for quantification were taken using an Aperio ScanScope AT Turbo (Aperio, Vista, CA). Follow up 40x images were taken for publication using a Nikon Eclipse 80i microscope and DS-Fi2 camera (Nikon, Tokyo, Japan). These 40x images were then stitched together using Photoshop’s photomerge feature. The images’ levels and color were enhanced in Photoshop.

Every 250 μm the ONL thickness was quantified by counting how many nuclei deep the layer was. If that location had an undulation present, the nearest area not containing an undulation was quantified instead. Quantification stopped 2000 μm away from the optic nerve and all 16 measurements for each eye were averaged together. Undulations were also counted. Within the plastic sections, a region was considered an undulation if there was a noticeable curvature in the ONL layer that moved away from Bruch’s membrane and then curved back towards it.

2.6. qRT-PCR

Immediately after the in vivo imaging (one week after the NaIO₃ injection) mice were euthanized with CO₂ asphyxiation followed by cervical dislocation. The OD eyes were enucleated and placed in cold PBS (ThermoFisher Scientific, Waltham, MA), and the cornea, iris, lens, and neural retina were dissected.. The neural retina was flash frozen on dry ice and stored at −80°C.

The neural retinas were then homogenized and RNA was isolated using Qiagen’s RNeasy kit (Qiagen, Valencia, CA) following the manufacturer’s protocol. The RNA quantity was determined with a NanoDrop spectrophotometer and was then reverse transcribed into cDNA using TaqMan Reverse Transcription Reagents (Applied Biosystems, Foster City, CA), following the manufacturer’s instructions. qRT-PCR was performed on a 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA) using the following TaqMan probes (Applied Biosystems, Foster City, CA): Gapdh (Mm99999915_g1), Hmox1 (Mm00516005_m1), and Rho (Mm00520345_m1).

All reactions were performed in technical triplicates and the data were analyzed using the comparative CT method (Schmittgen and Livak, 2008). ΔCT values were obtained by normalizing the CT values to the respective Gapdh CT values. Subsequently, ΔΔCT values were determined by normalizing the ΔCT value of each sample to the average ΔCT value of all untreated mice for each probe. The fold change was calculated by taking 2 to the negative power of the ΔΔCT value.

2.7. Statistical analysis

Statistical analysis was carried out in Prism 9.2 and 10.0 (GraphPad Software, Boston, MA). For analyses of the area of damage in cSLO images (Fig. 1A,B, 2A, S1), the distribution was skewed because many of the images had close to 0% or 100% damage. Because the distribution was not normal, the Kruskal-Wallis test, followed by Dunn’s multiple comparisons test, was used. For data that was normally distributed (Fig. 3C, 4E,F, 5), one-way analysis of variance (ANOVA) was performed and if found to be significant, post-hoc pairwise comparisons were made using Tukey’s multiple comparisons test.

Figure 1. Damage caused by 15, 20, and 25 mg/kg NaIO₃ in young mice.

(A) Quantification of the area of damage, defined as speckled autofluorescence, induced by 15 (n=10), 20 (n=86), and 25 (n=35) mg/kg NaIO₃ in young male mice, measured by percent of 102° field of view cSLO image containing damage. Each dot represents one mouse; mean ± SEM is reported. (B) Comparison of damage seen in cSLO images of young male (n = 35) and female (n = 26) mice injected with 25 mg/kg NaIO₃. Each dot represents one mouse; mean ± SEM is reported. (C) Proportions of the different categories of damage induced by 15 (male n=10), 20 (male n=86), and 25 (male n=35; female n=26) mg/kg NaIO₃. (D) cSLO images (blue autofluorescence (BAF), top; infrared autofluorescence (IRAF), middle) and vertical OCT images (bottom; left side of each image is inferior, right side is superior) of the four categories induced by low doses of NaIO₃ and saline-injected age-matched male control. All images except for the saline control are from male mice given 20 mg/kg NaIO₃. The set of images in each column come from the same mouse. The cSLO images cover a larger area than the OCT images, and the OCT image length is shown as the red line in the top left cSLO image.

