|Home | About | Journals | Submit | Contact Us | Français|
The pathogenic model of Alzheimer’s disease (AD) posits that aggregates of Aβ, a product of Aβ-precursor protein (APP) processing, cause dementia. However, alterations of normal APP functions could contribute to AD pathogenesis and it is therefore important to understand the role of APP. APP is a member of a gene family that shows functional redundancy as documented by the evidence that single knock-out mice are viable,whereas mice with combined deletions of APP-family genes die shortly after birth. A residue in the APP intracellular region, Y682, is indispensable for these essential functions of APP. It is therefore important to identify pathways that regulate phosphorylation of Y682 as well as the role of Y682 in vivo. TrkA is associated with both phosphorylation of APP-Y682 and alteration of APP processing, suggesting that tyrosine phosphorylation of APP links APP processing and neurotrophic signaling to intracellular pathways associated with cellular differentiation and survival. Here we have tested whether the NGF/TrkA signaling pathway is a physiological regulator of APP phosphorylation. We find that NGF induces tyrosine phosphorylation of APP, and that APP interacts with TrkA and this interaction requires Y682. Unpredictably, we also uncover that APP, and specifically Y682, regulates activation of the NGF/TrkA signaling pathway in vivo, the sub-cellular distribution of TrkA and the sensitivity of neurons to the trophic action of NGF. This evidence suggests that these two membrane protein’s functions are strictly interconnected and that the NGF/TrkA signaling pathway is involved in AD pathogenesis and can be used as a therapeutic target.
Genetic evidence points to a key role for APP processing in AD pathogenesis (J. Hardy and D. J. Selkoe, 2002; D. J. Selkoe and M. B. Podlisny, 2002). APP is a cell membrane protein that is sequentially cleaved by several enzymes. One of the APP-derived metabolites, Aβ, is believed to play a key role in AD pathogenesis; however, all attempts to validate this hypothesis clinically have failed so far. There are many possible reasons for these failures, including that other APP-derived metabolites cause the disease. Indeed, several other APP fragments have been implicated in triggering neurodegenerative processes; an NH2-terminal APP fragment interacts with DR6 to trigger axon pruning and neuron death (A. Nikolaev et al., 2009), and peptides derived from the APP intracellular domain can promote cell death both in vitro and in vivo (D. C. Lu et al., 2000; B. Passer et al., 2000; E. Bertrand et al., 2001; A. Madeira et al., 2005; L. Giliberto et al., 2008).
The intracellular region of APP, and specifically the tyrosine residue at position 682 (Y682, using the numbering of the APP695 isoform), has a critical role for developmental functions of APP (H. Li et al., 2010; A. P. Barbagallo et al., 2011) and in regulating APP processing (A. P. Barbagallo et al., 2010). Phosphorylation of Y682 has profound effects on APP protein/protein interactions. Some proteins interact with APP only when Y682 is phosphorylated, others only when this tyrosine is not phosphorylated (C. Russo et al., 2002; P. E. Tarr et al., 2002a; D. Zhou et al., 2004; R. Tamayev et al., 2009; D. Zhou et al., 2009). Given this crucial role of Y682, it is conceivable that phosphorylation of this residue has fundamental biological and pathological consequences. Therefore, it is significant to identify the mechanisms and pathways that regulate phosphorylation of Y682 in physiological conditions. We showed that over-expression of the nerve growth factor receptor TrkA produced phosphorylation of APP on Y682 and altered APP processing, suggesting that tyrosine phosphorylation of APP may functionally link APP processing and neurotrophic signaling to intracellular pathways associated with cellular differentiation and survival (P. E. Tarr et al., 2002b). The finding that interruption of NGF signaling in hippocampal neurons rapidly activates APP processing and causes neuronal apoptotic death further supported this link (C. Matrone et al., 2008; C. Matrone et al., 2009). In this study we asked whether the NGF/TrkA signaling pathway is a physiological regulator of APP phosphorylation. We find that NGF induces tyrosine phosphorylation of APP and, surprisingly enough unveils a vital role for APP in the activation and localization of the NGF/TrkA signaling pathway.
