IN-VIVO LONGITUDINAL CHANGES IN THICKNESS OF THE CANINE POSTNATAL RETINA

Abstract

The objectives of the present work were to assess by spectral domain optical coherence tomography (OCT) the changes in thickness of the outer nuclear layer (ONL), the ONL+ photoreceptor inner segment (IS), and the retinal thickness, as a function of age in the normal canine retina. OCT retinal scans extending from the edge of the optic nerve head (ONH) along the superior and inferior meridians were captured in both eyes of 17 normal dogs at age ranging from 4 to 119 weeks. The different parameters along the superior and the inferior regions were determined following manual segmentation using the Heidelberg Eye Explorer software. Changes in thickness with age were modeled using one-phase exponential decay models. In vivo OCT imaging results showed no interocular statistically significant differences in ONL, ONL + IS, and retinal thickness at any age. All three parameters were however found to be statistically significantly thicker in the superior vs inferior retina. A rapid thinning of the three layers occurs in both the superior and inferior retina between 4 and 12 weeks of age, before reaching a plateau at around 20 weeks of age. In conclusion, the ONL, ONL + IS, and retinal thickness of the normal canine retina decrease significantly during the first three postnatal months, and is likely attributed to an overall increase in the eye volume and tangential dispersion of the photoreceptor since early photoreceptor developmental cell death is very limited at that age. Establishment of the natural history of ONL, ONL + IS, and retinal thinning will allow a more accurate assessment of the progression of a retinal degenerative condition as well as facilitate the detection of positive rescue effect of novel retinal therapies evaluated in this large animal model.

1. INTRODUCTION

At birth, the canine retina is not fully developed (Acland and Aguirre, 1987 Aguirre et al., 1972; Gum et al., 1983) and continues to develop postnatally achieving structural maturity by 10 weeks of age (Acland and Aguirre, 1987). Histological studies have shown that at birth the retina is composed of an inner and outer neuroblastic layer, the latter contains the cell bodies of photoreceptor precursor cells (Ávila-García et al., 2012; Gum et al., 1983). During the first postnatal week, the outer plexiform layer (OPL) forms and separates the outer neuroblastic layer into inner and outer nuclear layers (ONL) (Aguirre et al., 1972). Photoreceptors subsequently undergo a centrifugal wave of maturation that extends from the edge of the optic nerve head (ONH) to the ora serrata such that by 6 (centrally) and 10 (peripherally) weeks of age photoreceptors are fully developed with adult-like inner and outer segments. Functionally, cone-mediated ERG responses are detectable before that of rods, and by 8 weeks of age ERG waveforms resemble that of adults (Acland and Aguirre, 1987; Gum et al., 1983).

Concomitantly to this postnatal development and maturation of photoreceptors, histological studies have shown a thinning of the canine ONL (Gum et al., 1983) that is considered to result from both ongoing postnatal developmental cell death (Beltran et al., 2006) and stretching due to eye growth (Tuntivanich et al., 2007).

Optical coherence tomography (OCT) has been a major advancement in clinical and experimental ophthalmology for over two decades, revolutionizing the ability to diagnose retinal diseases, monitor their progression, and response to therapeutic intervention. This non-invasive in vivo imaging technique allows high-resolution cross sectional imaging of the back of the eye and is used extensively to investigate structural changes in the outer retina, retinal nerve fiber layer, ONH, and choroid (Adhi et al., 2013). It is also used extensively in preclinical research as it allows longitudinal, noninvasive, and contactless evaluation of the retina with an axial resolution of novel OCT devices that now approaches cellular-level resolution.

A number of inherited retinal degenerations in dogs are known to cause photoreceptor loss and subsequently progressive thinning of the ONL (Miyadera et al., 2012) and are used as models to examine mechanisms of disease and test novel therapeutic approaches. In those diseases for which the onset of cell death occurs in the early postnatal period (Beltran et al., 2006; Downs et al., 2016; Santos-Anderson et al., 1980; Tuntivanich et al., 2009), it is currently unclear how much of the loss in ONL thickness at any given age is the result of the natural course of ONL thinning versus that caused by disease of photoreceptors and their nuclei loss. To address this question, establishing baseline values of ONL, ONL + inner segment (IS) and retinal thickness in early postnatal, juvenile, and adult retinas of wild-type dogs is necessary. This will provide unequivocal ability to estimate the extent of photoreceptor loss truly caused by disease and enable accurate assessment of ONL rescue following therapeutic intervention. In this study, we describe the natural course of retinal layer thinning that occurs postnatally in the normal canine eye using spectral domain OCT imaging, and show that a rapid decline occurs during the first 3 months of life. In addition, we established mathematical models for prediction of central ONL, ONL + IS, and retinal thickness in normal dogs of any given age.

