|Home | About | Journals | Submit | Contact Us | Français|
Amyloid-beta (Aβ)-induced changes in synaptic function in experimental models of Alzheimer disease (AD) suggest that Aβ generation and accumulation may affect fundamental mechanisms of synaptic plasticity. To test this hypothesis, we examined the effect of amyloid precursor protein (APP) overexpression on a well-characterized, in vivo, developmental model of systems-level plasticity, ocular dominance plasticity (ODP). Following monocular visual deprivation during the critical period, mice that express mutant alleles of amyloid precursor protein (APPswe) and Presenilin1 (PS1dE9), as well as mice that express APPswe alone, lack ocular dominance plasticity in visual cortex. Defects in the spatial extent and magnitude of the plastic response are evident using two complementary approaches, Arc induction and optical imaging of intrinsic signals in awake mice. This defect in a classic paradigm of systems level synaptic plasticity shows that Aβ overexpression, even early in postnatal life, can perturb plasticity in cerebral cortex, and supports the idea that decreased synaptic plasticity due to elevated Aβ exposure contributes to cognitive impairment in AD.
Amyloid-beta (Aβ) production and accumulation have been implicated in the pathogenesis of Alzheimer disease (Tanzi et al., 1987; Goate et al., 1991; Prasher et al., 1998; Nilsberth et al., 2001; Bentahir et al., 2006; Kumar-Singh et al., 2006; Rovelet-Lecrux et al., 2006). Exposure to soluble and fibrillar forms of Aβ is associated with synapse loss (Hsia et al., 1999; Walsh et al., 2002; Spires et al., 2005; Shankar et al., 2007), which may be mediated by a direct effect of Aβ on synaptic function: acute, exogenous application of oligomeric human Aβ to organotypic hippocampal cultures blocks long term potentiation (LTP) induction (Lambert et al., 1998; Chen et al., 2002; Walsh et al., 2002), and lowers the threshold for induction of long term depression (LTD; Kim et al., 2001; Shankar et al., 2008). Moreover, this synaptic activity may relate to a normal function of Aβ in synaptic physiology (Abramov et al., 2009).
These considerations suggest that elevation of Aβ levels in humans could directly perturb synaptic function independent of the aggregation of protein in amyloid plaques. Here we examine this possibility in the developing mouse visual cortex using a well-characterized paradigm of neural system plasticity, ocular dominance plasticity (ODP). ODP is triggered when one eye is closed or removed during a critical period of postnatal development (Gordon and Stryker, 1996; Tagawa et al., 2005; Smith et al., 2009; Tropea et al., 2009), thereby causing an increase in the responsiveness of neurons in primary visual cortex driven by visual stimulation of the open (or remaining) eye and a decrease in response to the closed eye. An expansion in the territory functionally connected to the open eye has also been observed (Tagawa et al., 2005). Although most robust during the critical period, similar phenomenon also occur in the adult (Sawtell et al., 2003; Tagawa et al., 2005; Fischer et al., 2007; Sato and Stryker, 2008). Here we report that APPswe-expressing transgenic mice demonstrate marked functional deficits in ODP that provide insight into the effect of Aβ on synaptic function in the context of intact neural circuits responding to physiologically relevant changes in input.
All experiments were carried out under national (United States National Institutes of Health) and institutional (Massachusetts General Hospital subcommittee for research animal care and IACUC at Harvard Medical School) guidelines. APP/PS1 mice (strain B6C3-Tg(APPswe,PSEN1dE9)85Dbo/J; Jankowsky et al., 2001), and mice expressing PSEN1dE9 alone (strain B6.Cg Tg(PSEN1dE9)S9 Dbo/J; Savonenko et al., 2005), were obtained from the Jackson Laboratory. Mice expressing APPswe alone, strain Tg2576 (Hsiao et al., 1996), were also used. Unilateral lid suture was performed at P28- the peak of the critical period- on transgenic and non-transgenic littermate mice; Arc induction was performed on P33, as described below. Monocular enucleation (ME) was performed on P32 mice to assay the Arc induction response in the absence of a prior period of unilateral deprivation or enucleation (Tagawa et al., 2005). ME was also performed at P22 to assess OD plasticity in visual cortex ipsilateral to the remaining eye 10 days later, with Arc induction performed on P33. Male and female mice were used in these experiments.
