Abstract
Visual impairments and retinal abnormalities occur in patients with Alzheimer’s disease (AD) and in mouse models of AD. It is important to know the visual ability of mouse models of AD to ensure that age-related cognitive deficits are not confounded by visual impairments. Using OptoMotry, the spatial frequency thresholds of male and female 3xTg-AD mice did not differ from their B6129SF2 wildtype controls between 1–18 months of age, but females had higher spatial frequency thresholds than males. However, the differences were quite small, and the visual ability of all mice was comparable to that of C57BL/6 mice.
INTRODUCTION
Patients with Alzheimer’s disease (AD) often develop visual impairments in color discrimination [1, 2], spatial contrast sensitivity [3–5], and backward masking [6], along with a decrease in electroretinogram responses [7] and visual evoked potentials [8, 9]. These visual deficits have been attributed to changes occurring throughout the visual pathway in AD [10, 11]. Since many anatomical and functional changes in the visual system have been documented in patients with AD, using the eye as biomarker for AD may be a non-invasive, inexpensive, early detection method for diagnosing AD [10–13]. Ocular biomarkers for AD could include any of the changes in the visual system that have been previously reported in AD patients, including changes in the retinal nerve fiber layer [14–16], retinal blood flow and vasculature [17, 18], and choroidal thickness [19, 20], as well as amyloid and tau deposition in the retina [21, 22]. Potential non-retinal biomarkers in the eye include pupillary reactions [23, 24], imaging the crystalline lens [25], and eye movements [26, 27]. It has been suggested that successful treatment of AD should start in the prodromal stages [28], therefore ocular biomarkers that can identify patients in the early stages of AD could have implications for the diagnosis of AD and for the development and evaluation of new therapeutic treatments.
Mouse models of AD are engineered to express mutant human AD genes for amyloid precursor protein (APP), and/or the presenilins (PS1 and PS2), and/or tau [29–34]. APP and PS mutations lead to excessive production of amyloid-β (Aβ), and Aβ plaques, and tau mutations lead to the expression of hyperphosphorylated tau and neurofibrillary tangles in the brain. Changes in the retina of AD model mice [35] include a significant overexpression of APP, loss of retinal neurons, glial reactivity, and the deposition of Aβ or hyperphosphorylated tau [36–41]. The 3xTg-AD mouse model of AD is homozygous for the PS1 mutation and the APPSwe and tauP301L transgenes, and develops both Aβ plaques and neurofibrillary tangles [42, 43]. Intracellular Aβ build-up occurs in the cortex as early as 3 months of age, and extends to the hippocampus between 3–6 months of age, followed by extracellular plaque formation at 6 months of age. Hyperphosphorylated tau aggregates are detected in the hippocampus at 6 months of age and tangle pathology becomes severe after 12 months of age [43]. The 3xTg-AD mouse shows deficits as early as 2 months of age in reference and working memory in the radial arm maze [44] and deficits in spatial learning and memory in the Barnes maze by 6.5 months of age [45].
Behavioral tasks such as the Barnes maze, Morris water maze, and radial arm maze require the use of visual-spatial cues, and visual impairments can confound the performance of mice in these tasks [46]. Since poor performance in these vision-based cognitive tasks could be the result of impaired vision rather than cognitive decline, it is important to determine whether the visual capabilities of animal models of AD decline during aging. The purpose of the present study was to determine whether 3xTg-AD mice show age-related changes in spatial frequency threshold (SFT) compared to B6129SF2 wildtype control mice, and whether there are sex differences in the SFT of these mice.
