An Intracellular Threonine of Amyloid-β Precursor Protein Mediates Synaptic Plasticity Deficits and Memory Loss

Mutations in Amyloid-ß Precursor Protein (APP) and BRI2/ITM2b genes cause Familial Alzheimer and Danish Dementias (FAD/FDD), respectively. APP processing by BACE1, which is inhibited by BRI2, yields sAPPß and ß-CTF. ß-CTF is cleaved by gamma-secretase to produce Aß. A knock-in mouse model of FDD, called FDDKI, shows deficits in memory and synaptic plasticity, which can be attributed to sAPPß/ß-CTF but not Aß. We have investigated further the pathogenic function of ß-CTF focusing on Thr668 of ß-CTF because phosphorylation of Thr668 is increased in AD cases. We created a knock-in mouse bearing a Thr668Ala mutation (APPTA mice) that prevents phosphorylation at this site. This mutation prevents the development of memory and synaptic plasticity deficits in FDDKI mice. These data are consistent with a role for the carboxyl-terminal APP domain in the pathogenesis of dementia and suggest that averting the noxious role of Thr668 is a viable therapeutic strategy for human dementias.


Introduction
Familial dementias are caused by mutations in APP [1] and genes that regulate APP processing. These include the PSEN1/2 genes, which code for the catalytic component of the gammasecretase, and the BRI2/ITM2b gene, whose protein product BRI2 binds APP and inhibits APP processing [1][2][3][4][5][6][7][8]. Cases caused by APP/PSEN mutations are classified as FAD and those caused by mutations in BRI2/ITM2b as FDD or Familial British dementia (FBD). The prevailing pathogenic model for these dementias posits that amyloid peptides trigger dementia. In AD, the amyloid peptide Aß is a part of APP; in FDD and FBD, the amyloidogenic peptides, called ADan and ABri respectively, are generated from the mutant BRI2 proteins [2,8]. FDD patients present mixed amyloid plaques containing both Ab and ADan. However, recent data suggest that these dementias share pathogenic mechanisms involving synaptic-toxic APP metabolites distinct from Ab [9,10].
In FDD, a 10-nucleotide duplication in the BRI2/ITM2B gene leads to the synthesis of a longer BRI2 protein [8]. In normal individuals, BRI2 is synthesized as an immature type-II membrane protein (imBRI2) that is cleaved at the C-terminus into mature BRI2 and a 23aa soluble C-terminal fragment [11]. In FDD patients, cleavage of the BRI2 mutant protein leads to the release of the longer ADan peptide [8]. To model FDD we generated FDD KI mice that like FDD patients [8], carry one wild type Bri2/Itm2b allele and the other one has the Danish mutation [12]. FDD KI mice develop synaptic and memory deficits due to loss of Bri2 protein, but do not develop amyloidosis [13]. BRI2 binds to APP and inhibits cleavage of APP by secretases [4][5][6][7]. Owing to the loss of BRI2, processing of APP is increased in FDD [14,15]. Memory and synaptic deficits of FDDKI mice require APP [14], and are mediated by sAPPß and/or ß-CTF produced during synaptic plasticity and memory acquisition. Inhibition of c-secretase, the enzyme that processes b-CTF to yield Aß, worsens memory deficits and is associated with an accumulation of ß-CTF [10,16,17]. In addition, caspase-9 in activated in FDD KI mice and caspase-9 activity mediates memory/synaptic plasticity deficits [18]. Overall, these results suggest that ß-CTF, rather than Aß, is a major toxic species causing dementia. Here, we have investigated further the pathogenic role of the carboxyl-terminal region of APP and especially the role of residue Thr 668 .