2. MATERIAL AND METHODS

2.1. Animals

Seventeen normal mesocephalic mongrel or beagle dogs (10 males and 7 females) ranging in age from 4 to 119 weeks, were enrolled in this study (Table 1). The dogs were bred and maintained at the University of Pennsylvania’s Retinal Disease Studies Facility (RDSF). Studies were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health, in compliance with the USDA’s Animal Welfare Act, Animal Welfare Regulations, and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. All protocols were approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania.

Table 1.

Animals (ID, status and gender) enrolled in the study: age (in weeks) at the time of in-vivo OCT scanning sessions

Dog ID	Breed	Status	Sex	OCT (weeks)
CFFCKM	Beagle	WT	F	34
CFICTB	Beagle	WT	M	25
CFICXX	Beagle	WT	M	24
CGBCAN	Beagle	WT	M	30, 36
CGBCDI	Beagle	WT	M	29, 36
CGBCGS	Beagle	WT	M	29, 35
N292	Mongrel	WT	M	4, 6, 8, 10, 11, 12, 14, 17, 19, 24, 36, 40, 44, 48, 52, 119
N293	Mongrel	WT	F	4, 6, 8, 10, 11, 12, 14, 17, 19, 24, 36, 40, 44, 48, 52, 119
N294	Mongrel	WT	F	4, 6, 8, 10, 11, 12, 14, 17, 19, 24, 36, 40, 44, 48, 52, 119
N300	Mongrel	WT	M	21
N301	Mongrel	WT	M	21
N302	Mongrel	WT	F	22
N308	Mongrel	WT	F	20
N321	Mongrel	WT	M	13
N325	Mongrel	WT	F	13
N334	Mongrel	WT	M	32
N335	Mongrel	WT	F	32

2.2. Non-invasive retinal imaging by optical coherence tomography

Pupils were dilated using atropine sulfate 1% and tropicamide 1% (both from Akorn Inc., Lake Forest, IL) and phenylephrine 10% (Paragon Biotech, Portland, OR) 3 times, 30 minutes apart in both eyes before the procedure. Dogs over 12 weeks of age were pre-medicated with atropine sulfate (0.02 mg/kg, Med-Pharmex Incorporated, Pomona, CA) and acepromazine (0.1–0.5 mg/kg, Phoenix, Clipper Distributing Company, St-Joseph, MO) subcutaneously 30 minutes before anesthesia. General anesthesia was induced using intravenous (IV) injection of propofol (2–6mg/kg, Zoetis, Kalamazoo, MI) and maintained with gas inhalation (isoflurane 2–3%, Akorn Inc., Lake Forest, IL). Dogs under 12 weeks of age were not pre-medicated but were either masked before intubation or induced with IV propofol and immediately followed by maintenance with gas inhalation (isoflurane 2–3%, Akorn Inc., Lake Forest, IL). The animals were positioned in sternal recumbency, an eyelid speculum was placed to keep the eye to be imaged open and exposed, two stay sutures (Vicryl 4–0, Ethicon Inc., Somerville, NJ) were placed at the limbus at the 2 and 10 o’clock positions, in order to enable rotation of the globe and obtain the desired position for the procedure. The cornea was kept moist by frequent irrigation with isotonic saline solution. Both eyes of each animal were imaged and used for ONL thickness analysis at each time point.