Figure 2. Damage caused by 20 and 25 mg/kg NaIO₃ in old mice.

(A) Quantification of the area of damage induced by 20 (male n=10, female n=7) and 25 (male n=14, female n=13) mg/kg NaIO₃, as measured by percent of 102° field of view cSLO image containing damage. Each dot represents one mouse; mean ± SEM is reported. (B) Proportions of the different categories of damage induced by 20 and 25 mg/kg NaIO₃ (same n as A). (C) cSLO images (BAF, top; IRAF, middle) and vertical OCT images (bottom) of the four categories induced by low doses of NaIO₃: none, sparse, patchy, and widespread. The representative “none” and “sparse” images are from male mice at 25 mg/kg NaIO₃, the “patchy” and “widespread” images are from female mice at 25 mg/kg NaIO₃, and the “saline control” images are from a saline-injected 24-month male mouse. The set of images in each column come from the same mouse. cSLO images cover a larger area than OCT images, with the OCT image length indicated by the red line in the top-left cSLO image.

Figure 3. OCT analysis of both sexes of young and old mice injected with 25 mg/kg NaIO₃.

(A) Representative horizontal and vertical OCT images (OCT-H and OCT-V respectively) of saline-injected 24-month-old male mouse and young male, young female, old male, and old female mice injected with 25 mg/kg NaIO₃, taken one week after injection. Each pair of images (horizontal and vertical) was captured from the same eye. White arrow shows an example of an undulation. (B) cSLO image showing the range of the OCT images. The white circle shows how far the central images in A extend. Peripheral images (Fig. S3) extended this range for measurements in C. Blue line: horizontal plane; red line: vertical plane. (C) Quantification of OCT images, measuring from the top of the INL to Bruch’s membrane, as shown in the brackets on the rightmost images in A. Measurements were taken every 50 pixels (approximately 109 μm). Mean ± SD is reported with the following distribution: young untreated males (n = 5), young male (n = 16), young female (n = 20), old untreated male (n = 5), old male (n = 14), and old female (n = 13). All locations were averaged together to get one data point per mouse for statistical analysis: Pairwise comparisons showed young control males, young males, and young females to be significantly different. Old male and female mice lacked significant distinction, but old females differed from untreated controls. Because the retinal thickness of untreated females is statistically equivalent to that of untreated males, only the data from untreated males is presented.

Figure 4. Retinal plastic sections of mice injected with 25 mg/kg NaIO₃.

(A–D) Representative images of plastic sections of (A) young male, (B) young female, (C) old male, and (D) old female mice injected with 25 mg/kg NaIO₃; mice were euthanized one week after injection. Scale bars represent 100 μm. (E) Quantification of the thickness of the ONL, measured by nuclei count. Nuclei were counted at 250 μm intervals, and the results were averaged. Each dot represents one mouse. Mean ± SEM is reported. (F) Quantification of the number of undulations present in each plastic section image. Each dot represents one image from different mice. (G) Undulations seen in IRAF cSLO image of an old female mouse injected with 25 mg/kg NaIO₃, imaged one week after injection. White box shows a magnified region with many curvilinear black lines representing undulations.

Figure 5. qRT-PCR analysis of retinal tissue.

(A-B) Fold change analysis of mRNA levels in the neural retina of (A) rhodopsin (Rho) and (B) heme oxygenase 1 (Hmox1). The letters above each group signify significant differences. Groups sharing the same letter indicate statistical similarity. Mean ± SEM is reported and each dot represents one mouse. Only one eye from each mouse was studied by qPCR.

The categorical data in Fig. 1C and 2C were not subjected to statistical analysis because of the failure to meet the assumption of the Chi-square test, which requires that the expected frequency of data be greater than 5 for at least 80% of the groups.

Mean ± standard error of the mean (SEM) was used for all graphs except for Fig. 3C, where mean ± standard deviation (SD) was used because the SEM was too small to be visualized. Significant values were marked as * (P < 0.05), ** (P < 0.01), *** (P < 0.001), and **** (P < 0.0001).

