Human amyloid precursor protein (hAPP) transgenic mice with high levels of amyloid-β (Aβ) develop behavioral deficits that correlate with the depletion of synaptic activity-related proteins in the dentate gyrus. The tyrosine kinase Fyn is altered in Alzheimer's disease brains and modulates premature mortality and synaptotoxicity in hAPP mice. To determine whether Fyn also modulates Aβ-induced behavioral deficits and depletions of synaptic activity-dependent proteins, we overexpressed Fyn in neurons of hAPP mice with moderate levels of Aβ production. Compared with nontransgenic controls and singly transgenic mice expressing hAPP or FYN alone, doubly transgenic FYN/hAPP mice had striking depletions of calbindin, Fos, and phosphorylated ERK (extracellular signal-regulated kinase), impaired neuronal induction of Arc, and impaired spatial memory retention. These deficits were qualitatively and quantitatively similar to those otherwise seen only in hAPP mice with higher Aβ levels. Surprisingly, levels of active Fyn were lower in high expresser hAPP mice than in NTG controls and lower in FYN/hAPP mice than in FYN mice. Suppression of Fyn activity may result from dephosphorylation by striatal-enriched phosphatase, which was upregulated in FYN/hAPP mice and in hAPP mice with high levels of Aβ. Thus, increased Fyn expression is sufficient to trigger prominent neuronal deficits in the context of even relatively moderate Aβ levels, and inhibition of Fyn activity may help counteract Aβ-induced impairments.
The amyloid precursor protein (APP) is expressed at high levels in synapses in which it appears to play roles in synaptic plasticity and neurite outgrowth (Turner et al., 2003). APP undergoes proteolytic cleavage resulting in several metabolites, including amyloid-β (Aβ) peptides (Turner et al., 2003). Abnormal accumulation and deposition of Aβ are pathological hallmarks of Alzheimer's disease (AD), which is associated with progressive cognitive decline and degeneration of synapses and neurons. However, the precise relationship between Aβ, synaptic impairments, and cognitive decline remains uncertain (Walsh and Selkoe, 2004).
APP/Aβ-induced neuronal and behavioral deficits have been investigated in several transgenic (TG) mouse models (Higgins and Jacobsen, 2003; Gotz et al., 2004; Kobayashi and Chen, 2005). Findings from many of these models, as well as from in vitro studies, are consistent with the hypothesis that small nonfibrillar assemblies of Aβ can aberrantly engage intracellular signaling pathways in neurons and thereby disrupt normal synaptic function (Lambert et al., 1998; Hsia et al., 1999; Klein et al., 2001; Verdier and Penke, 2004; Walsh and Selkoe, 2004). Several cell-surface receptors have been identified that might mediate these Aβ effects, including integrins (Sabo et al., 1995; Bi et al., 2002; Grace et al., 2002), nicotinic acetylcholine receptors (nAChRs) (Dineley et al., 2001), the p75 neurotrophin receptor (Yaar et al., 1997), and the receptor for advanced glycation end products (Yan et al., 1996; Arancio et al., 2004). Many of these receptors are located at the postsynaptic membrane, providing Aβ with a suitable vantage point from which to disrupt synaptic function.
Several signaling molecules that transduce receptor-activated signals are localized beneath the postsynaptic membrane and modulate cell-cell adhesion, receptor clustering, or the scaffolding of signaling molecules (Kennedy, 1997). One of these molecules is Fyn, a member of the Src family of tyrosine kinases (Thomas and Brugge, 1997). Strategically located at the postsynaptic density of glutamatergic synapses, it is in an excellent position to modulate the effects of Aβ on synapses. Indeed, Fyn activation occurs during engagement of integrins and glutamate receptors (Roskoski, 2004; Salter and Kalia, 2004), both of which have been implicated in Aβ-induced disruption of synaptic function (Klein et al., 2001; Grace et al., 2002; Williamson et al., 2002). Furthermore, the distribution and levels of Fyn are altered in AD brains (Shirazi and Wood, 1993; Ho et al., 2004), and the toxic effects of Aβ-derived diffusible ligands on murine hippocampal slices were blocked by the genetic ablation of Fyn (Lambert et al., 1998). Finally, ablation of Fyn decreased, whereas overexpression of Fyn increased, Aβ-induced synaptotoxicity and premature mortality in human APP (hAPP) mice (Chin et al., 2004). Together, these results suggest that Aβ may derange synaptic functions by aberrantly engaging Fyn-related signaling pathways.