Non-invasive retinal imaging using a combined confocal scanning laser ophthalmoscope (cSLO) and spectral-domain optical coherence tomography (SD-OCT) (Spectralis™ HRA/OCT, Heidelberg Engineering, Heidelberg, Germany) was performed either at a single time point in 11 dogs, or repeatedly (2 or more) in 6 dogs (Table 1, Supplementary Table 1). Near-infrared, blue reflectance and blue autoflorescence (with and without normalization) en face images from the fundus, using a 55° lens were acquired at each imaging session to confirm the normal retinal integrity. A 30° lens was then used to capture high resolution (20 Automatic Real-Time Tracking (ART)) single cross-sectional OCT scans in the central retina, extending 30° vertically from the ONH along the superior and inferior meridians (Fig. 1A). Using the device’s automatic eye tracking system, all images acquired during the first examination were used as reference images for those dogs having follow-up examinations. To adjust to the growth of the eyeball, the single scans were readjusted at different time points in order to maintain approximately the same location and to allow longitudinal evaluation of the changes overtime. Once acquired, they were subsequently used to measure the ONL, ONL + IS, and retinal thickness. The inner border of the ONL was defined as the transition zone between the ONL and the outer plexiform layer (OPL); and the external limiting membrane (ELM) defined the outer border of the ONL. The thickness of the ONL and the IS was measured from the same inner boundary to the outer limitation of the ellipsoid zone (EZ). The retinal thickness was measured from the inner limiting membrane (ILM) to the outer limitation of the EZ (Fig. 1 B).

Figure 1. Illustration of scans orientation and retinal layer segmentation.

(A) cSLO en face images showing in the central (extending up to 30° from the ONH) superior/tapetal and the inferior/non-tapetal retina the location (white arrows) of a 30° cross-sectional OCT B-scan. (B) Retinal thickness was measured from the ILM to the outer limit of the EZ; the ONL + IS from the OPL-ONL transition zone to the outer limit of the EZ; and of the ONL only from the OPL-ONL transition zone to the ELM.

Segmentation of the 3 layers described above along the central 30° cross-sectional OCT scans was done manually using the Heidelberg Eye Explorer program and the thickness were measured every 1.76° starting at the edge of the ONH. Measurements, in μm, were automatically provided by the Heidelberg Eye Explorer program and recorded except at sites located below a retinal blood vessel. A total of 10 to 16 measurements were acquired per OCT cross-sectional scan.

2.4. Statistical analysis

We assessed the symmetry of thickness measures (ONL, ONL + IS, and retinal thickness) between two eyes, and symmetry between inferior and superior regions at each time point using generalized linear models that account for inter-eye correlation; and correlation from repeated measures over time using generalized estimating equations (Zeger, SL Liang, KY, 1986). This analysis can accommodate the situation that some dogs had thickness measured at one time point only, while other dogs had thickness measured at more than one time point. We used the following one-phase exponential decay model (Leike, 2002) to describe the change of thickness over time.

$$\text{Y}\text{=}\left[\left({\text{Y}}_0-{\text{Y}}_{\infty }\right)\text{×}{\text{e}}^{\left(-\text{K}\text{×}\text{age}\right)}\right]+{\text{Y}}_{\infty }$$

Where Y₀ is the thickness (ONL, ONL + IS, or retinal thickness) at birth (age = 0 weeks), Y_∞ is the thickness value when age approaches infinity, K is the decay rate, and age is expressed in weeks. The decay rate is the decline, in log unit, of the thickness as a function of age. Its p value allows to assess whether the decay rate is significantly different from 0 or not.

We fitted one-phase decay model for superior thickness, inferior thickness, and the mean superior and inferior thickness for male and female dogs separately and combined using the non-linear mixed model (PROC NLMIXED) in SAS v9.4 (SAS Institute Inc., Cary, NC). The decay rates of thickness between male and female dogs were compared using two-sample t-test. Two-sided p <0.05 was considered to be statistically significant.

3. RESULTS

A total of seventeen dogs were enrolled in the study, and each dog had both eyes imaged. Eleven out of the seventeen dogs were imaged once, three were imaged twice and the other three were imaged sixteen times, thus totalizing 260 OCT B-scans for measuring ONL, ONL + IS, and retinal thickness (Table 1, Supplementary Table 1).

3.1. Outer nuclear layer thickness

When all-time points were combined, there was no significant difference between the mean (± SD) ONL thickness of the central (extending up to 30° from the ONH) superior retina in the right eye (59.5 ± 4.6 μm) and that of the left eye (58.1 ± 5 μm) (p=0.1); nor between the central inferior retina of the right eye (53.1 ± 8.1 μm) and that of the left eye (51.9 ± 7.6 μm) (p=0.14). As a result, measurements acquired from the right and the left eye of each dog were subsequently averaged, and a mean value reported. There was a statistically significant difference (p = 0.009) between the mean (± SD) ONL thickness of the superior retina (58.9 ± 4.6 μm) and the mean ONL thickness of the inferior retina (52.4 ± 7.7 μm).

