Lifelong Changes in the Choroidal Thickness, Refractive Status, and Ocular Dimensions in C57BL/6J Mouse

Abstract

Purpose: To investigate the changes in choroidal thickness (ChT), refractive status, and ocular dimensions in the mouse eye in vivo using updated techniques and instrumentation.

Methods: High-resolution swept-source optical coherence tomography (SS-OCT), eccentric infrared photoretinoscopy, and custom real-time optical coherence tomography were used to analyze choroidal changes, refractive changes and ocular growth in C57BL/6J mice from postnatal day (P) 21 to month 22.

Results: The ChT gradually increased with age, with the thickest region in the para-optic nerve head and thinning outward, and the temporal ChT was globally thicker than the nasal ChT. Retinal thickness remained stable until 4 months and subsequently decreased. The average spherical equivalent refraction error was −4.81 ± 2.71 diopters (D) at P21, which developed into emmetropia by P32, reached a hyperopic peak (+5.75 ± 1.38 D) at P82 and returned to +0.66 ± 1.86 D at 22 months. Central corneal thickness, anterior chamber depth, lens thickness, and axial length (AL) increased continuously before 4 months, but subsequently exhibited subtle changes. Vitreous chamber depth decreased with lens growth. ChT was correlated significantly with the ocular parameters (except for retinal thickness) before the age of 4 months, but these correlations diminished after 4 months. Furthermore, for mice younger than 4 months, the difference in the ChT, especially temporal ChT, between the two eyes contributed most to that of axial length and spherical equivalent refraction error.

Conclusions: Four months could be a watershed age in the growth of mouse eyes. Large-span temporal recordings of refraction, ocular dimensions, and choroidal changes provided references for the study of the physiological and pathological mechanisms responsible for myopia.

Myopia is a common eye disorder in humans.^1–3 It is predicted that approximately one-half of the world's population will suffer from myopia by 2050, and 9.8% will have high myopia.³ High myopia can be complicated by associated pathologies such as myopic macular degeneration, retinal detachment, and open-angle glaucoma, which can cause severe irreversible visual impairments and even blindness.⁴ However, the exact mechanisms underlying myopia remain unclear. Various animal models have been developed for myopia investigations.^5–9 Among these models, the mouse is used due to its strong operability, well-established methods of genetic manipulation,¹⁰ and, most important, the similar ocular manifestations to other animal models, including non-human primates, under conditions of lens-induced myopia¹¹ and form deprivation myopia.¹² 

Abundant human data indicate a strong correlation between myopia and eye growth.^13–15 As myopia progresses, the structure of the eye undergo abnormal changes. Excessive elongation in axial length (AL) is accompanied by decreased choroidal thickness (ChT) and scleral tissue remodeling.¹⁶ Evidence from multiple myopic animal models shows that optical defocus and/or the quality of retinal images can alter the normal course of emmetropization and influence ocular development.^17–19 The choroid may play an important role in this process. The choroid is known to supply oxygen and nutrients to the outer retina, absorb excess light, and modulate intraocular pressure.^20,21 Newly identified roles of the choroid in response to ocular defocus and scleral modulation, and therefore in eye growth regulation, have been proposed in recent years,^22,23 providing a new and promising direction for myopia research. 

However, the lack of efficient and reliable instrumentation for measurement of ocular biometric and refractive data in mouse eyes reduces the accuracy of phenotype monitoring during in vivo studies, rendering greater variability in the experimental results.^24–26 Therefore, exploring feasible methods for detecting mouse refractive status and ocular parameters and characterizing changes in these structures with age are critical for further investigations in myopia. 