Here, we tested the hypothesis that increased neuronal activity of Fyn can trigger and exacerbate Aβ-induced neuronal and behavioral deficits. Increasing Fyn activity in neurons elicited memory impairments in TG mice with moderate levels of hAPP/Aβ expression and significantly exacerbated their depletion of synaptic activity-related proteins, creating a striking phenocopy of functional and molecular impairments typically seen only in mice with much higher levels of Aβ. These results highlight the potential significance of pathogenic interactions between hAPP/Aβ and Fyn-related signaling pathways.
Materials and Methods
TG mice. hAPP TG lines J9 and J20 produce hAPP carrying the Swedish (K670N, M671L) and Indiana (V717F) familial AD mutations (hAPP770 numbering) directed by the platelet-derived growth factor (PDGF) β chain promoter (Rockenstein et al., 1995; Mucke et al., 2000). The lines were crossed for >10 generations onto a C57BL/6 background using NTG mice from The Jackson Laboratory (Bar Harbor, ME). Heterozygous FYN TG mice (line N8) overexpress wild-type mouse Fyn directed by the calcium/calmodulin-dependent protein kinase IIα promoter on a C57BL/6 background (Kojima et al., 1997, 1998). Unless indicated otherwise, mice were evaluated at 6-8 months of age. To decrease variability, only male mice were assessed in the Morris water maze; all other measurements were performed on gender-balanced groups. Mice were anesthetized and flush-perfused transcardially with PBS. Hemibrains were fixed in 4% phosphate-buffered paraformaldehyde or stored at -70°C. All experiments were approved by the Committee on Animal Research of the University of California, San Francisco.
Immunohistochemistry. Sliding microtome sections (30 μm) were avidin-biotin/immunoperoxidase stained using the following primary antibodies: anti-calbindin (1:15,000; rabbit polyclonal; Swant, Bellinzona, Switzerland), anti-Fos (1:10,000; rabbit polyclonal Ab-5; Oncogene, San Diego, CA), anti-dually phosphorylated extracellular signal-regulated kinase 1/2 (phospho-ERK1/2; 1:100; rabbit polyclonal; Cell Signaling, Beverly, MA), anti-Arc (1:8000; rabbit polyclonal; a gift from Dr. P. Worley, Johns Hopkins University, Baltimore, MD), and biotinylated anti-Aβ (1:500; mouse monoclonal 3D6; Elan Pharmaceuticals, South San Francisco, CA). Biotinylated goat anti-rabbit (1:200; Vector Laboratories, Burlingame, CA) was used as the secondary antibody. Diaminobenzidine was used as a chromagen. For the analysis of calbindin-immunoreactive (IR) structures, two coronal sections (300 μm apart) per mouse between -2.54 and -2.80 mm from bregma were selected. The integrated optical density (IOD) of immunostains was determined with BioQuant Image Analysis (R&M Biometrics, Nashville, TN) and averaged in two areas (0.04 mm2 each) of the molecular layer of the dentate gyrus and of the stratum radiatum of the CA1 region. Relative calbindin-IR levels were expressed as the ratio of IODs in the molecular layer and in the stratum radiatum of the same section. The mean ratio of nontransgenic (NTG) mice was defined as 1. The relative numbers of Fos-IR, phospho-ERK1/2-IR, and Arc-IR granule cells were determined by counting labeled cells in the granular layer in every 10th serial coronal section throughout the rostrocaudal extent of the granular layer. For plaque-load determination, the average percentage area of the hippocampus occupied by Aβ-IR deposits was determined in four coronal sections (300 μm apart) per mouse, as described previously (Mucke et al., 2000).
Vibratome sections (50 μm) were labeled with anti-glial fibrillary acidic protein (GFAP; 1:500; Chemicon, Temecula, CA) as described previously (Toggas et al., 1994; Mucke et al., 1995), followed by a biotinylated species-specific secondary IgG (1:200; Vector Laboratories). After development with diaminobenzidine, relative levels of immunoreactivity were determined densitometrically with a Quantimet 570C (Leica, Arcadia, CA).
