Spontaneously Myopic Guinea Pig: Model of Early Pathologic Myopia

Abstract

Purpose: To evaluate whether pigmented guinea pigs with spontaneous myopia present characteristic changes of pathologic myopia.

Methods: The fundus images of guinea pigs (3 weeks old) were graded according to fundus tessellation (FT) degree. Biometric parameters, including refraction, vitreous chamber depth (VCD), and axial length (AL), were measured at ages 21 and 43 days. Some of these animals were divided into three groups: hyperopic without FT (H w/o FT), myopic without FT (M w/o FT), and myopic with FT (M w/ FT). The horizontal and vertical radii of curvature of posterior sclera (R_P-H and R_P-V, respectively) and the radii of curvature and arc lengths of superior sclera (R_S and L_S, respectively), inferior sclera (R_I and L_I, respectively), nasal sclera (R_N and L_N, respectively), and temporal sclera (R_T and L_T) were evaluated by Fuji.

Results: The fundi were graded as type A or type B (both without FT), type C (mild FT), or type D (severe FT). The prevalence of FT was correlated with myopic refraction, longer VCD, and longer AL. Eyes of M w/FT animals had shorter R_P-H and R_P-V, longer R_S and R_T, and longer L_S and L_T than eyes of H w/o FT or M w/o FT animals. Refractions shifted toward hyperopia in eyes lacking FT, but not in eyes having FT. The changes in VCD were consistent with the changes in refraction. This relatively myopic shift in refraction and shortening of VCD were found only in myopic eyes with FT, but not in myopic eyes without FT.

Conclusions: Spontaneously myopic guinea pig eyes have a high prevalence of FT. Myopic eyes with FT presented characteristic signs of pathologic myopia.

Pathologic myopia is a subset of myopia that is accompanied by typical myopic fundus complications.¹ It is one of the most common causes of blindness, being responsible for 26.1% of blindness cases in the Chinese population,² but effective treatments are lacking for most of its fundus complications.^3–5 Thus, there is an urgent need to clarify the underlying mechanism in pathologic myopia so that new therapies can be developed. However, research on the mechanism of pathologic myopia is limited and has lagged behind, partly because of the lack of suitable animal models. 

The typical changes of pathologic myopia include continuous and excessive ocular axial elongation, often associated with posterior staphyloma and myopic maculopathy.¹ The International Photographic Classification and Grading System for Myopic Maculopathy, by the META-Analysis for Pathologic Myopia (META-PM) Study Group, uses color fundus photographs to grade myopic maculopathy into five categories: category 0 (C0), no myopic retinal lesions; C1, tessellated fundus only; C2, diffuse chorioretinal atrophy; C3, patchy chorioretinal atrophy; and C4, macular atrophy. Categories C2 or worse are further classified together as clinically significant myopic maculopathy.⁶ Fundus tessellation (FT), the condition in which the choroidal vessels can be observed clearly around the fovea and the arcade vessels, is the earliest sign of myopic maculopathy.⁶ A hospital-based retrospective study showed that FT can progress to diffuse atrophy (10.1%), lacquer crack formation (2.9%), or choroidal neovascularization (0.4% ), with a mean follow-up time of 12.7 years.⁷ In a population-based longitudinal study with a follow-up of 10 years, progression was observed in 19% of eyes with FT.⁸ In a longitudinal, observational cohort study involving 657 Chinese participants (spherical equivalent refractive error [SER] ≤ −6 D; age, 21.6 ± 12.2 years), 8% of eyes with FT had progressed to advanced myopic maculopathy at 2 years’ follow-up.⁹ A prospective, longitudinal study in an adult myopic population (SER ≤ −0.5 D) found that the 12-year cumulative incidence of myopic maculopathy reached 10.3% among 1504 myopic eyes without myopic maculopathy at baseline.¹⁰ The area under the curve (AUC) of FT in predicting myopic maculopathy reached 0.78, but, when combined with age, race, gender, and refractive error, it reached 0.86.¹⁰ Thus, FT has been established as a major predictor of myopic maculopathy, and intervention at the FT stage is critical to stop its progression. FT is the most common fundus change in high-myopic children,^11–13 and therefore it might also be an early sign of myopic maculopathy in young animals with pathologic myopia. 

There is no widely used animal model of pathologic myopia so far; this might be due to expectations that pathologic myopia develops slowly, because in humans its development generally requires decades. For example, the progression from FT to advanced myopic maculopathy in humans takes a decade,⁸ and posterior staphyloma is usually found in subjects more than 40 years old.¹⁴ Animal models of myopia have been developed in many species, including chicks,^15,16 tree shrews,¹⁷ guinea pigs,¹⁸ mice,¹⁹ and non-human primates;^20,21 however, these animals are sensitive to induction of myopia only at young ages,^15–22 and whether these older animals develop chorioretinal atrophy and posterior staphyloma is still unknown, although the eyes of chicks can develop can develop lacquer cracks after lid suturing for 8 weeks.²³ Recent studies in guinea pigs have found thinning of the retina and sclera in high myopia induced by lid suturing²⁴ and increased apoptosis in the choroid in lens-induced myopia.²⁵ These results suggest the existence of chorioretinal atrophy in myopic guinea pigs, but systematic studies on the fundus changes in myopic animals are still absent. 

A few genetic mutant animal models, such as mice with a genetic deficiency in low-density lipoprotein receptor-related protein 2 (Lrp2) and chicks with the retinopathy globe enlarged (rge) mutation, also have been reported to show pathologic myopia-like fundus changes. However, Lrp2-deficient mice showed rapid loss of retinal neurons at an early age (P5), and rge chicks have also shown corneal morphometric changes,²⁶ none of which is characteristic of pathologic myopia in humans.²⁷ More importantly, pathologic myopia in human beings is likely due to polygenic abnormalities and is affected by the visual environment,^28,29 so the monogenic myopia models of mice and chicks might not be appropriate models for pathologic myopia. 

