A Neurobehavioral Analysis of the Prevention of Visual Impairment in the DBA/2J Mouse Model of Glaucoma

Abstract

Purpose.: Timoptic-XE treatment was used to examine the relationship between age-related changes in intraocular pressure (IOP), retinal cell loss, visual ability, and neuronal labeling in the superior colliculus in the DBA/2J mouse model of pigmentary glaucoma.

Methods.: Mice were administered Timoptic-XE (0.0%, 0.25%, or 0.50%) daily from 9 weeks to 12 months of age. Visual ability and IOP were evaluated at 3, 6, 9, and 12 months of age. Mice from each group were then given intraocular injections of wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP), and estimates of the number of cells in the ganglion cell layer of the retina, WGA-HRP transneural labeling of cells, cell count, and cross-sectional area of Nissl-stained cells in the superior colliculus were obtained.

Results.: Mice treated with 0.50% and 0.25% Timoptic-XE maintained a high level of performance in behavioral vision tasks, while 12-month-old untreated mice (0.0% Timoptic-XE) exhibited impaired visual performance. Timoptic-XE therapy reduced IOP and cell loss in the ganglion cell layer of the retina and prevented somal shrinkage and the decrease in WGA-HRP transneural labeling in the superior colliculus that occurred in untreated mice at 12 months of age.

Conclusions.: This study provides a comprehensive assessment of the efficacy of Timoptic-XE in DBA/2J mice by correlating age-related visual system changes in the retina and brain with changes in IOP and visual ability. These results showed that reducing IOP not only rescued retinal ganglion cell atrophy but also restored visual function and altered patterns of neurodegeneration that occur with blindness.

Introduction

Glaucoma is the second leading cause of blindness after cataracts. ¹ Clinically, glaucoma is characterized by “cupping” of the optic nerve head and visual deterioration due to the progressive loss of retinal ganglion cells and axons. ^2,3 Pigmentary glaucoma is caused by the liberation of melanin pigment through the mechanical disruption of the posterior iris by the lens zonules. When pigment accumulates within the anterior segment structures and the trabecular meshwork, it blocks aqueous outflow, resulting in an increase in intraocular pressure (IOP) and subsequent damage to the optic nerve and retinal ganglion cell death. ^4,5 High IOP (>21 mm Hg) is the strongest and most consistent risk factor for glaucoma; and although not all patients exhibit ocular hypertension, the probability of retinal ganglion cell loss and optic nerve damage increases exponentially with higher IOP, thus all current glaucoma treatments involve the reduction of IOP. ^2,6 Timolol maleate, a nonselective β₁ and β₂ adrenergic receptor antagonist, applied topically to the eye, decreases IOP by reducing aqueous humor production by blocking of β-receptors on the ciliary epithelium. Timolol maleate (Timoptic) has been viewed as the gold standard therapy for glaucoma ⁷ and is used as the clinical benchmark in ocular hypotensive medication-registration trials. ^8,9  

DBA/2J mice develop a genetically determined form of pigmentary glaucoma that resembles iris stromal atrophy and iris pigment dispersion syndrome in humans. These conditions are due to two gene mutations. Iris stromal atrophy is induced by the b allele of tyrosine-related protein 1 gene (Tyrp1^b ), and pigment dispersion is caused by a premature stop codon mutation in the glycoprotein (transmembrane) nmb gene (Gpnmb^R150X ). ^10–12 DBA/2J mice exhibit elevated IOP, atrophic excavation of the optic nerve head, progressive loss of retinal ganglion cells, and visual impairment. ^12–14 At 3 to 4 months of age, these mice develop defects of the peripheral iris, pigment dispersion, and thickening of the iris at the pupillary border. Iris stromal atrophy begins by 6 months of age, and elevated IOP peaks at 6 to 9 months of age. ^12,15–18 The majority of mice display retinal ganglion cell loss between 10 and 12 months of age. ^13,18–20 Using the visual water task, we found that DBA/2J mice were able to complete visual discrimination and visual acuity tasks at 6 months but not at 12 months of age or older, while C57BL/6J mice, which do not develop age-related blindness, did not show such age-related reductions in visual ability. Thus the progression of retinal ganglion cell death has rendered DBA/2J mice functionally blind by 12 months of age. ¹⁴  

DBA/2J mice are responsive to conventional IOP-lowering therapies used in humans with glaucoma, ^21–24 but no systematic studies have evaluated the effects of these treatments on visual behavior, the retina, and the visual brain of these mice. The purpose of this study was to provide a comprehensive assessment of the efficacy of Timoptic-XE on behavioral, ocular, and neural measurements of glaucomatous damage from 3 to 12 months of age in the DBA/2J mouse model of pigmentary glaucoma. This experiment examined the longitudinal protective effect of Timoptic-XE in aging DBA/2J mice on visual behavior, intraocular pressure, retinal ganglion cell loss, cell size and cell number in the superior colliculus, and transneural labeling in the superior colliculus after intraocular injections of wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP). In this way, we examined the pattern of glaucoma-induced changes in the visual system of aging DBA/2J mice to determine if therapy to lower IOP is effective in protecting against the consequences of glaucoma in multiple aspects of the visual system. 

Methods

Subjects

The experimental protocol was approved by the Dalhousie University Committee on Animal Care. Mice were treated in accordance with the regulations set forth by the Canadian Council on Animal Care, following the guidelines established by the ARVO Statement for Use of Animals in Ophthalmic and Vision Research. Forty-eight DBA/2J mice (24 males and 24 females, JAX stock #000671) were obtained from The Jackson Laboratory (Bar Harbor, ME) at 6 weeks of age and housed in same-sex pairs in clear plastic cages (29.2 × 18.4 × 12.7 cm) with ad libitum Purina rodent chow (#5001; Purina, Aberfoyle, Ontario, Canada) and tap water. The colony room was maintained at 22 ± 2°C with a 12:12 hour reversed light:dark cycle (lights off at 9:45 AM). All behavioral testing was completed during the dark (active) portion of the light:dark cycle. Table 1 shows the number of male and female mice in each experimental group that completed behavioral testing at each age. 