As described in Tagawa et al. (2005), at P32, just prior to Arc induction, ME was performed (unless animals had received ME at P22). The mouse cage was placed in a ventilated container in total darkness for approximately 15 hours. The cage was then placed in bright, fluorescent lighting to permit a 30-minute period of vision, following which mice were euthanized. Brains were removed and flash-frozen in M-1 mounting medium (Shandon, ThermoScientific). 8–10 μM coronal cryosections through visual cortex were collected on Superfrost slides (Fisher Scientific) for subsequent in situ hybridization analysis.
As described in Schaeren-Wiemers and Gerfin-Moser (1993). Arc anti-sense probe template was derived by RT-PCR from mouse brain mRNA (Genbank ID: NM018790, nt 262..1338) and cloned into pBluescript KS (Stratagene). Digoxigenin-labeled antisense riboprobe was transcribed in vitro using T3 RNA polymerase (Roche).
Measurements of the width of visual cortex containing neurons responding to stimulation of the ipsilateral eye, as assessed by induction of mRNA for the immediate early gene Arc, were made by drawing a segmented line parallel to the cortical surface through the Arc-positive domain of layers 2/3, 4, 5 or 6 in primary visual cortex, using Image J software (NIH). The measurements from at least six sections were used for each mouse to determine the width of the responsive domain. Statistical analysis was performed using Statview software (Version 5.0.1; SAS Institute, Inc.). All measurements and analyses were performed blind to genotype.
Fresh frozen mouse brains were cryosectioned (8–10 μm) and mounted onto slides for subsequent staining. For immunohistochemistry, sections were fixed for 10 minutes in 4% paraformaldehyde in phosphate buffered saline, washed, blocked with 10% goat serum in Tris-buffered saline (TBS) and then incubated overnight with anti-APP/Aβ antibody 6E10 (1:1000; Covance). Slides were then stained with Cy3-conjugated anti-mouse IgG secondary (1:200; Jackson Immunoresearch) and coverslipped with DAPI-containing mounting medium (Vector Laboratories).
Brain homogenates were prepared from a P28 APP/PS1 mouse and a non-transgenic litter mate. Brain homogenates were sequentially fractionated in TBS, 1% Triton X-100 and 2% SDS. Immunoprecipitation with human APP/Aβ antibody 6E10 was performed with each fraction. Immunostaining of resultant blots was performed with 6E10 and with an anti-amino terminal Aβ antibody, 82E1 (IBL International).
In addition to Arc induction, we utilized intrinsic imaging as an independent means of assessing the integrity of plasticity responses in APP/PS1 and non-transgenic mice. Fifteen male and female mice underwent cranial window surgery in preparation for in vivo imaging of sensory responses in mouse primary visual cortex (8 APP/PS1 and 7 non-transgenic litter mates). For cranial window implant surgery, P22–P24 animals were anesthetized with isoflurane (1.2–2% in 100% O2). Dexamethasone was administered on the day of the procedure 2–6 hours prior to surgery (3.2 mg/kg, IM) and atropine administered at surgery onset (0.2 mg/kg, IP). Using aseptic technique, a headpost was secured in place using cyanoacrylate, dental acrylic and C&B Metabond (Parkell), and a 5 mm craniotomy was made over the left visual cortex (center ~2.5 mm lateral, 2.0 mm anterior to lambda) as described in Andermann et al. (2010). A 5 mm glass cranial window, consisting of an 8 mm coverslip cured to two 5 mm coverslips (Warner #1; total thickness: ~0.5 μm; thickness below skull: ~ 200 μm) using index-matched adhesive (Norland #71), was implanted. The window was secured in place using N-butyl cyanoacrylate and C&B Metabond, and the mice were allowed to recover for at least four days. Imaging of ipsilateral and contralateral retinotopic acuity was then preformed at age P28 (P27–P29; see below). In a subset of mice (N = 6 non-transgenic mice and 4 APP/PS1 mice), we also evaluated the effects of contralateral 4-day lid suture on binocular responses, beginning 2–4 hours after recovery from suture removal.