For each age, the mean ONL thickness per location along the central superior and inferior retina was calculated and represented as a spider graph (Fig. 2 A–C). The mean ONL thickness decreased rapidly between 4 and 12 weeks of age in both the superior and inferior retina.

Figure 2. Spider graphs along the central superior and inferior meridians measured by manual segmentation from cross-sectional OCT B-scans and one-phase exponential decay model fitting of the ONL thickness.

ONL thickness from 4 to 12 weeks of age (A), 13 to 36 weeks of age (B) and 40 to 119 weeks of age (C). For comparison, the ONL thickness at the 4 and 119 week time-points are represented on each of the three spider graphs. (D) Superior (dashed blue line) and inferior (black line) retinal ONL thickness with an exponential decay rate (SE) of −0.21 (0.02) Y_∞ of 57.1 μm and −0.16 (0.01) Y_∞ of 46.8 μm respectively, and (E) mean (superior + inferior) central ONL thickness with an exponential decay rate (SE) of −0.18 (0.01) and Y_∞ of 52.2 μm.

The decline in ONL thickness over time was best fitted by the one-phase exponential decay model. The fitted equations for the mean ONL thickness (μm) of the central superior and inferior retina; and that of the mean of the inferior and superior combined, as a function of age (in weeks), are detailed in Fig. 2 D–E, Table 2. There was no significant difference between the ONL thickness decline in males and in females (p = 0.26), the fitted equations are illustrated in Supplementary Fig. 1.

Table 2.

One-phase exponential decay model equations summary

Measure	Localization	Exponential Equations	Exponential Decay rate§ (SE)	P value*
ONL	Superior	(88.9 – 57.1) × e(−0.21 × Age) + 57.1	−0.21 (0.02)	<0.001
	Inferior	(97.4 – 46.8) × e(−0.16 × Age) + 46.8	−0.16 (0.01)	<0.001
	Mean (Sup + Inf)	(93.1 – 52.2) × e(−0.18 × Age) + 52.2	−0.18 (0.01)	<0.001
ONL + IS	Superior	(127.0 – 81.3) × e(−0.20 × Age) + 81.3	−0.20 (0.02)	<0.001
	Inferior	(130.9 – 70.6) × e(−0.18 × Age) + 70.6	−0.18 (0.01)	<0.001
	Mean (Sup + Inf)	(131.1 – 76.2) × e(−0.20 × Age) + 76.2	−0.20 (0.01)	<0.001
Total retinal thickness	Superior	(313.2 – 193.6) × e(−0.12 × Age) + 193.6	−0.12 (0.01)	<0.001
	Inferior	(300.1 – 146.9) × e(−0.18 × Age) + 146.9	−0.18 (0.01)	<0.001
	Mean (Sup + Inf)	(302.7–170.6) × e(−0.14 × Age) +170.6	−0.14 (0.01)	<0.001

^§The decay rate is the decline, in log unit, of the thickness as a function of age.

^*Its p value allows to assess whether the decay rate is significantly different from 0 or not.

3.2. Outer nuclear layer + IS thickness

When all-time points were combined, there was no significant difference between the mean (± SD) ONL + IS thickness of the central superior retina in the right eye (85.2 ± 7.1 μm) and that of the left eye (83.9 ± 7.5 μm) (p=0.2); nor between the central inferior retina of the right eye (76.7 ± 8.9 μm) and that of the left eye (75.4 ± 8.1 μm) (p=0.13). As a result, measurements acquired from the right and the left eye of each dog were subsequently averaged, and a mean value reported. There was a statistically significant difference (p = 0.02) between the mean (± SD) ONL + IS thickness of the superior (84.6 ± 6.9 μm) vs the inferior retina (75.8 ± 8.4 μm). For each time point, the mean ONL+IS thickness per location along the central superior and inferior retina was calculated and represented as a spider graph (Fig. 3 A–C). The mean ONL + IS thickness decreased rapidly between 4 and 12 weeks of age in both the superior and inferior retina.

Figure 3. Spider graphs along the central superior and inferior meridians measured by manual segmentation from cross-sectional OCT B-scans and one-phase exponential decay model fitting of the ONL + inner segment thickness.