Immunoprecipitation and Western blot analysis. A McIlwain tissue chopper was used to cut hemibrains into 450-μm-thick horizontal sections from which the dentate gyrus was microdissected on ice. For protein quantitations, dentate gyrus samples from each hemibrain were pooled and homogenized on ice in buffer containing 320 mm sucrose, 10 mm Tris-HCl, pH 7.4, 10 mm EDTA, 10 mm EGTA, 1% deoxycholate, 1 mm PMSF, Phosphatase Inhibitor Cocktails I and II (Sigma), and protease inhibitor mixture (Roche). Samples were then briefly sonicated on ice and centrifuged at 5000 × g for 10 min. Equal amounts of protein (determined by the Bradford assay) were resolved by SDS-PAGE on 4-12% gradient gels and transferred to nitrocellulose membranes. For immunoprecipitations, 70 μg of protein were incubated overnight at 4°C with 5 μl of goat-anti-Fyn polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Protein G-coupled magnetic beads were then added to precipitate antibody-bound proteins. After washing the beads, 2× sample buffer was added to liberate the antibody-protein complexes, and samples were resolved by SDS-PAGE. For analysis of phosphoproteins, membranes were labeled with anti-Tyr416, which recognizes active Fyn (1:1000; rabbit polyclonal; Cell Signaling), followed by incubation with HRP-conjugated goat anti-rabbit IgG (1:5000; Chemicon) secondary antibody. For analysis of total Fyn levels, blots were stripped and reprobed with anti-Fyn (1:1000; mouse monoclonal; BD Transduction Laboratories, San Jose, CA). The following antibodies were used for Western blot analyses and detected with species-appropriate HRP-conjugated secondary antibodies: anti-human APP (1:1000; Elan Pharmaceuticals), anti-striatal-enriched phosphatase (STEP; 1:5000; mouse monoclonal; Novus Biologicals, Littleton, CO), anti-α7 nAChR (1:1000; mouse monoclonal; Covance, Princeton, NJ), and anti-calbindin (1:15,000; rabbit polyclonal; Swant). Bands were visualized by ECL and quantitated densitometrically with Quantity One 4.0 software (Bio-Rad, Hercules, CA).
Novel environment. Mice were singly housed for 3 d before the experiment. Mice assigned to novel environment exploration were transferred to an adjacent room that was different in size, shape, light, and furnishing and placed into another cage. The new cage was kept uncovered and contained a different type of bedding as well as five novel objects. The mice were allowed to explore the new environment for 2 h, whereas the remaining mice were kept undisturbed in their home cages.
The activity of the mice and their interactions with the novel objects were quantified from video records during the first 10 min of each hour the mice spent in the new cage. Activities included ambulatory movements, rearing, sniffing, and digging. Overall activity was counted as the proportion of time the mice were engaged in any of these behaviors. An object interaction event was defined as any type of close exploratory activity with any of the five novel objects. Observers were blinded with respect to the genotype of mice.
At the end of the observation period, alternate mice assigned to the home cage or novel environment condition were taken to another room, deeply anesthetized, and perfused transcardially.
Morris water maze. The water maze pool (diameter, 122 cm) contained opaque water (22-23°C) with a platform (diameter, 10 cm) submerged 1.5 cm. A black and white striped mast (15 cm in height) was mounted on the platform for cued training sessions and removed for hidden platform sessions. For hidden platform sessions, the platform (14 × 14 cm) was submerged 1.5 cm and was not cued. Mice were trained to locate the cued platform (sessions 1-6) and the hidden platform (sessions 7-16) in two daily sessions (3.5 h apart), each consisting of two 60 s trials (cued training) or three 60 s trials (hidden training) with a 15 min intertrial interval. The platform location was changed for each cued platform session but remained constant in the hidden platform sessions. Entry points were changed semirandomly between trials. Twenty-four hours after completion of the hidden platform training, a 60 s probe trial (platform removed) was performed. Entry points for probe trials were in the west quadrant, and the target quadrant was the southeast quadrant. Performance was monitored with an EthoVision video-tracking system (Noldus Information Technology, Wageningen, The Netherlands).