Following recovery from headpost/cranial window implant surgery, we mapped retinotopic organization of ipsilateral and contralateral responses, and strength of responses, in awake, head-restrained mice at age P28 (P27–P29; Figure 4). To map visual cortical responses, we used in vivo epifluorescence imaging (Husson et al., 2007; Tohmi et al., 2009) to measure changes in autofluorescence signal. Autofluorescence produced by blue excitation (470 nm center, 40 nm band, Chroma) was measured through a green/red emission filter (longpass, 500 nm cutoff). Images were collected using a CCD camera (Sensicam, Cooke) at 2 Hz through a 5X air objective (0.14 NA, Mitituyo) using ImageJ acquisition software.
We presented visual stimuli using a 120 Hz LCD monitor (Samsung 2233RZ, 22″) that was calibrated with a spectrophotometer (Photoresearch PR-650). The monitor was positioned directly in front of the mouse, perpendicular to its body axis, at a distance of 21 cm. Stimuli were upwards-drifting square-wave gratings (16° square patch, 0.03 cycles/degree, 2 Hz, 80% contrast) similar to those used by Hofer et al. (2009). Stimuli were placed at elevations of approximately 20° above the horizontal meridian. Stimuli were presented for 5 s each, with 15 s of blank monitor screen (mean luminance) between trials. Stimuli could be presented to the ipsilateral (left) or contralateral (right) eye separately, in subsequent runs, by carefully positioning an opaque plastic visor in front of the non-stimulated eye (< 3 mm from the eye, just above the snout fur). Mice were not noticeably agitated by the placement of the visor. We first mapped ipsilateral retinotopic locations in primary visual cortex (V1) as follows: We presented stimuli at five horizontal retinotopic locations, each spaced 16 degrees apart (spanning the field from approximately −40° to 40° surrounding the vertical meridian), in pseudorandom order (12 trials/location). We then assessed which three contiguous stimulus locations were most effective at driving ipsilateral eye responses. The main stimulus protocol was then employed, consisting of pseudorandom presentation of stimuli at these three retinotopic locations. We presented 4–5 runs of stimulation per eye (60 trials per run total, 20 trials/stimulus location/run), with alternating stimulation of the ipsilateral or contralateral eye across runs to ensure that comparison of ipsilateral and contralateral responses over time was not influenced by non-stationary responses.
Visual responses consist of a small increase in the autofluorescence signal (corresponding to flavoprotein oxidization during increased metabolism; Tohmi et al., 2009) followed by a stronger negative signal (due to increased light absorption due to delayed increase in blood volume and deoxyhemoglobin concentration; Schuett et al., 2002). Thus, the response to a stimulus was computed as the percent change in fluorescence (% dF/F) between the average of 6 frames recorded during the first 3 seconds following stimulus onset and the average of all 20 frames recorded from 9 seconds to 19 seconds following stimulus onset. For presentation of imaging analyses, we have inverted the sign of the visual responses so that stronger responses correspond to more positive values. Imaging analyses were performed using Matlab software (Mathworks).
To test whether retinotopic responses to stimuli at different locations shifted in an orderly manner and were spatially localized in non-transgenic and APP/PS1 mice, we characterized the shifts in optical peak response location following stimulation at adjacent stimulus positions, as well as the width of optical response diameter, as follows: First, for each mouse, we obtained the one-dimensional response profiles along the line from peak response in V1 to peak response in lateromedial (LM) area, which has a binocular region adjacent and lateral to V1. This line was obtained manually by selecting the peak response in V1 and LM in response to contralateral eye stimulation at the rightmost stimulus location (see Figure 4A and B). The one-dimensional (1-D) response profiles for each stimulus condition were then estimated as the maximal response in a region from −1000 μm to 1000 μm perpendicular to the line between V1 and LM; this method minimizes the influence of vascular artifacts, which often had responses of the opposite sign to neuropil responses). The V1/LM border was defined as the midpoint between the peak responses in V1 and LM. All 1-D profiles were then low-pass filtered (465 μm cut-off). Peak V1 responses along the LM/V1 axis were then estimated for each stimulus position, in the region from 0 μm to 800 μm from the V1/LM border. For estimation of response bandwidth, we calculated the bandwidth at half of peak response (full width at half-maximum, FWHM), again using the response to the rightmost stimulus presented to the contralateral eye in P28 mice.
