The effect of positive lens defocus on ocular growth and emmetropization in the tree shrew

Abstract

Optical defocus influences postnatal ocular development in animal models. Induced negative lens defocus results in accelerated ocular elongation and myopia. Positive lens-induced defocus findings across animal models have been inconsistent. Specifically, in the tree shrew, positive lens-induced defocus has produced equivocal results. This study evaluated the response of the tree shrew to induced positive lens defocus. One treatment group wore positive lenses binocularly, which were increased in power sequentially from +2 to +4, +6, +8, and +9.5 D over 8 weeks. Other groups wore +4, +6, and +9.5 D lenses, respectively, for 8 weeks. Animals wearing zero-powered (plano) lenses served as controls. Refractive error and ocular dimensions were measured at the start of treatment and every week thereafter. Sequentially increasing positive lens power caused a relative hyperopia of +5.6 D ( p < 0.01) compared to the plano lens group (+10.9 ± 1.8 D vs +5.3 ± 0.5 D). Constant +4 D lens wear produced +6.9 D relative hyperopia, while +6 and +9.5 D lens wear did not induce hyperopia. Lens-induced defocus changes in refractive state were significantly correlated with vitreous chamber depth changes. The threshold for consistent responses to positive lens defocus in tree shrew was between +4 and +6 D. The results will enable targeted investigation of the efficacy of positive lens defocus in inhibiting myopic ocular growth.

Introduction

Ocular growth is closely regulated by the clarity of the image that is received on the retina (Wallman, 1993; Wallman, Gottlieb, Rajaram, & Fugate-Wentzek, 1987; Wallman & Winawer, 2004). Induced defocus from optical lenses has been demonstrated, in a variety of vertebrate species, to result in altered ocular growth in an attempt by the eye to attain functional emmetropia. More specifically, positive lenses, which bring the focus in front of the retina for a relaxed emmetropic eye (imposed myopic defocus), result in slowed ocular growth while negative lenses, which place the focus behind the retina in a relaxed emmetropic eye (imposed hyperopic defocus), result in accelerated ocular growth. Consequently, when the optical lenses are removed from in front of the eye, eyes wearing positive lenses are shorter and hyperopic (longsighted) and eyes wearing negative lenses are longer and myopic (shortsighted) (Schaeffel, Glasser, & Howland, 1988; Wallman & Adams, 1987). Compensation for imposed optical defocus is observed across several vertebrate species including fish (Kroger & Wagner, 1996), chicks (Irving, Sivak, & Callender, 1992; Schaeffel et al., 1988), guinea pigs (McFadden & Wallman, 1995), tree shrews (McBrien, Gentle, & Cottriall, 1999; Siegwart & Norton, 1993), and primates (Graham & Judge, 1999; Hung, Crawford, & Smith, 1995; Smith, 1998; Smith & Hung, 1999). 

In the most frequently studied model of ocular growth and refractive development, the chick, it has been demonstrated that the rate of ocular growth can be manipulated to bring about compensation for a wide range of lens-induced defocus (−10 D to +15 D), predominantly brought about by altered axial ocular dimensions. Less complete compensation was observed for lens powers beyond this range, indicating that the limits of compensation had been reached (Irving et al., 1992; Nevin, Schmid, & Wildsoet, 1998). There was also a considerable increase in the variability of responses for lens powers outside this range of compensation (Irving et al., 1992). In mammalian models, the reported findings have been more equivocal in the consistency of responses. In cats, the response to lens defocus was at best modest, with imposed optical defocus producing either no refractive changes (Nathan, Crewther, Crewther, & Kiely, 1984) or resulting in axial myopia, irrespective of the sign of defocus (Ni & Smith, 1989). These results were attributed to methodological issues since the use of contact lenses may have changed the corneal curvature, thus altering the imposed defocus. Guinea pigs respond to both positive and negative lenses in the appropriate direction for compensation of the imposed defocus; however, the range of reliable compensation observed for positive lenses was small (McFadden & Wallman, 1995). 

In addition to the important insights to the developmental biology of the eye, these findings are of particular interest in relation to the possibility of intervention in human refractive development, in particular the prevention of axial myopia. To this end, more recent data reported on the effects of imposed myopic defocus in chick induced by short durations of positive lens wear demonstrate markedly faster temporal effects than short periods of negative lens wear. Indeed short periods of positive lens wear effectively inhibit much longer periods of negative lens-induced hyperopic defocus in the chick (Winawer & Wallman, 2002; Winawer, Zhu, Choi, & Wallman, 2005; Zhu, Park, Winawer, & Wallman, 2005; Zhu, Winawer, & Wallman, 2003). The mechanisms underlying positive lens defocus may therefore be important in the context of prophylactic interventions for human myopia (Zadnik & Mutti, 1995). For this reason, translation of findings observed in the chick model to mammalian models will be an important step. 