Guinea pigs could be good candidates for establishing a pathologic myopia model. Because guinea pigs are mammals, the structure and development of their eyeballs are similar to those of human beings. Furthermore, guinea pigs are susceptible to form deprivation,¹⁸ defocus,³⁰ and near work-induced myopia,³¹ so they are widely used in myopia research. At the same time, as with human beings, at least one subpopulation of guinea pigs is known to develop spontaneous myopia.³² In this study, we evaluated whether pigmented guinea pigs with spontaneous myopia present characteristic changes of pathologic myopia such as chorioretinal atrophy, continuous and excessive ocular axial elongation, and signs of posterior staphyloma. 

Methods

Animals

The animal research in this study was approved by the Animal Care and Ethics Committee at Wenzhou Medical University, Wenzhou, China. Treatment and care of the animals were according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Male pigmented guinea pigs were obtained from the Animal Breeding Unit at Wenzhou Medical University at the age of 3 weeks. This strain of guinea pigs (Cavia porcellus) is the one we have used in our previous studies, since 2006.³² In this outbred strain, there is a significant proportion of spontaneous myopia (ranging from −15.67 to −1.5 D).³² All animals were raised on a cycle of 12 hours of illumination (on at 08:00 AM), provided by 300-lux, white-light LEDs (S-400-12; Lanmeng Electronic Co. Ltd., Shanghai, China), and 12 hours of darkness during the experimental period. 

Experimental Design

The experimental design is presented in Figure 1. The colored fundus images of 109 guinea pigs (218 eyes, 3 weeks old, anisometropia ≤ 2 D) were captured and graded according to fundus vascularity (degree of FT). The fundi were classified as type A, B, C, or D. Types A and B were without FT, whereas types C and D were with FT. Biometric parameters—refraction, anterior chamber depth (ACD), lens thickness (LT), vitreous chamber depth (VCD), and axial length (AL, from anterior corneal surface to anterior retinal surface)—were measured at age 21 days. Some of these guinea pigs were used to evaluate the changes in eyeball shape and myopia progression. These animals were divided into three groups: hyperopic without FT (H w/o FT), myopic without FT (M w/o FT), and myopic with FT (M w/ FT). Parameters of eyeball shapes, the horizontal and vertical average radii of curvature of posterior sclera (R_P-H and R_P-V, respectively), as well as the radii of curvature (R) and arc lengths (L) of superior (R_S, L_S), inferior (R_I, L_I), nasal (R_N, L_N), and temporal (R_T, L_T) sclera, were evaluated by ImageJ (National Institutes of Health, Bethesda, MD) using digital images of the excised eyes. To evaluate the development of myopia longitudinally, refraction and biometric parameters of 91 guinea pigs were measured at 21 and 43 days of age. 

Figure 1.

 

Experimental design. w/o FT, without fundus tessellation; w/ FT, with fundus tessellation; ICGA, indocyanine green angiography; EIR, eccentric infrared retinoscopy; ACD, anterior chamber depth; LT, lens thickness; VCD, vitreous chamber depth; AL, axial length; H w/o FT, hyperopic without FT; M w/o FT, myopic without FT; M w/ FT, myopic with FT; P-H, horizontal axis of posterior sclera; P-V, vertical axis of posterior sclera; S, superior; I, inferior; N, nasal; T, temporal.

Biometric Measurements

The methods of biometric measurements were as described in our previous studies.^30–33 Refraction was measured in darkness, without cycloplegia, by eccentric infrared retinoscopy (EIR) with custom software. A-scan ultrasonography (11-MHz frequency; Aviso Echograph, Class I, Type B; Quantel Medical, Cournon-d'auvergne, France) was used to measure the axial components of the eye (including ACD, LT, VCD, and AL) without anesthesia. The sound conduction velocities for converting measurements of time to length were taken as 1534 m/s for ACD, 1723.3 m/s for LT, and 1540 m/s for VCD, as described previously.³⁰ All of these measurements were performed during the daytime (9:00 AM to 6:00 PM), always by the same optometrist, who was blinded to the identity of the groups. 

Color Fundus Photography

Color fundus images were taken by a mouse fundus imaging system (Micro IV; Phoenix Research Labs, Pleasanton, CA, USA). Animals were anesthetized by intraperitoneal injection of phenobarbital sodium (0.006 g/L) and xylazine (0.0009 g/L), 0.5 mL per 175 g, and the cornea was moisturized and the pupil fully dilated by tropicamide and phenylephrine ophthalmic solution (H20044926; Kangye Pharmaceutical Co. Ltd., Handan, Henan, China). Fundus photographs of each eye were taken from the nasal, central, and temporal retinal regions. The thinner the choroid is, the lower incident light intensity is required to obtain the image with best effect in structure distinguishing. To standardize this process, we obtained the images at four incident light intensities (2000 lux, 5000 lux, 12,000 lux, and 20,000 lux) (Fig. 2) for eyes with thick choroids, but at only the two lower light intensities (2000 lux and 5000 lux) for eyes with thin choroids, all with the same exposure time and camera settings. Then we selected images with the most suitable light intensity (the choroidal vascular bed visible and clear) for grading. 

Figure 2.

 

Color fundus images of a guinea pig eye. Photographs were taken of temporal, central and nasal fundus with different light intensities (2000 lux, 5000 lux, 12,000 lux, and 20,000 lux) and are shown as imaged for identical exposure times.

The choroid usually can be clearly observed in the fundi of guinea pigs (Fig. 2). FT is defined as the condition in which well-defined choroidal vessels can be observed clearly around the fovea and arcade vessels according to Meta-PM.³⁴ Because guinea pigs have no fovea and retinal blood vessels, we used the optic disk as a reference instead. We classified the fundus of guinea pigs into four groups, according to the degree of FT (Fig. 3). In general, the fundus was found to be less pigmented superiorly than inferiorly, so the classification was based mainly upon the morphology of the more clearly imaged superior fundus. 

Figure 3.

 

The classification of fundus types in guinea pigs. The fundus was classified into four groups (A, B, C, or D), according to the structure of the choroidal vasculature (thickness of the choroid and morphology of its larger blood vessels). The top panel shows the colored fundus images, the areas in blue boxes are enlarged in the middle panel, and the bottom panel shows the ICGA images corresponding to the middle panel images. The red asterisk indicates the optic disk; the yellow arrow indicates large blood vessels; and the blue arrow indicates an area without blood vessels.