We obtained a measure of response strength for both ipsilateral eye (ipsi) or contralateral eye (contra) stimulation (Figure 5) in the binocular region of primary visual cortex (V1) by averaging the 1-D response profiles for all three stimulus types presented in the main stimulus protocol, in the region in V1 from 100 μm to 350 μm medial to the V1/LM border (Figure 5A, bottom panels, and Figure 5B). To obtain a measure of ocular dominance selectivity of this cortical region, i.e. selectivity for IPSI or CONTRA stimulation, we computed an ocular dominance selectivity index, ODI = (Rcontra − Ripsi)/(Rcontra + Ripsi). ODI values of 1 and −1 correspond to responses to contra or ipsi eye stimulation only, while ODI values equal to 0 indicate a perfectly balanced response to stimulation of either eye. We chose this index because it is less sensitive to changes in overall response strength than Rcontra/Ripsi (note that for non-transgenic mice, calculation of Rcontra/Ripsi in our sample resulted in values of ~ 2.5 and ~ 1 before and after 4 day monocular deprivation, respectively, which is consistent with results in Hofer et al., 2009).
For mice in which ODI indices could be directly compared in successive imaging sessions pre-MD (P27–29) and post-MD (P31–33), we ensured analysis of the identical cortical region in both sessions by aligning all images from the post-MD imaging session to the pre-MD imaging session using Matlab software (Mathworks). For all experiments, the duration of MD between imaging sessions was kept nearly constant (96 ± 2 hours).
To compare ODI indices from pre- and post-monocular deprivation (Figure 5C), we co-registered images obtained pre-MD and post-MD by applying the same vasculature-based affine transformation to all post-MD images using Matlab. This ensured estimates of responses from the identical location in cortex for pre-MD and post-MD sessions. All measurements and analyses were performed blind to genotype.
APP expression in the mice studied here (APP/PS1 mice) has been detected as early as 4 months of age (Garcia-Alloza et al., 2006). However, the critical period for ODP is even earlier, from P19 to P34 (Gordon and Stryker, 1996; Tagawa et al., 2005), prompting us to assess if detectable transgene and protein products are present at these ages in transgenic mice. Immunoprecipitation of sequentially-extracted fractions from P28 transgenic and non-transgenic mouse brain homogenates with an antibody that recognizes human APP and Aβ (6E10; Covance), followed by immunoblotting with 6E10 (Figures 1A and B) or an antibody recognizing the Aβ amino terminus (82E1; Figure 1C), reveals expression of full length APP as well as various proteolytic cleavage products, including Aβ (Figures 1B and C), in transgenic mice. Immunohistochemistry for human APP reveals expression in neurons throughout the brain, including visual cortex (Figure 1D and data not shown). These data provide evidence that human APP and Aβ can be detected in the brains of APP/PS1 mice as early as the critical period.
We next assessed the response of neurons in visual cortex of transgenic and non-transgenic mice to visual stimulation. Induction of the immediate early gene Arc was performed on P33 APP/PS1 and non-transgenic mice (Figures 2A and 2C). The method of Arc induction provides a convenient readout of the distribution of neurons functionally connected to the stimulated eye (Tagawa et al., 2005). As expected, Arc mRNA is induced in many neurons throughout V1 contralateral to the stimulated eye (Figures 2A and 2C), with a small domain located within the binocular region of V1 responding in ipsilateral cortex (Figures 2B and D). The pattern and width of induction is identical in non-transgenic and APP/PS1 mice (width of domain of layer 2/3 cells in visual cortex responding to ipsilateral eye stimulation in non-transgenic mice: 907 ± 13μm; APP/PS1 mice: 902 ± 15μm; p>0.79; non-transgenic layer 4: 1039 ± 46 μm; APP/PS1 layer 4: 1037 ± 61 μm; p=0.90; non-transgenic layer 5: 959 ± 62μm; APP/PS1 layer 5: 984 ± 59 μm; p=0.33; non-transgenic layer 6: 1050 ± 56 μm; APP/PS1 layer 6: 1079 ± 73 μm; p=0.30; n=10–13 mice in each group; Figures 2B, D and G). These data indicate that the basic spatial extent and laminar distribution of the visual cortical response to light is similar in juvenile APP/PS1 and non-transgenic mice, and suggest that the underlying anatomy of the geniculocortical projection is not influenced by the presence of the transgenes or their products.