Table 1. 

 

Number of Male (M) and Female (F) Mice in Each Drug Treatment Group That Completed Behavioral Testing at 3, 6, 9, and 12 Months of Age

Treatment	3 Months	6 Months	9 Months	12 Months
0.50% Timoptic-XE	7 M, 8 F	5 M, 6 F	3 M, 4 F	2 M, 1 F
0.25% Timoptic-XE	7 M, 7 F	5 M, 3 F	4 M, 2 F	1 M, 2 F
0.00% Timoptic-XE (0.60% Gelrite)	8 M, 8 F	7 M, 6 F	4 M, 4 F	2 M, 2 F
Total	22 M, 23 F	17 M, 15 F	11 M, 10 F	5 M, 5 F

Experimental Design

Mice were divided into three treatment groups, which were given 0.00%, 0.25%, or 0.50% Timoptic-XE (Merck Frosst Canada Ltd., Kirkland, Quebec, Canada), a sterile ophthalmic gel-forming solution, once a day from 9 weeks to 12 months of age. Each milliliter of Timoptic-XE (Tim) contained 3.4 mg timolol maleate (0.25% Tim) or 6.8 mg timolol maleate (0.50% Tim). The control solution (0.00% Tim) was an aqueous solution containing 0.6% Gelrite gellan gum (Sigma-Aldrich Canada Ltd., Oakville, Ontario, Canada), the inactive ingredient in Timoptic-XE that causes it to form a gel. Baseline IOP measurements were obtained for each mouse at 8 weeks of age, and drug treatment began at 9 weeks of age. Visual ability was evaluated at 3, 6, 9, and 12 months of age using the visual water task. Following vision tests at each age, IOP was measured in all mice, and then two males and two females from each treatment group were given intraocular injections of WGA-HRP. One week later these mice were sacrificed, and their eyes and brains were removed for further processing. 

This longitudinal study design allowed us to repeatedly measure the same animals as they aged, which reduced variability and also allowed us to chronicle the anatomical changes in the retina and brain over time and therefore correlate behavioral, retinal, and neural measurements within each individual mouse. Unfortunately, some mice died over the course of the experiment, which resulted in the small sample size at 12 months of age (Table 1). Based on our previous data with aging DBA/2J mice ¹⁴ we used GraphPad StatMate software (GraphPad Software, Inc., La Jolla, CA) to conduct a power analysis. This showed that with a sample size of 4 in each drug group, there was an 80% power to detect a difference between group means with alpha = 0.05. Thus even the reduced sample size in the oldest mice was sufficient to detect the large effect sizes that existed between our treatment groups with substantial power. 

Behavioral Measures of Visual Ability

The visual water task is a computer-based, two-alternative forced-choice swim task in which mice are trained to swim to a vertical sinusoidal grating in order to escape the water. ^14,25,26 Testing in the visual water task was conducted over 25 consecutive days: one day of pretraining (six trials) followed by eight days of testing in each of the visual detection, pattern discrimination, and visual acuity tasks. Visual detection testing consisted of training the mice over eight days (eight trials/day) to learn the location of a hidden platform, predicted by a sinusoidal vertical grating of 0.17 c/deg (S+) on a computer screen versus a computer screen displaying a homogeneous gray screen (S−). A trial was scored as correct if the mouse swam to the platform within 60 seconds and did not enter into the side of the apparatus displaying the S−. An error was scored if the mouse entered the S− arm at any point during the trial. If the mouse did not find the platform within the 60 seconds, it was guided to the platform and required to stay on the platform for ∼15 seconds, and the trial was scored as an error. Following an error trial, the mouse was given an additional 60-second trial with the same S+ and S− locations. The number of the correct trials (expressed as a percent correct for each day) was used as a measure of visual learning ability, which assesses how well the mouse learned to execute the task over the course of eight days. Criterion was set at 70% correct, and the day that a mouse reached criterion was an indication of how quickly it was able to learn the task. The percentage of correct responses on day 8 of testing served as a single measure of visual detection ability. The pattern discrimination task (eight days, eight trials/day) used the same procedure to test the ability of mice to discriminate between a vertical sinusoidal grating of 0.17 c/deg (S+) and a horizontal sinusoidal grating of 0.17 c/deg (S−). In the visual acuity task (eight days, eight trials/day), the spatial frequencies of the S+ for each trial ranged between 0.17 and 0.64 c/deg, while the S− was a gray screen. The same order of spatial frequencies was used each day in order to obtain a consistent threshold. The number of correct trials (expressed as a percent correct for each spatial frequency) was used to plot a frequency-of-seeing curve for each mouse, and a visual acuity threshold was obtained at the spatial frequency at which the mouse failed to achieve 70% correct responses. 

Intraocular Pressure

IOP was measured using the handheld TonoLab (Colonial Medical Supply, Franconia, NH), a rebound/impact tonometer designed to take six quick consecutive measurements and provide the calculated average in units of mm Hg. ²⁷ Mice were lightly anesthetized with 4% to 5% isoflurane (flow rate of 0.5–1 L/minute; Baxter Corp., Mississauga, Ontario, Canada) and maintained on 1% to 2% isoflurane during the procedure. Three readings were taken for each eye, with each reading representing the average of six successive measurements. These were averaged to give one IOP measurement per mouse that was used for all analyses. IOP measurements were recorded at approximately the same time of day (between 11 AM and 2 PM) during the course of the experiment. 