Elevated plus maze. The elevated plus-shaped maze consisted of two open arms and two closed arms equipped with rows of infrared photo-cells interfaced with a computer (Hamilton, Poway, CA) as described previously (Raber et al., 2000). Briefly, mice were placed individually into the center of the maze and allowed to explore the apparatus for 10 min. The number of beam breaks was recorded to calculate the amount of time spent and distance moved in each arm, as well as the number of times the mice extended over the edges of the open arms. The apparatus was cleaned with 0.25% bleach between the testing of each mouse to standardize odors.
Statistical analyses. Statistical analyses were performed with Statview 5.0 (SAS Institute, Cary, NC) or SPSS-10 (SPSS, Chicago, IL). Differences between means were assessed by Student's t test, Mann-Whitney U test, one-factor ANOVA, or two-factor ANOVA, followed by Tukey-Kramer or Fisher's PLSD post hoc tests. Correlations were assessed by simple regression analysis. The numbers of IR cells analyzed in the regressions are shown as relative values (compared with the mean of NTG mice) of the natural log of the total numbers of IR cells counted in each mouse.
To test for copathogenic effects between hAPP/Aβ and Fyn, we crossed heterozygous TG mice from a PDGF-hAPP line with moderate levels of neuronal hAPP/Aβ expression (J9, “hAPPlow”) (Hsia et al., 1999; Mucke et al., 2000; Chin et al., 2004) with heterozygous TG mice from a CamKIIα-FYN line (line N8) with increased neuronal levels of wild-type mouse Fyn (Kojima et al., 1997, 1998). The cross yielded four genotypes: NTG mice, singly TG hAPP mice, singly TG FYN mice, and doubly TG FYN/hAPP mice. These groups of mice were compared with heterozygous TG mice from a PDGF-hAPP line with higher levels of neuronal hAPP/Aβ expression (J20, “hAPPhigh”) (Mucke et al., 2000; Palop et al., 2003; Chin et al., 2004).
Increased expression of Fyn exacerbates hAPP/Aβ-induced depletion of calbindin and Fos
The calcium-binding protein calbindin-D28k regulates intracellular calcium levels and is important for long-term potentiation and for learning and memory (Molinari et al., 1996). It is markedly reduced in the dentate gyrus of many, but not all, hAPPhigh mice in a manner that strongly correlates with impaired performance in the Morris water maze (Palop et al., 2003). Calbindin levels in granule cells of the dentate gyrus were only slightly reduced in singly TG hAPPlow mice and unaffected in singly TG FYN mice compared with NTG controls (Fig. 1). In contrast, doubly TG FYN/hAPPlow mice had profound depletions of calbindin (Fig. 1A,B) that were similar in magnitude to those in age-matched singly TG hAPPhigh mice without the FYN transgene (Fig. 1C). These results were confirmed by Western blot analysis of dentate gyrus lysates, which demonstrated lower calbindin levels in FYN/hAPPlow mice than in NTG mice and singly TG controls (Fig. 1D, bottom). hAPP (Fig. 1D, top) and Fyn levels (Fig. 1D, middle) are shown to illustrate the similar extent to which these proteins are overexpressed in FYN/hAPPlow mice and in the corresponding singly TG controls.
Calbindin depletions in the dentate gyrus of hAPPhigh mice correlate tightly with reductions in Fos-positive granule cells (Palop et al., 2003). hAPPlow and FYN singly TG mice had fewer Fos-positive granule cells than NTG controls (Fig. 2A,C), although only FYN/hAPPlow mice had Fos reductions as severe as those in hAPPhigh mice (Fig. 2D). Decreases in Fos-positive granule cells correlated with calbindin reductions in mice of all TG genotypes in which Fos levels were reduced (Fig. 2B), suggesting that Fos and calbindin were affected by a common mechanism.