To evaluate critical period ODP in APP/PS1 mice, unilateral eyelid suture was performed on P28 mice. Mice experienced monocular deprivation (MD) for the subsequent 4 days, followed by Arc induction via stimulation of the non-deprived eye. ODP was then assessed by measuring the change in width of Arc mRNA induction in neurons located in layer 2/3 of visual cortex. As expected, following 4 days of MD, non-transgenic mice demonstrate an expansion in the domain of Arc mRNA induction relative to mice reared with normal binocular vision (Figures 2E and G), indicative of normal OD plasticity related to the non-deprived eye (Tagawa et al., 2005). Strikingly, APP/PS1 mice fail to demonstrate any expansion following a 4-day period of MD (p>0.15, n=10 in each group; Figures 2F and G).
Although APP/PS1 mice lack expected OD plasticity in response to a 4-day period of MD during the critical period, it remains possible that a stronger and longer period of monocular visual deprivation might elicit a response. Longer periods of deprivation and also monocular enucleation (ME) produce a stronger response in OD plasticity as compared to a 4-day period of MD (Horton and Hocking, 1998; Tagawa et al., 2005). APP/PS1 and non-transgenic mice underwent ME at P22, followed by Arc induction via the intact eye at P33. In non-transgenic mice, within visual cortex ipsilateral to the non-deprived eye, there is a robust expansion of the domain driven by visual stimulation, as assessed by Arc mRNA induction in layer 2/3 cells (Figures 3A and C). The ten-day period of monocular vision elicited strong OD plasticity in this assay in non-transgenic mice, consistent with published results (Tagawa et al., 2005; Syken et al., 2006; Datwani et al., 2009). In contrast, no significant increase in the width of the domain of Arc induction was observed in APP/PS1 mice even following this prolonged period of ME (p>0.81; Figures 3B and C). Thus, these observations further illustrate the disruption of ODP in the transgenic mice.
OD plasticity was also examined in an independent transgenic line that overexpresses APPswe alone. Using the strong protocol of 10-day ME, no significant expansion of territory belonging to the open eye was detected (n=10–16 mice per group; Figure 3C), implying that the absence of OD plasticity in these mice is not line-specific, nor is it dependent on the expression of mutant PS1. Consistent with this suggestion, transgenic mice that express mutant PS1 alone exhibit normal ODP, excluding the possibility that the defect in ODP observed in APP/PS1 mice is due to an APP-independent effect of mutant PS1 expression (Figure 3C).
The results of the Arc induction experiments above demonstrate that following short or long periods of monocular deprivation, there is a defect in ODP as indicated by a failure of the functional representation of the non-deprived eye to expand within visual cortex. To confirm this result and to compare responses to stimulation of each eye before and after MD in individual animals, OD plasticity was next examined in vivo using widefield imaging of intrinsic signals during localized visual stimulation. Optical intrinsic signal imaging is a well-established method for assessing retinotopic organization, the relative strength of cortical responses to stimulation of ipsilateral and contralateral eyes, as well as changes in response strength that occur following MD (Grinvald et al., 1986; Schuett et al., 2002; Cang et al., 2005; Hofer et al., 2009; Tohmi et al., 2009). The measured intrinsic signal is thought to be a result of neuronal activity and hemodynamic responses, both of which may be reduced by anesthesia and the behavioral state of the animal (Berwick et al., 2002; Niell and Stryker, 2010), prompting us to perform our analysis in awake mice that were habituated to head fixation.