No mice were excluded from the analyses. Mice were identified by their ID number only and during analysis researchers were masked with respect to the treatment versus control group.

3. Results

3.1. Dose dependence of NaIO₃-induced damage in young mice.

Young male mice (ages 3–5 months old) received an IP injection of 15, 20, or 25 mg/kg NaIO₃. One week later they were imaged with cSLO and OCT imaging systems. The extent of damage, defined as speckled autofluorescence, was quantified using both the cSLO’s blue autofluorescence (BAF) and infrared autofluorescence (IRAF) images. This damaged area was expressed as a percentage of the total retinal area captured by the 102° field of view (FOV) lens (Fig. 1A). The 15 mg/kg dose resulted in no damage. The 20 mg/kg dose showed variable damage, with about a third having no damage and a fourth having greater than 95% damage. The 25 mg/kg dose showed higher consistency with the majority of mice (83%) having greater than 95% damage.

Once the 25 mg/kg dose in male mice showed promise as a damage-inducing model, female mice were studied with that dose (Fig. 1B). No significant difference between the sexes was seen in the area damaged by the NaIO₃. A preliminary study of females given 20 mg/kg NaIO₃ showed variable damage similar to males at 20 mg/kg, but because the goal was to establish a model of consistent damage, further analysis of females at 20 mg/kg were halted (data not shown).

Four damage patterns were observed (Fig. 1C,D): ‘minimal’ (no damage or minimal enough that it wasn’t detected in the OCT images), ‘superior only’ (crescent shaped above optic disc), ‘moderate’ (damage circling optic disc, but not completely extending past the 102° FOV cSLO images), and ‘widespread’ (entire 102° FOV). Moderately and widely damaged retinas often had small undamaged areas directly around the optic disc.

Both eyes in each mouse were similarly affected. 91% of mice from all three doses had a difference of less than 5% damaged area between their right and left eyes (Fig. S1C). Mice with moderate or superior-only damage had greater variation between their eyes (average difference: 4.5%) than those with no or widespread damage (average difference: 0.3%).

3.2. Old male and female mice differ in extent of damage when given 20 and 25 mg/kg NaIO₃.

Old mice (18–24 months) received IP injections of 20 or 25 mg/kg NaIO₃, followed by cSLO and OCT imaging after one week. The extent of damage showed high variability, with female mice at the 25 mg/kg dose having larger damaged areas than males at either dose (Fig. 2A).

Interestingly, the damage patterns observed in old mice differed from those in young mice. These could be categorized as ‘none,’ ‘sparse’ (damage present but not necessarily in the superior region, see Fig. S2B for more examples), ‘patchy’ (patches of undamaged areas surrounded by prominent damaged regions, see Fig. S2C for more examples), and ‘widespread’ (Fig. 2B,C). Although both the young and old mice had widespread damage, this phenotype was different in the old mice: the IRAF in old mice appeared less speckled and the BAF of the young mice appeared more diffuse. Some undamaged eyes had small autofluorescent puncta in the BAF images (Fig. S2A), likely due to activated microglia or macrophages which are often associated with age and unrelated to NaIO₃ (Ferdous et al., 2021).

A significant proportion of male mice had no damage (70% at 20 mg/kg, 43% at 25 mg/kg) (Fig. 2A,B). The patchy damage, not seen in young mice, appeared across both sexes and doses. Females receiving the 20 mg/kg dose were most likely to develop this patchy phenotype. Females at 25 mg/kg had a greater chance of developing widespread damage compared to the other old groups. However, young mice given 25 mg/kg were more likely to develop the widespread phenotype than the old female mice (Fig. 2A,B and Fig. 1A,C).

Minimal variation was seen in the ‘none’ and ‘widespread’ groups between eyes (average difference: 0.2%), while ‘sparse’ and ‘patchy’ groups had higher variability (average difference: 9.4%; Fig. S1D). Age variation within the 18–24 month group did not impact the response to NaIO₃, so they were treated as a unified group in this study.

3.3. At the 25 mg/kg dose, young male mice have the thinnest retinas and young mice display the most consistent ONL thinning.

