Normal Ocular Development in Young Rhesus Monkeys (Macaca mulatta)

Abstract

Purpose

The purpose of this study was to characterize normal ocular development in infant monkeys and to establish both qualitative and quantitative relationships between human and monkey refractive development.

Methods

The subjects were 214 normal rhesus monkeys. Cross-sectional data were obtained from 204 monkeys at about 3 weeks of age and longitudinal data were obtained from 10 representative animals beginning at about 3 weeks of age for a period of up to 5 years. Ocular development was characterized via refractive status, corneal power, crystalline lens parameters, and the eye’s axial dimensions, which were determined by retinoscopy, keratometry, phakometry and A-scan ultrasonography, respectively.

Results

From birth to about 5 years of age, the growth curves for refractive error and most ocular components (excluding lens thickness and equivalent lens index) followed exponential trajectories and were highly coordinated between the two eyes. However, overall ocular growth was not a simple process of increasing the scale of each ocular component in a proportional manner. Instead the rates and relative amounts of change varied within and between ocular structures.

Conclusion

The configuration and contribution of the major ocular components in infant and adolescent monkey eyes are qualitatively and quantitatively very comparable to those in human eyes and their development proceeds in a similar manner in both species. As a consequence, in both species the adolescent eye is not simply a scaled version of the infant eye.

INTRODUCTION

In many species, neonates are typically more hyperopic than adults and as a group exhibit a greater amount of between subject variability in refractive errors. However, with time the two eyes of most neonates grow in a highly coordinated manner toward what is believed to be the optimal refractive state, typically a low degree of hyperopia, and the distribution of refractive errors becomes more leptokurtic (Hirsch & Weymouth, 1991, Sorsby et al., 1961).

Emmetropization is the process that guides eye growth toward this optimal refractive state. Emmetropization is an active process that is regulated by visual feedback associated with the eye’s refractive state (Norton & Siegwart, 1995, Smith, 1998, Wallman & Winawer, 2004, Wildsoet, 1997). The mechanisms responsible for the phenomenon of emmetropization are active beyond infancy and probably play a role through early adult life in maintaining the optimal refractive state and an appropriate interocular balance in refractive errors (Papastergiou et al., 1998, Siegwart & Norton, 1998, Troilo et al., 2000, Zhong et al., 2004). Individual differences in the operational properties of the mechanisms responsible for emmetropization and/or environmental signals operating via the mechanisms that regulate emmetropization are thought to contribute to the development of common refractive errors like juvenile onset myopia (Smith et al., 1999, Troilo et al., 2000, Wallman & Winawer, 2004, Wildsoet, 1997). Therefore, understanding the normal emmetropization process is an important step in understanding the pathophysiology of refractive errors.

Animal models have been invaluable in efforts to identify the manner in which the visual environment affects refractive development and to characterize the nature of the ocular changes that occur in response to the environment. Because of the close similarities between the anatomy of human and macaque monkey eyes, it is reasonable to expect that data derived from monkeys can be generalized to humans. However, detailed knowledge of normal ocular and refractive development in monkeys is needed to fully appreciate the implications of these animal studies for human refractive development. Ideally, one would like to know not only the normal course of refractive development in monkeys, but also the developmental courses for each of the eye’s major optical and axial components and the degree and nature of any interocular differences in development.

A number of studies have provided useful information on refractive development in monkeys. Young’s cross-sectional study (1964) of a large number of rhesus monkeys demonstrated that the distribution of refractive errors for rhesus monkeys was comparable to that for humans; in particular, in both species the distributions show considerable leptokurtosis for predominately adolescent populations and similar mean refractive errors. Thus, the outcome of emmetropization is analogous in humans and monkeys. In addition, both cross-sectional and longitudinal studies (De Rousseau & Bito, 1981, Greene, 1990, Kaufman et al., 1981, Kiely et al., 1987, Koretz et al., 1987, Young, 1964, Young & Leary, 1991) have provided developmental data on a variety of ocular components in monkeys with Bradley et al.’s study (1999b) providing the most thorough description to date of the changes that occur in the macaque monkey eye during early emmetropization. However, methodological concerns raise questions about the validity of some of the data in these studies. For example, many of the monkeys studied by Kiely et al. (1987) had the vision to one eye experimentally altered during the observation period and, thus, the data from these animals were potentially confounded by interocular treatment effects (Bradley et al., 1999a, Smith & Hung, 2000, Smith et al., 2002). In addition, limitations in the measurement range of the keratometer used by Bradley et al. (1999b) may have truncated the corneal curvature data from their youngest animals. Moreover, none of these studies reported data on the optical properties of the crystalline lens.

Knowledge of the developmental changes that take place in the crystalline lens and the relationship between these changes and the developmental trajectories of the eye’s other optical and axial components are critical (Mutti et al., 1998). Especially because the absolute power changes in the human lens are 2–3 times greater than those that take place in the cornea and the changes in lens power occur over a much longer age range (Mutti et al., 1998). In addition, it has been argued that alterations in axial growth associated with anomalous refractive errors affect the axial and optical properties of the lens and vice versa (Mutti et al., 1998, van Alphen, 1961). Several previous studies of crystalline lens development in young monkeys have shown that despite an increase in overall eye size, the crystalline lens becomes thinner at least during the first 4–5 years of life (De Rousseau & Bito, 1981, Denlinger et al., 1980, Koretz et al., 1987). However, the sampling intervals in these studies were too coarse to provide a detailed description of early lens development and other critical optical and axial properties of the eye were not measured.

The purpose of our study was to characterize the emmetropization process in normal rhesus monkeys in sufficient detail to support the development of schematic eye models for infant monkeys, to establish reference data for assessing abnormal refractive development in monkeys, and to allow quantitative comparisons between humans and monkeys of the time courses of development for each major ocular component.