Remtulla²⁷ and Hallett first described the optical component data of C57BL/6J mice. Schmucker and Schaeffel²⁸ developed a paraxial schematic eye model for growing mice. However, the reliability of these ocular dimension data is undetermined owing to the low sensitivity of in vitro measurements. To overcome the limited resolution of these measurement techniques in vitro, Schmucker used optical low coherence interferometry to measure ocular biometrics, but only AL, corneal thickness, and anterior chamber depth (ACD) were measured in living mice.²⁹ 

ChT has been studied in chicks,²³ guinea pigs,⁶ cynomolgus macaques³⁰ and rhesus monkeys,³¹ but it has rarely been studied in mice. Berkowitz et al. successfully measured the total ChT of C57BL/6J mice using magnetic resonance imaging (MRI).^32,33 Others have reported ChT in mice using a spectral domain optical coherence tomography (OCT) imager.³⁴ However, these methods are expensive and time consuming, and it hardly distinguished the choroidal capillary layer and large vessel layer. Moreover, the studies were limited by the age of the mice and the sample size. Recent technological advances in OCT imaging, particularly wide-field swept-source OCT (SS-OCT) with an A-scan rate as high as 400 kHz and an axial scan depth of 6 mm, have facilitated quantitative evaluation of the choroidal structure, providing an opportunity to study choroidal properties at higher resolutions in vivo. 

Accordingly, the present study aimed to obtain a comprehensive profile of changes in ChT, refractive status, and other ocular dimensions in C57BL/6J mice in vivo from postnatal day (P) 21 to 22 months old, using new techniques and instrumentation, and to investigate the potential relationships among ocular parameters. 

Materials and Methods

Animals and Rearing

The animal research was approved by the Animal Care and Ethics Committee at Peking University People's Hospital (Beijing, China). All treatment and care of the animals were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Black C57BL/6J wild-type mice were purchased from Beijing Weitong Lihua Experimental Animal Technology Co. Ltd. (China). All animals were screened before the experiment to exclude those with ocular abnormalities, such as microphthalmia, corneal opacity and congenital cataracts. Mice were raised in groups of three to six in standard mouse cages with a 12–12 hour day–night cycle (lights on at 8:00 am) at an ambient illuminance of approximately 500 lux during the daytime. The room temperature was maintained at 25°C. All animals were provided with abundant food and water. 

Experimental Protocol

This is a cross-sectional study. Seventy-two mice were assigned to 12 groups (n = 6 mice, 12 eyes). A series of ocular measurements were obtained for each group. In this study, we measured the refraction, axial components, and ChT of mice in sequence. Refractive status (recorded as spherical equivalent refraction error [SER]), ocular dimensions, including central corneal thickness (CCT), ACD, lens thickness (LT), vitreous chamber depth (VCD), AL, retinal thickness (RT), and ChT were comprehensively examined at 12 time points: P21, P23, P26, P28, P32, P39, P61, P82 days, 4 months, 10 months, 19 months, and 22 months. All data measured at each time point were examined on the same day. The experimental procedure was designed to ensure that all the mice survived throughout the entire experimental period. 

Refractometry

The refraction of each eye in alert animals was measured with an eccentric infrared photoretinoscopy, the PowerRefractor (Fig. 1B), according to a previously published procedure.²⁶ In brief, each mouse was dark adapted for 30 minutes. Eye drops (Santen Pharmaceutical Co. Ltd., Osaka, Japan; 0.5% tropicamide combined with 0.5% phenylephrine) were used to dilate the pupils. Both eyes were refracted in a dark environment (<3 lux). The mice were then placed on a restraining platform in front of the eccentric infrared photoretinoscopy calibrated as described previously.²⁶ Calibration was verified using lenses of increasing power, from −10 diopters (D) to +10 D placed in front of a mouse eye. The measured working distance is 56 cm. The mouse was positioned so that the first Purkinje image was in the center of the pupil. The data were recorded using software designed by Schaeffel et al.²⁶ Three independent measurements were taken per eye from each mouse, and the average of the three measurements was used for analysis. 

Figure 1.

 

Measurements of ChT, axial components, and refractive status. (A) A representative scanned image showing the vitreous chamber, retina, choroid, and sclera of the mouse eye using SS-OCT. The contents of the yellow boxes are enlarged to illustrate the details in the original and pseudocolor magnified images. The choroid was further divided into the capillary layer and large vessel layer. T1–T6: 0.15–0.90 mm on the temporal side; N1–N6: 0.15–0.90 mm on the nasal side. Scale bar, 15 µm. (B) Refraction measurement using eccentric infrared photoretinoscopy. A brightness distribution was observed in the pupil of the mouse eye during infrared photoretinoscopy (yellow frame). (C) A representative scanned image of a mouse eye from the anterior cornea to the posterior retina obtained using a custom real-time OCT instrument. The AL was defined as the distance between the apex of the cornea and the front surface of the retina and was calculated as the sum of the CCT, ACD, LT, and VCD.