Intraocular Eye Injections

Following intraocular pressure measurements, two males and two females from each treatment group (12 mice total) were given intraocular injections of wheat germ agglutinin conjugated to WGA-HRP to evaluate transneural transport from the optic nerve to the superior colliculus. WGA-HRP is a transneuronal tracer that is taken up by retinal ganglion cells, transported along the axons of the optic nerve, and transferred across the synapse to neurons in the superior colliculus. ^28–30 Mice were anesthetized with 4% to 5% isoflurane (flow rate of 0.5-1 L/minute) and maintained on 2% to 3% isoflurane during the procedure. An intravitreal dose of 2 to 3 μL WGA-HRP (40 mg/mL) (Vector Laboratories, Burlingame, CA) was injected into the right eye over ∼30 seconds using a 25 μL Hamilton syringe (Hamilton Co., Reno, NV). Following the procedure, an ophthalmic antibiotic ointment (Pentamycetin; Sabex, Boucherville, Quebec, Canada) was applied to the eye, and mice were allowed to recover in their home cages for seven days before histological processing began. 

Histological Processing

Mice were anesthetized with an intraperitoneal injection (0.1 mL/25 g) of pentobarbital sodium (Euthanyl, 65 mg/mL; Bimeda-MTC, Cambridge, ON) and perfused through the heart with 0.1 M phosphate buffer (PB; pH 7.4; Sigma-Aldrich, Oakville, Ontario, Canada) followed by 4% paraformaldehyde (Sigma-Aldrich) in 0.1 M PB. Mice were decapitated, their brains extracted, and their eyes enucleated. Brains were placed into 4% paraformaldehyde at 4°C, and eyes were stored in 0.1 M phosphate-buffered saline (PBS; Sigma-Aldrich) at 4°C until further processing. 

Retinal Histology and Quantification.

Using a stereomicroscope (Nikon SMZ-21; Nikon Instruments Inc., Melville, NY), the anterior segment of the eye and the lens were removed. The retina was then detached from the optic nerve and mounted onto a positively charged gelatin-coated glass slide with the ganglion cell layer facing up. Four cuts were made in the radius of the retina, which was flattened into four quadrants and allowed to dry for at least 24 hours before being histologically processed. Retinal flat mounts were stained for Nissl substance using a 0.1% cresyl violet (Sigma-Aldrich) solution to evaluate cell count in the ganglion cell layer. 

Stereological estimates of the number of cells in the ganglion cell layer of the retina were obtained using the optical fractionator probe within StereoInvestigator software (MicroBrightField Inc., Williston, VT), with the investigator blinded to the age and treatment group of each mouse. The fractionater method systematically partitions the sections of tissue into equal-sized volumes (dissectors) to provide unbiased estimates of population size and density of cells in the ganglion cell layer of the retina. ³¹ StereoInvestigator software (MicroBrightField Inc.) integrates a three-axis motor-driven specimen stage (Olympus BX51 Research Microscope; Olympus, Center Valley, PA) with a computer workstation via a digital camera (Optronics Microfire, Goleta, CA) to perform random systematic sampling on the entire retina. The optical fractionator probe provided a step-by-step workflow that assisted the user in determining the optimal sampling parameters, including the appropriate sizing and spacing of the optical dissectors, in order to achieve the desired coefficient of error (Scheaffer CE) of ∼5%. ^31–33  

The entire flat-mounted retina was delineated under 4× (0.13 NA) magnification, and the counting frame was randomly placed to mark the first area to be sampled. Approximately 300 counting frames were then systematically chosen and uniformly distributed by the software. The number of cells was counted under 40× (0.75 NA) magnification, with a top and bottom guard zone of 5 μm. Each dissector counting frame was 20 × 20 μm, which allowed approximately one to five cells within each frame. Only cells that exhibited a clearly defined border with a stained cytoplasm, pale nucleus, and darkly stained nucleolus were counted within each frame. The estimation of cell population in the ganglion cell layer of the retina was computed by StereoInvestigator software (MicroBrightField Inc.) and was calculated using the product of the total number of cells counted at each of the dissectors, the section sampling fraction (number of sections/total area), the area sampling fraction (area of section sampled/total area), and the thickness sampling fraction (section thickness/dissector height). The cell counts of the right and left eye for each mouse were averaged, and this number was used for all analyses. 

Brain Histology and Quantification.

After perfusion, brains were sectioned in the coronal plane at 35 μm using a cryostat (2800 Frigocut; Reichert-Jung, Depew, NY) and collected in two series. Sections containing the superior colliculus were processed for the immunohistochemical detection of WGA-HRP to evaluate transneural labeling, and alternate sections were Nissl stained using a 0.1% cresyl violet solution to quantify cell count and cell size. Brain sections stained for WGA-HRP activity were reacted according to the recommended protocol for the peroxidase localization of wheat germ agglutinin in neural tissue by Vector Laboratories (provided in the public domain by Vector Laboratories: http://www.vectorlabs.com/protocols.aspx). Free-floating tissue was incubated in biotinylated anti-wheat germ agglutinin (1:500 dilution in 0.1 M PBS) followed by an avidin-biotin-peroxidase complex (1:100, Vectastain Elite ABC reagent; Vector Laboratories). Then, using 3,3′-diaminobenzidine tetrachloride (DAB) as the chromogen (DAB Substrate Kit; Vector Laboratories), sections were incubated for 2 to 10 minutes until the desired stain intensity was visible on the superior colliculus. 

Stereological estimates of cell count in the superior colliculus and the cross-sectional area (μm²) of cells in the superficial layers of the right and left superior colliculus were obtained using the optical fractionator probe within StereoInvestigator software (MicroBrightField Inc.) and the nucleator probe within the optical fractionator probe (MicroBrightField Inc.), respectively, with the investigator blinded to the age and drug group of each mouse. The superior colliculus was encompassed in approximately 12 Nissl-stained sections (from ∼3.78–4.78 mm posterior to bregma; provided in the public domain by Allen Institute for Brain Science: http://mouse.brain-map.org/static/atlas); and the first section to be counted was randomly selected from sections 1 to 3, with subsequent sections being systematically selected following a constant sampling intensity of every third section (a total of four sections were counted for each mouse) for cell count, and every third section for cell size. 