However, in primates, while experiments have shown that anisometropic spectacle lenses do elicit different inter-ocular growth, the magnitude of refractive compensation is very limited compared to that observed in chicks (Graham & Judge, 1999; Hung et al., 1995). This was attributed to primates having a narrower effective operating range of the emmetropization mechanism and also to the view that higher amounts of imposed defocus (>±6 D) trigger visual development anomalies such as amblyopia (Smith & Hung, 1999). In particular, responses to large magnitudes of positive lens powers (>+6 D) have been reported to be inconsistent across different mammalian models (Wildsoet, 1997). The above studies highlight apparent species differences amongst mammalian models, and these inconsistencies in response to higher powered positive lenses hinder comparisons of results between chicks and mammalian models of refractive error development. 

In order to establish whether hyperopia occurs in response to higher powered positive lenses (>+6 D) in primate models, a “sequential” lens-rearing strategy where a small magnitude of defocus was imposed binocularly at the start of treatment and increased sequentially as compensation occurred was adopted in infant rhesus monkeys (Smith & Hung, 1999). This lens rearing strategy ensured small, constant, continual relative myopic defocus resulting in substantial amounts of absolute hyperopia. The binocular paradigm more closely mimicked the visual conditions experienced by monkeys in their natural viewing environment and avoided confounding inter-ocular interactions, as accommodation in primates is yoked (Smith & Hung, 1999). 

A study of optical defocus in the tree shrew, a well-established mammalian model of ocular growth and refractive development, has reported that positive lenses, especially of high powers (>+6 D), do not elicit hyperopia and result in myopia instead (Siegwart & Norton, 1993). It has been suggested that tree shrews may have a limited capacity to compensate for myopic defocus or indeed are unable to respond differently to the sign of defocus (Siegwart, Norton, & Robertson, 2003). The results of these studies on tree shrew raise the possibility that sign-dependent directional responses to imposed lens defocus are not a consistent finding across species. 

The aim of the present study was to examine whether tree shrews respond to binocular positive lenses by slowing ocular elongation, when the magnitude of defocus introduced was small and increased gradually, similar to the sequential lens strategy used by Smith and Hung (1999) in infant monkeys. A further aim was to assess whether sequential increases were the key to such responses or whether the combination of the lens power and optimal duration was more important in inducing a compensatory response in ocular growth. Thus, responses to a range of binocular positive lenses worn for a similar duration, without increasing lens power were also investigated to address this question. 

Methods

Experimental paradigms

Maternally reared tree shrews ( Tupaia belangeri) were used in this study. On the 14th day after eye opening, which occurred at the age of 19.7 ± 0.6 days (mean ± SEM) after birth, animals were removed from the maternal cage, allocated to positive and plano lens treatment groups, and housed in individual cages. Effective powers at the corneal plane for the positive powered lenses +2, +4, +6, +8, and +9.5 D were +2, +4.1, +6.2, +8.3, and +10 D, respectively, based on an average distance of 5 mm from the lens surface to the corneal apex. However, the lenses will be referred to by the optical power rather than the effective power at the corneal plane since the difference is at most 0.5 D. The average illumination at the floor of the cage was 265 lux. Food and water were available ad libitum. All animal procedures were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The responses to positive lens powers were examined in two studies described below: 

Sequential power increments of positive lenses

The sequential positive lens group ( n = 4 − 6) had positive lenses fitted binocularly in a head-mounted goggle (Siegwart & Norton, 1994) throughout the duration of the study. The lens powers were changed from low positive powers to higher positive powers during the course of the study. The lens change sequence is shown as a schematic in Figure 1A. The lowest power chosen was +2 D to ensure that defocus experienced was greater than the estimated depth of focus (estimated to be ±1.2 D) (Norton & McBrien, 1992). The duration of +2 and +4 D lens wear was 1 week each, whereas for +6, +8, and +9.5 D lenses it was 2 weeks each. In the plano lens group (n = 4 − 6); animals had zero-powered lenses fitted binocularly for the entire duration of the study (see schematic in Figure 1B). This group served to control for effects of lens wearing itself (like dust or scratches) and for any effects that periodic anesthesia and cycloplegia may have on ocular growth. 

Figure 1a, 1b

 

(A) The sequence of power increments of positive lens changes over the 8-week period of treatment. (B) The pattern of plano lens wearing in the control group over the 8-week period.

Constant power positive lens-wearing paradigm

Animals ( n = 17) were randomly assigned to three sub-groups that had +4 D ( n = 4), +6 D ( n = 6), +9.5 D ( n = 7) lenses fitted in goggles continuously for at least 6 weeks and up to 8 weeks. The plano lens group was used as the control group for this segment of the study as well. The variable sample sizes were due to detachment of the dental acrylic pedestals in some animals at later time points, as explained in a subsequent paragraph. 