The fundus images were taken by two trained examiners (YZ, ZZ) with an assistant (JL). To improve the reproducibility of grading, the images were randomly selected and graded independently by two trained graders (YZ, ZZ) masked to treatment, twice by each grader, with an interval of 1 week, so that the grader would not remember how that image was graded previously. When the evaluations did not match, the two graders would consult and discuss with another experienced ophthalmologist (SC). The match rate (intraclass correlation coefficient) between the two independent graders was 0.942. 

Indocyanine Green Angiography

Guinea pigs were anesthetized by inhalation of 4% to 5% isoflurane gas (R510-22; RWD Life Science, Shenzhen, China) in 99.9% oxygen (flow rate of 0.75–1.0 L/min) in a volume-limited ventilator (E-Z Systems, Palmer, TX, USA). Indocyanine green (H20055881; Dandong Yichuang Pharmaceutical Co., Liaoning, China) was freshly prepared before the injection (dilution, 2.5g/L; volume, 25 µL), and then was injected into an ear vein for indocyanine green angiography (ICGA) examination. Angiograms were taken with the SPECTRALIS HRA + OCT (Heidelberg Engineering, Heidelberg, Germany) using the ICGA + infrared reflectance (IR) mode. 

Classification of the Guinea Pig Fundus According to Fundus Morphology

There were different fundus morphologies in the guinea pigs, and some of the guinea pigs presented a morphology similar to that seen in human FT. Fundi without FT were graded as type A or type B, and fundi with FT were graded as type C (mild FT) or type D (severe FT). The detailed standard of grading was as follows: 

• • Type A—In the color fundus images, the central area of the posterior fundus was covered by intact choroid (Fig. 3A1). The larger blood vessels underneath were completely covered by choriocapillaris and small blood vessels, so that they could not be clearly identified in color fundus images even with the strongest light intensity (Fig. 3A2). Consequently, the ICGA image presented a uniform meshwork of small blood vessels (Fig. 3A3).
• • Type B—In the color fundus images, the fundus of this type was not as even and smooth as in type A. Larger blood vessels could be observed indistinctly at the peripheral fundus (Fig. 3B1), and small patches lacking blood vessels could be found between the larger blood vessels in both color fundus and ICGA images (Figs. 3B2, 3B3). Because the choriocapillaris and small blood vessel layers are thinner in this type, the blood vessel network appeared more distinct, and the mesh diameter appeared larger than in type A (Fig. 3B3).
• • Type C—The fundus of this type presented the appearance of mild FT. The larger blood vessels were clearly visible around the optic disk and thin flakes of choriocapillaris and small blood vessels, and small areas without blood vessels could be observed among the larger blood vessels (Figs. 3C1, 3C2). In the ICGA images, the larger blood vessels were enlarged and distorted; they sometimes had bead-like enlargements, and they formed a network with mesh diameter larger than that in type B (Fig. 3C3).
• • Type D—The fundus of this type showed severe FT. The choriocapillaris and small blood vessel layers were completely lost, leaving only distorted and enlarged blood vessels to be observed (Figs. 3D1, 3D2). In later-phase ICGA, the enlarged blood vessels appeared to form a loose network, and areas lacking blood vessels could be observed around the optic disk (Fig. 3D3).

Eyeball Shape Analysis

Guinea pigs were euthanized with excess pentobarbital sodium. After marking the cornea at the nasal and superior margin with a heated syringe needle, the eyeball was quickly enucleated. The extraocular muscle and connective tissue were removed under the microscope to make sure the outline of eyeball was exposed completely and smoothly. When it was positioned with the superior side up, the eyeball was adjusted to ensure that the nasal mark was parallel to the horizontal plane; when the eyeball was positioned with the nasal side up, the superior mark was verified to be parallel to the horizontal plane. The position of the eyeball was carefully adjusted to make sure the optic nerve head was just visible at the back of the eye. A microscope-connected camera (M620 F20; Leica Microsystems, Wetzlar, Germany) was used to take pictures of eyes when the edges of the eyeball were well focused. 

The parameters of eye shape were analyzed by Fuji software with the Kappa plug-in. The parameters analyzed included the average curvature (κ) of the posterior eyeball; a length of 6 mm of posterior sclera (centered upon the optic nerve head) was selected to analyze the average radius of curvature of the posterior eyeball, in both horizontal and vertical directions (R_P-H, R_P-V) (Fig. 4B). The arc length (L) and average radii of curvature of the superior (L_S, R_S), inferior (L_I, R_I), nasal (L_N, R_N), and temporal (L_T, R_T) eyeball were determined as follows: The outline of the eye, from the limbus to the edge of the optic nerve head, was traced manually, then the arc length and average curvature were calculated by Kappa, a curvature analysis program developed as a Fuji plug-in (Fig. 4B). Kappa allows a user to measure curvature in images in a convenient way by tracing an initial shape with a B-spline curve and then fitting that curve to image data with a minimization algorithm.³⁵ The radius of curvature is calculated as the reciprocal of curvature. Eyeball images were arranged in random order, and information regarding the corresponding animal's condition was masked. Each image was analyzed three times and then those three values were averaged. 

Figure 4.

 

Fundus morphology was strongly related to refraction. (A) ROC curve showing the sensitivity and specificity of the logistic regression between refraction and fundus grade; the closer the AUC is to 1, the better the model is for prediction of FT by the cut-off value. (B) Refraction–fundus scatterplot shows the distribution of different types of fundi versus refraction. Each colored dot represents one eye with a specific type of fundus (blue, type A; green, type B; yellow, type C; red, type D). The x-axis shows refraction of the eye, and the y-axis is dimensionless (points representing the same SER are spread vertically, so that individual points can be discriminated). The red number (−4 D) and the corresponding dashed black line show the cut-off value calculated by ROC analysis. (C1) Percentage of different fundus types in eyes with different refractive status (left) and percentages of different refractive statues in eyes with different fundus types (right). (C2) Cross-tabulation of data in C1 for χ² test. Eyes with SER ≤ −4 D had significantly higher prevalence of FT (types C and D) (C1), and FT was found mainly in animals with SER ≤ −4 D (B, C2). Black numbers in C2 indicate the numbers of eyes, and red numbers in brackets are the adjusted residuals; absolute values of adjusted residuals larger than 3 indicate significantly increased (positive value) or decreased (negative value) frequencies, as compared to expected frequencies. *SER ≥ 0 or −4 D < SER < 0 D compared with SER ≤ − 4 D (P < 0.05); &type A compared to type B (P < 0.05); $type A or type B compared to type C (P < 0.05); #type A, type B, or type C compared to type D (P < 0.05, χ², z-test adjusted by Bonferroni method).