Under 4× (0.13 NA) magnification, the three superficial layers (stratum zonale, stratum griseum superficiale, and the stratum opticum) of the right and left superior colliculus were delineated on each section as the area of interest, and approximately 150 counting frames with standard inclusion (right, upper) and extended exclusion lines (lower, left) as boundaries were counted in the four sections under 40× (0.75 NA) magnification in order to achieve a coefficient of error of ∼5% using Scheaffer CE and Gundersen CE with m = 1. Each counting frame was 20 × 20 μm, which allowed one to five cells within each frame, with a top and bottom guard zone of 5 μm. The estimation of cell population within the superior colliculus was computed by StereoInvestigator (MicroBrightField Inc.) software and was calculated by the product of the total number of cells counted at each of the dissectors, the section sampling fraction (number of sections/total area), the area sampling fraction (area of section sampled/total area) and the thickness sampling fraction (section thickness/dissector height). Using the same 20 × 20 μm counting frame and parameters, cell size was measured at high magnification with a 100× (1.4 NA) oil objective on four sections of the superior colliculus to achieve the desired CE of ∼5% (Scheaffer CE and Gundersen CE with m = 1). To estimate of the cross-sectional area of each cell, the nucleator probe first marked the nucleolus inside the nucleus as a reference point and then created four lines from the nucleolus that intersected the cell membrane. Only cells that exhibited a clearly defined border with a stained cytoplasm, pale nucleus, and stained nucleolus were counted within each frame for cell count and cell size. 

Stereological estimates of the number of cells in the superior colliculus exhibiting WGA-HRP reaction product in the left superior colliculus (contralateral to the injected right eye) were obtained using the same methods and parameters described for cell count in Nissl-stained sections, with the exception that only ∼75 counting frames (20 × 20 μm) were necessary to achieve the desired coefficient of error of ∼5% (Scheaffer CE and Gundersen CE with m = 1). Only cells in which the entire somatic profile was clearly defined were included in the analysis. The estimation of the number of cells exhibiting transneural labeling of WGA-HRP within the left superior colliculus was computed by StereoInvestigator software (MicroBrightField Inc.) and was calculated by the product of the total number of cells counted at the each of the dissectors, the section sampling fraction (number of sections/total area), the area sampling fraction (area of section sampled/total area), and the thickness sampling fraction (section thickness/dissector height). 

Statistical Analyses

Sex differences were analyzed for all behavioral, ocular, and neural measurements; but because there were no significant sex differences, data from male and female mice of each drug group were pooled for all analyses. For the visual water task, the percentage of correct responses was analyzed across days for the visual detection and pattern discrimination tasks and across spatial frequencies for the visual acuity task using a 3 (drug treatment) × 8 (day or spatial frequency) between-within repeated measures ANOVA. Age effects within each drug group were analyzed using a 4 × 8 (age × days or spatial frequency) repeated measures ANOVA. One-way ANOVAs were used to analyze drug group differences at each age and age differences within each drug group in the percentage of correct responses on day 8 for the visual detection and pattern discrimination tasks and visual acuity threshold. Post hoc analyses were conducted using Fisher's protected least significance difference (PLSD) tests. Effects of drug treatment within each age group and age-related changes within each drug treatment group in IOP, cell count, cell size, and WGA-HRP-labeled cells were analyzed using one-way ANOVAs with Fisher's PLSD post hoc tests. Pearson product–moment correlations were used to examine the relationships between measures of visual ability, IOP, and cell count in the retinal ganglion cell layer and measures of transneural labeling, cell number, and cell size in the superior colliculus. Only the data from the age at which the mouse were sacrificed was used for the correlational analyses (one measurement per mouse). All statistical analyses were performed using Statview 5.0 (Abacus Concepts Inc., Berkeley, CA). 

Results

Behavioral Measures of Visual Ability

When mice were 3, 6, and 9 months of age, all three drug treatment groups were able to reach the criterion of 70% percent correct in the visual detection task (Figs. 1A–D). At 12 months of age, mice receiving 0.50% and 0.25% Tim reached 70% correct, but mice receiving 0.00% Tim did not. Mice receiving 0.50% Tim did not show any significant age-related changes in visual detection performance at 12 months of age, whereas mice receiving 0.25% Tim showed a significant decrease in percent correct responses at 12 months of age (P = 0.0013) and mice receiving 0.00% Tim performed at 50% percent correct (chance) at 12 months of age (P = 0.01). There were no significant differences between drug treatment groups in the percentage of correct responses on day 8 of visual detection testing at 3, 6, or 9 months of age; but at 12 months of age, mice receiving 0.00% Tim had a significantly lower percentage of correct responses [F(2,7) = 33.512, P = 0.0003] than mice receiving 0.50% (P < 0.0001) and 0.25% Tim (P = 0.0016). Mice receiving 0.25% Tim also performed worse than mice receiving 0.50% Tim at 12 months of age (P = 0.0263; Fig. 2A). 

Figure 1. 

 

Mean (±SEM) percentage of correct responses on each of the eight days of testing for mice receiving 0.50%, 0.25%, and 0.00% Timoptic-XE in the visual detection task at (A) 3, (B) 6, (C) 9, and (D) 12 months of age and the pattern discrimination task at (E) 3, (F) 6, (G) 9, and (H) 12 months of age. Frequency-of-seeing curves showing visual acuity thresholds in spatial frequency (c/deg) for mice receiving 0.50%, 0.25%, and 0.00% Timoptic-XE at (I) 3, (J) 6, (K) 9, and (L) 12 months of age.