Axial Component Measurements

Mice were anesthetized via intraperitoneal injection of 2.5% tribromoethanol (350 mg/kg, Sigma-Aldrich, St. Louis, MO, USA) before measurement.³⁴ Eye drops containing 0.5% tropicamide combined with 0.5% phenylephrine (Santen Pharmaceutical Co. Ltd.) were instilled in both eyes to ensure cycloplegia. A custom real-time OCT instrument was used to obtain images (Fig. 1C), which is an SS-OCT system with a center wavelength of 1060 nm, a spectral bandwidth of 40 nm, an A-scan speed of 10 kHz, and a MEMS-VCSEL laser (HSL-1, Santec Pharmaceutical Co. Ltd.). An achromatic lens was used as the objective lens (AC300-100-C, Thorlabs, Newtson, NJ, USA) with a working distance of 88 mm and an imaging field of view of 20 mm × 20 mm. The axial resolution of the system in tissue was 10 µm and the maximum imaging depth was 50 mm. During scanning, the optical axis of the incident ray passed vertically through the vertex of the cornea. The real-time images were displayed on a vertical image monitoring screen. A researcher grasped and oriented the animal's eye to ensure that both irises were horizontal and symmetrical. Another researcher operating the machine was responsible for ensuring that the center of the incident ray was located at the vertex of the cornea and recording the measurement data. The OCT scans lasted approximately 2 seconds; 50 images were saved and the best quality images were selected for analysis. Three independent measurements were taken for each eye, and the axial components are the average of three measurements. 

The original OCT images were used to calculate the corrected geometrical lengths from the optical path lengths. The biometric data included CCT, ACD, LT, VCD, and AL. Two researchers (TT and RC) manually measured the ocular biometrics using ImageJ software (https://imagej.net/ij/). In this study, AL was defined as the distance between the apex of the cornea and the anterior surface of the retina and was calculated as the sum of the length of the cornea, anterior chamber, lens, and vitreous chamber.^34-–37 For this OCT instrument, the pixel density was 0.015 mm/pix. The following formula was used to calculate the length of each axial component in millimeters:  

\begin{equation*}L = \frac{{P \times PDensity}}{{RI}},\end{equation*}

where L is the length of each axial component, P is the number of pixels of this component measured on the original OCT image, PDensity is the pixel density of the machine, and RI is the refractive index of each component; these values were 1.4015 for the cornea, 1.3336 for the aqueous humor, y = 0.0005 age in days + 1.557 for the lens, and 1.3329 for the vitreous fluid in mice.³⁵ 

Choroidal Thickness

The choroidal measurements were taken during the same period of the day (from 17:00 to 21:00) to avoid potential influences of diurnal variations. Mice were anesthetized with intraperitoneal injection of 2.5% tribromoethanol (350 mg/kg, Sigma-Aldrich), as previously described.³⁴ After the pupils were dilated with eye drops containing 0.5% tropicamide combined with 0.5% phenylephrine (Santen Pharmaceutical Co. Ltd.), mice were positioned gently on a homemade restraining platform in front of an ultra-widefield SS-OCT instrument (BM-400 K BMizar, TowardPi Medical Technology, Beijing, China). The scan rate of this instrument was 400,000 A-scans/second, and the wavelength was 1060 nm. The axial resolution was 3.8 µm, and the transverse resolution was 10 µm, enabling the capture of high-resolution fundus images. The scanning plane was positioned to pass through the center of the cornea and optic nerve head (ONH), ensuring consistency and accuracy in each measurement. Radial scanning at intervals of 10° obtained images with the following dimensions: vertical 12 mm × horizontal 12 mm × scan depth 6 mm. In this measurement mode, we can measure images with a maximum of 12 mm in the vertical and horizontal radial directions, as well as images with a maximum depth of 6 mm in the axial direction. Images quality was graded from 1 to 10 by the SS-OCT platform, with only images graded 8 or greater were used in the present study. Imaging of each mouse was completed within 10 minutes to prevent opacification of the cornea and lens. After the examination was completed, the eyes were washed with physiological saline, and levofloxacin eye drops were applied to prevent infection. Sections parallel to the horizontal meridian of the eye were selected for measurement. 