Mice were presented with drifting gratings positioned at one of three horizontal locations on an LCD monitor (each stimulus spanning 16° of the visual field; see Methods), to assess retinotopic organization as well as to determine the coordinates of the binocular region of visual cortex that produced maximal responses to stimulation of the ipsilateral eye (Figures 4A–C). To ensure that cortical responses to visual stimulation following cranial window installation were comparable in non-transgenic and APP/PS1 mice, cranial window installation was performed on P22–24 and mice were allowed to recover from surgery for 5–6 days before imaging at P27–29 (Figure 4A). No significant difference was observed comparing the magnitude of cortical responses to visual stimulation of ipsilateral or contralateral eye in APP/PS1 mice and non-transgenic mice (dF/F non-transgenic ipsilateral stimulation: 0.19 ±0.05%; APP/PS1 ipsilateral stimulation: 0.22 ± 0.03%; p=0.661; non-transgenic contralateral eye stimulation: 0.39 ± 0.05%; APP/PS1 contralateral eye stimulation: 0.30 ± 0.03%; p=0.112; t-test; n=7–8 per group; ± S.E.M.), suggesting that transgene expression does not affect the overall responsiveness of the visual cortex following window placement. The magnitude of shifts in peak response location following a shift in visual stimulus location did not differ between non-transgenic and APP/PS1 mice (Figure 4D; 2-tailed ranksum test, p>0.28; see Methods), nor was there any difference in the spatial extent of responses to a localized 16° stimulus (Figure 4E; 2-tailed ranksum test, p>0.4). No significant differences (t-tests, p>0.05) were observed at P28 between non-transgenic and APP/PS1 mice in retinotopic spacing (Figure 4D), response width (Figure 4E), ipsilateral eye response amplitude (Figure 5C) or contralateral eye response amplitude (Figure 5C), suggesting that APP/PS1 mice have grossly intact visual responses and cortical retinotopic organization.
Mice then underwent a 4-day period of MD, after which imaging of responses to both deprived and non-deprived eye stimulation was again conducted (Figure 5A). Non-transgenic mice demonstrate the expected, weakened response to stimulation of the deprived, contralateral eye (pre-MD, 0.39±0.05; post-MD 0.21±0.04; p<0.01, t-test; Figures 5B and 5C). Although the increase in the magnitude of responses to the ipsilateral, non-deprived eye is not statistically significant (pre-MD, 0.194 ± 0.048; post-MD, 0.222 ± 0.028; p=0.1435; Figures 5B and 5C), there is a decrease in the ocular dominance index (ODI) following MD (non-transgenic ODI pre-MD, 0.428 ± 0.065; non-transgenic ODI post MD, −0.027 ± 0.100; p=0.005; Figure 5D). Prior to MD, APP/PS1 mice demonstrate a smaller ratio of contralateral to ipsilateral response strengths (APP/PS1 C/I ratio, 1.49 ± 0.21; non-transgenic C/I ratio, 2.58 ± 0.43; p=0.035, t-test; Figure 5C, see discussion below). In addition, APP/PS1 mice demonstrate no significant change in ODI measurements following MD (APP/PS1 ODI pre-MD, 0.160 ± 0.063; APP/PS1 ODI post-MD, 0.333 ± 0.084; APP/PS1 ODI pre-MD vs post-MD, p=0.28; pre-MD ODI non-transgenic vs APP/PS1, p=0.03; post-MD ODI non-transgenic vs APP/PS1, p=0.04; Figure 5D). Following MD in APP/PS1 mice, responses to stimulation of the deprived, contralateral eye decrease in magnitude (pre-MD, 0.30±0.03; post-MD, 0.19±0.03; p=0.015; Figure 5C), suggesting that weakening of responses to the deprived eye is intact, however, the magnitude of responses to stimulation of the non-deprived eye also decrease (pre-MD, 0.22±0.03; post-MD, 0.09±0.02; p=0.015; Figure 5C). In sum, the decrease in ODI present in non-transgenic mice does not occur in APP/PS1 mice following 4-day MD (non-transgenic, −0.455 ± 0.109; APP/PS1, 0.146 ± 0.108, p=0.0057, t-test), confirming that APP/PS1 mice have reduced ocular dominance plasticity (Figure 5D).