Refraction, ocular dimensions and general procedures

In vivo optical and structural data were collected using retinoscopy and A-scan ultrasonography at the start of the treatment period and every week thereafter for 8 weeks, while the animal was anesthetized using ketamine (90 mg/kg) and xylazine (10 mg/kg). One percent tropicamide was used to dilate the pupil and as a mild cycloplegic. While tropicamide is a weaker cycloplegic, it was preferred against stronger cycloplegic agents such as atropine that have been shown to affect ocular growth (McBrien, Moghaddam, & Reeder, 1993; Raviola & Wiesel, 1985; Tigges et al., 1999). Refractive data were corrected for the working distance and for the effective power at the cornea. However, the small-eye artifact (Glickstein & Millodot, 1970) was not corrected for. Corneal curvature measurements were also taken in a subset of animals from the different groups at the start and end of treatment periods with a modified Bausch and Lomb one-position keratometer. 

Lenses were cleaned regularly, and during cleaning, the animals had their goggles removed and were placed in a dark nesting box, typically for 2–3 min. Also, the lenses were assessed, and any obviously scratched lenses were replaced with new ones. All animals were checked four to five times daily during the light cycle to ensure that the goggles and lenses were in place. Due to the relatively long-term treatment (8 weeks) and routine handling of animals to clean lenses, the dental acrylic pedestal occasionally detached. This occurred during handling of the animal to remove or replace the goggle. Whenever possible, the dental acrylic pedestal was rebuilt under anesthesia. However, when this was not possible, the experiment was terminated and this resulted in fewer numbers of animals towards the later time points in all experimental groups. 

Statistical analysis

Statistical analysis was conducted using the Graphpad statistical software (Graphpad software, Inc. USA). Since the treatment was binocular, there was the likelihood of high correlation between observations from the right and left eyes of an animal, which would lead to a falsely high statistical significance value if data were pooled (Ray & O'Day, 1985). Analyses of the results were carried out separately for the right and left eyes of animals in each group, and the statistical comparisons between one eye alone have been presented. Unpaired t-tests were employed for comparison of positive and plano-lens-treated groups at specific time points. Two-way ANOVA with Bonferroni post hoc tests were employed to assess differences between the two groups at the various time points. All group data have been presented as means with the standard error of the mean (mean ± SEM) unless otherwise specified. 

Results

Sequential power increments of positive lenses

The difference between the right and left eyes of binocularly treated animals was not significant in either the sequential positive lens or plano lens group. The statistical results of comparisons between the right eyes in the two groups were similar to that between the left eyes and as such statistical comparisons will be made just between the right eyes of each of the treatment groups, although some figures will show both right and left eye data. 

Refractive state measured at the start of treatment was not significantly different between the right eyes of animals in the sequential positive and plano lens groups, respectively (+10.4 ± 0.2 D vs +10.0 ± 0.3 D, unpaired t-test, p = 0.22, Figure 2). Statistically significant differences in ocular refraction were observed following 1 week of treatment, with the sequential positive lens-treated eyes demonstrating significant relative hyperopia compared to the plano-lens-treated eyes. This relative hyperopia was observed for the positive lens-treated group compared to the plano-lens-treated group at all subsequent time points thereafter (ANOVA, p < 0.0001; Bonferroni post hoc test, p < 0.01, Figure 2). The standard deviations for the positive lens-treated group ranged between 0.2 and 0.9 D (average standard deviation of mean of right and left eyes of animals across time points) compared to a range of 0.1–0.3 D in the plano-lens-treated group. The ocular refraction of the plano lens group was not significantly different to normal untreated tree shrews from an earlier study (Norton & McBrien, 1992) at comparable time points. 

Figure 2

 

The eyes of sequential positive lens-treated tree shrews were significantly hyperopic relative to age matched control tree shrews wearing zero (plano) power lenses from the first week of treatment (P < 0.01) onwards. This was the case for both right and left eyes separately or the mean refraction. The refractive state of plano-lens-wearing animals was not significantly different to age-matched normal animals from an earlier study (Norton & McBrien, 1992).

The sequential positive lens group showed a relatively stable refractive status from initiation of treatment. The cumulative refractive change in the right eyes of sequential positive lens-treated animals over the entire duration was +0.5 D. In contrast, a consistent reduction in hyperopia was observed in the plano lens group every week, with a cumulative reduction in hyperopia of 4.7 D (change per week −0.6 ± 0.1 D, mean ± SEM) over the duration of the study. The sequential positive lens group had greater variability in refraction across animals compared to the plano lens group, with across-animal standard deviation (average of standard deviation from the means of right eye values across the entire duration of the study) of 1.1 D and 0.7 D, respectively, in the two groups (variance ratio test, p < 0.001). 