One week after injecting 25 mg/kg NaIO₃, OCT images centered on the optic nerve were captured (Fig. 3A). To extend the range of the OCT system’s 45° FOV, temporal, nasal, superior, and inferior images were also taken (Fig. 3B, white circle shows optic-nerve-centered image coverage, blue and red lines extending outside the white circle indicate the peripheral image range; see Fig. S3 for example peripheral images, which have the optic disc on the edge of each image). In order to analyze more of the retina, these peripheral images were used to measure retinal thickness (Fig. 3C). Because the ONL boundaries were obscured in the NaIO₃-affected OCT images, measurements were taken from the top of the INL to Bruch’s membrane (see brackets on the right side of Fig. 3A).

Young mice of both sexes had very consistent retinal thinning (Fig. 3C, left column, SD reported). The nasal and inferior regions showed more variability due to 8% of mice experiencing superior-only damage, which also partially affected the temporal region. This increased variability in the inferior and nasal regions was due to the lack of damage in these regions in 8% of mice, making them different from the majority of mice. Young female mice had significantly less damage, with an average retinal thickness of 126 μm, compared to 95 μm for young males (150 μm in untreated young males). The most pronounced difference between young male and female mice was observed in the ONL, which was much thinner in males (Fig. 3A).

In contrast to the young mice, the OCT images of the old mice showed less retinal thinning (an average thickness of 139 μm for old males, 134 μm for old females, and 149 μm for untreated old males; Fig. 3C, right column, SD reported). Old mice often had ONL undulations, visible as fluorescent ‘peaks’ in the OCT images (Fig. 3A, see white arrow in old female column as an example). These undulations were more common in old females but were also observed in some old males.

3.4. Plastic sections of retinas damaged by 25 mg/kg NaIO₃ show young males have the thinnest retinas and all other groups, especially old females, have retinal undulations.

To understand how the ONL responds to 25 mg/kg NaIO₃, plastic sections were studied. Three eyes from each of the four groups (young male, young female, old male, and old female) were chosen based on their cSLO images, selecting those that best represented the overall characteristics of the group (Fig. 4A–D). To focus on the response to NaIO₃, mice with widespread damage were used unless not enough were available (one sparse-damaged eye was used in the old male group and one patchy-damaged eye was used in the old female group).

ONL thickness was quantified by counting nuclei per vertical row every 250 μm and averaging the counts per mouse (Fig. 4E). On average, young female, old male, and old female groups were 7 nuclei thick (compared to a typical 10–11 nuclei in undamaged mouse retinas (Knott et al., 2011)). Young male mice were significantly different, with an average thickness of 4 nuclei.

Undulations of the ONL were observed in all groups except young males (Fig. 4F). Old females had the most undulations, significantly differing from young males. The undulations were prominent in histology sections (Fig. 4C,D, S4), visible in OCT images (Fig. 3A), and even seen in IRAF cSLO images as dark curvilinear patterns (Fig. 4G). Additionally, old males, old females, and to a lesser extent, young females exhibited multilayering due to RPE migration (Fig. S5), a feature also observed in human AMD patients (Zanzottera et al., 2016). This multilayering appeared as large clumps of pigmented cells.

3.5. mRNA analysis shows decreased retinal health and increased oxidative stress in response to 25 mg/kg NaIO₃.

One week after injecting 25 mg/kg NaIO₃, mRNA levels in the neural retina were assessed (Fig. 5). Rhodopsin (Rho) mRNA, which encodes for a photoreceptor-specific protein involved in light sensing (Nathans, 1992), was used to measure photoreceptor health (Fig. 5A). All groups, except old males, showed significantly lower Rho levels compared to untreated young males.

The decrease in Rho in young males was likely due to photoreceptor death, as suggested by histological data (Fig. 4A,E). Other groups experienced a similar Rho decrease despite having more photoreceptor nuclei (Fig. 4B–E), suggesting that their decrease in Rho came from rod stress rather than rod death. Mice with normal Rho mRNA levels had minimal damage in in vivo images.