MATERIALS AND METHODS

Subjects

Data are presented for 214 infant rhesus monkeys (Macaca mulatta). The animals were obtained at 1 to 3 weeks of age and housed in our primate nursery that was maintained on a 12-hour light/12-hour dark lighting cycle. The details of the nursery care for our infant monkeys have been described previously (Hung et al., 1995, Smith & Hung, 1999). All of the rearing and experimental procedures were reviewed and approved by the University of Houston’s Institutional Animal Care and Use Committee and were in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Cross-sectional data on refractive error, corneal power, axial dimensions and crystalline lens phakometry were obtained from both eyes of all subjects at about 3 weeks of age (mean = 24.5 ± 4.7 days). Longitudinal biometric data on refractive and ocular development were obtained from a subset of 10 infants that were selected randomly. For the animals in the longitudinal group, measurements were made every 2–4 weeks during the first year of life and then at 1–12 month intervals thereafter. All animals exhibited clear media and showed no signs of ocular pathology. Data on refractive error and the eye’s axial dimensions have been reported for most of the animals in the cross-sectional group and 5 of the monkeys studied longitudinally (Hung et al., 1995, Smith et al., 2003, Smith et al., 2005, Smith & Hung, 1999, Smith & Hung, 2000)

Biometric Measurements

Many of the methods that were used to assess refractive development in the infant monkeys have been described in detail previously (Smith & Hung, 1999). To make these measurements, the monkeys were anesthetized with intramuscular injections of ketamine hydrochloride (15–20 mg/kg) and acepromazine maleate (0.15–0.2 mg/kg) and topically instilled 0.5% tetracaine hydrochloride. Cycloplegia was achieved by topically instilling 2 drops of 1% tropicamide, 20–30 minutes before performing any measurement that would potentially be affected by the level of accommodation. While the measurements were being taken, the eyelids were gently held apart by a custom made speculum and the tear film was maintained by the frequent application of a commercially available ocular irrigation solution.

The refractive status of each eye, both the spherical and cylindrical components, was measured along the pupillary axis by two experienced investigators using a streak retinoscope and averaged (Harris, 1988). An eye’s refractive error was defined as the mean spherical-equivalent, spectacle-plane refractive correction (11.5 mm vertex distance).

Anterior radius of curvature of the cornea was measured with a hand-held keratometer (Alcon Auto-keratometer; Alcon Systems Inc, St Louis, MO) and/or a videotopographer (EyeSys 2000; EyeSys technologies Inc, Houston, TX). We have previously shown that both instruments provide repeatable and comparable measures of corneal curvature in infant monkeys (Kee et al., 2002). It was assumed that the cornea was effectively a single spherical refracting surface and total corneal refracting power was calculated using an assumed image space refractive index of 1.3375.

The axial dimensions of the eye, including anterior chamber and vitreous chamber depths, lens thickness, and the sum of these, axial length, were measured with a high-frequency A-scan system using a focused, 30-MHz polymer transducer (Panametrics, Waltrham, MA) digitized at 100 MHz (model 8100 A/D board; Sonix, Springfield, VA). The transducer was coupled to the eye using a closed, water-filled interface. A three-axis positioner mounted on a slit lamp base was used to align the transducer to simultaneously maximize the echoes from the major optical components and the vitreous-retinal interface. Eight to 10 readings were recorded and averaged. The average velocities for ultrasound in human eyes were used to calculate intraocular distances (Shammas et al., 1998).

The curvatures of the anterior and posterior lens surfaces were measured by video phakometry (Mutti et al., 1992). Specifically, the equivalent radii of curvature for the anterior and posterior surfaces were derived by measuring the apparent sizes of Purkinje Images I, III, and IV produced by the collimated light from two point sources that were optically imaged at infinity. The angle between the light sources and the CCD camera system (Cohu 6415 camera with a 55 mm, F1.4 lens on a 2X teleconverter) was fixed at 20 deg. During the measurements, the camera and source lights were positioned on opposite sides of the eye’s approximate optical axis resulting in a lateral separation (for measurements of the vertical meridian) of Purkinje images I, III, and IV in the center of the pupil. The camera was focused on each of the Purkinje images separately. The camera’s telecentric optical system minimized angular magnification effects of small focusing errors. Video images were stored via a frame grabber and imaging software was used to measure the sizes of the digitized images. Data were obtained for the 45, 90, and 135 deg meridians and then averaged. At least 2 clear frames were measured for each image. The equivalent radii for the lens surfaces were determined by comparing the sizes of the Purkinje images to the catoptric images obtained from a series of precision ball bearings (Mutti et al., 1992). With knowledge of the eye’s refractive error, corneal power, and axial dimensions, the anterior radius of curvature for the crystalline lens was calculated by paraxial ray tracing. For the posterior lens surface, an iterative procedure was used to find the posterior lens radius of curvature and equivalent refractive index of the lens which produced agreement between the measured refractive error and that calculated from ocular component values. The equivalent lens power was then calculated using the thick lens formula (Keating, 2002).

Statistical Analysis

A one-sample Kolmogorov-Smirnov Test of Composite Normality was used to test whether the refractive error and ocular components were normally distributed for the infant rhesus monkeys, i.e., the first measurements for all monkeys. A paired t-test was used for interocular comparisons.

In order to describe the longitudinal changes in individual ocular components, a locally weighted regression scatter plot smoothing method (LOESS) was used to generate developmental curves and to describe growth data for refractive error, corneal power, the eye’s axial dimensions, and the properties of the crystalline lens. LOESS is a nonparametric smoothing algorithm that allows data to express itself in a trend without initial mathematical assumptions. LOESS was most applicable for our monkey data because the data were irregularly spaced and there were variable numbers of observations at each point in time (Mose et al., 1992). The LOESS statistical analyses were conducted using S-plus 6 software (Insightful Corporation, Seattle, WA).