Hippocampal and medial septal neurons were prepared from embryonic day 17/18 (E17/E18) mice, as previously reported (C. Matrone et al., 2008). The hippocampus and septum was dissected in Hanks’ balanced salt solution buffered with Hepes and dissociated via trypsin/EDTA treatment. 106 cells were plated on 3.5-cm dishes pre-coated with poly-l-lysine. After 2 days of culturing in neurobasal medium with B-27 supplement and glutamax, cytosine arabinofuranoside was added to reduce glial proliferation. Neurons were exposed to NGF (50 ng/ml) for 48 h at 3–4 days after plating.
Dorsal Root (DRG) and Sympathetic Cervical ganglia (SCG) were collected from either WT or APPYG/YG mice (P0-P3). 30.000 cells were plated on slide and neurons were cultured for 3 days in NGF-containing DMEM plus serum. The anti-mitotic agent cytosine arabinose (3 ug/ml) was added 12hrs after plating. To determine the response to NGF, the number of neurons, stained by a specific neuronal marker (NeuN, Chemicon), was assessed.
Mice of either sex were on a C57BL/6 background. Mice were handled according to the Ethical Guidelines for Treatment of Laboratory Animals of Albert Einstein College of Medicine. APPYG/YG generation and genotyping has been described (A. P. Barbagallo et al., 2010). Genotyping for the APP-/- allele was performed as described in the Jackson Laboratory WEB site.
Transverse hippocampal slices (400 μm) were transferred to a recording chamber where they were maintained at 29°C and perfused with artificial cerebrospinal fluid (ACSF) continuously bubbled with 95% O2 and 5% CO2. The ACSF composition in mM was: 124 NaCl, 4.4 KCl, 1 Na2HPO4, 25 NaHCO3, 2 CaCl2, 2 MgSO4, and 10 glucose.
Equal amounts (10-30μg) of proteins were separated on 4-12% Bis-Tris SDS-PAGE gels or 4-20% Tris-Gly gels (Invitrogen), blotted onto PVDF membranes (Millipore) and incubated overnight with the appropriate primary antibody. The antibodies used were: mouse monoclonal α-APP (22C11) from Chemicon; α-αtubulin antibody from Sigma. α-pTrKA was kindly provided by M. Chao; α-TrkA, which detects basal levels of Trk receptors, α-pAKT and α-AKT were from Cell Signaling.
Hippocampal neurons were fixed for 20 min in PBS containing 4% formaldehyde and 4% sucrose, permeabilized with 0.1% Triton X-100 (20 min, 20°C), and processed for single labelling with monoclonal αAchE (Chemicon). Secondary antibodies coupled to Alexa dyes (488 or 594) were from Molecular Probes (Invitrogen). Nuclei were visualized by staining with Hoechst dye 33258 (1 μg/ml) (Sigma). Digital images were obtained with an Olympus BX51 microscope (100X oil and 60X oil objectives) equipped with a Spot Diagnostic Instruments camera and collected with Spot image analysis software. Controls were performed either by omitting the primary antibody or by preincubating the primary antibody with the corresponding peptide.
For DRG and CVS double immunofluorescence for APP (Y188) and TrkA (clone B3) primary antibodies were from Abcam and Santa Cruz Biotec, respectively. Secondary antibodies coupled to Alexa dyes (488 and 594) were from Molecular Probes (Invitrogen). Slides were examined with a confocal laser scanning microscope (Leica SP5, Leica Microsystems, Wetzlar, Germany) equipped with four laser lines: violet diode emitting at 405 nm, argon emitting at 488 nm, helium/neon emitting at 543 nm and helium/neon emitting at 633 nm. Confocal acquisition modality was multitrack and all the settings were maintained constant across different cases. For production of figures, brightness and contrast of images were adjusted. Cells (derived from four independent experiments) from CVS and DRG of both WT and APPYG/YG mice were first visually studied at the fluorescence microscope. Only cells displaying an healthy appearance: characterized by a regular cellular and nuclear pattern (the latter examined in the Hoechst channel), and a medium-high immunofluorescence intensity of both TrkA and APP were selected for confocal acquisition. For each cell particular care was taken to sample the equatorial plane of the cell, indentified through a fast scanning along the z-axis and by selecting the focal plane where the nucleus displayed the larger diameter.