Statistical Analysis

All data were verified to be normally distributed, and the descriptive statistics were calculated as mean ± standard error of the mean (M ± SEM). Ocular biometry or changes of ocular biometry between two groups were compared by independent t-test. When there were three groups, the data were compared by one-way ANOVA. Bonferroni corrections were applied in post hoc analysis when the requirement of homogeneity variance was satisfied, or the Brown–Forsythe test and Tamhane corrections were applied in post hoc analysis when the data distributions did not meet the homogeneity variance criteria. Ocular biometry parameters of the same animals, before and after treatment were compared by paired t-test. Spearman correlation analysis was used to analyze the correlation between the biometric parameters with fundus morphology. The receiver operating characteristic (ROC) curve was applied to evaluate the sensitivity and specificity of SER, VCD, and AL for the prediction of FT, and the corresponding AUC and cut-off values were calculated. The frequencies of different types of fundi were compared by χ² test. Values of P < 0.05 were considered to be statistically significant. SPSS Statistics 16.0 (IBM, Chicago, IL, USA) was used for the statistical analyses. 

Results

Fundus Morphology Was Strongly Related to Refraction

Initial refractions of these guinea pigs ranged from hyperopic to myopic, in parallel with the range from type A to type D fundus, and the refractions of eyes with FT (types C and D) were significantly more myopic than those of eyes without FT (types A and B) (Table 1). Spearman correlation analysis showed that there was a significant correlation between refraction and fundus grade (Spearman's rank correlation coefficient R_s = −0.520; P < 0.001). The ROC curve showed that refraction was reliably correlated with FT (AUC = 0.8155; sensitivity, 81.48%; specificity, 82.93%) (Fig. 4A, Table 2) with the cut-off refraction of −4 D. Thus, in this study, guinea pigs with SER ≤T−4 D were considered to have high myopia, and those with −4 D < SER < 0 D were considered to have mild myopia. 

Table 1.

 

Biometrics of Guinea Pigs (21 Days Old) With Different Types of Fundi (M ± SEM)

 

Table 2.

 

ROC Curve Analysis of the FT Prediction With Biometrics for Each Eye of Guinea Pigs (M ± SEM)

 

A scatterplot of refraction versus fundus type also showed that highly myopic animals had a high prevalence of FT (61%) (Fig. 4B). The distribution of fundus types was strongly related to refraction (χ² = 86.690; P < 0.001, χ²). Only 8% of hyperopic and 5% of mildly myopic eyes were associated with FT type C (Fig. 4C1); this was significantly lower than the percentages associated with type A (hyperopic, 37%; mildly myopic, 29%) and type B (hyperopic, 56%; mildly myopic, 67%) (P < 0.05, χ², z-test) (Fig. 4C2). Among highly myopic animals, the percentages of types C and D increased to 26% and 35%, respectively, but the percentage of type A decreased to 4% and type B to 35% (Figs. 4C1, 4C2). In contrast, only 6% and 23% of animals with type A or type B fundi, respectively, were highly myopic, whereas the percentage of highly myopic animals increased to 66% and 100% in animals with type C or type D fundi, respectively (P < 0.05, χ², z-test) (Figs. 4C1, 4C2). 

It is worth noting that the animals with FT were screened and selected so as to increase the sample size in the type C and type D groups; thus, the distribution of refractions in this study does not reflect the distribution of refraction in the naturally outbred population. 

Fundus Morphology Was Strongly Related to VCD

Spearman correlation analysis showed that there was a significant correlation between VCD and fundus grade (R_s = 0.585; P < 0.001), and the VCD of eyes with FT was significantly longer than that of eyes without FT (Table 1). The ROC curve showed that VCD was well correlated with, and therefore likely predictive of, FT (AUC = 0.8530; sensitivity, 68.52%; specificity, 92.07%) (Fig. 5A) with VCD cut-off at 3.37 mm. For highly myopic eyes (SER ≤ −4.0 D, according to the cut-off value of refraction by ROC analysis), the AUC of the ROC curve for VCD increased to 0.8922 (sensitivity, 66.67%; specificity, 100%) (Fig. 5D), with the VCD cut-off point at 3.50 mm (Table 2). 

Figure 5.

 

Fundus morphology was highly related to VCD. (A, D) ROC curve for VCD of all eyes (A) or only those with SER ≤ − 4 D (D). (B, E) VCD–fundus scatterplots for all eyes (B) or only those with SER ≤ −4 D (χ²) show the distribution of different types of fundi versus refraction. Each colored dot represents one eye with a specific type of fundus (blue, type A; green, type B; yellow, type C; red, type D). The x-axis shows the VCD of this eye, and the dashed black line and red number (3.50 mm) show the cut-off value calculated by ROC analysis. (C1, F1) The percentages of different fundus types in eyes with longer VCD (cut-off VCD = 3.37 mm for all eyes and 3.50 mm for those ≤ − 4 D, left), and the percentage of eyes with short or long VCD in eyes with different fundus (right). (C2, F2) Cross-tabulation of data in C1 for χ² test. VCD was a sensitive and specific parameter for predicting FT (A, D), and FT was found mainly in eyes with longer VCD (C, F). At the same time, the prevalence of FT (C, F) was significantly higher in eyes with VCD longer than cut-off VCD. Black numbers in C2 and F2 are the number of eyes and the red numbers in brackets are adjusted residuals. *Longer VCD compared with shorter VCD (P < 0.05); $type A or type B compared to type C (P < 0.05); #type A, type B, or type C compared to type D (P < 0.05, χ² test, z-test adjusted by Bonferroni method).