Nonlinear regression analysis was used to compare the rates of ocular development between humans and rhesus monkeys. Nonlinear regression analysis requires a predetermined model. Four distinct growth tissue patterns have been described for human development (e.g., lymphoid tissue, neural tissue, reproductive tissue, and general growth; (Cameron, 2002)). Ocular growth follows the developmental pattern for neural tissue, starting with an initial value at birth and gradually increasing to an asymptote around 18 years of age with a constantly decreasing rate of growth (Cameron, 2002). It should be noted that no arbitrary mathematical model will fit the data perfectly. However, based on a mathematical model of emmetropization, Carroll (1982) (Carroll, 1982) has argued that the growth of individual ocular parameters can be described by a set of exponential curves. Preliminary inspection of longitudinal growth data revealed that the data from both human and monkey eyes approximated an asymptotic regression model. Asymptotic regression models are used to describe a response y that approaches a horizontal asymptote as x approaches +∞ and can be described by the following formula:

$$y(x)={\mathrm{\Phi }}_1+({\mathrm{\Phi }}_2-{\mathrm{\Phi }}_1)exp(-exp({\mathrm{\Phi }}_3)x)$$

where y represents the ametropia or the ocular component in question, x represents age, exp(a) = e^a, Φ₁ is the asymptote as x → ∞, Φ₂ is y(0) and Φ₃ is the logarithm of the rate constant that is used to calculate the time required to reach 50% of the asymptotic value (t_0.5 = log2/exp (Φ₃)).

Schematic Eyes

Schematic eyes were constructed with the refractive errors and ocular component values obtained via our biometric measurements using the methods described by Bennett and Rabbetts (1989). Specifically, to construct the model eyes it was assumed: 1) that the refractive indices of the crystalline lens and the aqueous and vitreous humors were homogeneous; 2) that the optical surfaces were spherical and coaxially aligned and 3) that the cornea was a single spherical refracting surface separating air from the aqueous humor. Schematic eye models were constructed for young rhesus monkeys at 3 weeks and 3 years of ages using the average data from the longitudinal group.

RESULTS

At the initial measurement, there were no statistically significant differences in the average spherical-equivalent refractive errors, the surface curvatures of the cornea and lens, or the axial dimensions of any ocular compartment between the two eyes of our infant monkeys (Student t-test, p values ranged from 0.06 for equivalent lens index to 0.92 for refractive error). Consequently, except for the interocular comparisons that include the results from both eyes, only data for the right eyes are reported.

As observed previously in pig-tailed macaques by Young & Leary (1991), male rhesus monkeys generally had larger eyes than females. While there were no significant differences in refractive error (male +4.29 ± 1.60 D, n = 111; female +4.11 ± 1.28 D, n = 90; p = 0.38) or anterior chamber depth (male 2.57 ± 0.26 mm, female 2.58 ± 0.30 mm, p = 0.76) between males and females at the first measurement session, male rhesus monkeys had longer axial lengths (14.57 ± 0.45 vs. 14.38 ± 0.46 mm, p = 0.0036), longer vitreous chambers (8.63 ± 0.28 vs. 8.50 ± 0.29 mm, p = 0.0017), thicker crystalline lenses (3.37 ± 0.18 vs. 3.29 ± 0.17 mm, p = 0.0013) and flatter cornea curvatures (+61.53 ± 1.65 vs. +62.97 ± 1.55 D, p < 0.0001). However, to facilitate comparisons with previous studies, which usually combined data from both males and females, and to increase our sample size, the data for both genders were combined.

Refractive errors and ocular components in infant monkeys

Figure 1 shows the frequency distributions for refractive error, corneal power and the eye’s axial dimensions for all 214 monkeys at 24.5 ± 4.7 days of age. Our infant monkeys exhibited a broad range of predominantly hyperopic ametropias (−0.06 to +9.25 D). The average refractive error was +4.22 ± 1.46 D and the distribution of refractive errors was well described by a normal Gaussian distribution (solid line; one-sample Kolmogorov-Smirnov Test of composite normality; also see Table 1). Similarly, the frequency distributions for corneal power, total axial length and the axial dimensions of the anterior chamber, crystalline lens, and vitreous chamber did not differ significantly from normal distributions, implying that early in life, refractive error, corneal power and the eye’s axial dimensions were randomly distributed in the population of infant monkeys (Figure 1, Table 1).

Figure 1.

Frequency distributions for spherical-equivalent refractive error (A), average corneal refracting power (B), axial length (C), vitreous chamber depth (D), lens thickness (E) and anterior chamber depth (F) obtained for the 214 infant monkeys at 24.5 ± 4.7 days of age. The open bars represent the data for the 10 normal monkeys that were followed longitudinally and the filled bars represent data for all our other infant monkeys prior to initiating any experimental treatment. The solid lines represent the best fitting normal distributions. The insert in the upper right corner of panel A shows the distribution of refractive errors for the 10 longitudinal monkeys at about 1 year of age.

Table 1.

Refractive Error and Ocular Components Comparision for Cross-sectional and Longitudinal Subjects

		Age (days)	Refractive Error (D)	Corneal Power (D)	Axial Length (mm)	Vitreous Chamber Depth (mm)	Lens Thickness (mm)	Anterior Chamber Depth (mm)
Cross-sectional Group (n=204) except Cornea Power (n=139)	Range	14 to 39	−0.06 to + 9.25	58.29 to 67.43	13.25 to 15.59	7.78 to 9.40	2.55 to 3.91	1.85 to 3.37
	Mean ± SD	24.54 ± 4.71	4.17 ± 1.52	61.83 ± 1.87	14.49 ± 0.47	8.58 ± 0.29	3.34 ± 0.20	2.58 ± 0.28
	Normality Test p value	N/A	p > 0.15	p >0.15	p > 0.15	p > 0.15	p > 0.15	p > 0.15
Longitudinal Group (n=10) 3 weeks age	Range	17 to 24	1.56 to 5.50	61.82 to 65.09	13.68 to 15.05	8.18 to 8.80	3.12 to 3.54	2.20 to 2.86
	Mean ± SD	21.4 ± 1.90	3.78 ± 1.23	62.64 ± 1.60	14.50 ± 0.41	8.45 ± 0.22	3.33 ± 0.14	2.70 ± 0.20
Longitudinal Group (n=10) 1 year age	Range	325 to 368	1.19 to 3.06	53.62 to 57.17	16.39 to 17.57	9.84 to 10.99	3.46 to 3.70	2.87 to 3.30
	Mean ± SD	343.00 ± 13.25	1.83 ± 0.59	54.49 ± 1.10	17.14 ± 0.35	10.48 ± 0.31	3.59 ± 0.09	3.06 ± 0.13

The initial measurements for the 10 infant monkeys that were followed longitudinally are represented by the open bars in Figure 1. There were no obvious differences in refractive error or any of the measured ocular components between the subgroup of monkeys that was followed longitudinally and the larger group of normal infants, i.e. the 10 monkeys in the longitudinal group were representative of the entire colony (see Table 1).