MATERIALS AND METHODS
Seventy-four 3xTg-AD mice (34 male and 40 female) and 80 B6129SF2 WT mice (40 male and 40 female) were tested from 1–18 months of age. All mice were bred in-house from parents originally purchased from the Jackson Laboratory (Bar Harbour, ME, 3xTg-AD, JAX#004807; B6129SF1/J, JAX #101043). After weaning at 21 days of age, mice were housed in same-sex groups of 2–4 littermates, or were single-housed in plastic cages (18.75×28×12.5 cm) with a PVC tube (4 cm diameter×7 cm length) for enrichment, woodchips for bedding, and metal wire covers. They had ad libitum access to rodent chow (Purina #5001) and tap water, and were housed in a colony room (22±2°C) on a reversed 12 : 12 h light:dark cycle with lights off at 10am. All testing was completed during the dark phase of the light:dark cycle. Some of the mice were only tested once (n = 58) and some were tested multiple times as they aged (n = 89). All experimental procedures were conducted in accordance with the guidelines of the Canadian Council on Animal Care and were approved by the Dalhousie University Committee on Laboratory Animals (#15-099).
All mice were tested in an OptoMotry apparatus (CerebralMechanics Inc.), a computer-based, non-invasive method for testing various visual thresholds in untrained, unrestrained mice. Mice were placed one at a time on a small, elevated platform within an enclosed black Plexiglas box with a hinged door. Visual stimuli were sine-wave gratings that were projected as a virtual cylinder on four computer monitors, which were positioned around the mouse [47, 48]. When the cylinder wall ‘rotates’, the mouse fixates on the stimuli and follows the grating with reflexive head movements that coincide with the rotation of the cylinder. When the mouse no longer tracks the stimuli, it is determined that the mouse can no longer ‘see’ the grating, therefore the visual SFT has been surpassed. The SFT for each eye is measured independently by controlling the direction of the rotation of the stimuli, with a clockwise rotation measuring the left eye, and a counter clockwise rotation measuring the right eye. For the current project, the spatial frequency of the sine-wave gratings was increased until head movements were no longer detectable. Thresholds were measured for both eyes, and then averaged for statistical analyses, as no significant differences between eyes were determined. A univariate ANOVA was used to analyze the effects of genotype and sex on SFT but because age groups were not categorically distinct, a multiple-level linear regression was run using age, genotype, sex, and repeated measure condition (if the mouse was tested multiple times as they aged, or only once) as potential predictors of SFT. The program SPSS was used for statistical analyses.
RESULTS
SFTs in cycles per degree (cpd) across the lifespan are shown for male (Fig. 1A) and female (Fig. 1B) 3xTg-AD and B6129SF2 WT mice. There were no significant genotype differences in SFT (F (1,150) = 1.393) between 3xTg-AD (0.3971 cpd±0.0229) and WT mice (0.3988 cpd±0.0113) but females (0.4070 cpd±0.01354) had a significantly higher SFT (F (1,150) = 7.615, p < 0.007) than males (0.4028 cpd±0.0131). There was not a significant interaction between genotype and sex (F (2,150) < 1.0).
Spatial frequency thresholds, in cycles per degree (cpd), for male (A) and female (B) 3xTG-AD and B6129SF2 WT mice across the lifespan. For both A and B, age on testing day is on the abscissa and spatial frequency threshold (cpd) is on the ordinate, with closed triangles representing B6129SF2 animals, and open circles 3xTg-AD animals.
A multiple-linear regression analysis indicated that the best model for predicting SFT included sex and age (F (2,151) = 6.930, p < 0.001, R2 = 0.084). Predicted SFT was equal to SFT = 0.396 + 0.007 (sex) – 0.00002111 (age), where sex is coded as 1 = male, 2 = female, and age is measured in days. Therefore, the SFT of female mice was 0.007 cpd higher than male mice and SFT decreased by 0.00002111cpd for each day of age. Mouse genotype and repeated measures testing were not significant predictors of SFT in this model.
DISCUSSION
To determine whether 3xTg-AD mice had age-related changes in their visual ability compared with B6129F2 WT controls, their SFTs were measured across the lifespan. There were no significant differences between genotypes but female mice had a significantly higher SFT than males. Age and sex were significant predictors and accounted for 7.2% of the variance in SFT. However, the differences between sexes and the decline in SFTs as mice aged were very small. In fact, the SFTs of all mice were comparable to those of C57BL/6J mice [48], indicating that visual ability in the 3xTg-AD model is “normal” for a mouse.