The VCD–fundus scatterplot showed that animals with VCD longer than the VCD cut-off point (≥3.37 mm) had a high prevalence of FT (74%) (Fig. 5B). The distribution of fundus types was related to VCD (χ² = 104.499; P < 0.001, χ²). Of the eyes with VCD < 3.37 mm, only 11% had FT (type C, 10%; type D, 1%) (Fig. 5C1), whereas of eyes with VCD ≥ 3.37 mm, 74% had FT (type C, 26%; type D, 48%) (Fig. 5C1). The percentages of type A and type B were significantly larger than those of type C and type D in animals with VCD < 3.37 mm, and were significantly lower in eyes with VCD ≥ 3.37 mm (P < 0.05, χ², z-test) (Fig. 5C2). In animals with type A or type B fundi, only 6% and 9% were with VCD ≥ 3.37 mm, respectively. These percentages increased to 45% and 96% in animals with type C or type D fundi, respectively (P < 0.05, χ², z-test) (Fig. 5C2). 

Among the eyes with high myopia, the prevalence of FT reached 97% in those with longer VCD than the cut-off value (≥3.50 mm) (Fig. 5E). The distribution of fundus types was also related to VCD in highly myopic animals (Fisher's exact test = 26.043; P < 0.001, χ²). Of highly myopic eyes with VCD < 3.50 mm, 39% had FT (type C, 16%; type D, 23%), whereas of those with VCD ≥ 3.50 mm, 97% had FT (type C, 43%; type D, 54%) (Fig. 5F1). Of highly myopic eyes with type A or type B fundi, 0% and 4%, respectively, had VCD ≥ 3.50; these prevalences increased to 63% and 60%, respectively, in eyes with type C or type D fundi (P < 0.05, χ², z-test) (Fig. 5F2). 

Fundus Morphology Was Highly Related to AL

Spearman correlation analysis showed that there was a significant correlation between AL and fundus grade (R_s = 0.587; P < 0.001), and AL was significantly longer in eyes with FT than in eyes without FT (Table 1). The AUC of the ROC curve for AL was 0.8841 (sensitivity, 77.78%; specificity, 88.41%), with the cut-off AL at 8.07 mm (Table 2). In eyes with high myopia (SER ≤ −4.0 D), the AUC of the ROC curve for AL reached 0.9180 (sensitivity, 82.05%; specificity, 86.36%) (Table 2), with the cut-off AL of 8.12 mm. 

The AL–fundus scatterplot showed that most eyes (59%) with FT had AL longer than the cut-off value (≥8.07 mm) (Fig. 6B). The distribution of fundus types was related to AL (χ² = 96.652; P < 0.001, χ²). Only 6% and 1% of eyes with short AL (AL < 8.07 mm) had type C or type D fundi, respectively, with frequencies that were significantly lower than those of types A and B (P < 0.05, χ², z-test) (Figs. 6C1, 6C2). In contrast, 67% of eyes with long AL had FT (type C, 31%; type D, 38%) (Fig. 6C1), significantly higher than the frequencies of types A and B (P < 0.05, χ², z-test) (Fig. 8C2). Only 2% and 16% of eyes with types A and B fundi, respectively, had long AL, whereas the percentages of long AL were significantly higher in type C (66%) and type D (92%) eyes (P < 0.05, χ², z-test) (Fig. 6C2). 

Figure 6.

 

Fundus morphology was strongly related to AL. (A, D) ROC curve for AL of all eyes (A) or only those with SER ≤ −4 D (D). (B, E) AL–fundus scatterplots for all eyes (B) or for those with SER ≤ −4 D (E) show the distribution of different types of fundi versus refraction. Each colored dot represents one eye with a specific type of fundus (blue, type A; green, type B; yellow, type C; red, type D); the x-axis shows the AL of this eye; the dashed black line and red number show the cut-off value calculated by ROC analysis. (C1, F1) The percentages of different fundus types in eyes with longer AL (cut-off AL = 8.07 mm for all eyes and 8.12 mm for SER ≤ −4 D, left), and the percentage of eyes with short or long VCD in eyes with different fundus (right). (C2, F2) Cross-tabulation of data in C1 for the χ² test. AL was a sensitive and specific parameter in predicting FT (A, D). FT was mainly found in eyes with AL longer than the cut-off AL (C, F). At the same time, FT was significantly more prevalent in eyes with longer AL (C, F). Black numbers in C2 and F2 are the numbers of eyes, and red numbers in brackets are adjusted residuals. *Longer AL compared with shorter AL (P < 0.05); &type A compared to type B (P < 0.05); $type A or type B compared to type C (P < 0.05); #type A, type B, or type C compared to type D (P < 0.05, χ² test, z-test adjusted by Bonferroni method).

The AL–fundus scatterplot showed that most eyes (82%) with FT had AL longer than the cut-off value (≥8.12 mm) in high-myopic eyes (Fig. 6E). The distribution of fundus types in highly myopic eyes was also related to AL (Fisher's exact test = 27.304; P < 0.001, χ²). The 29% of eyes with shorter AL (AL < 8.12 mm) had FT (type C, 16%; type D, 13%); this percentage increased to 88% (type C, 35%; type D, 53%) in those with longer AL (AL ≥ 8.12 mm) (Figs. 6F1, 6F2). Also, 0% and 29% of eyes with type A or type B fundi, respectively, had AL ≥ 8.12 mm, and the percentage increased significantly to 74% and 84%, respectively, in those with type C or type D fundi (P < 0.05, χ², z-test) (Figs. 6F1, 6F2). 

Out-Pouching of Posterior Sclera and Expansion of Superior and Temporal Sclera in Myopic Eyes With FT

The radii of curvature and arc lengths of posterior, superior, inferior, temporal, and nasal eye walls were measured to assess whether the myopic eyes with FT presented early signs of posterior staphyloma (Fig. 7A). Myopic eyes with FT had more myopic refractions than those without FT and had correspondingly longer VCD and AL. For H w/o FT versus M w/o FT versus M w/ FT, SERs were 4.40 ± 0.35 D versus −5.16 ± 1.28 D versus −8.97 ± 0.41 D, respectively (P < 0.001); VCDs were 3.15 ± 0.02 mm versus 3.25 ± 0.02 mm versus 3.44 ± 0.05 mm, respectively (P < 0.001); ALs were 7.92 ± 0.03 mm versus 7.98 ± 0.03 mm versus 8.25 ± 0.04 mm, respectively (P < 0.001, Brown–Forsythe) (Figs. 7C–7E). There were no significant differences in ACD (P = 0.852, F = 0.161, one-way ANOVA) or LT (P = 0.337, F = 1.117, one-way ANOVA). 