The refractive differences were predominantly due to differences in the vitreous chamber depth, with the vitreous chamber significantly shallower in the sequential positive lens-treated group compared to the plano-lens-treated group, although multiple comparison testing identified significant differences at only the third, fourth, fifth, and seventh week measurement points (ANOVA, p < 0.0001; Bonferroni post hoc test, p < 0.05, Figure 3A). The differences in refractive state between the two groups at each time point were well correlated with the vitreous chamber depth differences ( r = 0.77; r ² = 0.58, p < 0.01). Differences in refractive error were also well correlated with changes in axial elongation over the course of the study. Positive lens treatment caused a reduced mean rate of axial elongation (ANOVA, p < 0.0001, Figure 3B) compared to plano-lens-treated animals. 

Figure 3a, 3b, 3c, 3d

 

(A) Vitreous chamber depth for tree shrews treated with positive lenses of sequentially increasing power compared to tree shrews wearing plano (zero) power lenses for 8 weeks. Due to the rapidly increasing lens thickness during postnatal ocular development (see Figure 3D), the vitreous chamber depth (back of crystalline lens to front of retina) in tree shrews decreases in depth during this phase of development. However, the decrease is significantly slower ( p < 0.01) in positive lens-treated animals compared to animals wearing plano lenses from the end of week 1 onwards to week 7. (B) Axial length (anterior cornea to front of retina) is significantly shorter in the animals wearing positive lenses compared to animals wearing plano lenses from week 2 to week 7 inclusive. (C) There was no significant difference in the mean anterior chamber depth of right eyes (front of cornea to front of crystalline lens) at any time point between the positive lens-wearing animals and the plano power wearing animals. (D) No significant differences in the mean lens thickness of right eyes was observed between positive lens-wearing animals and plano lens-wearing animals over the treatment period.

The anterior chamber depth and lens thickness measures were not significantly different in the right eyes of the positive lens and plano lens groups at any measurement period throughout the study (ANOVA, p = 0.41, Figure 3C; and p = 0.16, Figure 3D, respectively). The lens thickness steadily increased with time in both groups (ANOVA, p < 0.0001, Figure 3D). These developmental increases in the lens thickness contributed to the reduction in the vitreous chamber depth. This increase in lens thickness in both treatment groups was similar to an earlier study (Norton & McBrien, 1992) on normal ocular growth in the tree shrew. 

The corneal radius measurements collected prior to treatment (on the 14th day after eye opening) from the right eyes of the plano lens, sequential positive lens, +4 D, +6 D, and +9.5 D groups were combined ( n = 20) and considered as baseline data. These data were not significantly different from the mean corneal radius taken from the right eyes ( n = 5) of untreated age-matched normal tree shrews in an earlier study of normal ocular growth (Norton & McBrien, 1992) (3.32 ± 0.04 mm vs 3.34 ± 0.05 mm, unpaired t-test, p = 0.80). A reduction of corneal curvature was observed following 8 weeks of treatment in both the sequential positive lens and plano lens groups, similar in degree to that for normal animals collected an earlier study (Norton & McBrien, 1992). Also, no significant differences were observed in the corneal radius measurements between the sequential lens group and the plano lens group (n = 4 each group) following 8 weeks of treatment (3.52 ± 0.02 mm vs 3.46 ± 0.06 mm, unpaired t-test, p = 0.33). 

In summary, sequential positive lens-treated eyes demonstrated relative hyperopia and slowing of axial ocular growth in contrast to the normal decrease in hyperopia that occurred during emmetropization in tree shrews wearing plano lenses or normal tree shrews from earlier studies (Norton & McBrien, 1992). 

Sustained positive lens defocus: Optical threshold for compensatory responses

From previously reported findings on the tree shrew model of ocular growth and refractive development, it was not possible to determine if the equivocal responses to positive lens defocus were due to the amount of defocus used, the limited duration of the treatment period, or a combination of both. While the above results from the present study, utilizing a sequentially increasing positive lens paradigm, clearly demonstrate consistent compensatory responses to positive lens-induced defocus in tree shrew, they do not provide information on the maximum threshold level of positive lens blur that induces a compensatory response or duration of treatment required to get an appropriate response. 

To address these specific issues, 3 additional sub-groups of tree shrews were treated with constant positive lens defocus binocularly using either +4 D, +6 D, or +9.5 D lenses and monitored for up to 8 weeks. The binocular plano-lens-wearing group from the sequential lens study acted as the comparison control group for this part of the study. 

The refractive changes observed between the right and left eyes were not significantly different, and only right eye data will be statistically compared, although both right and left eye data will be presented graphically. 