Figure 2. 

 

Mean (+SEM) percentage of correct responses on day 8 of testing in the visual detection task (A) and pattern discrimination task (B) and visual acuity thresholds (C) at 3, 6, 9, and 12 months of age for mice receiving 0.50%, 0.25%, and 0.00% Timoptic-XE. At 12 months of age, mice receiving 0.00% Tim had a significantly lower percentage of correct responses in the visual detection and pattern discrimination tasks and a significantly lower visual acuity threshold than mice receiving 0.50% and 0.25% Tim. *Indicates a significant difference of <0.05.

At 3, 6, and 9 months of age, mice in all three treatment groups reached 70% correct in the pattern discrimination task; but at 12 months of age, mice in the 0.00% Tim group were not able to reach this criterion (Figs. 1E–H). Mice receiving 0.50% Tim showed an age-related decrease in the percentage of correct choices in the pattern discrimination task [F(3,31) = 9.583, P = 0.0001] as performance at 12 months of age was significantly worse than at 3 (P = 0.0370), 6 (P < 0.0001), and 9 months of age (P = 0.0370). The 0.25% Tim mice performed significantly worse at 12 months of age [F(3,27) = 7.494, P = 0.0008] than at 6 (P = 0.0398) or 9 months of age (P = 0.0009); and 0.00% Tim mice showed a decrease in percent correct as they aged [F(3,37) = 5.458, P = 0.0033], with performance at 12 months of age being significantly worse than at all three younger ages (P < 0.01). There were no significant differences between drug treatment groups in the percentage of correct responses on day 8 of the pattern discrimination task at 3, 6, or 9 months of age, but at 12 months of age mice receiving 0.00% Tim performed worse than those receiving 0.50% (P = 0.0063) and 0.25% Tim (P = 0.0123; Fig. 2B). 

There were no significant differences between drug groups in the percentage of correct responses in the visual acuity task at 3, 6, or 9 months of age (Figs. 1I–L), but at 12 months of age, the 0.00% Tim mice did not show a measurable visual acuity; their performance remained at ∼50% for all spatial frequencies tested. There were no significant age-related changes in the percentage of correct responses in the visual acuity task in mice receiving 0.50% or 0.25% Tim; but mice receiving 0.00% Tim showed a significant decrease in percent correct as they aged [F(3,36) = 5.120, P = 0.0047], with performance at 12 months being significantly worse than at younger ages (P < 0.05). Based on these data, mice receiving 0.00% Tim had a significantly lower visual acuity threshold at 12 months of age than mice receiving 0.50% or 0.25% Tim (P < 0.0001; Fig. 2C). 

Intraocular Pressure

At 2 months of age there were no significant differences in IOP between prospective drug groups; but at 3 months of age, mice receiving 0.00% Tim had significantly higher IOP than mice receiving 0.50% or 0.25% Tim (P < 0.01), and mice receiving 0.25% Tim had significantly higher IOP than mice receiving 0.50% Tim (P = 0.0372), demonstrating the dose-related effect of Timoptic-XE even at young ages (Fig. 3). Between 3 and 6 months of age, there were significant increases in IOP in the 0.50% [F(4,38) = 13.097, P < 0.0001], 0.25% [F(4,33) = 9.092, P < 0.0001], and 0.00% [F(4,44) = 26.553, P < 0.0001) Tim groups, but there was no significant difference between treatment groups. At 9 months of age, IOP of mice receiving 0.00% Tim increased further (P = 0.0033) and was significantly higher than in mice receiving 0.50% (P = 0.0015) or 0.25% Tim (P = 0.0010). At 12 months of age, the 0.00% Timoptic-XE mice had higher IOP than the other two groups, but this did not reach significance due to the small sample size [F(2,7) = 2.143, NS]. 

Figure 3. 

 

Mean (+SEM) intraocular pressure (mm Hg) for mice receiving 0.50%, 0.25%, and 0.00% Timoptic-XE at 2, 3, 6, 9, and 12 months of age. At 3 months of age, 0.00% Tim mice had significantly higher IOP than mice receiving 0.50% or 0.25% Tim, and mice receiving 0.25% Tim had significantly higher IOP than mice receiving 0.50% Tim. Between 3 and 6 months of age, there were significant increases in IOP in the 0.50%, 0.25%, and 0.00% Tim groups, but there was no significant difference between treatment groups. At 9 months of age, IOP of mice receiving 0.00% Tim was significantly higher than in mice receiving 0.50% or 0.25% Tim. *Indicates a significant difference of <0.05.

Cell Count in the Ganglion Cell Layer of the Retina

There were no significant differences in cell count in the ganglion cell layer of the retinas between the right and left eye in any of the three treatment groups, indicating that the intraocular injection of WGA-HRP given to the right eye had no effect on cell survival and was not responsible for differences in cell count in the ganglion cell layer between drug treatment groups. All mice showed a significant decrease in cell count in the ganglion cell layer as they aged (Fig. 4), with 9- and 12-month-old mice having significantly fewer cells than 3- and 6-month-old mice in the 0.50% Tim [F(3,9) = 11.502, P = 0.0020], 0.25% Tim [F(3,8) = 6.502, P = 0.0154], and 0.00% Tim groups [F(3,10) = 26.821, P < 0.0001]. There were no significant differences in cell count between drug treatment groups at 3, 6, or 9 months of age, but 12-month-old mice receiving 0.00% Tim had significantly fewer cells in the ganglion cell layer [F(2,7) = 10.523, P = 0.0078] than mice receiving 0.50% (P = 0.0049) or 0.25% Tim (P = 0.0078). 

Figure 4. 