Fyn and hAPP/Aβ decrease the activity of ERK1/2 in the dentate gyrus
The expression of Fos and other immediate-early genes is controlled by transcription factors, the nuclear translocation and transcriptional activity of which are regulated by synaptic activity-dependent kinases such as ERK1/2 (Kelleher et al., 2004; Sweatt, 2004). To test whether the reduction in Fos might be related to impaired ERK1/2 activity, we examined the expression of dually phosphorylated (activated) ERK1/2 in granule cells. Similar to the results obtained for Fos, the number of phospho-ERK1/2-positive granule cells was significantly reduced in singly TG hAPPlow mice and most severely affected in FYN/hAPPlow mice (Fig. 3A-C). Once again, the reductions in phospho-ERK1/2-positive granule cells in FYN/hAPPlow mice matched those in hAPPhigh mice (Fig. 3D). The numbers of Fos-positive and phospho-ERK-positive granule cells were correlated in all groups of TG mice, suggesting that these proteins are coregulated (Fig. 3B).
Combined expression of Fyn and hAPP/Aβ impairs induction of Arc expression after exploration of a novel environment
ERK1/2 also regulates the expression of the immediate-early gene product Arc (Waltereit et al., 2001; Kremerskothen et al., 2002; Ying et al., 2002). Arc is highly enriched in dendrites of hippocampal neurons where it binds to cytoskeletal elements (Lyford et al., 1995) and has been characterized for the accumulation of its mRNA and protein in segments of dendrites that were recently active (Steward et al., 1998). Exploration of a novel environment has been shown to induce Arc expression in dentate granule cells (Guzowski et al., 1999; Pinaud et al., 2001; Chawla et al., 2005). Novel environment exploration strongly increases Arc expression in granule cells of the dentate gyrus in NTG mice but not in hAPPhigh mice (Palop et al., 2005). In contrast, singly TG hAPPlow and FYN mice showed normal Arc induction in this paradigm, although their basal Arc levels were lower than those of NTG littermates (Fig. 4A,D). Notably, FYN/hAPPlow mice had more prominent reductions in basal Arc levels and actually showed a decrease in Arc expression instead of an increase after environmental exploration (Fig. 4A,B,D). No significant differences were found in locomotion or object interactions among the four genotypes (data not shown). All groups of mice showed similar basal levels of Arc and comparable Arc induction in the pyramidal cell layer of CA1 (Fig. 4A,E) and in the neocortex (Fig. 4C,F), although the induction was slightly attenuated in hAPPlow and FYN/hAPPlow mice. The selective deficit in basal and induced Arc expression in the dentate gyrus of doubly TG FYN/hAPPlow mice is very similar to that observed in singly TG hAPPhigh mice (Palop et al., 2005), suggesting that Fyn and hAPP/Aβ act synergistically to impair Arc expression in a brain region-specific manner.
Cooperation of Fyn and hAPP/Aβ impairs spatial memory retention and alters emotional behavior
Because the molecules that were reduced in the dentate gyrus of FYN/hAPPlow mice are critically involved in the formation or consolidation of memories (Molinari et al., 1996; Atkins et al., 1998; Blum et al., 1999; Guzowski et al., 2000; Guzowski, 2002; He et al., 2002), we next assessed the spatial learning and memory of mice in the Morris water maze. NTG, hAPPlow, FYN, and FYN/hAPPlow mice were able to learn both the cued (platform location visible) and the spatial (platform hidden) components of this task (Fig. 5A). However, when retention of spatial memory was tested in a probe trial 24 h after the last training trial, only FYN/hAPPlow mice were impaired, showing no preference for the target quadrant based on comparisons of the time they spent in each quadrant (Fig. 5B) and of their swim paths (Fig. 5C). hAPPhigh mice have similar probe trial deficits, although many of these mice also show deficits in task acquisition (Palop et al., 2003).
To determine whether Fyn and hAPP/Aβ may interact also in brain regions other than the dentate gyrus, we examined the behavior of the mice in the elevated plus maze, a paradigm that provides putative measures of emotionality and exploratory behaviors that are mostly independent of the hippocampus (Dawson and Tricklebank, 1995; Wall and Messier, 2001). The coexpression of FYN and hAPPlow also affected the performance of mice in the elevated plus maze. FYN/hAPPlow mice spent significantly more time in the open arms of the maze and poked their noses over the edges of the open arms more often than hAPPlow mice, FYN mice, and NTG controls (Fig. 6A,B). FYN/hAPPlow mice showed no significant preference for the closed arms over the open arms of the maze, in contrast to the other three groups of mice (Fig. 6C). These behavioral alterations of FYN/hAPPlow mice are remarkably similar to those of hAPPhigh mice in this test (Fig. 6D-F).