Tree shrews wearing +4 D lenses binocularly ( n = 4) developed significant hyperopia relative to the plano-lens-wearing control animals. Refractive error measurements of the two groups differed significantly from each other from the second week of treatment onwards (ANOVA p < 0.0001; Bonferroni post hoc test, p < 0.05, Figure 4A). The effect of constant +4 D binocular lens wear was comparable to that observed with sequential increases of positive lens powers in producing a stable relative hyperopic refractive error throughout the 8-week monitoring period. A shallower vitreous chamber was the main structural correlate responsible for the relative hyperopia observed. The vitreous chamber depth was significantly shallower than the vitreous chamber depth of the plano-lens-wearing group from 1 week of treatment onwards (ANOVA p < 0.0001; Bonferroni post hoc test, p < 0.05, Figure 4B). There were no significant differences in the mean anterior chamber depth or lens thickness between the right eyes of the +4 D and plano lens groups, (ANOVA, p > 0.05). The corneal radius of curvature was not significantly different between the right eyes of the +4 D and plano lens groups at the final measurement time point (3.52 ± 0.02 mm vs 3.46 ± 0.06 mm, unpaired t-test, p > 0.05). 

Figure 4a, 4b, 4c, 4d

 

Constant positive lens wear for a period of 8 weeks. (A) Constant binocular wearing of +4 D lenses prevents the normal reduction of hyperopia seen in normal animals or animals wearing plano lenses. (B) The structural correlate to the relative hyperopia induced by +4 D lenses is a shallower vitreous chamber depth. (C) Constant +6 D binocular lens wear initially is less effective at inhibiting the normal reduction in hyperopia with development and from 5 weeks induced anisometropia in eyes wearing +6 D lenses with some eyes moving into relative myopia. (D) The refractive change was well correlated to vitreous chamber depth differences ( p < 0.05). (E) Constant +9.5 D lens wear only partially delayed the reduction in hyperopia seen in plano-lens-wearing animals, and after 5 weeks there was a shift towards myopia and significant differences in the two eyes of +9.5 D lens-wearing animals. (F) Changes in refractive error were well correlated with changes in vitreous chamber depth.

The refractive changes in the right and left eyes of individual animals in the +6 D lens group ( n = 6) showed greater variability. The standard deviation of the mean of right and left eyes ranged between 0.2 and 2.9 D over the 8 weeks of treatment. The fewer animals towards the end of the study in the +6 D group ( n = 2 each at the 7th and 8th week time point; due to some animals dislocating their pedestal after 6 weeks) made assessment of treatment effects problematic at these time points. The constant wear +6 D binocular lenses induced relative hyperopia in the right eyes of animals in this group ( n = 6) compared to the right eyes of plano-lens-wearing controls, across all measurement time points just reached significance (ANOVA, p < 0.05, Figure 4C). However, multiple comparison tests did not reveal significant differences between the mean refractive measurements at any individual time point during the study. Some animals exhibited relative myopic changes from baseline and large degrees of anisometropia (approx. 4 to 5 D of anisometropia) from 4 weeks of +6 D lens wear onwards. No animal in the sequential positive or the constant +4 D lens-wearing group demonstrated anisometropia. The vitreous chamber depth measurements were not significantly different between the right eyes of the constant +6 D lens-wearing group and the plano control group at any time point (ANOVA, p = 0.93, Figure 4D). Similar to findings in other groups, the anterior chamber depth and the lens thickness did not differ between the plano and +6 D lens groups (ANOVA, p = 0.14 and p = 0.26, respectively). Overall, the axial length measurements did not differ between the right eyes of the +6 D and the plano lens groups (ANOVA, p = 0.96; Bonferroni post hoc test, p = 0.96). Also, the corneal radius of curvature was not significantly different between the right eyes of the +6 D and the plano lens groups following 8 weeks of treatment (3.45 ± 0.01 mm vs 3.46 ± 0.06 mm, unpaired t-test, p = 0.91). 

The refractive error in the right and left eyes of animals constantly wearing binocular +9.5 D lenses showed even greater variability, particularly at later time points (average standard deviation of mean of right and left eye refractions ranging between 0.4 and 5.4 D across time points), although similar in pattern to that observed in the +6 D group. The refraction in the right eyes of animals treated with +9.5 D lenses ( n = 7) was not significantly different from the plano lens group (ANOVA, p = 0.53, Figure 4E). On average, the +9.5 D lens-treated eyes were more myopic than the plano-lens-wearing eyes towards the end of the treatment period. The vitreous chamber depth showed no significant differences between the right eyes of the +9.5 D constant wear and the plano lens group (ANOVA, p = 0.68, Figure 4F). The anterior chamber depth was not significantly different between the right eyes of +9.5 D and plano-lens-wearing animals at any time point (ANOVA, p = 0.14). While positive lens-wearing eyes had a significantly greater lens thickness than plano-lens-wearing animals (ANOVA, p < 0.05), no significant difference was found at any individual time point in post hoc comparisons (Bonferroni post hoc test, p > 0.05). No significant differences were observed between the axial length measurements in the right eyes of the two groups (ANOVA, p = 0.42). No significant differences in the corneal radii of curvature of the +9.5 D group compared to the plano lens group following 8 weeks of treatment was observed (3.50 ± 0.04 mm vs 3.46 ± 0.06 mm, unpaired t-test, p = 0.61). 

