Normal Amplitude of Electroretinography and Visual Evoked Potential Responses in AβPP/PS1 Mice

INTRODUCTION

Alzheimer’s disease (AD) is the most common neurodegenerative disorder worldwide and its prevalence will drastically increase as the population ages [1]. The most disabling manifestations of AD are cognitive defects; however, sensory systems including vision are also affected [2, 3]. Besides impairments in visual tests [4 –7], declined electroretinography (ERG) and visual evoked potential (VEP) responses have been reported in AD patients [8, 9]. Anatomically, a pioneering study showed optic nerve degeneration from AD patients’ postmortem samples [10]. Later studies reported retinal ganglion cell degeneration [11, 12] and retinal amyloid-β (Aβ) plaques [13]. Several optical coherence tomography imaging studies suggestthinning of retinal layers, particularly the retinal nerve fiber layer [9 , 14–16].

In animal models, the AβPPswe/PS1dE9 (AβPP/PS1) mice, a common transgenic mouse model of AD-related brain amyloidosis, has been reported to manifest retinal Aβ plaque formation similar to human AD patients [13]. Aβ accumulation in the retina has been found in other amyloid plaque forming AβPP and/or PS1 transgenic lines as well [17 –19]. Aβ deposits have been detected in the AβPP/PS1 mouse retina in postmortem immunohistochemistry [13 , 21] and also using in vivo imaging [13]. Few previous studies have reported ERG alternations in aged AβPP/PS1 mice [20, 22]. Moreover, a recent follow-up study found a significant age-related impairment in visual discrimination between young adult (5–8 months) and aged (20–26 months) mice, which was slightly more severe in AβPP/PS1 mice than in wild-type littermates [23]. These results are very alarming concerning the validity of the transgenic AβPP mouse models for the cognitive decline in AD in general, since the consistently reported age-dependentimpairment in spatial memory in AβPP transgenic mice as assessed by Morris swim task [24, 25] could equally well result from age-related visual impairment. Therefore, we compared the function of the entire visual pathway between AβPP/PS1 mice of our colony and their wild-type littermates at an age (13–15 months) when spatial memory deficit and brain amyloidosis are fully manifested [25].

MATERIALS AND METHODS

Male AβPPswe/PS1dE9 mice (AβPP/PS1 mice, n = 22 in total) and wild-type littermates (n = 24 in total) at the age of 13–15 months were used in the study. Male mice were chosen because male C57BL/6J mice show better visual contrast sensitivity [26] and spatial memory in the Morris water maze task [27] than females. The mice were bred in our local colony at the University of Eastern Finland. The line was originally generated as C3HxC57BL/6J hybrid (John Hopkins University, Baltimore, MD, USA; [28]) but had been back-crossed to C57BL/6J for 16 generations. Rd1 mutation has been screened in our mouse line and is absent (Leinonen et al., unpublished data). Visual testing was done by optimized ERG and VEP recordings derived from protocols published earlier [29 –31]. The ERG responses (AβPP/PS1: n = 10, wild-type: n = 9) to brief luminance flashes were recorded in dark- and light-adapted conditions to probe rod- and cone-mediated vision, respectively. The VEPs were recorded for pattern stimuli in constant luminance. Only a subset of recorded mice (n = 27 recorded, n = 12 accepted) were accepted to final analysis according to VEP profile and histology criteria illustrated in Fig. 2A-C. The visual acuity and contrast threshold derived from pattern VEPs has been shown to correspond with the values obtained by learned visual tasks and optokinetic reflex in mice [26, 32], and recently a direct link was established between VEP and behavioral response magnitude to visual stimulation [33]. More detailed description of the methods can be found in our Supplementary Material. The research was conducted under the EU directive 2010/63/EU for animal well-being using protocols approved and monitored by the Animal Experiment Board of Finland (ESAVI-2013-002477).

RESULTS

In this study, we recorded visual sensory function by electrophysiology from the eye and the brain in response to standardized stimuli [29 –31]. The ERGs were recorded under isoflurane anesthesia whereas the VEPs were recorded in awake, head-fixed mice. The amplitude of ERG responses did not differ between genotypes neither in dark-adapted (rod-pathway) nor in light-adapted (cone-pathway) conditions (Fig. 1A,B,C,F). Surprisingly at non-linear phase of intensity-response function (i.e., at low stimulus intensities), the AβPP/PS1 mice had faster dark-adapted ERG kinetics (Fig. 1D, see Supplementary Table 1). In addition, the frequency of oscillatory potentials in dark-adapted ERG were faster in AβPP/PS1 mice (Fig. 1E; U-test: z = –2.22, p <  0.05; wild-type: mean±SEM = 107.2±1.2 ms, AβPP/PS1: mean+SEM = 112.2±1.6 ms). At light-adapted conditions, the ERG response was equally fast between genotypes (Fig. 1G). Similar heart and breathing rates between genotypes advocated similar response to anesthesia (Fig. 1H). The VEP response was similar between genotypes over changing pattern size (Fig. 2F,G) and contrast (Fig. 2I,J; U-test: p >  0.1 for all stimulus modalities and parameters). Sample tests showed normal retinal morphology in AβPP/PS1 mice (Fig. 1I). The brain Aβ burden was verified at the end of the VEP experiments (see Fig. 2D for a representative example). The ERG statistics can be found in Supplementary Table 1.