Increased Fyn expression does not affect amyloid deposition
We previously used ELISA to demonstrate that neither increasing nor ablating Fyn expression affects hippocampal Aβ levels in 6- to 7-month-old mice (Chin et al., 2004). To determine whether the exacerbation of hAPP/Aβ-induced neuronal and behavioral alterations by Fyn could stem from changes in the deposition of Aβ, we compared Aβ deposits (plaque load) in hAPPlow and FYN/hAPPlow mice. At 6-8 months, the age at which all other biochemical, immunohistochemical, and behavioral experiments were performed, no plaques were detected in the hippocampus of these mice (Fig. 7A, top). By 15-16 months of age, hAPPlow and FYN/hAPPlow mice had comparable hippocampal plaque loads (Fig. 7A,B) (p = 0.13; Mann-Whitney U test). No correlation was found between plaque load and calbindin levels in either hAPPlow or FYN/hAPPlow mice (Fig. 7C) at this age.
Increased astroglial injury response in FYN/hAPPlow mice
Astrocytes upmodulate the expression of GFAP in response to a variety of neural injuries (Eddleston and Mucke, 1993; Eng et al., 2000). We analyzed this sensitive indicator of neuronal distress at 6-8 months of age. Hippocampal GFAP levels in hAPPlow and FYN mice were similar to those of NTG controls, whereas FYN/hAPPlow mice had significantly higher levels of GFAP labeling (Fig. 7D). This increase in GFAP expression in FYN/hAPPlow mice without plaques was similar in magnitude to that in age-matched hAPPhigh mice with plaques (Fig. 7E).
Regulation of Fyn activity in hAPPhigh and FYN/hAPPlow mice
To determine whether hAPP/Aβ actually increases Fyn activity in vivo, we immunoprecipitated Fyn from dentate gyrus lysates of hAPPhigh mice and examined the levels of Tyr416 phosphorylation with a phospho-specific antibody. Phosphorylation at this site activates Fyn and is indicative of kinase activity (Nguyen et al., 2002; Roskoski, 2004; Santuccione et al., 2005). In hAPPhigh mice with relatively preserved levels of calbindin (higher than the average of all hAPPhigh mice), levels of active Fyn were similar to those in NTG mice. In contrast, hAPPhigh mice with prominent depletions of calbindin (levels lower than the average of all hAP-Phigh mice) exhibited decreased levels of active Fyn (Fig. 8A). Notably, our previous studies demonstrated that the extent of calbindin depletion in hAPPhigh mice correlates tightly with the severity of learning and memory deficits in the Morris water maze test (Palop et al., 2003).
Next, we examined whether levels of active Fyn were also different in FYN and FYN/hAPPlow mice. Although levels of total Fyn were similar in these groups (Figs. 1D, 8B, bottom blot), levels of active Fyn were strikingly lower in FYN/hAPPlow mice than in FYN mice (Fig. 8B).
Concurrent increases in STEP and α7 nAChR in hAPPhigh and FYN/hAPPlow mice
What regulatory mechanisms might underlie the surprising downmodulation of Fyn in hAPPhigh and FYN/hAPPlow mice? To begin to address this question, we measured the 61 kDa form of STEP, which associates with and dephosphorylates Fyn at the regulatory Tyr416 (Nguyen et al., 2002). STEP is present at postsynaptic densities of glutamatergic synapses and interacts with both src kinase family members and NMDA receptors, opposing src-mediated upregulation of NMDA receptors (Pelkey et al., 2002). STEP levels were indeed significantly increased in the dentate gyrus of hAPPhigh mice relative to NTG controls (Fig. 9A). Because Aβ has recently been shown to activate STEP in cultured neurons via α7 nAChRs (Snyder et al., 2005), we measured α7 nAChR levels in the dentate gyrus of our mice. The α7 nAChR levels were approximately twice as high in hAPPhigh mice as in NTG controls (Fig. 9B), indicating that Aβ engages this pathway also in vivo.