Discussion

The principal finding of the present study was that, using a sequential increasing lens power paradigm, the tree shrew responds to positive lens-induced defocus (myopic defocus) with appropriate compensatory changes in ocular axial length and refractive state. Specifically, induced binocular positive lens defocus causes a slowing of the axial ocular growth rate in the tree shrew eye relative to age-matched controls. The causative structural component is a shallower vitreous chamber depth than age-matched controls. These changes in ocular growth result in a significant relative hyperopia when compared to control age matched tree shrews wearing plano lenses binocularly or normal age-matched controls. When the above data are combined with previous results in tree shrew showing compensatory changes to lens-induced hyperopic defocus (Cottriall & McBrien, 1996; Siegwart & Norton, 1993; Siegwart et al., 2003), it demonstrates that tree shrews show sign dependent compensatory changes in ocular growth and refractive state to induced optical defocus. The bi-directional responses in ocular growth to positive and negative lens defocus in the tree shrew are qualitatively similar to the responses seen in other animal models such as the chick (Irving et al., 1992; Schaeffel et al., 1988), guinea pigs (McFadden & Wallman, 1995), rhesus monkeys (Hung et al., 1995), and marmosets (Graham & Judge, 1999). This finding is significant since it demonstrates categorically that bi-directional ocular growth responses to lens defocus occur in this common mammalian model of human refractive error development and provides a consistency of findings in vertebrate models to date. 

When the current data are viewed in combination with previous findings on lens-induced defocus in other animal models of ocular growth and refractive development, there is a consistent pattern of compensatory changes to optical defocus from both avian and mammalian models including nonhuman primates (see Figure 5) (Graham & Judge, 1999; Hung et al., 1995; Irving et al., 1992; McFadden & Wallman, 1995; Schaeffel et al., 1988; Siegwart & Norton, 1993). The eye is able to alter its growth rate to either accelerate normal developmental axial elongation in response to hyperopic defocus or decelerate its ocular elongation in response to myopic defocus, to effectively render the eye “emmetropic” when wearing the lens. From Figure 5, it can be seen that the operating range of this “emmetropization” response varies between species, with the widest range found in the chick model and the narrowest operating range found in nonhuman primates. In keeping with this, it is not surprising to find that the tree shrew, a species close to the primate line, also has a restricted operating range. 

Figure 5

 

Comparison of the operating range of the emmetropization responses in some commonly used animal models. Chicks demonstrate the widest range of compensation whereas the mammalian models of tree shrew, guinea pig, and monkey show a restricted operating range for compensation.

The level of defocus experienced by tree shrews receiving sequential positive lens treatment can be estimated based on their refractive state and viewing distances available to them. Tree shrews were hyperopic (∼+10 D) at the start of treatment (15 days after eye opening), a value which includes the artifact of retinoscopy (∼+5 D in tree shrews) observed in small eyes (Glickstein & Millodot, 1970), implying an “effective hyperopia” of approximately +5 D at the start of treatment (Norton & McBrien, 1992). Since intentionally no attempt was made to control or restrict the viewing distance of animals, tree shrews received varying defocus information depending on the lenses worn and on the proportion of the day spent looking at the objects at different distances within their visual environment. Binocular +2 D lens wear would have partially corrected the initial hyperopia and the subsequent use of +4 D lenses after 1 week would have rendered the eyes nearly emmetropic for distant objects. All near objects in the tree shrews' visual environment would then be expected to cause less hyperopic defocus than previously and thus require less accommodation to be clearly observed. Since tree shrew eyes wearing +4 D lenses would have been close to functionally emmetropic for distant objects, this might be expected to cause a slowing of ocular growth since the process of emmetropization would have been signaled as near complete. From the data presented in Figure 3, it is apparent that slowing of ocular growth (significantly shallower vitreous chamber depth) occurred at the 2nd week measurement session after 1 week of binocular +2 D and 1 week of binocular +4 D positive lens wear. From week 3, lens power was increased to +6 D for 2 weeks and then +8 D and +9.5 D each for a further 2 weeks. At the end of the 2nd week, the measured hyperopic refractive state had increased to over +11 D for the sequential positive lens-treated group. Thus, the introduction of +6 D binocular lenses would, taking into account the small artifact of retinoscopy of ∼+5 D, still produce close to a “functional emmetropia” for the tree shrews when viewing distant objects, explaining the stable refractive state. By week 5 of sequential positive lens wear, the introduction of +8 D lenses would have produced a small amount of myopic defocus for distant objects (∼2 D), clear images at 50 cm, and with the use of ocular accommodation, clear images at closer viewing distances would have been possible. The introduction of +9.5 D lenses (effective power at the cornea of +10 D) would have produced ∼4 D of myopic defocus for distance viewing. It is noteworthy that both the measured hyperopia and the vitreous chamber depth difference reduced with respect to the plano lens group in the last week of lens wear. Clear vision would have been easily attainable while wearing +2 D, +4 D, +6 D, and +8 D when viewing objects of interest within the visual environment of tree shrews such as food, water, cage meshwork, and tree shrews in nearby cages. These objects were at varying near distances ranging from 2 m to 5 cm. When wearing the +9.5 D lenses, most objects of interest within the visual environment would also have been in focus (closer than 30 cm); however, viewing more distant objects of interest such as animals in other cages would have induced some myopic defocus. 