Longitudinal Changes in Refractive Error and the Major Ocular Components

The refractive errors for the right eyes and the interocular differences in refractive error are plotted as a function of age for each of the monkeys in Figure 2. The thick solid lines represent the LOESS growth curves. In Figure 2A, the LOESS function for refractive error decreased relatively rapidly from a refractive correction of +3.30 D at 17 days of age to +1.58 D at 500 days of age and then more slowly to an asymptote of +1.19 D at around 1500 days of age. Refractive error reached 50% of the four-year decrease by about 250 days of age and the 90% level by 1000 days of age. By 1 year of age, the distribution of refractive errors had also narrowed considerably (see the insert in the upper right corner of Figure 1A).

Figure 2.

Spherical-equivalent refractive error (A, B, C) and the degree of anisometropia (D; absolute value) plotted as a function of age for individual animals. Panels A and D include the data from all 10 of the monkeys in the longitudinal group (symbols connected by dotted lines in A; filled symbols in D) and all of the monkeys in the cross-sectional group (small open diamonds. Note that ages for the animals in the cross-sectional group have been represented as negative values to facilitate comparisons between the cross-sectional and longitudinal groups). Panels B and C show longitudinal refractive-error data for 4 representative monkeys (filled symbols). In panels A–C, the thick solid lines represent the LOESS growth curves generated for the monkeys in the longitudinal group.

LOESS functions have been used previously to summarize longitudinal changes in refractive error in infant monkeys (Bradley et al., 1999b) and in this study the calculated functions provided a reasonable description of the longitudinal changes in refractive error for some individual monkeys. However, several monkeys showed changes in refractive error that did not follow the simple monotonic trends represented by the LOESS function. For example, Monkeys ORS and RIS (Figure 2B) showed the most obvious departures from the LOESS growth curve. These animals were hyperopic (ORS: +4.88D, RIS: +2.81D) at three weeks of age and maintained their hyperopia throughout the observation period without any sign of emmetropization. Although the failure of emmetropization in Monkey ORS and its departure from a normal course in Monkey RIS were obvious, there was nothing in the early histories of these animals, other than their refractive errors, that distinguished them from the other infant monkeys. It is interesting that emmetropization did not proceed as expected in Monkeys ORS and RIS because the dimensions for all of the ocular components for these monkeys were well within the range of values for the other monkeys that exhibited emmetropization.

In other monkeys, the LOESS function adequately captured the course of emmetropization after about 100 days of age, e.g., Monkeys QUI and PAG (Figure 2C), but frequently underestimated the dramatic early changes seen in many monkeys. In this respect, linear regression analysis of the changes in refractive error and axial length during the first 6 months of life indicated that the early rates of emmetropization and axial elongation for individual animals were directly related to the initial ametropia. Specifically, infant monkeys with higher initial degrees of hyperopia exhibited faster reductions in hyperopia (r² = 0.54, p = 0.02) and faster axial elongation rates (r² = 0.36, p = 0.07). A similar phenomenon has been observed in human infants (Mutti et al., 2005).

Throughout the observation period, the refractive errors for the two eyes of a given animal were well matched (Fig 2D). In the cross-sectional group, the absolute levels of anisometropia varied between 0 and 1.25 D with an average of 0.15 ± 0.18 D. Only 4 out of 204 available comparisons of anisometropia exceeded 0.50 D (open diamonds, note, the cross sectional data are shown as negative age values to facilitate comparisons). Out of the 255 available comparisons for the longitudinal group (filled dots) the degree of anisometropia exceeded 0.50 D on only 7 occasions. There was a tendency to find larger interocular differences during the early period of rapid emmetropization (e.g., the first 200 days). However, in the longitudinal group, the degree of anisometropia at around 20 days of age was not significantly higher than that observed at 1500 days of age (p = 0.34). Hence, at any time during the first 4–5 years of life of a rhesus monkey, interocular differences in refractive error greater than 0.50 D (i.e., > mean ± 2 SDs) are potentially significant.

Between 3 weeks and 4 years of age, the eye’s axial length increased by an average of 3.90 mm from 14.68 mm to 18.58 mm (Figure 3A) with 50% of the four-year growth occurring by 231 days of age, roughly around the same time that refractive error has completed 50% of its four-year changes. After about 800 days of age, axial length had completed 90% of its four-year growth and the rate of axial growth was relatively constant for the rest of the observation period. With the exception of one monkey (monkey QUE, open hexagons), the calculated LOESS function provided a reasonable description of the axial growth for individual animals (Note, Monkey QUE also had a very steep cornea (Figure 5), but exhibited normal emmetropization. QUE’s short axial length may reflect compensating growth associated with emmetropization). Increases in vitreous chamber depth and anterior chamber depth accounted for 82% and 22% of the four-year increase in axial length, respectively, and the LOESS growth curves for both the anterior and vitreous chambers were qualitatively similar in shape to that for overall axial length. In contrast, the changes in lens thickness followed a very different growth trajectory. According to the LOESS estimates, from 21 to 280 days of age, lens thickness increased by an average of 0.06 mm. Subsequently, there was a decrease in lens thickness so that by the end of the observation period, the average lens thickness was 0.15 mm thinner than it was at 3 weeks of age. However, the average lens thickness of monkeys at 1500 days of age was not significantly thinner than that at 21 days of age (Two- sample t-test, p = 0.11).

Figure 3.

Axial length (A), vitreous chamber depth (B), lens thickness (C) and anterior chamber depth (D) plotted as function of age for individual monkeys. The open diamonds represent the cross-sectional data and are represented as negative age values to facilitate comparison with the longitudinal data (open symbols connected by dotted lines). The thick solid lines represent the LOESS growth curves generated for the monkeys in the longitudinal group.

Figure 5.

Corneal power (A) and absolute interocular differences in corneal power (B) plotted as a function of age for individual monkeys. In panels A and B, the open diamonds represent the cross-sectional data and are represented as negative age values to facilitate comparison with the longitudinal data (A, open symbols connected by dotted lines; B, filled circles). The thick solid line in panel A represents the LOESS growth curve generated for the monkeys in the longitudinal group.