Finally, we examined whether STEP and α7 nAChRs were also increased in the dentate gyrus of FYN/hAPPlow mice. Compared with NTG and FYN mice, FYN/hAPPlow mice had approximately twofold higher levels of STEP and fivefold higher levels of α7 nAChRs (Fig. 9C-E). STEP levels were also increased in singly TG hAPPlow mice, although to lesser extent than in FYN/hAPPlow mice (Fig. 9C,D).
The current study demonstrates that increased neuronal expression of Fyn can promote hAPP/Aβ-dependent neuronal and behavioral deficits that may be relevant to the cognitive decline in AD. Elevating neuronal expression of Fyn in hAPPlow mice with relatively minor neuronal deficits created a striking phenocopy of the more severe deficits seen in hAPPhigh mice.
In considering the different possibilities in which Fyn and Aβ might interact, it is not unreasonable to speculate that Fyn actually mediates the detrimental effects of Aβ. However, our measurements of Fyn activity suggest that Fyn could also act as a separate copathogenic factor that converges onto similar or identical downstream targets. Compared with NTG controls, hAPPhigh and FYN/hAPPlow mice had more severe neuronal and behavioral deficits than hAPPlow and FYN mice. Yet, Fyn activity was lower in hAPPhigh mice than in NTG controls and was also lower in FYN/hAPPlow mice than in FYN mice. These findings suggest the involvement of complex regulatory mechanisms that may become engaged to protect neurons against synergistic effects of Aβ and Fyn.
In line with this interpretation, we found significant increases in levels of STEP, a phosphatase that inactivates Fyn, in both hAPPhigh and FYN/hAPPlow mice. STEP opposes the phosphorylation of src family kinase substrates such as Tyr1472 of the NR2B subunit of NMDARs (Pelkey et al., 2002). Interestingly, hAPPhigh mice have significant decreases in levels of phospho-Tyr1472 of NR2B (Palop et al., 2005), consistent with the increased STEP levels and suppressed Fyn activity we detected in these mice in the current study. Aβ-induced activation of STEP also led to reduced phospho-Tyr1472 of NR2B in cultured neurons, resulting in endocytosis of the receptor and reduced NMDAR-mediated signaling (Snyder et al., 2005). The study by Snyder et al. (2005) also demonstrated that Aβ can recruit STEP via α7 nAChR-dependent signaling. Our identification of abnormally high α7 nAChR levels in both hAPPhigh and FYN/hAPPlow mice high-lights the potential functional relevance of this pathway in vivo. Indeed, increased levels of nAChRs and STEP in hAPPhigh and FYN/hAPPlow mice were associated with depletions of synaptic activity-related proteins that are known to correlate tightly with deficits in learning and memory (Palop et al., 2003, 2005). Unlike the majority of membrane-bound receptors, nAChR expression is increased after receptor stimulation (Fenster et al., 1999), supporting the notion of increased signaling through α7 nAChRs in hAPPhigh and FYN/hAPPlow mice. Aβ and Fyn activities may also converge on other cell-surface molecules, including integrins and NMDARs, and these possibilities deserve to be tested in future studies.
Independent of the precise relationship between Aβ and Fyn-related signaling pathways, the results presented here clearly demonstrate that increased neuronal Fyn expression can critically exacerbate neuronal and behavioral impairments in the context of even relatively moderate increases in Aβ production. The similarity of impairments in FYN/hAPPlow and hAPPhigh mice was essentially complete with respect to molecular alterations and somewhat less complete with respect to behavioral deficits. In the water maze test, FYN/hAPPlow mice were impaired only in retention of spatial memory assessed by the 24 h probe trial, which provides a sensitive measure of hippocampus-dependent function (Lipp and Wolfer, 1998; Wolfer et al., 1998). In contrast, hAPPhigh mice were also impaired in task acquisition (Palop et al., 2003). Thus, Fyn signaling may play a more prominent role in hippocampus-dependent memory retention deficits than in acquisition deficits. The possibility of such dissociations is highlighted by our previous study (Chin et al., 2004), which demonstrated that hAPP/Aβ-dependent premature mortality and loss of synaptophysin are modulated by Fyn, whereas aberrant axonal sprouting is not.