ONL + IS thickness from 4 to 12 weeks of age (A), 13 to 36 weeks of age (B) and 40 to 119 weeks of age (C). For comparison, the ONL + IS thickness at the 4 and 119 week time-points are represented on each of the three spider graphs. (D) Superior (dashed blue line) and inferior (black line) retinal ONL thickness with an exponential decay rate (SE) of −0.20 (0.02) Y_∞ of 81.3 μm and −0.18 (0.01) Y_∞ of 70.6 μm respectively, and (E) mean (superior + inferior) central ONL + IS thickness with an exponential decay rate (SE) of −0.20 (0.01) and Y_∞ of 76.2 μm.

The decline in ONL + IS thickness over time was best fitted by the one-phase exponential decay model. The fitted equations for the mean ONL + IS thickness (μm) of the central superior and inferior retina; and that of the mean of the inferior and superior combined, as a function of age (in weeks), are detailed in Fig. 3 D–E and Table 2.

3.3. Retinal thickness

When all-time points were combined, there was no significant difference between the mean (± SD) retinal thickness of the central superior retina in the right eye (210.4 ± 22.8 μm) vs the left eye (208.5 ± 27.9 μm) (p=0.42); nor between the central inferior retina of the right eye (160.9 ± 23.5 μm) vs the left eye (160.6 ± 19.3 μm) (p=0.89). As a result, measurements acquired from the right and the left eye of each dog were subsequently averaged, and a mean value reported. There was a statistically significant difference (p = 0.02) between the mean (± SD) retinal thickness of the superior retina (210.2 ± 24.2 μm) compared to the inferior retina (160 ± 20.9 μm).

For each time point, the mean retinal thickness per location along the central superior and inferior retina was calculated and represented as a spider graph (Fig. 4 A–C). The mean retinal thickness decreased rapidly between 4 and 12 weeks of age in both the superior and inferior retina.

Figure 4. Spider graphs along the central superior and inferior meridians measured by manual segmentation from cross-sectional OCT B-scans and one-phase exponential decay model fitting of the retinal thickness.

Retinal thickness from 4 to 12 weeks of age (A), 13 to 36 weeks of age (B) and 40 to 119 weeks of age (C). For comparison, the retinal thickness at the 4 and 119 week time-points are represented on each of the three spider graphs. (D) Superior (dashed blue line) and inferior (black line) retinal ONL thickness with an exponential decay rate (SE) of −0.12 (0.01) Y_∞ of 193.6 μm and −0.18 (0.01) Y_∞ of 146.9 μm respectively, and (E) mean (superior + inferior) central retinal thickness with an exponential decay rate (SE) of −0.14 (0.01) and Y_∞ of 170.6 μm

The decline in the retinal thickness over time was best fitted by the one-phase exponential decay model. The fitted equations for the mean retinal thickness (μm) of the central superior and inferior retina and that of the mean of the inferior and superior combined, as a function of age (in weeks), are detailed in Fig. 4 D–E and Table 2.

4. DISCUSSION

Spectral domain OCT has become a standard in vivo imaging technique for assessment of intraocular structures, in particular of the retina. In this study, we confirmed by non-invasive imaging the age-related changes in ONL, ONL + IS, and retinal thickness previously observed by histology and provide mathematical models to predict the ONL, ONL + IS, and retinal thickness of the healthy central canine retina based on the animal’s age. Our results indicate that there is a concomitant thinning of the 3 layers described above that occurs predominantly between 4 and 12 weeks of age, followed by a slower decline before reaching a plateau at around 20 weeks of age (Fig. 2–4 D, E). Thereafter, there are no further changes associated with aging up to 119 weeks of age.

The superior fundus of the retina presents a very characteristic retinal layer that is not present in other species such as primates and humans, the tapetum lucidium (Ollivier et al., 2004). This specialized choroidal anatomical structure, located between the RPE and the choroid, is highly reflective and contributes to the largest and most intense signal on the OCT. This makes the differentiation between the RPE and the tapetum, very difficult in the superior/tapetal region of dogs. Therefore, we decided to consider the outer limitation of the ellipsoid zone to determine the outer boundary of the ONL + IS, and retinal thickness measurements.

Values of ONL thickness measured by OCT are available for normal adult beagles (Hernandez-Merino et al., 2011), and adult cats (Gekeler et al., 2007) but no in depth measurements in early post-natal animals have been reported so far in either species. Our study, now provides data for dogs during this period.