Thr 668 of APP Mediates Object Recognition Deficits found in FDD KI Mice
Recent findings suggest that products of BACE1-processing of APP (predominantly ß-CTF) trigger several pathological features related to human dementias both in a mouse model of FDD [10,16] and human neurons derived from familial and sporadic AD [9]. Thus, we decided to probe in more details the pathogenic function of the carboxyl-terminal region of APP, focusing on the intracellular Thr 668 residue (following the numbering of the APP 695 isoform). The phosphorylation status of Thr 668 either creates or destroys docking sites for intracellular proteins that interact with APP [19][20][21][22]. In addition, phosphorylation at Thr 668 is increased in AD cases [23] suggesting potential pathogenic implications. We generated mice expressing APP with a Thr 668 Ala mutation, called APP TA [24]. Western blot analysis of hippocampal synaptosomes from either APP WT/WT or APP TA/TA mice shows that the Thr 668 Ala mutation abolishes phosphorylation at Thr 668 ( Figure 1a).
Thus, the APP TA mice are an ideal genetic tool to study the role of Thr 668 and its phosphorylation in the pathogenesis of dementia. To this end, we utilized FDD KI mice, which develop severe agingdependent memory and synaptic plasticity deficits that first become measurable at ,5 months of age [13]. Most importantly, these deficits are prevented when FDD KI mice lack one allele of APP, reducing the APP protein load [14], and require production of APP ß-CTF [10,16]. Thus, since memory and synaptic deficits of FDD KI mice are dependent on endogenous APP, we can test the pathogenic role of Thr 668 by introducing this APP mutation on the FDD KI background.
By crossing FDD KI /APP TA/WT to APP TA/WT mice we generated littermates of the following 6 genotypes: WT, FDD KI , FDD KI /APP TA/TA , FDD KI /APP TA/WT , APP TA/TA and APP TA/ WT . To test memory, six-month-old mice were subjected to the novel object recognition (NOR) task, which is a non-aversive task that relies on the mouse's natural exploratory behavior. Open field studies showed that FDD KI , FDD KI /APP TA/TA , FDD KI /APP TA/WT , APP TA/TA and APP TA/WT mice have no defects in habituation and locomotor behavior, sedation, risk assessment and anxiety-like behavior in novel environments (Figure 1b and c). During the training session, mice of all genotypes spent the same amount of time exploring the two identical objects during the training phase ( Figure 1d). The following day, when a novel object was introduced, FDD KI spent the same amount of time exploring the two objects as if they were both novel to them, while the WT, APP TA/TA , and APP TA/WT mice still spent more time exploring the novel object ( Figure 1e). Notably, FDD KI /APP TA/TA and FDD KI /APP TA/WT mice behaved like the WT mice and explored preferentially the novel object (Figure 1e), demonstrating a prevention of the defect of the FDD KI mice. We subjected the mice to the NOR task at 9 months, and also at 12 months to confirm that this is a true prevention of deficits and not a delay. We found similar data to the data at 6 months with the FDD KI mice showing no preference between the two objects on the second day, while the FDD KI /APP TA/TA , FDD KI /APP TA/WT , APP TA/TA , APP TA/WT mice all behaved similar to the WT mice (Figure 1f and 1g). These data confirm that memory is impaired in FDD KI mice upon aging in an ethologically relevant, non-aversive behavioral context; remarkably, development of this deficit is fully prevented by changing the Thr 668 residue on the intracellular region of APP to an Alanine.

Thr 668 of APP Mediates Short-term Memory Deficits Found in FDD KI Mice
To further test memory, WT, FDD KI , FDD KI /APP TA/TA , FDD KI /APP TA/WT , APP TA/TA , APP TA/WT mice were subjected at 5.5 months of age to the radial arm water maze (RAWM) task, a spatial working memory test that depends upon hippocampal function [25]. This task tests short-term memory, which is the memory affected in early stages of AD. The six genotypes were required to learn and memorize the location of a hidden platform in one of the arms of a maze with respect to spatial cues. WT, APP TA/TA , and APP TA/WT mice were able to acquire (A) and retain (R) memory of the task. FDD KI mice showed severe abnormalities during both acquisition and retention of the task (Figure 2a), confirming that FDD KI mice have severe impairment in shortterm spatial memory for platform location during both acquisition and retention of the task. This defect was due to a deficit in memory per se and not to deficits in vision, motor coordination or motivation because testing with the visible platform showed no difference in the swimming speed and the time needed to find the platform between the FDD KI and WT mice (Figure 2c and d). Both the FDD KI /APP TA/TA and the FDD KI /APP TA/WT mice showed no defects in the memory test (Figure 2a), showing that mutating the intracellular APP residue Thr 668 to an alanine prevented the RAWM deficit of FDD KI mice, and confirming the data seen in NOR. To ensure that this was not simply a delay of the deficit, the mice were re-tested at 9 months in the RAWM task, and once again the FDD KI /APP TA/TA and the FDD KI /APP TA/WT mice did not show the deficit seen in the FDD KI mice (Figure 2b).