The axial dimensions of the two eyes of a given monkey were well matched shortly after birth (Figure 4). For example, in the cross-sectional group the absolute interocular differences in axial length at three weeks of age ranged from 0 to 0.45 mm with an average difference of 0.11 ± 0.09 mm. Although the interocular differences in axial dimensions generally decreased with age, the average interocular differences for total axial length and for the individual axial compartments at 21 days of age were not significantly different from those at 1 year of age (Two sample t-test, axial length p = 0.71, anterior chamber p = 0.75; lens thickness p = 0.17; vitreous chamber p = 0.38).

Figure 4.

Interocular differences in axial length (A), vitreous chamber depth (B), lens thickness (C) and anterior chamber depth (D) plotted as a function of age for individual animals. The open diamonds represent the cross-sectional data and are represented as negative age values to facilitate comparison with the longitudinal data (filled circles).

From the LOESS curve in Figure 5A, the average corneal power was +63.08 D at 3 weeks of age and declined in a rapid monotonic fashion to an asymptote of about +53.75 D at about 800 days of age (Figure 5A); 50% of the total decrease in power took place by 84 days of age and 90% took place by about 295 days. With the exception of monkey QUE (open hexagons), the LOESS function provided a good description of the longitudinal changes in corneal power for individual animals. In the cross-sectional group, the absolute interocular differences in corneal power ranged from 0 to 1.59 D with an average of 0.42 ± 0.33 D at 3 weeks of age. At 1 year of age the interocular differences in corneal power were significantly smaller than those at 21 days of age (p = 0.04).

Both the anterior and posterior lens radii increased in a monotonic fashion with age (Figure 6). The LOESS growth curve indicated that the anterior lens radius of curvature increased from 5.65 mm at 21 days of age to 8.96 mm at 1500 days with 50% of the increase taking place by 318 days of age and with 90% of the increase being completed by 1000 days of age. The posterior lens surface, which was steeper at all ages, exhibited smaller absolute changes in radius. Between 21 and 1500 days of age, the posterior lens radius increased from 4.38 to 6.50 mm with 50% of the increase taking place by 340 days of age and with 90% of the increase being completed by 1065 days of age.

Figure 6.

Anterior lens radius of curvature (A), posterior lens radius of curvature (B), equivalent lens power (C) and equivalent lens index (D) plotted as a function of age for the 10 monkeys in the longitudinal group. See Figure 3 for details.

The overall power changes in the lens were substantially larger (about 21 D) than those observed for the cornea. Specifically, the equivalent lens power decreased from +56.03 D at 21 days of age to +35.19 D at 1500 days of age with 50% of these changes taking place by 332 days of age and with 90% of the increase being completed by 975 days of age (Figure 6C). However, it is important to note that the growth trajectory for equivalent lens power had not reached an asymptote by the end of the observation period. The calculated equivalent lens index showed a more complicated growth function (Figure 6D). The equivalent index increased from 1.4799 to 1.4837 between 21 and 232 days of age; however this change was not significant (Two sample t-test, p = 0.086). Thereafter, we observed a decline in equivalent index to 1.4712 at 1500 days of age, but again, the difference between the average index at 223 days and 1500 days was not statistically significant (Two sample t-test, p=0.10).

The relative growth rates for all of the major ocular components are shown in Figure 7. For comparison purposes, the local slopes of the LOESS functions for each ocular component were calculated for successive 10-day intervals and expressed in terms of the average daily changes at fixed age points. As illustrated in the top panel, the eye achieves a relatively stable refractive state before any of the major ocular components stop growing, i.e., very early in life, the emmetropization process achieves and then subsequently maintains the presumed optimal refractive state despite significant ongoing changes in the eye’s optical power and overall axial length. This pattern of results is in agreement with the idea that there is still active regulation of ocular components and their growth after emmetropia is initially achieved. Of the eye’s two major optical components, the changes in the equivalent power of the crystalline lens were larger and faster and these changes in lens power continued over a much longer period of time than the changes in total corneal power. All of the major individual axial components exhibited different growth rates. In particular, the vitreous chamber exhibited the highest elongation rate and after about 600 days of age was largely responsible for any subsequent increases in overall axial length. Initially, the anterior lens surface flattened at a faster rate than the posterior surface, thus increasing the asymmetries between the anterior and posterior lens surface curvatures. However, at the end of the observation period, the radii of curvature for the anterior and posterior surfaces were increasing at relatively similar rates. It is important to note that there were clear differences in the age-dependent changes in growth rate for the various ocular components and that these differences in growth rate clearly demonstrate that the eye does not grow in a simple proportional manner. As a consequence, the adolescent monkey eye is not simply a scaled version of the infant monkey eye. To further illustrate the point that different segments of the eye develop at different rates, we plotted the relative longitudinal changes in individual ocular components as a function of age. The changes were normalized to unity at the beginning of the observation period, i.e. about 3 weeks of age. Figure 8 shows that while the anterior and vitreous chambers expanded by approximately the same relative amount and speed, axial length showed a smaller, slower expansion and lens thickness showed no increases, but instead a small decrease. The relative decrease in corneal power was smaller and was completed earlier than that for crystalline lens power, again demonstrating that within the anterior segment, the cornea and lens each have their own development trajectories. Even within the crystalline lens itself, the anterior lens curvature and the posterior lens curvature showed different growth courses with the anterior lens curvature flattening proportionally more and at a relatively faster pace than that of the posterior lens surface.

Figure 7.

Longitudinal growth rate curves for refractive error, corneal power, and lens power (A), axial length, anterior and vitreous chamber depth and lens thickness (B), and anterior and posterior lens radii of curvature (C) for the 10 monkeys in the longitudinal group.

Figure 8.

Relative changes in refractive error and the individual ocular components during the observation period. The data were normalized to the values of the LOESS functions at 21 days of age. A. Axial dimensions including axial length, anterior and vitreous chamber depths, and lens thickness. B. Corneal power and equivalent lens power. C. Anterior and posterior lens radii of curvature and equivalent lens index.