For immunoprecipitations, protein samples were added to Dynabeads-ProteinG, according to the procedure described by the manufacturer (Invitrogen), and eluted with 0.1 M citrate buffer (pH 2.3). pH was adjusted by adding Tris/HCl 2 M. Immunoprecipitation with α-pTyrosine antibody was performed with pY99 or pY20 antibodies (30 μg/100 μL of Dynabeads-ProteinG) from Sigma. For TrkA IP we used α-TrkA antibody (clone B3) from Santa Cruz.
Values are expressed as mean ± SE. Statistical analysis was performed with ANOVA, followed by the Newman–Keuls test. Statistical significance was accepted at the 95% confidence level (P < 0.05).
To test whether NGF triggers phosphorylation of APP on Tyr residues in physiological conditions, we treated primary hippocampal neurons with 50 ng/ml of NGF for the indicated time periods (Fig. 1A). Lysates from treated neurons were immuno-precipitated with an α–APP antibody, and precipitants were analysed by WB analysis with an α–pTyr antibody (α-pY). As shown in Figure 1A, we detected a protein of molecular mass similar to mature APP that is immunoprecipitated by the α–APP antibody and is phosphorylated on Tyr residue(s). This phosphorylation is detectable 10min after NGF exposure and persists for at least 24hrs (Fig. 1A). In reciprocal experiments, we found that APP is immunoprecipitated by the α–pTyr antibody ((α-pY) from samples treated with NGF (Fig. 1B). The evidence that the TrkA receptor is phosphorylated upon NGF treatment indicates that the NGF/TrkA signaling pathways is properly activated in this experimental setting (Fig. 1B). In addition, the phosphorylation was very rapid and long-lasting indicating a direct and sustained NGF effect and further supporting the specificity of the NGF/TrkA signaling involvement. These results strongly suggest that NGF treatment induces APP phosphorylation on tyrosine(s) in primary neuronal cells.
It is likely that the NGF signaling in primary hippocampal neuronal cultures (Fig. 1A-B) is due to the cholinergic afferent population, that is considered one of the most responsive NGF population (Siegel GJ et al., 1999) and accounted for approximately 20-25% of the overall population in our experimental conditions (Fig. 1C). We speculated whether such NGF dependent APP phosphorylation may occur also in hippocampal slices, where most of neuronal connections are preserved, and in septal nuclei slices (data not shown), that are considered to be the major one’s cholinergic output of the CNS (J. Hartikka and F. Hefti, 1988). 10 min of NGF exposure induced APP phosphorylation on Tyr residues in hippocampal slices as well as in septum slices. Again, a band corresponding to mature APP is immunoprecipitated by the α-pY antibody in samples treated with NGF and this phosphorylation is synchronous to the activation of TrkA (Fig. 1D). Activation of TrkA is necessary to mediate Tyr-phosphorylation of APP, since the TrKA inhibitors CEP-2563 and K-252a prevented phosphorylation of APP (Fig. 1E and 1F). Thus, NGF/TrkA signaling mediate Tyr-phosphorylation of APP in two anatomical regions known to be involved in Alzheimer pathogenesis. Altogether the results indicate that activation of TrkA by NGF produces phosphorylation of APP on one or more tyrosine residues under physiological conditions. Although it is conceivable that TrkA, a receptor with tyrosine kinase activity, directly phosphorylates APP, phosphorylation of APP may also be mediated by other kinases that are activated by TrkA.