Thr 668 of APP Mediates Synaptic Deficits Found in FDD KI Mice
The FDD KI mice have compromised long-term potentiation (LTP) [13], a long-lasting form of synaptic plasticity that is thought to be associated with learning and memory. Like for memory, the LTP deficit of FDD KI mice are prevented when APP protein levels are halved [14], and by inhibiting processing of APP by BACE1 (also known as b-secretase) [10,16]. Thus, we tested if this one amino acid change in APP could also prevent the synaptic plasticity defect found in the FDD KI mice. To this end, we investigated synaptic transmission and plasticity using the Schaeffer collateral pathway in hippocampal slices from WT, APP TA/TA , FDD KI and FDD KI /APP TA/TA mice. As expected, LTP was reduced in FDD KI mice compared with WT littermates ( Figure 3). Strikingly, the APP TA/TA point mutation prevented LTP impairments in FDD KI mice ( Figure 3). Taken together, these findings provide compelling genetic evidence that APP and BRI2 functionally interact, and that the synaptic and memory deficiencies due to loss of Bri2 function require the APP intracellular residue Thr 668 .

Discussion
In this manuscript, we have pinpointed an intracellular residue of APP that is required for memory and synaptic plasticity deficits. FDD KI mice allow for a genetic analysis of pathogenic pathways on a genetic background that is congruous to the human disease. We showed that haplodeficiency in APP prevented all FDD KI mice's deficits at all ages. Now we take this further by showing that mutation in just one residue of APP, the intracellular amino acid Thr 668 , can also prevent the memory and synaptic deficits.
We studied the functional relevance of Thr 668 of APP because APP p Thr 668 is enriched in AD patients [23], suggesting a pathogenic role for phosphorylation at this residue, and because it has profound effects on APP protein/protein interactions and APP biology. For example, Thr 668 phosphorylation impairs APP/ Fe65 interaction [20,21] but promotes Pin1 binding [22]. In addition, this phosphorylation regulates trafficking of APP and APP derived metabolites [26]. Previous studies in mice suggested a protective role for phosphorylation of Thr 668 in the pathogenesis of AD by showing that Pin1 decreases APP processing and Aß production by binding APP phosphorylated on Thr 668 [27]. However, analysis of the APP TA mice has shown that preventing phosphorylation by mutating Thr 668 into an Ala does not change Aß levels in vivo [28,29].
If Aß were a major neuro-toxic peptide in dementia, FDD KI / APP TA/TA mice should either have deficits comparable to FDD KI mice based on the evidence that the Thr 668 Ala mutation does not change Aß levels [28,29], or should present with a worsened phenotype based on the hypothesis that binding of Pin1 to APP p Thr 668 reduces Aß levels [27]. Instead, we have found that the Thr 668 Ala mutation on one or both alleles of APP prevents all the memory and synaptic deficits found in the FDD KI mice. This is seen in a short-term memory test, such as the RAWM task, and also in an ethologically relevant, non-aversive behavioral context, such as the NOR task. The memory deficits were prevented at their start and no deficits could be found even as late as 9-12 months of age. The same was true for the synaptic plasticity at 12 months old. The FDD KI mice show strong synaptic defects in the Schaffer collateral pathway, however, FDD KI /APP TA/TA the mice showed no such deficits.
In this context, it is worth noting that mutation of another phosphorylated amino acid present in the APP intracellular region, namely Tyr 682 , results in a different (almost opposite) phenotype. This tyrosine is comprised in the intracellular 682 YENPTY 687 sequence of APP, a docking region for numerous APP-binding proteins that regulate processing and functions of APP [19,[30][31][32][33][34][35]. Phosphorylation of Tyr 682 is consequential. Some proteins, such as Grb2 [36], Shc [37,38], Grb7 and Crk [39] interact with APP only when Tyr 682 is phosphorylated; others, like Fe65, Fe65L1 and Fe65L2 only when this tyrosine is not   [40], suggesting that phosphorylation-dephosphorylation on Tyr 682 modulates APP functions. To test the in vivo function of Tyr 682 we have created mice with Tyr 682 replaced by a Gly. This knock-in mutation alters the function of APP in memory formation, development/aging [41,42] and changes APP processing, leading to a significant decrease in Aß levels [29]. Thus, while the Thr 668 Ala mutation on APP, which does not reduce Aß production, prevents memory deficits of FDD KI mice, the Tyr 682 Gly mutation, which reduces Aß production, causes cognitive defects on its own [41]. These data show that the intracellular region of APP has a fundamental role in memory formation, a role that is not linked to Aß.
New evidence points to b-derived metabolites of APP, especially ß-CTF, as the synaptic-toxic APP fragments mediating synaptic and memory impairments. The data presented here suggest that the synaptic-toxic activity of ß-CTF requires Thr 668 (Figure 4a-c). It is possible that this synaptic-toxic activity necessitates or is enhanced by phosphorylation of Thr 668 (Figure 4d-f), which is abolished by the Thr 668 Ala mutation. It is interesting to note that this mutation does not alter essential biological functions of APP during development [24], suggesting that targeting the role of Thr 668 , and perhaps its phosphorylation, in dementia may be an effective and safe therapeutic approach to dementias. Since the APP TA mutation prevents memory and synaptic deficits in heterozygosis, a partial reduction of the noxious pathogenic functions mediated by Thr 668 will be therapeutically efficient.