Schematic eyes of young rhesus monkeys

Based on the average refractive errors and ocular parameters of our rhesus monkeys at 3 weeks and 4 years of age, two schematic eye models were constructed. Full morphometric information for the models along with the locations of the cardinal points is presented in Table 2. Figure 9 shows the scaled schematic representations of the monkey model eyes at 21 and 1745 days of age, including the locations of the calculated principal planes, nodal points and focal points. The schematic eyes for human infants (Lotmar, 1976) and adults (Bennett & Rabbetts, 1989) are also presented for comparison purposes.

Table 2.

Schematic Eyes for Young Rhesus Monkeys and Human Schematic Eye

		21 days (N=10)	1745 days (N=4)	Human Baby (Lotmar 1976)	Human Schematic Eye (Bennett & Rabbetts)
Radii of curvature
cornea	r1	5.64	6.29	6.80	7.80
crystalline: first surface	r2	5.19	8.43	5.00	11.00
crystalline: second surface	r3	−4.24	−5.87	−3.70	−6.48
Axial separations
depth of anterior chamber	d1	2.31	3.40	2.60	3.60
thickness of crystalline lens	d2	3.43	3.35	3.70	3.70
depth of vitreous chamber	d3	8.41	11.93	11.00	16.79
overall axial length		14.15	18.67	17.30	24.09
Mean refractive indices
air	n1	1.0000	1.0000	1.0000	1.0000
aqueous humour	n2	1.3360	1.3360	1.3360	1.3360
crystalline	n3	1.4742	1.4470	1.4300	1.4220
vitreous humour	n4	1.3360	1.3360	1.3340	1.3360
Surface powers
cornea	F1	59.59	53.44	49.41	43.08
crystalline: first surface	F2	26.61	13.17	18.80	7.82
crystalline: second surface	F3	32.59	18.92	25.41	13.28
Equivalent powers
crystalline	FL	57.18	31.52	42.97	20.83
eye	Fo	106.37	78.33	85.00	60.00
Equivalent focal lengths of eye
first (PF)	fo	−9.40	−12.77	−11.76	−16.67
second (P′F′)	f′o	12.56	17.06	15.72	22.27
Distances from corneal vertex
first principal point	A1P	1.64	1.58	1.76	1.51
second principal point	A1P′	2.01	1.87	2.09	1.82
first nodal point	A1N	4.80	5.87	5.71	7.11
second nodal point	A1N′	5.16	6.16	6.04	7.42
first principal focus	A1F	−7.76	−11.18	−10.01	−15.16
second principal focus	A1F′	14.57	18.92	17.81	24.09
Refractive state (D)	K	3.69	1.65	1.27	1.20

Figure 9.

Schematic eye model for rhesus monkeys at 21 (A) and 1745 days of age (C), new born human infants (B, modified from (Lotmar, 1976), and young adult humans (D, from (Bennett & Rabbetts, 1989)). The positions of the anterior and posterior focal points (F, F′), principal planes (P, P′), and nodal points (N, N′) are shown in each schematic eye.

Although there are obvious differences between the absolute sizes of human and monkey eyes, the positions of the cardinal points relative to the eye’s overall axial length are quite similar. As can be seen in Table 3, monkey and human infants exhibited similar position ratios for all of the cardinal points and the age-dependent changes in the relative positions for all of the cardinal points were also similar. For example, the position of the first principal point relative to total axial length in rhesus monkeys decreased from 11.6% at 3 weeks to 8.5% at 4 years of age while for humans it decreased from 10.2% in infants to 6.3% in adults.

Table 3.

Relative Cardinal Positions to the Axial Length

	21 days (N=10)	1745 days (N=4)	Human Baby (Lotmar 1976)	Human Schematic Eye (Bennett & Rabbetts)
A1P	11.6%	8.5%	10.2%	6.3%
A1P′	14.2%	10.0%	12.1%	7.6%
A1N	33.9%	31.4%	33.0%	29.5%
A1N′	36.5%	33.0%	34.9%	30.8%
A1F	−54.8%	−59.9%	−57.9%	−62.9%
A1F′	103.0%	101.4%	102.9%	100.0%

Comparisons between monkey and human growth curves

For various ocular parameters, nonlinear regression functions were fitted to our monkey data and to a collection of human data. For comparison purposes, ametropias were expressed as the eye’s optical error in order to ensure that the function increased in a positive direction with age. As shown in Figure 10, the exponential function adequately fit the human refractive error data with a t_0.5 value (the time required to reach 50% of the asymptotic value) of 276 days. For the monkeys, the exponential function also adequately fit the data with a t_0.5 value of 213 days. The ratio between the t_0.5 values for monkeys and humans was 1:1.3, roughly 1:1, indicating that emmetropization operates in a similar time scale in humans and monkeys.

Figure 10.

A Comparison of refractive-error development between humans (A) and rhesus monkeys (B). The human data were obtained from Zadnik et al. (2003), Garner et al. (1995), Wood et al. (1995), Saudners (1995), Pennie et al. (2001), Atkinson et al. (1996), Mayer et al. (2001), Edwards (1991), Thompson (1995), Sorsby et al. (1961), Larsen (1971a), Garner et al. (1990), Wood et al. (1996), Gwiazda et al. (1993). The thick solid lines represent the best fitting nonlinear regression curves.

Figures 11 and 12 compare the longitudinal human and monkey data for axial length and corneal power and anterior chamber and vitreous chamber depth, respectively. In each case the exponential functions fit the data for both monkeys and humans well. For both humans and monkeys the t_0.5 values were distinct and different for each ocular component; however, the ratio of the t_0.5 values between monkeys and humans were relatively constant. For example, for humans, the t_0.5 values for axial length, corneal power, anterior chamber depth and vitreous chamber depth were 584, 251, 815 and 384 days, respectively. The corresponding t_0.5 values for monkeys were 196, 75, 258 and 133 days respectively, which yielded monkey-to-human t_0.5 ratios of 1:3.0, 1:3.4, 1:3.2 and 1:2.9 for axial length, corneal power, anterior chamber depth and vitreous chamber depth, respectively. In each case, the major ocular component in the monkey eye matured approximately 3 times faster than the analogous component in the human eye.

Figure 11.