Figure 7.

 

Out-pouching of posterior sclera and expansion of superior and temporal sclera in highly myopic guinea pig eyes with FT. (A) Examples of eyes for H w/o FT (left), M w/o FT (middle), and M w/FT (right). (B) Schematic diagram of radius of curvature and arc length analysis. The bright green outline of the eyeball (sclera, indicated by green arrow) shows the area that was analyzed. Arc lengths and average radii of curvature of nasal (N), temporal (T), superior (S), and inferior (I) sclera were measured from the outer edge of the limbus (black arrow) to the edge of the optic nerve (orange arrow). R_P-H and R_P-V were measured along a line 6 mm long, passing through the center of the optic nerve (approximately 3 mm to each side of the optic nerve). (C–E) Refraction (C), VCD (D), and AL (E). (F) Differences in scleral radii of curvature in different regions of the eye. (F1) Average radius of curvature. (F2) Ratio of eyeball radius of curvature differences in M w/ FT when compared to H w/o FT (black) and M w/o FT (blue). (F3) Differences in radius of curvature between horizontal and vertical posterior sclera (P_V–H), nasal and temporal sclera (N–T), and superior and inferior sclera (S–I). (G) Differences in arc length in different regions of the eye. (G1) Average arc length. (G2) Ratio of arc length changes in M w/ FT compared to H w/o FT (black) and M w/o FT (blue). (G3) Differences between arc lengths in nasal and temporal sclera (N–T) and superior and inferior sclera (S–I). *P < 0.05, **P < 0.01, ***P < 0.001. Brown–Forsythe, post hoc with Tamhane if without homogeneity of variance; one-way ANOVA, post hoc with Bonferroni if with homogeneity of variance.

The growth of myopic eyes with FT varied with the fundus region. R_P-H and R_P-V in M w/ FT were significantly shorter in myopic eyes with FT than in eyes without FT. For H w/o FT versus M w/o FT versus M w/ FT, R_P-H values were 5.34 ± 0.07 mm versus 5.41 ± 1.00 mm versus 4.14 ± 0.12 mm, respectively (P < 0.001); R_P-V values were 5.10 ± 0.06 mm versus 5.06 ± 0.09 mm versus 4.03 ± 0.16 mm, respectively (P < 0.001, Brown–Forsythe) (Fig. 7F1). R_S, R_T, and R_I were significantly larger in eyes with FT than in eyes without FT. For H w/o FT versus M w/o FT versus M w/ FT, R_S values were 3.87 ± 0.06 mm versus 3.91 ± 0.03 mm versus 4.15 ± 0.07 mm, respectively (P = 0.001); R_T values were 3.71 ± 0.04 mm versus 3.66 ± 0.03 mm versus 3.94 ± 0.05 mm, respectively (P < 0.001); R_I values were 3.72 ± 0.05 mm versus 3.79 ± 0.04 mm versus 3.95 ± 0.07 mm, respectively (P = 0.006, Brown-Forsythe), but not in R_N (P = 0.153, Brown–Forsythe) (Fig. 7F1). The scleral radii of curvature in eyes with FT were shorter by approximately 20% to 30% in the posterior sclera and were longer by approximately 3% to 7% in the superior, temporal, inferior, and nasal regions than in eyes without FT (Fig. 7F2). In eyes without FT, R_P-V was generally shorter than R_P-H, whereas R_N was larger than R_T; in eyes with FT, however, this difference was absent. For R_P-V versus R_P-H, H w/o FT, P = 0.003; M w/o FT, P = 0.007; M w/ FT, P = 0.401 (paired t-test), with significantly higher variance in differences between R_P-V and R_P-H, R_N, and R_T (coefficient of variation [CV]: H w/o FT vs. M w/o FT vs. M w/ FT: R_P(V-H), 134.5% vs. 76.4% vs. 315.9.8%, respectively ; R_N-T, −61.5% vs. −72.3% vs. −183.9%, respectively) (Fig. 7F3). 

Only L_S (arc length, superior) and L_T (arc length, temporal) increased significantly in eyes with FT. For H w/o FT versus M w/o FT versus M w/ FT, L_S values were 9.18 ± 0.15 mm versus 9.09 ± 0.11 mm versus 9.61 ± 0.14 mm, respectively (P = 0.022), and L_T values were 7.82 ± 0.12 mm versus 7.72 ± 0.16 mm versus 8.60 ± 0.20 mm, respectively (P = 0.001, Brown-Forsythe) (Fig. 7G1). The differences in arc length were greatest in temporal sclera (Fig. 7G2). In eyes without FT, L_N (arc length, nasal) was consistently longer than L_T, and L_S was longer than L_I (arc length, inferior; P < 0.001 for both H w/o FT and M w/o FT, paired t-test) (Fig. 7G3). In eyes with FT, whereas L_S and L_T increased selectively, the difference between L_N and L_T decreased, and the difference between L_S and L_I increased significantly. For H w/o FT versus M w/o FT versus M w/ FT, L_S–L_I values were 1.31 ± 0.11 mm versus 1.31 ± 0.16 mm versus 1.75 ± 0.14 mm, respectively (P = 0.032, F = 3.765, one-way ANOVA); L_N–L_T values were 1.80 ± 0.09 mm versus 1.66 ± 0.20 mm versus 0.94 ± 0.31 mm, respectively (P = 0.012, Brown–Forsythe) (Fig. 7G3). 