APP contains three tyrosine residues in the cytoplasmic tail (Y653, Y682, and Y687 of the APP695 isoform). The latter two are included in the APP-Y682ENPTY687 sequence, which is instrumental for the association with the PTB domains of the APP interacting proteins identified to date (G. D. King and R. Scott Turner, 2004). We, and others, have previously shown that over-expression of TrkA as well as constitutively active form of the tyrosine kinase Abl induces phosphorylation of APP on Y682 (N. Zambrano et al., 2001; P. E. Tarr et al., 2002b), suggesting that this tyrosine is the phosphorylation target of the NGF/TrkA signaling pathway. To test this hypothesis we have taken advantage of an APP Y682G knock-in (KI) mouse, in which Y682 has been mutated into Glycine, to understand the physiological functions of Y682 in vivo (A. P. Barbagallo et al., 2010). Hippocampal slices prepared from either APPYG/YG mice or Wild Type (WT) littermates were treated with NGF. The APPYG/YG mutant is not phosphorylated on Tyr 10 min after NGF treatment (Fig. 2A), suggesting that Y682 is either the phosphorylation target of the NGF/TrkA signaling pathway or that it is necessary for phosphorylation of other Tyr residues of APP by TrkA. However, and surprisingly, we found that TrkA phosphorylation induced by NGF was absent in APPYG/YG mice (Fig. 2A, bottom panel). Lack of TrkA phosphorylation did not depend on reduced TrkA expression by APPYG/YG mice (Fig. 2B) and suggests that TrkA signaling is impaired by the mutation at Y682. This prediction was confirmed by the finding that NGF fails to activate down-stream signaling molecules, such as Akt, in APPYG/YG mice (Fig. 2B).
Next we asked whether APPYG/YG mutation abolishes NGF/TrkA signaling by either a loss or a gain of function mechanism. We answered this question by analyzing APP null mice. Of note, the Y682G mutation acts like a null allele when the essential function of APP in development are analyzed in vivo (H. Li et al., 2010; A. P. Barbagallo et al., 2011). Consistent with this, we found that NGF-dependent TrkA activation is also impaired in APP-/- hippocampal slice, indicating that APP plays an important role in TrkA signaling and that this function requires Y682 (Fig. 2C). Moreover, no significant differences were observed in p75 expression and processing (data not shown).
We than tested whether the role of APP in TrkA signaling had functional consequences. NGF exerts a trophic activity on DRG and CVS neurons, which express high levels of TrkA (Table 1). Notably, DRG and CVS neurons isolated from APPYG/YG mice were insensitive to this trophic function of NGF (+NGF) and an extensive neuronal loss was assessed when compared to the corresponding +NGF samples from WT mice (table 1).
To dissect the molecular and biochemical mechanism by which APP regulates TrkA signaling, we tested whether APP and TrkA physically interact in vivo, under physiological condition. To this end, we prepared protein samples from septum either of WT, APPYG/YG or APP-/- mice. Samples were immunoprecipitated with an α-TrkA antibody and analyzed by WB with an α-APP antibody. As shown in Fig 3A, a fraction of endogenous TrkA is complexed to endogenous APP in mouse brain. Interestingly, the APPYG/YG mutant does not interact with TrkA although APPYG/YG mice express normal amounts of APP, suggesting an essential role for Y682 in the APP/TrkA interaction. However the lack of a selective α–TrkA antibody fails to rule out the possibility that also TrkB may be involved in these events.
Next, we determined whether APP regulates the sub-cellular distribution of TrkA. For this, we prepared from both WT and APPYG/YG mice primary neuronal cells derived from the DRG (Fig. 3) and SCG (not shown). In WT mice, at low magnification, TrkA immunofluorescence was distributed on neuronal cell bodies and at high intensities along neuritis. At high magnification, TrkA immunofluorescence, localized in vesicles of small and homogeneous size, showed a preferential distribution on the cellular membrane or immediately beneath (Fig. 3B,D; arrows). Of particular interest, TrkA staining was increased in proximity, and along, the neuritic domains and displayed a more grainy or beaded appearance with the tendency to organize in clusters (Fig. 3B; arrows). These clusters filled the neuritis and sometime was possible to observe a certain number of TrkA positive vesicles organized in a row suggesting an anterograde transport through cytoskeletric structures (Fig. 3C,D). The distribution pattern of APP was slight different than that of TrkA. APP immunopositive vesicles appeared less structurally defined and of more variable size (Fig. 3B-D). As for TrkA, also APP appeared to increase in immunoreactivity in proximity and along the neuritic domains and showed a preferential membrane distribution (Fig. 3B-D; arrows). The increase and accumulation of both APP and TrkA immunoreactivity in proximity and along neuritis (Fig. 3B) suggest the presence of an intense trafficking toward the cellular peripheral structures (axonal terminals, spines), where both proteins may exert their primary functions. Colocalization between TrkA and APP appeared to be selectively confined to the proximity and along neuritic domains (Fig. 3B-D; asterisks).