Although structural changes in the retina in AD mice have been documented [36–41], the present study is one of the few that investigates the functional vision of AD mice. Leinonen et al. [49] found little evidence of altered vision in APP/PS1 mice, as these mice had normal amplitude of electroretinography and visual evoked responses to patterned stimuli at 13–15 months, an age when spatial memory deficits and plaque formation are fully manifested in these mice. This result contrasts with others that have observed abnormal electroretinography and visual evoked responses in these mice at 13 months [50], 11 and 24 months [51], and 12 to 19 months [41] of age.
We previously found that 20–26-month-old APP/PS1 mice had impaired visual acuity, resulting in deficits in visuo-spatial learning and memory in the Morris water maze, but not in conditioned odor preference or conditioned taste aversion memory tasks [52]. However, the method used to measure visual ability in APP/PS1 mice differs from that used in the present study. The visual water box is a learning-based swimming task that measures cortical vision and produces higher visual threshold values than OptoMotry [48, 53]. Measurement of vision by OptoMotry uses the optokinetic tracking response, which is a reflexive head movement in the direction of moving visual stimuli and is dependent on the function of subcortical visual pathways [48]. Thus, visual thresholds obtained from the two methods cannot be directly compared since they rely on different visual pathways and, in the case of the visual water task, require other skills to complete the task that are independent of vision, such as swimming ability, learning, and memory. Since these parameters can also be affected in aging AD model mice, and could confound the results in the visual water task, we recommend using OptoMotry to measure visual thresholds in mice with motor, learning, or memory deficits.
There has been little work documenting the age-related changes in the visual system of 3xTg-AD mice. The 3xTg-AD mice have significantly more Aβ42 peptide in the retina than WT control mice, but the Aβ42 levels were much lower than in the 5xFAD and Tg2576 mouse models of AD [54]. Retinal macroglial changes, such as Müller cell and astrocyte activation, were found in 9-month-old 3xTg-AD mice [37], and phosphorylated tau accumulation was observed in retinal ganglion cell soma, dendrites, and intraretinal axons in 3xTg-AD mice as early as 3 months of age, preceding its aggregation in the brain [55]. The results of the present study demonstrate that although age-related structural changes may be occurring in the retina, visual ability is still normal at older ages in the 3xTg-AD mice. In contrast, many genotype and sex differences have been documented at different ages in the 3xTg-AD mouse model in other behavioral tasks. For example, 3xTg-AD mice show a greater acoustic startle response and greater PPI of startle than WT mice between 6 and 18 months of age [56] and 6-month-old 3xTg-AD females but not males show age-related deficits in olfactory acuity [57]. In terms of motor behavior tests, 3xTg-AD mice perform better than WT mice in the rotarod test of motor-coordination and motor learning but are worse than WT mice on the balance beam between 2 and 15 months of age [58, 59]. Female 3xTg-AD mice often outperform males on motor tasks but this is due to their lower body weight. On visuospatial cognitive tasks, 3xTg-AD mice perform worse in reference and working memory in the 8 arm radial arm maze than WT mice from 2 to 15 months of age and males make more errors than females [44]. The 3xTg-AD mice also perform worse than WT mice on learning and memory in the Barnes maze at 6 months of age [45]. The present study demonstrates that these genotype differences in behavioral tasks that rely on visual cues are not due to differences in visual ability. Although we found that female mice of both genotypes had significantly higher SF thresholds than males, this difference of 0.007 cpd between the sexes is so small that it is unlikely to result in a difference in behavioral outcome on vision-based tasks. Therefore, based on our results measuring SFT using OptoMotry, 3xTg-AD mice can be tested in visual based behavioral tasks without the concern that performance is confounded by visual deficits.