Myopia Develops More Continuously and Excessively in Eyes With FT Than in Eyes Without FT

Refractions shifted to hyperopia during the treatment period from day 21 to day 43 in eyes without FT; for type A, from 0.95 ± 0.77 D to 2.3 ± 0.64 D (P = 0.003), and for type B, from −2.3 ± 0.48 D to −1.0 ± 0.78 D (P = 0.040, paired t-test) (Fig. 8A1). The same did not hold true for eyes with FT; for type C, from −6.4 ± 0.66 D to −7.1 ± 0.48 D (P = 0.240), and type D, from −7.6 ± 0.42 D to −8.0 ± 0.47 D (P = 0.397, paired t-test) (Fig. 8A1). From the beginning to the end of the treatment period, the changes in VCD were in parallel with the changes in refraction. In eyes without FT, VCD became significantly less: type A, from 3.14 ± 0.02 mm to 3.08 ± 0.02 mm (P < 0.001), and type B, from 3.25 ± 0.02 mm to 3.19 ± 0.02 mm (P < 0.001, paired t-test) (Fig. 8B1). In eyes with FT, VCD did not become significantly less; type C, from 3.29 ± 0.07 mm to 3.30 ± 0.08 mm (P = 0.527); type D, from 3.48 ± 0.03 mm to 3.47 ± 0.04 mm (P = 0.704, paired t-test) (Fig. 8B1). AL increased significantly in eyes with all fundus types (P < 0.001, paired t-test). 

Figure 8.

 

Continuous and excessive VCD elongation and large myopic shift in refraction in eyes with FT. (A1) Refractions in the type A and type B groups shifted toward hyperopia during the period from day 21 to day 43, but those in the type C and type D groups did not. (B1, C1) The changes in VCD were consistent with refraction (B1), although AL in all groups continued elongation (C1). (A2, B2, C2) In myopic animals, the refractions in those without FT (M w/o FT) also shifted toward hyperopia, whereas refractions in those with FT (M w/ FT) did not. The changes of refraction from d21 to d43 were significantly different in these two groups (A2), whereas the changes in VCD (B2) and AL (C2) were parallel to the changes in refraction between those two groups. ∆ indicates changes in refraction, VCD, and AL (d43–d21). *P < 0.05; **P < 0.01; ***P < 0.001, paired t-test; #P < 0.05, independent t-test.

Among myopic eyes, continuous and excessive myopia development was observed only in those with FT. The refractions of myopic eyes without FT shifted toward hyperopia, from −3.35 ± 0.38 D to −1.92 ± 0.78 D (P = 0.033, paired t-test) (Fig. 8B1), whereas those of eyes with FT did not, shifting from −7.26 ± 0.38 D to −7.68 ± 0.39 D (P = 0.298, paired t-test) (Fig. 8B1). The developmental changes of refraction in these two groups were significantly different (1.41 ± 0.67D vs. −0.51 ± 0.38D; P = 0.027, independent t-test) (Fig. 8A2). The VCDs of myopic eyes without FT decreased consistently during this period (from 3.25 ± 0.01 mm to 3.19 ± 0.03 mm; P = 0.002, paired t-test) (Fig. 8B2), whereas the VCDs of those with FT did not (from 3.42 ± 0.04 mm to 3.42 ± 0.04 mm; P = 0.944, paired t-test) (Fig. 8B2). The changes in VCD in these two groups were significantly different (−0.02 ± 0.00 mm vs. −0.00 ± 0.00 mm; P = 0.022, independent t-test) (Fig. 8B2). The ALs of myopic eyes, both with and without FT, increased significantly from day 21 to day 43 (P < 0.001, independent t-test) (Fig. 8C2); however, the change in AL in myopic eyes with FT was significantly larger than the changes of those without FT (0.03 ± 0.00 mm vs. −0.04 ± 0.00 mm; P = 0.019, independent t-test) (Fig. 8C2). 

Discussion

Pathologic myopia is characterized by continuous and excessive development of myopic refraction, axial elongation, posterior staphyloma, and myopic maculopathy. In this study, we observed signs of early stages of pathologic myopia in spontaneously myopic young guinea pigs. Over half (61%) of the eyes with high myopia (SER st−4 D) presented FT, which was associated with myopic refraction and increases in VCD and AL. In myopic eyes with FT, the increase in curvature of the posterior eyeball and selective expansion of the sclera in temporal and superior quadrants may be early signs of posterior staphyloma or its equivalent in guinea pigs. We also found that, although the refractions of eyes without FT shifted toward hyperopia and VCD became shorter during development, the refractions and VCDs of myopic eyes with FT remained unaltered. Thus, the eyes of myopic guinea pigs with FT presented more continuous and excessive myopia development than those of animals without FT. 

FT in Myopic Guinea Pigs

The spontaneously developing FT found in this study in guinea pigs was strongly related to myopic refraction and increases in VCD and AL, as is FT in early-stage pathologic myopia in humans. With ROC analysis, the cut-off refraction that predicted FT in guinea pigs was found to be ≤−4 D, the cut-off for VCD was 3.37 mm, and the cut-off for AL was 8.07 mm. These findings are reminiscent of the relationship between a higher degree of FT and longer AL that has been found by clinical studies in China,³⁶ Singapore,¹² and Japan.³⁷ 

Guinea pigs are sexually mature at 8 weeks of age; therefore, the ages of the 3- to 4-week-old animals used in this study correspond roughly to 5 to 8 years of age in humans. The prevalence of FT in young guinea pigs is similar to that in children. The prevalence of FT was 61% in young guinea pigs with SER ≤r−4 D but was only 8% or 5%, respectively, in guinea pigs with hyperopia or low myopia. In human beings, FT is also the most common fundus change in high-myopic children. The prevalence of FT in highly myopic adolescents has been reported to vary over a range of 50% to 94.3% in different studies,^13,38,39 a wide range that might be due to differences in age, ethnicity, and inclusion criteria in the different studies. In contrast, FT has been found to be rare in children with no myopia or with low myopia; for example, the prevalence of FT in children 10.29 ± 0.60 years old having low myopia (SER > −3 D) was found to be very low (3.1%).³⁸ These results indicate that FT might be an early sign of pathologic myopia in guinea pigs. Because of the young age of guinea pigs in this study, we rarely observed myopic maculopathy more severe than FT in our study, and it has not been reported in previous studies, either. In human beings, advanced myopic maculopathy is also uncommon among highly myopic children. Kobayashi et al.¹¹ found no choroidal neovascularization or geographic atrophy in 46 children (1–8 years old, 80 eyes) with high myopia. Koh et al.¹² observed FT in 18 out of 21 eyes of young Chinese men (age 21.8 ± 1.3 years) with SER < −10.0 D. In a study of highly myopic adolescents, He et al.¹³ found that the prevalence of FT reached 50%, whereas that of diffuse chorioretinal atrophy was only 0.08%. However, it is worth noting that, although clinically significant myopic maculopathy (worse than C2) was found to be uncommon among younger participants, the risk of clinically significant myopic maculopathy was increased disproportionately in children with early-onset high myopia.⁴⁰ 