The ChT was defined as the distance between Bruch's membrane and the choroid-sclera interface. The distances from the measurement points to the ONH were identified automatically by the built-in software, and T1-T6 (0.15 mm, 0.30 mm, 0.45 mm, 0.60 mm, 0.75 mm, and 0.90 mm, respectively, on the temporal side) and N1-N6 (0.15 mm, 0.30 mm, 0.45 mm, 0.60 mm, 0.75 mm, and 0.90 mm, respectively, on the nasal side) were labeled. The thicknesses of the retina, choroidal capillary layer and choroidal large vessel layer were manually measured on the pseudocolor images of the selected sections (Fig. 1A). The total ChT was derived by summing the thicknesses of the choroidal capillaries and large vessels. 

Statistical Analysis

All statistical analyses were performed using R (version 4.3.0; The R Foundation for Statistical Computing, Vienna, Austria) and the SPSS statistical software package (version 22.0, IBM Corp., Armonk, NY, USA). Differences between the two eyes of the same animals in each age group were analyzed statistically for refractive status, axial component, RT, and ChT (paired-sample t test). One-way ANOVA was used to determine overall effects for refractive status, axial components, RT, and ChT in the different age groups for both eyes. Spearman's correlation coefficient calculation and locally estimated scatterplot smoothing regression analyses were performed to further evaluate the relationships between refractive status and each of the ocular biometric parameters. The factors that contributed to binocular differences in the SER, AL and ChT were extracted by principal component analysis (PCA). The information on the functions and R packages used in the analyses is listed in the Supplementary Table S1. All continuous variables are expressed as mean ± standard deviation. A P value of less than 0.05 indicated statistical significance. 

Results

The mean values and standard deviations of refractive error and ocular biometrics at different ages are shown in Supplementary Table S2. There were no significant differences in refractive error, corneal thickness, ACD, LT, VCD, AL, or RT between the right eye and left eye in most age groups of the mice. With the exception of a very few regions where ChT differed, most ChT was not significantly different between the right and left eyes (Supplementary Tables S3–S5). Therefore, the average values for both eyes were analyzed for correlation and regression analyses. 

Profile of Mouse ChT at Different Ages

The representative SS-OCT images show an age-related increase in the total ChT on both the nasal and temporal sides of mouse eyes (Fig. 2A). As shown in the heatmaps in Figures 2B, 2C, and 2D, the choroidal capillaris and large vessel layers, as well as the total choroid, globally thickened as the mice aged (ANOVA, P < 0.05). At any age, the thickness of the choroidal large vessel layer and total choroid varied with distance to the ONH; these layers were the thickest in the para-optic disc region (T3 and N3, 0.45 mm to the ONH on the temporal and nasal side), but the choroidal capillaris exhibited no changes. The line charts in Figures 2B, 2C, and 2D demonstrate that these trends in choroidal changes were consistent when the left and right eyes were considered separately. In addition, the line charts intuitively showed a globally thicker thickness of the choroidal large vessels and total choroid on the temporal side than on the nasal side in both eyes. Detailed ChT data are provided in Supplementary Tables S3, S4, and S5. 

Figure 2.

 

Profile of ChT in mouse eyes from postnatal day 21 to month 22 measured using high-resolution SS-OCT. (A) Representative single SS-OCT images of ChT at different ages. Scale bar, 15 µm. (B–D) Average thickness of the choroidal capillary layer (B), choroidal large vessel layer (C) and total choroid (D) at different measurement positions. Heatmaps (top) show the average thickness of all the tested eyes; line charts show the mean ± standard deviation of the right eye (middle) and left eye (bottom) measurements. T1–T6: 0.15–0.90 mm to the ONH on the temporal side; N1–N6: 0.15–0.90 mm to the ONH on the nasal side.