Although singly TG FYN and hAPPlow mice had no behavioral alterations, both groups showed some depletion of synaptic activity-dependent proteins in granule cells of the dentate gyrus. Although these depletions were less severe than those in hAPPhigh and FYN/hAPPlow mice, they demonstrate that increased neuronal expression of Fyn is sufficient to elicit molecular alterations similar to those induced by moderate levels of hAPP/Aβ. These findings also suggest that our molecular outcome measures may be more sensitive than the behavioral tests we used in this study.
Because reductions in Fos, Arc, and ERK1/2 can contribute to deficits in learning and memory (Molinari et al., 1996; Atkins et al., 1998; Blum et al., 1999; Guzowski, 2002; He et al., 2002), it is possible that the neuronal network can compensate for the relatively subtle reduction of these factors in FYN and hAPPlow mice but is overcome by their more severe depletion in hAPPhigh and FYN/hAPPlow mice, resulting in behavioral decompensation. Consistent with this hypothesis, Arc induction was normal in FYN mice and in hAPPlow mice, suggesting that activity-regulated gene expression was normal. However, there was no induction of Arc in FYN/hAPPlow mice that explored a novel environment. In fact, these mice had even fewer Arc-positive granule cells than home-caged FYN/hAPPlow mice, suggesting a paradoxical response of these cells or of the network to which they belong.
FYN/hAPPlow mice also exhibited altered behavior in the elevated plus maze in a manner similar to hAPPhigh mice. This test is used to measure emotional behaviors that are mostly independent of the hippocampus (Dawson and Tricklebank, 1995; Wall and Messier, 2001). Thus, the alterations observed in FYN/hAPPlow mice in this test suggest that activities of Fyn and hAPP/Aβ may interact also in brain regions other than the dentate gyrus. In addition, they suggest that this interaction can contribute not only to memory deficits, but also to alterations in emotional behavior, which afflict both hAPP mice and AD patients (Terry et al., 1991; Cummings and Kaufer, 1996; Lee et al., 2004).
Although increased production of Aβ may be sufficient to cause early-onset autosomal-dominant AD (Tanzi and Bertram, 2005), the much more frequent cases of late-onset “sporadic” AD are likely caused by the combination of different genetic and environmental risk factors (Mayeux, 2003; Kamboh, 2004; Tanzi and Bertram, 2005). It will be interesting to determine whether some of these risk factors elevate Fyn activity, because our findings suggest that this process could synergize with Aβ to induce neuronal dysfunction. Indeed, increases in neuronal Fyn expression triggered hAPP/Aβ-dependent behavioral deficits and worsened the depletion of multiple factors involved in synaptic plasticity.
It is noteworthy in this context that 6- to 8-month-old FYN/hAPPlow mice had not yet formed amyloid plaques but in most molecular and behavioral measures were as severely impaired as plaque-bearing hAPPhigh mice of the same age. This finding extends to the molecular level the hypothesis that Aβ-induced behavioral deficits in hAPP mice are primarily plaque independent (Holcomb et al., 1999; Hsia et al., 1999; Westerman et al., 2002; Palop et al., 2003). Indeed, our results are most consistent with the hypothesis that Fyn promotes the pathogenic effects of small nonfibrillar Aβ assemblies (Lambert et al., 1998), although interactions with alternative hAPP metabolites have not yet been excluded. Studies are needed to determine whether therapeutic manipulations of Fyn or related pathways could protect the aging brain against pathogenic Aβ assemblies and AD-related neurological decline.
This work was supported by fellowships from the John Douglas French Alzheimer's Foundation (J.C.) and the Academy of Finland (J.P.) and by National Institutes of Health Grants NS41787 and AG022074 (L.M.). We thank N. Kojima for Fyn transgenic mice; P. Worley for the anti-Arc antibody; G.-Q. Yu, X. Wang, and H. Ordanza for technical assistance; and D. Murray McPherson for administrative assistance.
Correspondence should be addressed to Dr. Lennart Mucke, Gladstone Institute of Neurological Disease, 1650 Owens Street, San Francisco, CA 94158. E-mail:.
J. Puoliväli's present address: Cerebricon Ltd., Microkatu 1, P.O. Box 1188, 70211 Kuopio, Finland.
Copyright © 2005 Society for Neuroscience 0270-6474/05/259694-10$15.00/0