 

(A) Mean (+SEM) cell count in the ganglion cell layer of the retina for mice receiving 0.50%, 0.25%, and 0.00% Timoptic-XE at 3, 6, 9, and 12 months of age. All mice showed a significant decrease in cell count in the ganglion cell layer as they aged, with 9- and 12-month-old mice having significantly fewer cells than 3- and 6-month-old mice. Twelve-month-old 0.00% Tim mice had significantly fewer cells in the ganglion cell layer than mice receiving 0.50% or 0.25% Tim. Representative pictomicrographs of Nissl-stained flat-mounted retinas, showing the ganglion cell layer at 40X magnification from 3- and 12-month-old mice treated with 0.50% Timoptic-XE (B, E), 0.25% Timoptic-XE (C, F), and 0.00% Timoptic-XE (D, G). *Indicates a significant difference of <0.05.

Cell Count and Cell Size in the Superior Colliculus

There were no significant differences between drug treatment groups in cell count in Nissl-stained sections of the superior colliculus at any age, and no significant age-related changes in cell count in the superior colliculus in mice receiving 0.50%, 0.25%, or 0.00% Timoptic-XE (Fig. 5A). Mice receiving 0.50% Tim did not show any age-related changes in cross-sectional area of cells in the superior colliculus, but mice receiving 0.25% Tim [F(3,8) = 4.087, P = 0.0494] and 0.00% Tim [F(3,11) = 10.834, P = 0.0013] showed significant decreases in cross-sectional area of cells in the superior colliculus at 12 months of age (Fig. 5B). 

Figure 5. 

 

Mean (+SEM) (A) cell count and (B) cross-sectional area (μm²) of cells in the superior colliculus at 3, 6, 9, and 12 months of age for mice receiving 0.50%, 0.25%, and 0.00% Timoptic-XE. Mice receiving 0.25% Tim and 0.00% Tim showed significant decreases in cross-sectional area of cells in the superior colliculus at 12 months of age. *Indicates a significant difference of <0.05.

Transneural Labeling in the Superior Colliculus

Drug treatment groups did not differ in the number of WGA-HRP–labeled cells in the superior colliculus at 3 (Figs. 6B–D) or 6 months of age. However, at 9 months of age, significant differences in the number of labeled cells [F(2,6) = 5.207, P = 0.0488; Fig. 6A] indicated that mice receiving 0.50% Tim had significantly more transneural labeling than mice receiving 0.00% Tim (P = 0.0210). At 12 months of age, mice receiving 0.00% Tim showed significantly less transneural labeling [F(2,6) = 5.248, P = 0.0481] than mice receiving 0.50% (P = 0.0284) or 0.25% Tim (P = 0.0339) (Figs. 6E–G). Mice receiving 0.50% or 0.25% Tim did not show any age-related changes in transneural labeling, while mice receiving 0.00% Tim showed a significant decrease in the number of WGA-HRP–labeled cells as they aged [F(3,11) = 10.834, P = 0.0013], with measures at 9 and 12 months of age being significantly lower than at 3 and 6 months of age. 

Figure 6. 

 

(A) Mean (+SEM) count of WGA-HRP–labeled cells in the superior colliculus at 3, 6, 9, and 12 months of age for mice receiving 0.50%, 0.25%, and 0.00% Timoptic-XE. At 9 months of age, 0.00% Tim mice had significantly less transneural labeling than 0.50% Tim mice; and at 12 months of age, these mice showed significantly less transneural labeling than both treated groups. Mice receiving 0.50% or 0.25% Tim did not show any age-related changes in transneural labeling, while mice receiving 0.00% Tim showed a significant decrease in the number of WGA-HRP–labeled cells as they aged. Representative pictomicrographs of 35 μm coronal sections at 4X magnification, showing WGA-HRP labeling of the superior colliculus from 3- and 12-month-old mice receiving 0.50% Timoptic-XE (B, E), 0.25% Timoptic-XE (C, F), and 0.00% Timoptic-XE (D, G). Boxed areas in E, F, and G show the approximate location at which magnified 40X pictomicrographs were taken for 12-month-old mice receiving 0.50% Timoptic-XE (H), 0.25% Timoptic-XE (I), and 0.00% Timoptic-XE (J). *Indicates a significant difference of <0.05.

Correlations between Behavioral, Ocular, and Neural Measurements

The correlation matrix between behavioral, ocular, and neural measures using mice from all four ages with complete data (N = 38) is shown in Table 2. All three behavioral measures of visual ability (percentage of correct responses on day 8 of the visual detection and pattern discrimination task and visual acuity threshold) were significantly positively correlated, while IOP was significantly negatively correlated with each measure of visual ability, cell count in the ganglion cell layer of the retina, and number of WGA-HRP–labeled cells in the superior colliculus. Cell count in the ganglion cell layer of the retina was significantly positively correlated with performance in all three vision tasks, as well as cell size in the superior colliculus and number of WGA-HRP–labeled cells in the superior colliculus, and was negatively correlated with IOP. Cell count in the superior colliculus was not correlated with any measure of visual ability, but cell size in the superior colliculus was significantly correlated with percent correct on day 8 of the visual detection task, visual acuity threshold, and cell count in the ganglion cell layer of the retina. The number of WGA-HRP–labeled cells in the superior colliculus was significantly positively correlated with all three measures of visual ability and cell count in the retina, but significantly negatively correlated with IOP. 

Table 2. 

 

Pearson Product-Moment Correlations Showing the Relationship between Behavior, Ocular, and Neural Measures in Mice Aged 3, 6, 9, and 12 Months

	% Correct PD 8	VA Threshold	IOP	RGC Count	SC Cell Count	SC Cell Area	WGA-HRP Labeling
% Correct VD 8	0.779****	0.533****	−0.328*	0.489***	0.070	0.444***	0.475***
% Correct PD 8	-	0.484***	−0.442***	0.569****	0.077	0.230	0.528****
VA threshold		-	−0.388**	0.401**	0.194	0.367*	0.434***
IOP			-	−0.561****	−0.271	−0.201	−0.438***
RGC count				-	0.021	0.492***	0.382**
SC cell count					-	0.109	0.265
SC cell area						-	0.099

 VD, visual detection; PD, pattern discrimination; VA, visual acuity; RGC, retinal ganglion cell; SC, superior colliculus.