Changes in Refraction and Ocular Dimensions

The changes in refractive error in the mice as measured with infrared photoretinoscopy is shown in Supplementary Table S2 and Figure 3. The refraction and ocular dimensions change significantly between day 21 and month 22 of age (ANOVA, all P < 0.05). The average refraction on day 21 was −4.64 ± 2.71 D. The refractive error of myopia decreased rapidly, reaching emmetropia by day 32 (+0.16 ± 2.59 D). Then, the refractive error of hyperopia increased rapidly and reached a hyperopic peak of +5.75 ± 1.38 D at approximately day 82. After day 82, the measured refractive error stabilized for up to 4 months. After 4 months, the refraction of hyperopia gradually decreased, and by 22 months, the refractive error returned to +0.66 ± 1.86 D. 

Figure 3.

 

Changes in refractive status and growth curves of axial components in the normal C57BL/6J mice from 21 days to 22 months. The data are shown as the mean ± standard deviation.

Developmental data on other ocular parameters, including CCT, ACD, LT, VCD, AL, and RT are shown in Supplementary Table S2 and Figure 3. The average CCT significantly increased from 0.127 ± 0.007 mm on day 21 to 0.148 ± 0.010 mm at 4 months. The average ACD significantly increased from 0.253 ± 0.022 mm on day 21 to 0.396 ± 0.025 mm at 4 months. The LT increased rapidly between day 21 (1.690 ± 0.027 mm) and day 39 (1.989 ± 0.013 mm). After 39 days, the LT continued to increase slowly until 4 months. Because the LT increased markedly, the VCD decreased accordingly from 0.645 ± 0.050 mm on day 21 to 0.544 ± 0.032 mm at day 82, and remained stable after 82 days. The RT (measured at 3 mm to the ONH) did not change significantly between day 21 and 4 months (P > 0.05), with an average of 0.243 ± 0.009 mm. The AL and ACD increased at a very slow rate, and the RT decreased continuously over months 4 through 10. 

Relationships Among the ChT, Refractive Status, and Ocular Parameters

Before 4 months of age, the choroidal capillaris layer, large vessel layer, and total ChT showed substantial correlations with the SER, AL, CCT, ACD, LT, and VCD; these correlations were slightly greater in the temporal regions (Fig. 4). Increases in the CCT, ACD, LT, VCD, and AL were associated significantly with age (all P < 0.01). LT had the highest correlation with AL, followed by ACD, CCT, RT, and VCD. 

Figure 4.

 

Correlations among ChT, age, SER, and ocular dimensions in mice younger (top) or older (bottom) than 4 months of age. Only the statistically significant results are shown. The numbers within the bubbles represent Spearman's correlation coefficient (ρ). T1–T6: 0.15–0.90 mm to the ONH on the temporal side; N1–N6: 0.15–0.90 mm to the ONH on the nasal side.

After 4 months of age, these correlations became much more subtle. LT had the highest correlation with AL, followed by ACD and VCD, whereas CCT and RT had no correlation with AL (Fig. 4). 

Correlation Between Refractive Status and Ocular Biometrics

For mice aged 21 days to 4 months, the refractive error had the strongest correlation with LT (ρ = 0.82; P < 0.001), followed by ACD (ρ = 0.79; P < 0.001), and AL (ρ = 0.77; P < 0.001). Refraction also correlated significantly with the CCT (ρ = 0.52; P < 0.001) and VCD (ρ = −0.40; P < 0.001). The refractive error did not correlate with RT (P > 0.05). The refractive error correlated with the thickness of the choroidal capillaries (P < 0.01), choroidal large vessels (P < 0.01), and the total choroid (P < 0.01). 

For mice older than 4 months of age, refractive error correlated only with LT (ρ = −0.55; P = 0.02); CCT, ACD, VT, VCD, AL, and RT did not significantly correlate with refractive error. The refractive error correlated with the total choroid (T3; ρ = −0.75; P < 0.05) and choroidal large vessels (T6; ρ = −0.49; P = 0.04); no correlations were observed between refractive error and ChT in other regions (all P > 0.05) (Fig. 5). 

Figure 5.