In addition to this marked abnormality in plastic response to 4-day MD, APP/PS1 mice also appear to have a more subtle defect. As described above, we observe a decreased ratio of contralateral to ipsilateral response strengths even with normal rearing up to P28. This change in the contralateral to ipsilateral eye ratio is reminiscent of the alteration recently described in dark-reared mice (McCurry et al., 2010), suggesting that APPswe/Aβ expression may subtly interfere with activity-dependent establishment of normal contralateral to ipsilateral eye binocularity even in the absence of monocular deprivation.
We tested the hypothesis that overexpression of APP disrupts fundamental mechanisms of synaptic plasticity in the intact nervous system. The major finding of this study is that dramatic deficits in synaptic plasticity can be detected in the cerebral cortex of APPswe-expressing transgenic mice in vivo. We demonstrate a defect in ocular dominance plasticity in the visual cortex during the peak of the critical period, i.e., in the 4th and 5th weeks of life. Following either a brief period of MD or an extended period of ME, the normal expansion of territory functionally connected to the open eye fails to occur in APPswe-expressing transgenic mice. This result, demonstrated using two complementary methods of assessing ODP, cannot be explained by a defect in the baseline functional organization of visual responses in APPswe-expressing mice or by an alteration in Arc transcription; APPswe-expressing transgenic mice reared with normal visual experience demonstrate a normal pattern of Arc induction and have similar visual retinotopic acuity as that of non-transgenic mice (Figures 2A–D and Figure 4). The comparable width of the domain of visual cortex responsive to the ipsilateral eye in normally-reared APPswe-expressing mice argues against a defect in maturation of visual system circuitry; previous studies using Arc induction in analyses of responses at earlier ages have shown that the domain of cortex responsive to ipsilateral eye stimulation is much wider at P17 and refines with maturation to the smaller size observed when testing animals at P28–P33 (Tagawa et al., 2005). The C/I ratio in normally-reared APP/PS1 mice is lower than that of non-transgenic mice, suggesting that there may be subtle developmental physiological changes that affect the establishment of normal binocular balance, however, there does not appear to be a serious maturational defect in APP-expressing transgenic mice. In addition, mice expressing APPswe alone (Tg2576; Hsiao et al., 1996) are known to perform normally on tasks that require visual function as late as 10 months of age, further suggesting that these mice do not have a defect in vision per se at one month of age.
A number of studies have looked at Arc expression in APP-expressing transgenic models and have demonstrated decreases (Wegenast-Braun et al., 2009), or increases (Grinevich et al., 2009; Perez-Cruz et al., 2011) in APP-expressing mice. We interpret these differences as the result of a number of variables that differ between each of the reported studies, including brain regions analyzed, strains and ages of mice studied, and behavioral assays performed. Our current analysis suggests that our readout of neuronal activity, Arc mRNA expression following a 30-minute light exposure, is not affected by the expression of APP or Aβ in APP-expressing mice at one month of age. It is also important to emphasize that, in contrast to the above studies, we examined induction of Arc expression in response to a specified visual stimulus.
The defect in ODP in visual cortex may be the earliest evidence of synaptic dysfunction in APPswe-expressing mice described to date. Transgenic mice are typically tested at much later ages and in some studies, cognitive deficits are not detected until mice begin to accumulate amyloid plaques (Hsiao et al., 1996; Chen et al., 2000; Billings et al., 2005). Memory deficits have been demonstrated prior to plaque formation in 3-month old mice, but these defects were not detected at 2 months of age (D’Amelio et al., 2011). Our data suggest that even though synaptic dysfunction may be present at much earlier ages, compensation and functional redundancy at synapses or circuits responsible for learned behaviors may mask the basic synaptic dysfunction produced by constitutive expression of APP and Aβ. However, as we show here, this defect in synaptic plasticity can be revealed by an acute challenge in the form of MD or ME.