Heme oxygenase 1 (Hmox1), a gene that encodes for an antioxidant whose transcription is driven by Nrf2 activation, was used to measure oxidative stress (Fig. 5B) (Keyse et al., 1990). Young males and old females had the highest Hmox1 levels, with young males significantly higher than young females and old males. Similar to the Rho mRNA analysis, mice with Hmox1 levels similar to the control group displayed little to no damage in cSLO and OCT images.

4. Discussion

NaIO₃ is an oxidant useful to vision scientists because of its retina-specific damaging effects, rapid induction of damage, and convenient administration. However, the progressive nature of AMD, characterized by its gradual onset, contrasts with the rapid damage induction caused by NaIO₃. Consequently, therapies that exhibit potential for protecting the retinas of AMD patients may not demonstrate effectiveness in the NaIO₃ model, inadvertently leading researchers to prematurely discontinue their investigations based on misleading outcomes. It is thus important to reduce the damage caused by the NaIO₃ as much as possible while ensuring that the damage caused by the oxidant remains largely consistent.

At a dose of 25 mg/kg, administered via IP injection, NaIO₃ induces consistent damage in young (3–5 month old) male and female mice. At this dose, 90% of mice experienced widespread or moderate areas of damage throughout the entire retina, which creates a solid platform for testing therapies. Although cSLO images showed that both sexes had a similar proportion of mice developing the widespread area of damage, OCT images, plastic sections, and qRT-PCR analysis demonstrated that male retinas were consistently more thinned than female retinas. While the present study did not include functional assays, the observed thinning of the outer nuclear layer (ONL) and decreased levels of Rho mRNA suggest a potential loss of vision. NaIO₃-induced degeneration similar to that demonstrated in our study has previously correlated with diminished optokinetic tracking (OKT) and electroretinogram (ERG) responses (Chowers et al., 2017).

Estrogen levels may play a role in the difference between the sexes. Estrogens have antioxidant properties (Bhavnani et al., 2001) and it has been shown that women given postmenopausal hormone therapy have a decreased risk of late-stage neovascular AMD (Feskanich et al., 2008). Furthermore, tamoxifen, a selective estrogen receptor modulator, has been demonstrated to protect the retina from oxidative stress caused by light damage (Wang et al., 2017). Because female mice experience less retinal thinning than male mice, female mice injected with 25 mg/kg NaIO₃ provide a greater chance for therapies to demonstrate their protective effects. A next step in this research would involve injecting young female and male mice with 25 mg/kg NaIO₃, while also administering a therapeutic agent, and assessing the consistency of response to the therapeutic in each sex.

These findings build upon prior research, which noted that young male retinas experienced greater thinning than female retinas (Schnabolk et al., 2020). In contrast, another study reported a greater area of damage in RPE flat mounts in females than males (Yang et al., 2021). We did not see any significant difference in area of damage on cSLO images between the sexes of young mice, but that could have been because 25 mg/kg IP was high enough to overwhelm any potential differences or because RPE flat mounts are a more sensitive approach for assessing RPE damage. Yang et al. focused on measuring RPE health while the present study primarily assessed photoreceptor health. In light of these observations, it can be concluded that in the NaIO3 model, RPE cells are damaged more in females and photoreceptors are damaged more in males. This distinction could arise from the different modes of cell death in the NaIO3 model: RPE cells predominately undergo necroptosis and photoreceptors mainly undergo apoptosis (Hanus et al., 2016). Estrogen, which has protective effects against apoptosis (Mo et al., 2013), may protect photoreceptors while having limited protective effects on RPE cells.

Old mice responded differently to NaIO₃ than young mice. Although the results showed a pronounced impact of NaIO₃ on the old mice, the old mice were less likely to be affected by the NaIO₃ when compared to the young mice. Furthermore, a contrasting trend emerged between the sexes, with the old male mice showing lower susceptibility to the NaIO₃’s effects than the old female mice. The old male mice were more likely to have no damage at all from either 20 mg/kg or 25 mg/kg NaIO₃ and old female mice injected with 25 mg/kg NaIO₃ had significantly greater area of damage than the old male mice. Where cSLO-defined damage did occur, both male and female old mice exhibited ONL thinning to a similar extent as observed in the young female mice. The damaged retinas of the old mice were more likely than retinas of the young mice to exhibit the striking phenotypes of ONL undulations and multilayering resulting from RPE migration.