 

Locally estimated scatterplot smoothing regression analysis for refractive status (SER) in mice younger or older than 4 months. The marginal bar diagrams in each plot demonstrate the data distribution of the corresponding parameter. Red lines: estimates in younger mice; blue lines: estimates in older mice; grey regions: 95% confidence intervals. T3, T6: 0.45 mm and 0.90 mm from the ONH on the temporal side; N3, N6: 0.45 mm and 0.90 mm to the ONH on the nasal side. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Factors Contributing to Binocular Differences

PCA describes the relationships between multiple factors using a small number of variables. An initial requirement for PCA is that the original variables to be correlated strongly. Figures 4 and 5 show that all correlation coefficients in mice younger than 4 months of age were greater than 0.3, indicating that these variables were suitable for PCA. The PCA data include differences in ChT and ocular dimensions between both eyes of all the mice younger than 4 months. Figures 6A, 6C, and 6D illustrate the 33rd and 66th percentiles of binocular disparity of AL, T3, and N3 of the total choroid, which were divided into minimum, medium, and maximum variance, and Figure 6B illustrates the 50th percentile of binocular disparity of the SER, which was divided into minimum and maximum variance. Figure 6A shows that VCD and temporal ChT had the greatest effect on AL. Figure 6B shows that VCD, AL and temporal ChT (T3 of total choroid and choroidal large vessels) had the greatest effect on SER. Figure 6C shows that AL and VCD had the greatest effects on T3 of the total choroid. Figure 6D shows that the SER and ocular dimensions had no significant effect on N3 in the total choroid, owing to the high degree of overlap among the three clusters according to the PCA. 

Figure 6.

 

Factors that contributed to binocular differences in AL (A), SER (B), and ChT (C, D) were extracted by PCA. (Left) Three-dimensional visualization of the PCA. (Right) Two-dimensional visualization of the PCA. The gray arrows show the original variable axes projected onto the principal components. a: total choroid; b: choroidal large vessels; c: choroidal capillaris. T3 and T6: 0.45 mm and 0.90 mm to the ONH on the temporal side; N3 and N6: 0.45 mm and 0.90 mm to the ONH on the nasal side.

Discussion

In this study, we investigated the normal changes in refractive status and ocular dimensions in mice in vivo over a broad range of ages (postnatal 21 days to 22 months) and determined the relationship between refraction and ocular dimensions at different growth stages. Moreover, our study is the first to describe the growth profile of ChT in mice using high-resolution SS-OCT, providing an opportunity to investigate choroidal characteristics at different ages and choroidal changes during the process of emmetropization in the mouse eye in vivo. 

Compared with higher primates and humans, mice have a shorter life cycle; therefore, the pattern of organ development throughout the body is not the same as that of higher primates and humans. In general, mice begin to open their eyes at postnatal day 12 to 14, are weaned at approximately postnatal day 21, reach sexual maturity at postnatal days 40 to 60, and die naturally by approximately 99 weeks.²⁸ As mice age, the refraction and ocular dimensions of the eye grow accordingly. 

The refractive development of animals such as tree shrews,³⁸ guinea pigs,³⁹ and monkeys⁴⁰ at the time of eye opening is characterized by high hyperopia, which rapidly decreases in the early postnatal period of visual exposure, during which the animals undergo emmetropization. Unlike these species, mice are myopic at postnatal day 21 and become more hyperopic with age.^26,28 Two recent studies using eccentric infrared photoretinoscopy reported that young mice were myopic, and gradually developed hyperopia and finally became stable thereafter.^24,25 These findings suggest that refractive development in mice is not exactly the same as that in other mammalian species. Consistent with the findings of a study by Zhou et al.,²⁵ we found that C57BL/6J mice were myopic at postnatal day 21. Then, the myopic refraction continued to decrease, with the mice undergoing emmetropization by postnatal day 32, and rapidly progressed to hyperopia within 1 week, reaching a hyperopic peak at approximately 82 days. This developmental pattern might be attributed to the increasing lens size and decreased VCD, resulting in a hyperopic shift in refraction. An increasing corneal curvature may also be involved in this process,^24,25 Zhou et al.²⁵ reported that the corneal curvature of mice increased from day 22 (1.316 ± 0.011 mm) to day 95 (1.546 ± 0.012 mm), and changes in corneal curvature can also affect refraction; unfortunately, we did not measure corneal curvature in this study owing to the unavailability of the required equipment. Between 3 and 4 months of age, the refraction of the mice remained stable, but hyperopic refraction gradually decreased as they aged, and by 22 months, the refraction returned to +0.66 ± 1.86 D. One possible reason is that the RT became progressively thinner, and the AL further elongated. In contrast, as in other mammalian species, the LT and lens density of mice increase with age, which results in an increased refractive index of the lens,⁴¹ causing myopic drift in refraction. 