In APPYG/YG mice, at low magnification, TrkA immunofluorescence resembled the distribution pattern observed in WT mice. However, at high magnification, several marked differences were evident. The TrkA prevalent membrane distribution showed in WT mice was lost in APPYG/YG mice in favor of a more intracytoplasmatic, clearly perinuclear, distribution pattern (Fig. 3 E). The loss of an organized distribution pattern was also clearly evident when analyzing the neuritic domains where a marked reduction of TrkA immunofluorescence can be observed. Very often, only few labeled grains, distantly interspaced, and organized in a row, could be followed (Fig. 3 G): suggesting the disruption of the TrkA cytoskeletric transport. As for TrkA, APP immunoreactivity was highly decreased from the cell membrane and neuritic domains, while appeared to increase in the intracellular and perinuclear regions (Fig. 3E-G). Colocalization pattern between TrkA and APP appeared to be switched from the proximity and along neuritic domains (Fig. 3B-D; asterisks) in WT mice to the intracellular and perinuclear regions (Fig. 3E-G, asterisks) in APPYG/YG mice. Altogether these results suggest that in APPYG/YG mice both TrkA-APP interaction and cellular distribution are affected.
This study was initiated to determine whether NGF/TrkA promote phosphorylation on APP on tyrosine residues under physiological conditions, and to determine whether Y682 is the primary target. We answered the first question conclusively by showing that NGF treatment triggers APP Tyr-phosphorylation via activation of TrkA. To address the second question we studied the APP-KI mutant APPYG/YG, which bears a mutation in Y682. However, we could not determine whether Y682 is the TrkA phosphorylation target since NGF/TrkA signaling is strongly reduced in APPYG/YG hippocampal samples. This observation, and the finding that TrkA signaling is also compromised in APP null hippocampal samples, unveiled an unexpected role of APP in NGF/TrkA signaling. The evidence that neuronal cells from APPYG/YG mice are insensitive to the trophic function of NGF underlined the functional relevance of this role. TrkA and APP interact in brain, and this association is dependent on Y682. Immununofluorescence studies suggest that APP regulates peripheral export of TrkA and that this function is impaired in APPYG/YG mutant mice. Indeed, localization of both TrkA and APP is altered in APPYG/YG neurons. This evidence further underlines the functional relationship between APP and TrkA, although it does not clarify whether the localization of these proteins is changed because the interaction is abolished or whether the interaction is abolished because the localization of APP is altered by the YG mutation. Since plasma membrane TrkA is accessible to NGF, the data suggest that APP regulates NGF/TrkA signaling by regulating the surface levels of TrkA.
This evidence, together with previous findings involving the NGF/TrkA signaling axis in the regulation of APP processing (P. E. Tarr et al., 2002b; C. Matrone et al., 2008; C. Matrone et al., 2009), suggest that these two membrane protein’s functions are strictly interconnected and that the NGF/TrkA signaling pathway may be involved in the de-regulation of APP processing that causes AD and be used as a therapeutic target to correct such a de-regulation. It would also be important to test whether the APP mutation causing Familial Alzheimer disease, alter the NGF/TrkA signaling pathway and whether such alterations have a pathogenic role in AD.
This research was supported by: European Molecular Biology Organization (ASTF264-2010) to CM; FIRB (2010-2012) from the Italian Ministry of Higher Education and Scientific Research, IT to PC; Alzheimer Disease Research Grants (A2003-076 to L.D. and NIRG-10-173876 to APB); grant from NIH (AG033007 to L.D) and Thome Foundation (to L.D.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. We thank Dr Jhon Sutachan, Dr Rosalia Castro Perez (from Skirball Institute) for their support and technical assistance in neuronal cell culture preparation. We greatly appreciate administrative support of P Papa.
Conflict of Interest: None