As its name suggests, a large contributing factor of AMD is age itself. Regardless of the presence of disease, the retina has many structural and molecular changes as it ages. Some of these aging effects include decreased number of synapses, remodeled dendrite architecture, altered gene expression, damaged DNA, and heightened oxidative stress (Jarrett and Boulton, 2012; Samuel et al., 2011). Considering the multitude of known and unknown changes associated with aging, we hoped that inducing oxidative stress using NaIO₃ in older mice could better model AMD. However, at the doses studied (20 and 25 mg/kg), the older mice displayed considerable variability in the extent of damage induced. Young mice provide a reproducible platform for drug testing but the degeneration in old mice is more patchy and variable, making them less suitable for drug testing. However, given that age and oxidative stress play roles in AMD, it would be interesting in the future to investigate the reasons for age-related differences in the NaIO₃ damage model.

NaIO₃ has several routes of administration. IP and IV injections are most common, but intravitreal and subretinal injections have also been used. The impact of each administration route differs; for instance, IV injections tend to induce more severe damage compared to IP injections at equivalent doses. In this study, only IP injections were used. The observations regarding sex and age differences may vary based on the administration route employed.

The results of this study not only demonstrates the use of 25 mg/kg IP NaIO₃ as a model for testing therapeutics and the differences of response by young and old mice, but also clearly shows that there is a strong difference between the sexes in both young and old mice. In 2015, the NIH released a statement mandating that scientists include both sexes in any research funded by the NIH (National Institutes of Health, 2015). This study confirms the validity of the NIH’s mandate, particularly in the context of NaIO₃. Further investigation of the mechanisms of retinal degeneration should include both sexes. Young female mice injected with 25 mg/kg NaIO₃ do provide a low-dose oxidative stress model to test novel therapeutics, but it is important to follow up any pilot study with experiments that include both males and females. Additionally, while both sexes exhibit consistent damage at 25 mg/kg, there might be an unexpected variability in their responses to therapies. Factors such as the estrous cycle in females could potentially introduce inconsistencies in their responses. Thus, when conducting therapeutic studies, further consideration should be given as to which sex is more suitable for the study.

As with any model of a multifactorial disease such as AMD, NaIO₃ toxicity does not directly represent AMD’s complete underlying molecular mechanism. Because of that, any findings related to therapeutic interventions tested in this model have questionable direct translational relevance to AMD. However, it is clear that oxidative stress plays a key role in the development of AMD and as such it is important to provide a reliable, rapid, and straightforward method for modeling oxidative stress in the retina. By reducing the IP dose to 25 mg/kg and using young female mice for the initial screening, researchers can effectively generate a consistent, moderate amount of damage. This approach creates a potent tool for determining which potential therapies should proceed to further testing across a diverse range of available AMD models.

Highlights.

• 25 mg/kg intraperitoneal NaIO₃ is the lowest dose that induces consistent damage

• Young male mice have more retinal thinning than females given 25 mg/kg IP NaIO₃

• Old mice show different damage pattern, with males having less damage than females

Abbreviations:

AMD

age-related macular degeneration

ART

automatic real-time

BAF

blue autofluorescence

cSLO

confocal scanning laser ophthalmoscope

ERG

electroretinogram

FOV

field of view

Hmox1

heme oxygenase 1

INL

inner nuclear layer

IP

intraperitoneal

IRAF

infrared autofluorescence

IV

intravenous

NaIO3

sodium iodate

NIA

National Institute of Aging

NIH

National Institutes of Health

OCT

optical coherence tomography

OKT

optokinetic tracking

ONL

outer nuclear layer

ANOVA

analysis of variance

Rho

rhodopsin

ROS

reactive oxygen species

RPE

retinal pigment epithelium

SD

standard deviation

SEM

standard error of the mean

SD-OCT

spectral-domain optical coherence tomography

UHR

ultra-high resolution

UWF

ultrawide field