Previous studies have reported that the growth of mouse AL is divided into rapid and slow periods.^25,28,29 A recent MRI study showed that the mouse eye grows in three distinct phases, a rapid growth (11 µm/d) lasts until postnatal day 40, followed by a reduced rate of growth (3 µm/d) until postnatal day 67, and a very slow pace of growth (2 µm/d) afterward. A similar growth pattern in AL was found in tree shrews,³⁸ rhesus monkeys,³¹ and humans.⁴² In the present study, mouse AL grew rapidly until postnatal day 39, followed by slow growth until 4 months of age. After 4 months, the AL continued to increase at a very slow rate. 

We found that the mouse CCT increased gradually from 21 days to 4 months and subsequently stabilized. Schmucker and Schaeffel^28,29 also reported a small increase in corneal thickness in growing C57BL/6J mice using frozen sectioning or the optical low coherence interferometry technique. In contrast, OCT data of mouse CCT revealed no age-related changes from 22 to 102 days of age.²⁵ This discrepancy is probably owing to differences in methodology, measurement accuracy of the instrument, and/or manual measurement errors. 

The mouse ACD increased from postnatal 21 (0.253 ± 0.022 mm) to 61 days (0.392 ± 0.033 mm). After that period, the ACD gradually changed until 4 months of age. The optical low coherence interferometry technique revealed that the mouse ACD increased significantly from days 22 (0.260 ± 0.019 mm) to 67 (0.370 ± 0.013 mm), and then changed slowly afterward.²⁵ Tkatchenko et al.²⁴ also found that the mouse ACD increased from day 21 (0.361 ± 0.001 mm) to 67 (0.381 ± 0.003 mm) as measured by MRI. Similar growth patterns were reported in other mammalian species.^31,38–40 Furthermore, we found that the ACD further deepened at a very slow rate in older mice, whereas in humans, the older population had a shallower ACD than a younger population.⁴³ 

The mouse LT significantly increased from day 21 to 61 and continued to increase slowly by 4 months of age. Similar findings were reported by two recent studies.^24,25 Unlike tree shrews’ lenses, which continue to grow throughout the life cycle,³⁸ mouse lenses remain stable after 10 months. Before day 82, the VCD decreased as the lens became thicker. After that, the mouse VCD exhibited a minimal change until 22 months. The age-related changes in mouse LT and VCD are not quite the same as in tree shrews and higher primates. In the tree shrew, there was a rapid increase in VCD during the first 15 days of visual exposure, followed by a decrease in VCD owing to the significant increase in LT and slight increase in AL. In Macaca rhesus monkeys, LT continued to increase for 12 months, followed by a decrease in LT from 12 to 22 months, whereas the VCD continued to increase.³¹ Thus, the first 4 months of life may be the main period during which refractive status and ocular dimensions develop in mice. Eye development in mice is characterized mainly by lens thickening and axial elongation, accompanied by an increase in the ACD and a shallower VCD. These developmental characteristics of the refractive media are different from those of humans and may be responsible for the differences in refractive development between mice and humans. Therefore, it is important to consider the effect of age on eye structure when selecting mice for experimental eye studies. 

Similar to the study by Zhou et al.,²⁵ we did not observe obvious age-related changes in RT from 21 days to 4 months of age. As mice aged, the RT decreased significantly from 4 to 22 months. An advanced age was associated with a decreased outer RT, which may be due to the thickness of the cone sheath becoming atrophic with age. Similar findings have been observed in humans.^44–47 Aging is associated with decreased areas of dendritic and axonal arbors and a decreased density of cells and synapses in the retina, with thinning of retinal layers, including the ganglion cell layer, inner plexiform layer, inner nuclear layer, outer plexiform layer, and RPE.⁴⁵ Unfortunately, owing to the resolution of the measurements used in this study, we were unable to determine which layer of the retina thins with age in the mouse eye. 