 * P < 0.05.

 ** P < 0.02.

 *** P < 0.01.

 **** P < 0.001.

Discussion

This study provides a comprehensive assessment of the efficacy of a conventional anti-glaucoma medication (Timoptic-XE) on behavioral, ocular, and neural measurements of glaucomatous damage from 3 to 12 months of age in the DBA/2J mouse model of pigmentary glaucoma. Long-term Timoptic-XE therapy administered from 9 weeks to 12 months of age prevented deficits in visually dependent behavioral tasks and reduced IOP and cell loss in the retina. In addition, Timoptic-XE therapy prevented the decrease in transneural labeling and somal shrinkage in the superior colliculus that occurred in untreated control mice. Because we measured age-related changes in these parameters, we can determine the progression of glaucomatous changes and therefore speculate on their consequences for the development of the disease. 

As in humans, Timoptic-XE was successful in reducing IOP in both young and old DBA/2J mice (Fig. 3). IOP increased significantly from 3 to 6 months of age in all mice (from ∼14 mm Hg to 21 mm Hg), and this higher IOP was stable in both 0.50% and 0.25% Timoptic-XE mice at 9 and 12 months of age. IOP in 0.00% Timoptic-XE mice was higher at 9 and 12 months of age than it was at 6 months of age. Thus, it appears that the doses of Timoptic-XE used in this study did not protect against the gradual increase in IOP that occurred before 6 months of age but prevented the subsequent increase in IOP at 9 months of age. This subsequent elevation of IOP may be particularly important for the induction of cell death in the ganglion cell layer of the retina that occurred in 0.00% Timoptic-XE mice at 12 months of age. 

Intraocular pressure was significantly negatively correlated with cell count in the ganglion cell layer of the retina (Table 2), suggesting that elevated intraocular pressure is related, either directly or indirectly, to retinal ganglion layer cell death. Mice in all three drug groups showed an age-related increase in cell loss in the ganglion cell layer of the retina at 9 and 12 months of age; however, this effect was greatest for 0.00% Timoptic-XE mice (Fig. 4). Cell loss in the retinal ganglion cell layer in 0.00% Timoptic-XE mice was maximal at 12 months of age following a period of high IOP from 6 months of age. However, cell loss in the retina was not significantly different between drug treatment groups at 9 months of age, despite IOP being significantly higher in 0.00% Timoptic-XE mice. This dissociation between elevated IOP and cell death at 9 months of age could be indicative of an IOP-independent mechanism causing retinal ganglion cell death; alternatively, it could suggest that a window of time exists between IOP elevation and cell death in the retina, when IOP reduction therapy may preserve retinal cell survival. In our study, the prevention of IOP elevation was sufficient to prevent further retinal cell loss at 12 months of age in mice treated with 0.50% and 0.25% Timoptic-XE. It may therefore be possible to start Timoptic-XE therapy at 6 months of age and still reduce IOP and retinal cell loss at 12 months of age. Because both retinal ganglion cells and displaced amacrine cells were stained in the ganglion cell layer by cresyl violet, it was not possible to determine which type of cells was more affected by increased IOP. However, since amacrine cell number is not affected in glaucoma, at least in DBA/2J mice, ^16,19,34 the loss of cells in the ganglion cell layer can be attributed exclusively to the retinal ganglion cell loss. ^13,35  

Retinal ganglion cells undergo a progression of structural changes preceding complete degeneration. These changes include somal shrinkage, truncated and misdirected dendrites, abnormally thin axons, deficits in anterograde and retrograde labeling, and downregulation of retinal ganglion cell-specific genetic programs. ^34,36–39 Retrograde transport from the superior colliculus to the retinal ganglion cells in DBA/2J mice is compromised by 12 months of age, and this deficit precedes retinal ganglion cell axon and somal loss. ^34,37 Disruption of anterograde transport from the retina to the superior colliculus occurs by 8 to 10 months of age and progresses from distal to proximal, appearing in the colliculus first, followed by appearance in more proximal secondary targets and then in the optic tract before transport finally fails in the retina. ³⁶  

The results of the WGA-HRP transneural labeling indicate that distal transport loss is predegenerative and may represent a therapeutic target, as there may be a window of time before complete retinal ganglion cell degeneration and thus irreversible vision loss occur. ⁴⁰ Our results provide support for this hypothesis because differences in the anterograde transport of WGA-HRP to the superior colliculus between mice treated with different doses of Timoptic-XE emerged at 9 months of age, three months earlier than significant differences between treatment groups in retinal cell loss (Fig. 6). Transneural labeling of WGA-HRP in the superior colliculus was significantly reduced in control mice at 9 and 12 months of age, but this effect was prevented in 0.50% and 0.25% Timoptic-XE mice, which did not show such an age-related decrease in transneural labeling in the superior colliculus. 