The use of these higher plus powered lenses may not have necessitated a change from the slowed rate of ocular growth already triggered by wearing +2 D and +4 D lenses. However, in previous studies on tree shrews, using moderate to high-powered positive lenses (+5 D and +10 D) from the start of treatment did not result in significant relative hyperopia and consequent slowing of ocular growth (Siegwart & Norton, 1993; Siegwart et al., 2003) despite providing clear vision at some near planes. Thus, the small amount of “relative” myopic defocus introduced by lower powered lenses at the start of the treatment period in the current study may have played a critical role in inducing slowed ocular growth responses (seen as a decrease in vitreous chamber depth). In addition, the sequential paradigm would have ensured that the defocus was detectable (i.e., greater than the depth of focus) and within the maximum limit of compensation with every lens change in the current study. 

Also of relevance to previous findings that have failed to show a compensatory response to positive lens defocus is the developmental stage at which positive lens wear was initiated. In the present study, induced defocus was initiated at 15 days after eye opening, the time when tree shrew ocular development is most susceptible to environmentally imposed changes on ocular growth and refractive state (McBrien & Norton, 1992). The timing of initiating optical defocus would appear to be important in the consistency of response to myopic defocus (plus lenses) as a previous study on tree shrew reported, in abstract form, that where myopic defocus had been initiated at a later time point in development (24 days after eye opening) inconsistent responses have been observed (Siegwart & Norton, 1993; Siegwart et al., 2003). Thus, not only the level of defocus is important in inducing appropriate compensatory responses but also the point in postnatal ocular development when defocus is initiated. This has important implications as to the timing of when any intervention treatments might be applied in studies to prevent myopia in adolescent children. 

Studies conducted in some species have demonstrated that positive defocusing lenses result in absolute hyperopia (i.e., relatively greater hyperopia than at the start of treatment or than in the contralateral control eye). In the present study, step-like increases in absolute hyperopia commensurate with increases in positive lens powers were not observed, unlike observations in infant monkeys (Smith & Hung, 1999). A notable difference in the current study is that myopic defocus was not provided from the commencement of treatment. In positive lens-treated animals, normal developmental increases occurred in lens thickness and, to a lesser degree, the ocular axial length as occurs in normal tree shrews (Norton & McBrien, 1992), implying that there is neither a cessation of normal ocular growth nor a physical shortening of the ocular axial length. This, as suggested in a previous study (Siegwart et al., 2003), may mean that the minimum rate of preprogrammed ocular growth limits the capacity of the tree shrew eye to demonstrate absolute hyperopia. While an increase in choroidal thickness could facilitate initial compensation to positive lenses, the choroidal thickness has been reported either to be unaffected (McBrien, Lawlor, & Gentle, 2000) or to contribute only to about 1 D change in the refractive state (Norton & Kang, 1996) during recovery from induced myopia in the tree shrew. Although the resolution of the ultrasound system used in the present study (10 MHz) was insufficient to allow reliable estimation of the individual choroidal thickness changes, it is unlikely, based on the above studies, to be a major factor in compensation in the tree shrew. This is unlike observations in the chick model, where significant increases in choroidal thickness occur during positive lens wear, allowing a rapid increase in hyperopia during compensation (Wallman et al., 1995; Wildsoet & Wallman, 1995). 

While sequential increases of positive lens powers caused slowing of the rate of ocular elongation and prevented the normal developmental decrease in hyperopia, the stabilization of initial hyperopia was similar to that seen when binocular moderate-powered positive lenses were worn constantly for a longer period of time in monkeys (Hung et al., 1995; Smith & Hung, 1999). Since there were no step-like increases in hyperopia concurrent with the changes in lens power in the current study, it was unclear whether the changes seen in the tree shrew eye were related to the sequential increases in lens power or simply to the effect of wearing positive lenses for a longer duration than in previous reported studies. Also, the sequential lens strategy would not allow estimation of the upper threshold for positive lens compensation in the tree shrew since the change in magnitude of defocus would always have been within the limit of compensation. 