The mammalian choroid, which is composed of abundant vessels and stromal tissue, has three vascular sublayers: the innermost layer, which contains the choriocapillaris; the middle layer, which contains medium-sized vessels (Sattler's layer) and the outermost layer, which contains large vessels (Haller's layer).^20,48 Previous studies have reported that the subfoveal ChT in humans is negatively associated with advancing age and increasing AL and that the subfoveal ChT is the thickest on the superior side, followed by the foveal area and the temporal, inferior, and nasal sides.^49,51,52 The choroidal large vessel layer thickness in human is thinnest at a younger age, then gradually increases and peaks at 41 to 50 years of age, followed by a slight change with age. The subfoveal ChT and choriocapillaris-medium choroidal vessel layer thickness peak at 11 to 20 years of age, and then decrease slowly with age.⁴⁹ The age-related decrease in vascular components was dominated by a decrease in choriocapillaris and medium choroidal vessels.⁴⁹ It is speculated that part of the choriocapillaris layer may occlude and detach with age, which is associated with the occurrence of age-related macular degeneration. 

The mouse choroid is divided into two layers: the choriocapillaris and the large vessel layer. Linne et al.⁵⁰ and Jeong et al.¹¹ used a spectral domain OCT imager to measure ChT; however, the data of the choroid provided by this method are limited. Berkowitz et al.³² attempted to detect the choroid in mice via MRI, but the choriocapillaris layer and large choroidal vessel layer were difficult to distinguish owing to insufficient resolution. Interestingly, unlike age-related changes in human choroids,⁵¹ we found an age-related increase in the total ChT on both the nasal and temporal sides of the mouse eye. The choriocapillaris and large vessel layer, as well as the total choroid, globally thickened as the mice aged. The thickness of the large vessel layer and total choroid varied with distance to the ONH; these layers were the thickest in para-optic disc regions, but the choriocapillaris did not exhibit the same trend. Additionally, we found a globally greater thickness of the large vessel layer and total choroid on the temporal side than on the nasal side in both eyes. Moreover, AL had a significant effect on the ChT, especially the temporal ChT, in mice. All of these thickness distributions of mouse choroids are similar to those of human choroids. 

Our study shows that refraction in mice aged 21 days to 4 months correlates significantly with LT, ACD, AL, CCT, and VCD but not with RT. Similar findings were reported by Zhou et al.³⁹ Moreover, refraction was correlated significantly with the thicknesses of the choriocapillaris, large vessel layer, and total choroid. PCA revealed that VCD had the greatest effect on refractive development. Similar results were found in humans⁵² and other animal models of myopia such as chickens,⁵³ guinea pigs,³⁹ and monkeys.⁵⁴ Additionally, ChT, especially temporal ChT, has an important role in refractive development and axial elongation, suggesting that changes in the choroid might drive ocular growth in mice intrinsically. 

There are several limitations of this study. In this study, each age group included a different set of animals. In subsequent studies, we will perform longitudinal follow-up studies on the same group of mice, thus increasing the accuracy of the results. In addition, future studies should measure corneal curvature in mice and observe associations with refractive and choroidal changes. Finally, we will further expand the sample size to increase the accuracy and credibility of the study results. 

In conclusion, the mouse choroid has a similar thickness distribution to the human choroid, but has slightly different growth and development patterns. Four months could be a watershed age in the growth of the mouse eye, before which refraction continues to develop with ocular dimensions, and the choroid might play an essential role in ocular growth in mice. This study explored refractive changes and ocular growth over the lifespan of mice, and detailed data from the choroid provides references for the study of the physiological and pathological mechanisms responsible for myopia in mice. 

Acknowledgments

Supported by National Natural Science Foundation of China (Grant Nos. 82371087 and 82171092), Capital's Funds for Health Improvement and Research (No. 2022-1G-4083), and the National Key R&D Program of China (No. 2020YFC2008200 and No. 2021YFC2702100). 

Disclosure: T. Tang, None; C. Ren, None; Y. Cai, None; Y. Li, None; K. Wang, None; M. Zhao, None