A comparison of axial length (A) and corneal radius development (B) between rhesus monkeys and humans. The human axial length data were obtained from Zadnik et al. (2003), Sorsby et al. (1961), Larsen (1971d), Garner et al. (1990), Hellstrom et al. (1997). The human corneal curvature data were obtained from Zadnik et al. (2003), Sorsby et al. (1961), Garner et al. (1990), Garner et al. (1995), Inagaki (1986), Asbell et al. (1990), Woodruff (1971), Borish (1970). The thick solid lines represent the best fitting nonlinear regression curves.

Figure 12.

A Comparison of vitreous chamber (A) and anterior chamber depth development (B) between rhesus monkeys and humans. The human vitreous chamber data were obtained from Zadnik et al. (2003), Garner et al. (1995), Larsen (1971b). The human anterior chamber data were obtained from Zadnik et al. (2003), Sorsby et al. (1961), Larsen (1971a), Garner et al. (1995). The thick solid lines represent the best fitting nonlinear regression curves.

DISCUSSION

Infant monkey eyes achieve the optimal refractive state quickly and then maintain relatively stable refractive errors despite rapid and sizeable changes in the eye’s major optical and axial components. With the exception of lens thickness, all of the eye’s major axial and optical components exhibited monotonic growth patterns with corneal and crystalline lens powers following exponential decays while the anterior chamber, vitreous chamber, and total axial length follow exponential growth trajectories. However, overall ocular growth was not a simple process of increasing the scale of each ocular component in a proportional manner. Instead the rates and relative amounts of change varied within and between ocular structures. For example, within the crystalline lens the changes in the anterior surface curvature were larger and took place faster than those for the posterior surface and the crystalline lens exhibited much larger changes in total power that occurred over a much longer time scale than the changes in corneal power. Despite the complex nature of ocular growth, each of the ocular components was generally well matched in the two eyes at all ages, especially refractive error.

Comparisons with previous monkey studies

Figure 13 compares the findings of this study with the data available from previous studies of early ocular development in macaque monkeys. Qualitatively, many aspects of the developmental changes are similar across studies; however, there are obvious quantitative differences. For example, all studies report that the average neonates are moderately hyperopic and that emmetropization occurs quickly and is largely complete by about 500 days. But, there are differences in the degree of hyperopia observed during early development and the end point or target refractive error of the emmetropization process.

Figure 13.

Comparisons between the longitudinal data from our monkeys for refractive error (A), axial length (B), corneal power (C), and lens thickness (D) with data from previous studies of macaque monkeys. The symbols legends in each plot indicate the source of the data (Bradley et al.,(1999b); Young & Leary, (1991), Kiely et al.,(1987), De Rousseau & Bito,(1981), Denlinger et al.,(1980), Koretz et al.,(1987)).

We have previously shown that our retinoscopy measures are highly correlated with autorefractor measures (Smith & Hung, 1999), but that our retinoscopy measures in adolescent monkeys were on average +1.4 ± 0.6 D more hyperopic than subjective measures of refractive errors (Ramamirtham et al., 2001). Thus, there are hyperopic biases in our refractive error measures and although our measures suggest that the target for emmetropization was a low degree of hyperopia, it is likely that the true end point was essentially emmetropia. Since the other studies of early refractive development in monkeys also employed retinoscopy to measure refractive state and were presumably affected by the same hyperopic biases, the observed refractive-error differences can not be explained by obvious instrumentation differences. However, we used tropicamide to produce cycloplegia whereas both Bradley et al. (1999b) and Young & Leary (1973) (based on their previous publications) used cyclopentolate, which has generally been shown to result in more hyperopic refractive error measurements in human children (Egashira et al., 1993, Lovasik, 1986, Mutti et al., 1994). Differences in the cycloplegic agent could account for much of the difference between our findings and Bradley et al.’s data (1999b), which were also obtained from rhesus monkeys. It is tempting to attribute the even higher amounts of hyperopia reported by Young and Leary for infant pig-tailed macaques (Macaca nemistrina) to interspecies differences. However, that does not seem reasonable since the same lab has previously reported much lower degrees of hyperopia for a larger population of pigtailed macaques (Young, 1964, Young & Leary, 1973) and that rhesus monkeys were usually more hyperopic than pig-tailed macaques (Young, 1964). Presumably the higher hyperopic errors reported by Young and Leary (1991) reflect individual peculiarities in the relatively small number of monkeys in their longitudinal study.

There was very good quantitative agreement in the axial length measures from the rhesus monkeys obtained in this study and those reported by Bradley et al.(1999b) (Figure 13B). Both the time course and the absolute values for axial length were in good agreement. The data from Kiely et al. (1987) show that cynomologus monkeys exhibited slightly shorter axial lengths than rhesus monkeys at all ages, but that the time course for axial growth for cynomologus monkeys was very comparable to that for rhesus monkeys. In agreement with the Young lab’s previous cross-sectional studies of adult macaque monkeys (Young & Leary, 1973), Young and Leary’s longitudinal data indicate that pigtailed macaques have significantly larger eyes at all ages than other macaque monkeys and that pig-tailed macaques exhibit a faster and possibly more prolonged increase in axial length than other macaque monkeys.

As illustrated in Figure 13C, the available longitudinal data show that for all macaques, corneal power stabilized at a younger age (about 200 days of age) than either refractive error or axial length. There were, however, substantial between study differences in the initial cornea powers for neonates and the pattern of power changes observed shortly after birth. In the present study, we found much higher corneal powers in neonates than previous studies and that there was subsequently a rapid, monotonic decrease in power. In contrast, Bradley reported that corneal power was stable for about 4 weeks after birth (Bradley et al., 1999b). The discrepancy between our study and their study could be due in part to the fact that we did not start our measurements until about 3 weeks of age. However, we also found decreases in corneal power as a function of age in our cross-sectional data that extended to infants as young as 2 weeks of age. It seems likely that limitations in the measurement range of the keratometer employed by Bradley et al. (1999b) truncated the data obtained from their youngest animals resulting in lower average corneal powers and an absence of age-related changes until the corneal powers fell within the effective measurement range of their keratometer.