Our data suggest that APP-expressing transgenic mice are unable to generate a normal neocortical plastic response to the loss of input from one eye. In WT mice, studies have shown that OD plasticity in response to MD occurs in 2 phases (Frenkel and Bear, 2004; Kaneko et al., 2008; Smith et al., 2009). First there is a weakening of the cortical response to the deprived eye, followed later by strengthening of responses driven by stimulation of the non-deprived eye. The defect observed here in APP-expressing transgenic mice following MD or ME involves a failure of the non-deprived eye to expand its territory within visual cortex, as well as a defect in the mechanism of open-eye strengthening. The expansion of the responding domain to stimulation of the ipsilateral, non-deprived eye, following a 4-day period of MD, was observed in non-transgenic mice, but not in APP-expressing mice, using Arc induction, a technique that captures the cortical response- at cellular resolution- to the complex visual stimulation presented to the mouse by allowing it to move freely within its home cage while exposed to light for 30 minutes. Strengthening of responses to the non-deprived eye was not observed in non-transgenic mice in our optical imaging experiments, or those of others (McCurry et al., 2010). With lower spatial resolution, optical imaging provides a way of assessing the relative magnitude of responses using a stimulus much less likely to stimulate the cortex to the same degree as the stimulus provided in Arc induction. A larger proportion of the cells would have to respond to produce a measurable change by intrinsic imaging, which may be more evident following longer periods of MD (Kaneko et al., 2008; Tohmi et al., 2009). Our imaging analysis was restricted to responses to stimuli positioned 0–30° from the vertical meridian, focusing our analysis on neuronal populations most likely to exhibit a plastic response following MD (Gordon and Stryker, 1996). Using optical imaging, we see abnormally weakened responses to the non-deprived eye following a 4-day period of MD in APP/PS1 mice, confirming an abnormality in the plastic response to the non-deprived eye. It is worth noting that mice lacking TNFα or TNF receptors have a similar OD plasticity defect, with deprived eye weakening remaining intact in these mutant mice, consistent with the proposal that weakening and strengthening involve distinct cellular and molecular mechanisms (Kaneko et al., 2008). Our observations raise the possibility that APPswe/Aβ expression may alter the same mechanisms of synapse strengthening. Loss of function studies suggest that APP plays a role in neurite pruning and neuromuscular junction development (Wang et al., 2005; Nikolaev et al., 2009; Wang et al., 2009). In addition, Aβ exposure has been shown to affect NMDA receptor function and calcineurin-dependent calcium signaling pathways (Hsieh et al., 2006; Shankar et al., 2007; Wu et al., 2010) crucial for intact ocular dominance plasticity during the critical period (Sawtell et al., 2003; Yang et al., 2005). We postulate that dysregulation of NMDA receptor activity or calcineurin signaling may play a role in the ODP defect in APPswe-expressing mice.
The current data suggest that overexpression of APPswe, and by implication, Aβ, can have a marked effect on synaptic plasticity in the intact brain. Amyloid plaques exert complex effects on their local microenvironment, including increased glial activity, oxidative stress, disruption of calcium homeostasis, morphologic changes in dendrites and axons and alterations in activity-dependent mRNA translocation and translation (Spires et al., 2005; Kuchibhotla et al., 2008; Meyer-Luehmann et al., 2009). By contrast, the current model system of OD plasticity reveals dramatic effects of Aβ in a well-studied form of systems-level synaptic plasticity in the absence of plaques (Garcia-Alloza et al., 2006), thus affording an opportunity to identify some of the physiologic changes that occur at the earliest stages of the disease process (Caselli et al., 2009). In addition to having potential relevance to Down syndrome developmental abnormalities, we speculate that these data support a role for Aβ accumulation in causing diminished plasticity responses consequent to lesion in the adult; if so, synaptic reserve may be increasingly compromised as the disease progresses and decrease patients’ ability to withstand accumulating CNS insults. The robust nature of the ODP experimental paradigm represents a useful approach to understanding synaptic dysfunction and provides a way to test the efficacy of therapeutic agents in stopping or reversing the very early effects of Aβ on synaptic function.
We thank Anthony Moffa for assistance with data collection and Sergey Yurgenson and Aleksandr Vagodny for technical contributions. We thank members of the Shatz, Hyman, and Reid labs for advice and discussion. This work was supported by NIH (Grants K08 NS069811 to C.M.W., R01 EY018742 to R.C.R., DP2 OD006662 to M.W.A.), by the Dr. and Mrs. E.P. Richardson Jr. Fund for Neuropathology (C.M.W.), by a fellowship from the Ludcke Foundation and Pierce Charitable Trust (M.L.A.), and by the Mathers Charitable Foundation (C.J.S.).