Continuous and Excessive Myopia Development and Axial “Elongation” in Myopic Guinea Pigs With FT

The typical changes of pathologic myopia include continuous and excessive development of myopic refractive error and ocular axial elongation.¹ In guinea pig eyes without FT, refraction shifted toward hyperopia during the period from day 21 to day 43, and VCD consistently decreased. However, refraction and VCD were unaltered in eyes with FT, and AL increased more in myopic eyes with FT than in those without FT. These results demonstrate that spontaneous high myopia in guinea pig eyes with FT is characterized by continuous and excessive increases in myopic refraction and axial length, much as in human pathologic myopia. 

Changes of Eye Shape in High-Myopic Guinea Pigs

In human pathologic myopia, continuous and excessive ocular axial elongation can lead to posterior staphyloma,¹ and posterior staphylomas generally develop only after 40 years of age.¹⁴ Kobayashi et al.¹¹ found only one eye (1.3% of the study population) with posterior staphylomas in children 1 to 8 years old, and that belonged to a 7-year-old child with −16.5 D myopia. These results suggest that the incidence of posterior staphyloma in young guinea pigs’ eyes will be very low. Posterior staphyloma is an out-pouching of the sclera, resulting in curvature steeper than that of the surrounding eye wall.^6,41 For this reason, we assessed the shape of the eyeball to detect any early signs of posterior staphyloma formation, and we found that the curvature of the posterior eyeball increased and the temporal and superior quadrants of eyeball selectively expanded in myopic guinea pigs with FT, but not in myopic guinea pigs without FT. 

Data regarding the histopathological features and the mechanism responsible for posterior staphyloma in humans are inconsistent. Jonas et al.⁴² reported a localized Bruch's membrane defect and scleral thinning in all staphylomatous regions, but without changes in RPE density or choriocapillaris thickness and density. However, other studies found that posterior staphyloma was closely related to reductions in choriocapillaris perfusion and subfoveal choroidal thickness, as well as thinning of the retina and sclera.^43,44 As we have found in the present study, the selective expansion of the sclera in the posterior and temporal–superior regions of guinea pig eyes with FT might be an early sign of posterior staphyloma. The association of FT with thinning of the choroid suggests that the mechanisms underlying regionally selective expansion of the guinea pig eyeball could be similar to those in human posterior staphylomas, which are linked to thinning of the choroid.⁴⁵ Studies of older animals are needed to determine whether guinea pigs develop staphyloma. 

The regionally selective out-pouching and expansion of the posterior and temporal–superior regions of sclera in guinea pig eyes with FT are somewhat different from those in human staphylomas, which are not confined to specific regions of the eye.⁴¹ These differences between the locations of excessive scleral expansion in guinea pig eyes versus posterior staphylomas in human eyes could be due to the differences in eye placement and viewing direction in the two species. In guinea pigs, the visual stimuli viewed by the temporal–superior retina are likely to be near targets (ground and food), whereas those viewed by the nasal–inferior retina are biased toward distant targets (sky and sideways); in human beings, the projections of visual stimuli to different retinal regions are much more uniform. 

Limitations of This Guinea Pig Model of Pathologic Myopia

Only young guinea pigs were used in this study, and young myopic guinea pigs with FT can serve only as a model for early signs suggestive of incipient pathologic myopia. Thus, it will be important for future studies to assess whether FT will progress to severe myopic maculopathy and whether classical posterior staphyloma will develop in aged guinea pigs. In any case, however, FT is the first sign of myopic maculopathy, so early detection and intervention are critical for preventing progression of pathologic myopia. This model can be valuable also for revealing the underlying mechanisms of FT progression and so could contribute to the development of treatments for pathologic myopia. 

The grading of fundus structure was subjective and thus depended on the skills of the examiner. Measures were taken to improve the accuracy and repeatability of these assessments in that the examiners were trained, the images to be graded were masked with respect to refractive error and ocular dimensions, and both examiners graded these images twice. Second, the appearance of FT depends on the physiologic background pigmentation of the eye, and a few guinea pigs with very dark fundi were excluded from this study. 

The present study describes only the microanatomy of the fundus, eyeball shape, and refractive changes in myopic eyes with FT. These results indicate, in a novel animal model, an association or correlation between the myopia and fundus changes typical of human pathologic myopia. Further experiments designed to test hypotheses of underlying pathogenetic mechanisms will be needed to clarify whether there is a causal relationship between them. 

In conclusion, the eyes of young, spontaneously myopic guinea pigs have a high prevalence of FT, which is associated with higher myopia, and substantial increases in VCD and AL. Myopic guinea pig eyes with FT present characteristic signs of pathologic myopia, such as continuous and excessive myopia progression and ocular axial elongation, plus regionally selective excessive expansion of sclera at the posterior pole and in the temporal–superior quadrants. These results indicate that naturally myopic guinea pig eyes could be a useful model of pathologic myopia with early myopic maculopathy, in which the causal mechanisms could be identified and for which preventive or corrective therapies could be devised. 

Acknowledgments

The authors thank William K. Stell, PhD, MD, professor emeritus, Department of Cell Biology and Anatomy, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada, for thoughtful discussions and for critical reading and editing of the manuscript. 

Supported by grants from the National Natural Science Foundation of China (82271121, 82025009, 81830027, 81970833, U20A20364, and 81790641) and the Department of Science and Technology of Zhejiang Province (2021C03102 ). 

Disclosure: Y. Zhang, None; W. Tang, None; J. Liang, None; X. Zhou, None; S. Chen, None; Z. Zhi, None