Therefore, different groups of tree shrews fitted with binocular positive lenses worn constantly over a similar 8-week duration were investigated. Animals wearing +4 D lenses for a similar duration as the sequential lens group resulted in relative hyperopia compared to plano-lens-wearing animals. The refractive course, the structural change, and the variability in response seen with constant +4 D binocular lens wear were similar to the sequential positive lens group, effectively producing a slowing of ocular growth in tree shrews. This implies that responses were for relative alteration of defocus per se, provided the lens powers or increments were within the range of compensation for positive lenses in the tree shrew. Since these lenses approximately corrected the hyperopia measured at the start of treatment, this did not necessitate any further alterations in the refractive state resulting in stable hyperopia. 

Animals wearing +6 D binocular lenses from day 15 were estimated to experience a modest amount of myopic defocus (≈2–3 D) for distant objects and approximately emmetropia for intermediate viewing distances of 0.25 to 0.5 m. However, consistent relative hyperopia was not observed in this group of animals. Three out of six animals (50%) developed relative hyperopia and a reduced rate of elongation, while the other three animals showed either significant decreases in hyperopia or absolute myopia. Also, there was no evidence of a slowed rate of vitreous chamber elongation in contrast to the +4 D lens-wearing group. An important observation was that the normally yoked changes in refraction seen between right and left eyes of an animal were disrupted in the 3 animals whose refractive state moved to myopia, with significant amounts of anisometropia developing (see Figure 4). It is therefore reasonable to conclude that +6 D lenses are beyond the upper limit of compensatory growth responses to myopic defocus in tree shrews of this age. Similar poor compensatory responses have been observed when higher powered positive lenses were used from the start of intervention in other mammalian models such as the guinea pig, marmoset and rhesus monkey (Graham & Judge, 1999; McFadden & Wallman, 1995; Smith & Hung, 1999). 

In the animals wearing binocular +9.5 D lenses, it is estimated that approximately 5 diopters of myopic defocus is induced for viewing distant objects. Wearing +9.5 D lenses from the start of treatment caused increased variability between the two eyes of individual animals and between animals. The large variation in the responses between the two eyes of individual animals is indicative of a breakdown of the yoking between the eyes, presumably due to the large magnitude of defocus. This effect is more clearly demonstrated in Figure 6, which plots the refractive error ( Figure 6A) and vitreous chamber depth ( Figure 6B) of the right and left eyes of 3 animals wearing binocular +9.5 D lenses for the 8-week duration of treatment. Figure 6 highlights the marked variation of up to 8 diopters between the right and the left eyes of an individual animal and that the structural cause of these intra-animal refractive difference is due to intra-animal differences in the vitreous chamber depth. This could reflect a conservative strategy whereby one of the eyes does not alter its refractive state since some near vision is possible, while the other eye develops relative myopia due to the substantive myopic defocus, similar to the altered fixation pattern seen with monocular positive lens wear in monkey (Smith, 1998). These changes indicate failure in compensation to myopic defocus of this magnitude, possibly initiating a default rate of ocular elongation, which would have the consequence of increasing the magnitude of myopic defocus for both distant and near objects. Similar results have been reported when high-powered lenses were used in the chick and rhesus monkey (Irving et al., 1992; Smith & Hung, 1999), where there was less complete compensation with increased variability. 

Figure 6a, 6b

 

For the three animals in the +9.5 D lens treatment group who completed the full 8 weeks of binocular lens wear, the refractive error (A) and vitreous chamber depth (B) of the right and left eyes are plotted over the 8 weeks. After the first few weeks of +9.5 D lens wear, there was a breakdown in the normally high concordance in refractive error and vitreous chamber depth observed between right and left eyes of an individual animal. In particular, animals pos 4 and pos 5 demonstrate over 8 D difference after 8 weeks of treatment, the structural cause of which is marked intra-animal differences in vitreous chamber depth (B).

Thus, the maximum positive lens power that reliably produced slowed ocular elongation, and a relative hyperopia in the tree shrew was +4 D with the upper threshold for appropriate compensatory growth response to positive lens defocus in the range of +4 to +6 D. 

The results of the present study are important because they demonstrate that compensation to positive lenses (myopic defocus) occurs in this well-established mammalian model of ocular growth, refractive error, and emmetropization. Furthermore, this study has characterized the ocular growth responses to a range of positive lens powers, thus establishing the maximum lens power that produces relative hyperopic responses in this model. The data from the present study form the basis for future studies to assess the possible beneficial effects of positive lens wear as an optical intervention to prevent the development or progression of myopia in adolescents. 

Acknowledgments

The authors wish to thank Dr. Andrew Jobling and Dr. Ravikanth Metlapally for their expert assistance with some of the animal procedures used in this study. This work has been presented in part at the Association for Research in Vision and Ophthalmology (ARVO) conference, 2005, Florida, USA. This work was supported by The National Health and Medical Research Council, Australia. Grant # 251557. 

Commercial relationships: none. 

Corresponding author: Neville A. McBrien. 

Email: to:[e-mail]. 

Address: Department of Optometry and Vision Sciences, University of Melbourne, Parkville, Victoria 3010, Australia.