Mouse Handling
The animals used for these studies were backcrossed to C57Bl6/ J mice for at least 14 generations. Mice were handled according to the Ethical Guidelines for Treatment of Laboratory Animals of Albert Einstein College of Medicine. The procedures were described and approved in animal protocol number 200404. The Institutional Animal Care and Use Committee (IACUC) approved this protocol. IACUC is a federally mandated committee that oversees all aspects of the institution's animal care and use program, facilities and procedures. The regulations of the USDA and PHS require institutions using animals to appoint an IACUC. The members of the IACUC are appointed by the Dean of Albert Einstein College of Medicine of Yeshiva University (Einstein).

Electrophysiology and Behavior
Only male mice were used to avoid variations due to hormonal fluctuations during the estrous female cycle, which influence severely behavioral and electrophysiological tests.

Spatial Working Memory
A six-armed maze was placed into white tank filled with water (24-25uC) and made opaque by the addition of nontoxic white paint. Spatial cues were presented on the walls of the testing room. At the end of one of the arms was positioned a clear 10 cm submerged platform that remained in the same location for every trial in 1 d but was moved approximately randomly from day to day. On each trial, the mouse started the task from a different randomly chosen arm. Each trial lasted 1 min, and errors were counted each time the mouse entered the wrong arm or needed more than 10 s to reach the platform. After each error, the mouse was pulled back to its starting position. After four consecutive acquisition trials, the mouse was placed in its home cage for 30 min, then returned to the maze and administered a fifth retention trial. The scores for each mouse on the last 3 days of testing were averaged and used for statistical analysis.

Visible Platform Testing
Visible platform training to test visual and motor deficits was performed in the same pool as in the RAWM; however, the arms of the maze were removed. The platform was marked with a black flag and positioned randomly from trial to trial. Time to reach the platform and speed were recorded with a video tracking system (HVS 2020; HVS Image).

Open Field and Novel Object Recognition
After 30 min to acclimate to the testing room, each mouse was placed into a 40 cm640 cm open field chamber with 2 ft high opaque walls. Each mouse was allowed to habituate to the normal open field box for 10 min, and repeated again 24 hours later, in which the video tracking system (HVS 2020; HVS Image) quantifies the number of entries into and time spent in the center of the locomotor arena. Novel object recognition was performed as previously described [43]. Results were recorded as an object discrimination ratio (ODR), which is calculated by dividing the time the mice spent exploring the novel object, divided by the total amount of time exploring the two objects.

Electrophysiology
Transverse hippocampal slices (400 mm) were transferred to a recording chamber where they were maintained at 29uC and perfused with artificial cerebrospinal fluid (ACSF) continuously bubbled with 95% O 2 and 5% CO 2 . The ACSF composition in mM was: 124 NaCl, 4.4 KCl, 1 Na 2 HPO 4 , 25 NaHCO 3 , 2 CaCl 2 , 2 MgSO 4 , and 10 glucose. CA1 field-excitatory-postsynaptic potentials (fEPSPs) were recorded by placing both the stimulating and the recording electrodes in CA1 stratum radiatum. For LTP experiments, a 30 min baseline was recorded every minute at an intensity that evoked a response approximately 35% of the maximum evoked response. LTP was induced using a tetaburst stimulation (four pulses at 100 Hz, with bursts repeated at 5 Hz and each tetanus including one ten-burst train). Responses were recorded for 90 min after tetanization and plotted as percentage of baseline fEPSP slope.  (a and b), Due to loss of BRI2 protein, APP processing is increased during synaptic transmission and memory acquisition in FDD leading to increased production of ß-CTF. This event compromises synaptic plasticity and memory acquisition leading to memory deficits. (c), Thr 668 is essential for the pathogenic role of ß-CTF, as shown by the evidence that mutating this residue into an Ala prevents development of memory/synaptic deficits. (d-f), Phosphorylation of Thr 668 may be required or facilitate the synaptic-toxic role of ß-CTF, since the Thr 668 Ala mutation prevents phosphorylation. doi:10.1371/journal.pone.0057120.g004

Statistical Analysis
All data are shown as mean 6 s.e.m. Experiments were performed in blind. Statistical tests included two-way ANOVA for repeated measures and t-test when appropriate.