DISCUSSION

We found little evidence of altered vision (except subtle changes in dark-adapted ERG kinetics, see discussion below) among AβPP/PS1 mice compared to wild-type littermates by means of ERG and VEP. The ERG amplitude was similar between genotypes over the range of increasing stimulus intensity both in dark- and light-adapted conditions. The VEPs for stationary patterned stimuli at constant luminance did not differ between the genotypes over the range of changing pattern size or contrast. Our data indicates that visual acuity and contrast threshold are normal in this AD mouse line at an age when we consistently observe spatial memory impairment in the Morris swim task [25, 34].

Our results contrast with the previous reports suggesting decreased ERG amplitudes and increased VEP latencies in AβPP/PS1 mice [19 , 22]. Some differences in the study details may account for these differences. First, a common denominator in the previous studies is the use of ketamine-xylazine (KX) anesthesia, which may raise problems as KX cocktail evokes strong hyperglycemia and acidosis in mice [35, 36]. Indeed, we have observed a high mortality and seizure rate in transgenic animals when using KX, whereas isoflurane is well tolerated (Leinonen et al., unpublished data). Therefore, isoflurane has been recommended instead of KX for ERG recordings in mice [38]. In addition, KX anesthesia affects NMDA and adrenergic receptors, which may lead to differential drug effects in AβPP/PS1 mice known to have altered NMDA/adrenergic system [36, 37]. Second, the previous reports are somewhat inconclusive since they do not show intensity-response function, which is considered a standard in visual stimulation protocols [29, 39]. Third, our AβPP/PS1 mice are practically of pure C57BL/6J background after extensive back-crossing while the commercially available AβPPswe/PS1dE9 (Jackson Laboratories) used in most other studies are C57BL/6JxC3 H hybrids. The C3 H line carries the retinal degeneration 1 gene [41] that should be excluded by PCR screening. Nonetheless, independent of the background gene, all the AβPPswe/PS1dE9 mice show age-related amyloid plaque formation and spatial memory impairment approximately at the same age.

The ERG is a sensitive method for detecting changes in retinal processing. By manipulating light adaptation and stimulus parameters combined to specific analysis methods, the ERG can give insight to anatomy and function of almost all retinal cell classes [29]. The a-wave arises mainly from photoreceptor activation while the b-wave marks bipolar cell depolarization. Specific retinal ganglion cell dysfunction can be detected by ERG utilizing the scotopic threshold response (i.e., responses for light intensities smaller than –3.60 log cd’s/m² in darkness [41]) or photopic negative response [42]. The cellular origins of the oscillatory potentials (OPs), mainly overlying the b-wave of ERG, are not well known [43]. However, evidence suggests their involvement in inhibitory feedback circuits of the retina, employing primarily amacrine cells [44]. Most waveform parameters were similar between the genotypes except that dark-adapted b-waves seemed to peak at different times. Unexpectedly, AβPP/PS1 mice had shortened dark-adapted b-wave latency at low stimulus intensities. We also found slightly increased frequency of OPs in AβPP/PS1 mice. However, these changes in ERG kinetics were so small that they unlikely influence the behavioral outcome, especially under photopic light conditions. Indeed, cortical responses for visual stimuli (VEP) were normal at photopic condition. The ERG findings may be related to impaired cholinergic innervation of the retinas in AβPP/PS1 mice. Acetylcholine is known to modulate retinal function and a recent study showed that acetylcholine triggers GABA release onto rod bipolar (RB) cells modulating the signal transmission at the RB-AII amacrine cell synapse under low light conditions [45]. Impaired cholinergic innervation in the retina may partly account for visual disturbances in AD. We do not have data yet on cholinergic innervation in the AβPP/PS1 mouse retina, but decline in cholinergic neurotransmission can be detected in these mice already at 7 months of age [37].

In conclusion, this study provides a relieving message to AD research community showing that the stationary visual pattern stimuli yields normal responses at the visual cortex of male AβPP/PS1 mice at the age when behavioral studies are typically conducted. The subtly hastened kinetics in dark-adapted ERGs and the increased frequency of OPs are interesting findings that may demonstrate inadequate cholinergic innervation in the AβPP/PS1 mice retinas and warrant furtherinvestigation.