Intraocular pressure, retinal ganglion cell count, and transneural labeling in the superior colliculus were significantly correlated with visual ability in behavioral tasks (Table 2). At 3, 6, and 9 months of age, DBA/2J mice receiving 0.00% Timoptic-XE were able to reach the criterion of 70% correct in the visual detection and pattern discrimination tasks; they had a definable visual acuity threshold that did not differ from that in mice receiving 0.50% and 0.25% Timoptic-XE and was similar to that in untreated mice with normal vision. ²⁶ However, at 12 months of age, mice receiving 0.00% Timoptic-XE performed at chance (50%) levels in all three vision tasks (Fig. 1), which is similar to the performance of mice with retinal degeneration in these tasks. ²⁶ These results indicate that retinal cell dysfunction and/or loss is extensive enough to render mice receiving 0.00% Timoptic-XE functionally blind by 12 months of age. ¹⁴  

By combining histological examination with behavioral evaluation, we have provided a method for quantifying the ocular effects of glaucomatous damage by psychophysical testing and also identified the time point at which the amount of retinal cell loss resulted in total visual impairment, which is crucial for determining the point at which vision loss is salvageable and possibly reversible with the appropriate treatment regime. Although retinal cell loss showed an age-related decrease that was significant by 9 months of age in all drug groups, mice were still able to successfully detect visual stimuli as well as they could at younger ages. Thus visual ability was still functional despite significant retinal ganglion cell loss at 9 months of age. This finding is particularly noteworthy because it provides an insight into the structural–functional relationship of retinal cells. In humans and primates with glaucoma, a relatively large proportion (50%–60%) of retinal ganglion cells must be lost before visual field thresholds in clinical perimetry measurements exceed the normal range and reach statistical significance. ^41,42  

In the mouse, the superior colliculus is the main target of primary optic projections as it receives direct input from the retina, with most (70%–90%) retinal ganglion cells projecting to its three superficial layers: the stratum zonale, stratum griseum superficiale, and the stratum opticum. ^43–45 In the present study, we did not find any significant differences in cell count in the superior colliculus between drug treatment groups at any of the ages tested (Fig. 5A) or within drug groups as mice aged. However, we found that cell size in the superior colliculus was significantly smaller in 0.25% and 0.00% Timoptic-XE mice at 12 months of age compared to 3 to 9 months of age (Fig. 5B), and these age-related changes in cell size were not significant in 0.50% Timoptic-XE mice. Shrinkage of neurons is a well-known cellular response to injury, and persistence of the pathogenic state can lead to neuronal death. ⁴⁶ It is possible that the cells in the superior colliculus in 12-month-old 0.25% and 0.00% Timoptic-XE mice were undergoing a progression of structural changes, such as somal shrinkage, that preceded complete degeneration and therefore that cell loss in the superior colliculus would be evident at older ages in these mice. 

To our knowledge this is the first study that demonstrates changes in the visual centers of the brain in aging DBA/2J mice as they go blind. In other rodents, cell loss has been demonstrated in the superior colliculus in adult mice following N-methyl-D-aspartate (NMDA)-induced retinal damage, ⁴⁷ and a decrease in the volume of the superior colliculus and the density of the retinocollicular projection has been demonstrated after partial loss of retinal ganglion cells induced by intravitreal injections of kainic acid in adult rats. ⁴⁸ In addition, evidence for reorganization of the retinocollicular projection has been observed in adult rats following partial crush of the optic nerve, with surviving retinal ganglion cell axons from all quadrants of the retina redistributing their projections to the rostromedial region of the superior colliculus. ⁴⁹  

Glaucoma may involve transsynaptic or transneuronal degeneration, which involves the transmission of the disease from sick to healthy neurons by synaptic connections along anatomical and functional neural pathways. ^50,51 In humans and primates, approximately 90% of retinal ganglion cells project to the lateral geniculate nucleus (LGN), the first major visual target center in the brain and the predominant relay station of information to the primary visual cortex. ⁵¹ In humans with glaucoma, LGN neuron density is decreased ⁵² and the optic nerve, LGN, and visual cortex are degenerated, with the degree of degeneration correlating with inferior rim loss of the optic nerve head and visual field deficits. ⁵³ LGN atrophy in glaucoma patients compared to normal controls was shown in a magnetic resonance imaging (MRI) study, ⁵⁴ and visual function loss in glaucoma patients was correlated with alterations in the blood oxygen level–dependent (BOLD) functional MRI response in the primary visual cortex, suggestive of retinotopic reorganization of the primary visual cortex in response to glaucomatous damage. ⁵⁵ Therefore, the measured visual field loss in glaucoma patients is not only representative of optic nerve damage and retinal ganglion cell loss but is also indicative of the damage to the visual pathways in the brain, as proposed by the transsynaptic theory of glaucoma. ^50,56 Thus, the significant correlations between visual behavior (percent correct on day 8 of the visual detection task and visual acuity threshold) and cell size in the superior colliculus in the present study (Table 2), may indicate that the age-related decline in visual ability in DBA/2J mice is representative of neuronal degeneration occurring in the superior colliculus. 

In conclusion, we have documented the natural age-related progression of behavioral, ocular, and neural changes in the DBA/2J mouse model of glaucoma and demonstrated that the administration of Timoptic-XE from 9 weeks to 12 months of age was successful in preventing glaucomatous damage. This study is the first to correlate natural age-related visual system changes in the retina and brain with changes in visual ability in DBA/2J mice, thereby evaluating a protective strategy not only on its ability to rescue retinal ganglion cell layer atrophy but also on its potential to restore visual function and alter patterns of neurodegeneration that occur with blindness. The DBA/2J mouse is a powerful experimental tool for elucidating, under rigorous experimental conditions, the cellular, molecular, and genetic mechanisms that underlie human glaucoma. Because these mice display a clinical phenotype similar to that of humans with glaucoma and are responsive to the same therapy, they are a valuable resource for testing new glaucoma treatment regimens for humans. 

Acknowledgments

We thank Rhian Gunn, Ashley Whittaker, Bryan Daniels, Harjit Seyan, and Stephen Whitefield for their technical assistance. Anatomical studies were carried out in the Cellular & Molecular Digital Imaging facility at Dalhousie University. 

Footnotes

 Supported by a grant from the Natural Sciences and Engineering Research Council of Canada (REB) and a Doctoral Research Award from the Canadian Institutes of Health Research (AAW).

Footnotes

 Disclosure: A.A. Wong, None; R.E. Brown, None