Our results showed that the ONL in the inferior retina was significantly thinner thanin the superior retina. However, findings from another group, did not reveal such difference (Hernandez-Merino et al., 2011). In that study measurements were made along horizontal OCT B-scans located in the central superior temporal and the central inferior temporal quadrant, parallel to the tapetal - non tapetal junction, and reported values were the average of 5 measurements along each scan. By comparison, our study evaluated OCT B-scans that extended vertically as far as 30° from the edge of the ONH both along the superior and inferior meridians, and averaged between 10 and 16 measurements along each scan. The differences found may be methodological, or reflect topographic differences as the same areas were not imaged in both studies.

A recent study (Ofri and Ekesten, 2019), reported the total retinal, outer retinal, and nerve fiber layer thickness in juvenile (20 weeks of age), mature (6.5 ± 2.1 years) and elderly (10.8 ± 1.2 years) female beagles. These results showed that there was no interocular difference, that the retina was thinner ventrally, and thinner in older animals. In the present study, we conducted measurements in much younger animals (as early as 4 weeks of age) and observed early changes in retinal thickness that occurred well before 20 weeks of age. The prominent reduction seen between 4 and 12 weeks of age occurs simultaneously with the rapid growth of the canine eye and the fast expansion of its axial globe length (Tuntivanich et al., 2007). Minimal post-developmental photoreceptor apoptosis measured by TUNEL assay occurs between 4–6 weeks of age in the normal canine retina (Beltran et al., 2006), thus the thinning of the ONL thickness observed between 4 and 12 weeks of age is likely caused, instead, by increase in eye size, that results in stretching of the retina and tangential dispersion of photoreceptors. These changes with aging are present equally in males and females.

Retinitis pigmentosa is a large group of non-allelic inherited retinal disorders that typically cause progressive degeneration of rods and then cones, and can manifest variable spatio-temporal patterns of abnormal ONL thinning (Beltran et al., 2006; Downs et al., 2016). Characterizing the natural history of photoreceptor cell loss in both clinically-relevant animal models and patients has been critical for establishing proof of concept of novel therapeutic strategies and targeting these to the optimal regions of the retina. A number of these therapies have been developed and tested in canine models, and changes in ONL thickness have been frequently used as an outcome measure of intervention by comparing ONL thickness at the endpoint to that measured or estimated at the time of delivery of the therapy (Beltran et al., 2015, 2007; Cideciyan et al., 2018). In light of this current study, such approach for evaluating the treatment effect of interventions delivered in young (< 12 weeks) dogs likely underestimates the extent of the photoreceptor rescue effect. Assuming that a similar level of physiological ONL thinning occurs during the first twelve weeks of life in WT and mutant retinas, the established mathematical models will now provide a way of more accurately quantifying the level of rescue of ONL thickness attributed to therapeutic intervention by comparing the ONL thickness at the endpoint to that of an age matched WT animal (Fig. 5).

Figure 5. Adjustment of ONL rescue post therapy delivered to the young canine retina considering the natural course of ONL thinning.

Schematic Illustration of the natural course of ONL thinning in WT retinas (blue curve), in mutant retinas undergoing a rapid onset of retinal degeneration (red curve), and in mutant retinas post therapy (green curve). The black two-sided arrows represent the differences in ONL thickness of mutant retinas post therapy with (thick arrow) and without (thin arrow) consideration of the physiological thinning of the ONL that occurs during the first 3-month of life.

In summary, we have established the kinetics of ONL, ONL + IS, and retinal thinning that occurs postnatally in the normal canine retina. These results and mathematical models will be valuable to accurately assess the extent of rescue following photoreceptor-targeted therapies in canine models of photoreceptor degeneration.

Supplementary Material

1

Supplementary Fig. 1. One phase exponential decay model fitting of (A) the superior, (B) the inferior, and (C) the mean (superior + inferior) ONL thickness per sex and their corresponding equations. The exponential decay rate (SE) and corresponding p value for the females and males respectively are: −0.20 (0.02), P <0.001 and −0.30 (0.08), P = 0.007 for the superior; −0.16 (0.01), p<0.001 and −0.14 (0.02), P<0.001 for the inferior and; −0.18 (0.01) P<0.001 and −0.15 (0.02) P<0.001 for the mean (Sup + Inf).

2

Supplementary table 1. Distribution of animals enrolled in the study per age (in weeks)

Highlights.

• Changes by OCT of the ONL, ONL + IS and retinal thickness of the canine retina.

• Thinning of retinal thickness occurs early in life.

• The superior retina is thicker than the inferior.

• The decline was modeled fitting a one-phase decay model.