There is general agreement between studies that lens thickness decreases during the first 4–5 years of life in infant monkeys (Figure 13D). Although in our study the thinning of the lens that we observed between about 1 and 4 years of age was not statistically significant, the magnitude of this change (0.21mm) was comparable to the decreases in lens thickness obtained by ultrasonography by De Rousseau and Bito (1981) for rhesus monkeys over a similar age range. Denlinger et al.(1980), using an optical phacometer also reported decreases in lens thickness for the same population of rhesus monkeys over the same time period and Koretz et al. (1987) reported similar results using slitlamp Scheimpflug photography. Thus, lens thickness does not contribute to the overall increase in axial length in young macaque monkeys.

Comparisons between monkeys and humans

In many respects the infant rhesus monkey eye is a scaled version of an infant human eye. The relative contributions of the three major axial components to total axial length are similar (monkeys vs. humans; anterior chamber depth: 16.3% vs. 15.0%; lens thickness: 24.2% vs. 21.4%; vitreous chamber: 59.4% vs. 63.6%). The cornea provides 55% and 58% of the total refracting power in infant monkey and infant human eyes, respectively. The anterior lens radii are 22% and 35% longer than the posterior lens radii in monkeys and humans, respectively. As a consequence, the relative positions of the principal points and nodal points are comparable in infant monkeys and infant humans (Table 3).

Refractive and ocular development proceeds in a similar manner in both rhesus monkeys and humans. In particular, in both species different segments of the eye develop at different rates and exhibit different magnitudes of change during growth (Pennie et al., 2001). As a consequence, adolescent eyes are not simply scaled versions of infant eyes. For example, during early development, both human and monkey eyes exhibit exponential increases in anterior and vitreous chamber depths and relatively slower and smaller decreases in lens thickness. As a consequence, the relative contributions of the anterior and vitreous chambers to total axial length increase with age. Comparisons between the relative growth trajectories indicate that, as in humans, the anterior chamber in monkeys matures faster than the vitreous chamber and that the relative contribution of the vitreous chamber to overall axial length increases with age. Both humans and monkeys exhibit exponential decreases in corneal power and lens power. Between infancy and adolescence, corneal power decreases by about 14% and 10% in monkeys and humans, respectively. However, in both humans and monkeys the decreases in lens power take place over a longer period of time and are larger than the corneal power changes in both relative and absolute dioptric terms (Mutti et al., 1998). For example, in both monkeys and humans, the power of the crystalline lens decreases by about 50% during early maturation eventually contributing between 35–40% of the eye’s total refracting power. As a consequence, the principal planes move forward during development in both monkeys and humans reflecting the relative increase in the contribution of the cornea to the eye’s total refracting power. In addition, within the crystalline lens there are asymmetries in growth with the anterior surface curvature decreasing proportionally more than that of the posterior surface in both monkeys and humans.

In both monkeys (Young & Leary, 1991) and humans(Larsen, 1971b, Larsen, 1971c, Larsen, 1971d, Zadnik et al., 2003), male infants have relatively larger eyes than females. Specifically in both species, it has been consistently demonstrated that male infants have flatter corneas, thicker crystalline lenses, and longer vitreous chambers and axial lengths. Although some human studies report that human males also have deeper anterior chambers (Zadnik et al., 2003), our study in monkeys and other human studies (Larsen, 1971a) have failed to find any gender related differences in anterior chamber depth. No gender related differences in spherical-equivalent refractive error have been observed in either human (Larsen, 1971a) or monkey infants.

Given that humans and monkeys have different life spans and different absolute rates of maturation, deriving meaningful age equivalents between species is important for establishing homologies and when extrapolating animal data to humans. A variety of strategies have been used to establish relative age equivalents for monkeys and humans. For example, based on differences in life span, Torczynski (1979) estimated that 1 year for a monkey is equivalent to 3 years for a human. Based on the relative rates for the development of visual acuity and contrast sensitivity, it has been estimated that monkeys mature 4 times faster than humans (Boothe et al., 1980, Teller et al., 1978). Comparisons of the age-dependent reductions in accommodation suggest that the monkey eye ages 3 times faster than the human eye (Bito et al., 1982, Smith & Harwerth, 1984). Age equivalents estimated from qualitative comparisons of the relative rates of axial elongation indicate that eye growth falls within this range with Kiely et al. (1987) indicating that cynomologus monkey eyes mature 3 times faster than humans and Bradley et al. (1999b) reporting that rhesus monkey eyes mature 4 times faster. Our quantitative comparisons of the rate constants fit to the longitudinal data for corneal power, total axial length, and anterior and vitreous chamber depths indicate that although these different ocular properties develop at different rates in humans and monkeys, the rhesus monkey eye matures approximately 3 times faster than the human eye.

Interestingly, the relative rates of emmetropization are not very different in humans and monkeys. The rate constant for the exponential functions fit to the refractive-error data indicate that emmetropization in monkeys is only 1.3 times faster than in humans (Figure 10). Given that good optical images are critical for early visual system development in both monkeys and humans and that the eye’s refractive state is a product of regulated ocular growth, it may not be surprising that there are smaller differences in the rates of emmetropization than for overall eye growth. Figure 14 compares the course of emmetropization between humans and several species of animals that are commonly used in refractive error research. There are clear and obvious differences in the initial degrees of hyperopia (Note that the data have not been corrected for any potential artifact associated with retinoscopy) and the target refractive errors between species. However, each species demonstrates a rapid reduction in refractive error to what is presumed to be the optimal ametropia for that species. Moreover, despite substantial differences in the relative ocular growth rates between these different species, the dramatic rapid phase of emmetropization is completed and near optimal refractive errors are achieved shortly after birth in each species. However, the strong qualitative and quantitative agreement between ocular growth and refractive development in rhesus monkeys and humans reinforces the use of rhesus monkeys in experiments designed to understand human refractive development.

Figure 14.

Comparisons of normal emmetropization between humans, monkeys, chicks, marmosets and tree shrews. The symbol legend indicates the source of the data (monkey, Bradley et al., 1999b; marmoset, Graham & Judge, 1999; chick, Li et al.,1995; chick, Wallman & Adams, 1987; humans, Edwards, 1991; humans, Wood et al., 1995; human, Thompson (from Saunders, 1995); human, Saunders, 1995; human, Atkinson et al., 1996).