Introduction
Amyloid-β-protein (Aβ), the key component of senile plaques in Alzheimer's disease (AD) neuropathology, was first isolated from meningovascular amyloid deposits in AD and Down's syndrome [
1,
2], and has also been reported to be the subunit of the plaque amyloid [
2‐
4]. The current amyloid hypothesis of AD states that the imbalance between Aβ generation and Aβ clearance is the basis of AD neuropathogenesis. Aβ is generated from amyloid precursor protein (APP). Specifically, APP is first hydrolyzed by β-secretase to generate a 99-residue membrane-associated C-terminus fragment (APP-C99) [
5‐
8]. APP-C99 is further cleaved to release a ~4-kDa peptide, Aβ, and the amyloid precursor protein
i ntra
c ellular
d omain (AICD). This cleavage is achieved by an unusual form of proteolysis in which the protein is cleaved within the transmembrane domain (at residue +40 or +42) by γ-secretase [
9‐
11]. α-secretase cleaves the majority of APP in the middle of the Aβ region of APP. This cleavage will preclude Aβ generation, lead to the release of a large ectodomain (α-APPs), and leave behind a carboxy-terminus fragment of 83 amino acids (APP-C83) in the membrane. γ-Secretase cleaves APP-C83 to produce p3, an amino-terminally truncated form of Aβ [
12,
13], [see review in [
14]].
The cleavage of the APP cytoplasmic tail by γ-secretase generates AICD, which contains the strongly conserved YENPTY-motif. The YENPTY-sequence is a consensus motif for the binding of adaptor proteins that possess a phosphotyrosine-binding domain (PTB) present in several APP adaptor proteins, such as X11, Fe65, ShcC, Numb, Dab and JIP families [see review in [
15]]. We have previously reported that RNAi knock-down of X11α, ShcC and Fe65 in H4 human neuroglioma cells lower Aβ levels [
16,
17].
Dab (encoded by gene
DAB), the PTB-containing APP adaptor protein, can bind to and interact with the YENPTY-motif of APP [
18,
19]. Dab has been reported to function as an adaptor molecule in signal transduction process [
20,
21]. Numb (encoded by gene
NUMB) is known to interact via its PTB domain with APP [
22,
23]. A recent study [
24] also suggest that high levels of Notch, another substrate of γ-secretase, can reduce levels of Numb and Numblike.
To date, the effects of reduced expression of Dab and Numb on APP processing and Aβ production, the key components of AD neuropathogenesis, have not been assessed. For this purpose, we established RNAi knock-down of Dab and Numb in H4 human neuroglioma cells overexpressing full-length (FL)-APP (H4-FL-APP cells) and C-99 (H4-APP-C99 cells), and evaluated the effects of RNAi-mediated knock-down of Dab and Numb on APP processing and Aβ levels.
Discussion
Aβ accumulation resulting from imbalance between Aβ generation and clearance in brain is a foundation of AD neuropathogenesis [[
1], see review in [
26]]. Aβ is produced via serial proteolysis of APP by two proteases, β-secretase and γ-secretase [
5‐
8]. Several APP adaptor proteins [see review in [
15]] have previously been shown to affect APP processing and Aβ production following overexpression [
27‐
35]. Recent studies showed that RNAi-mediated knock-down of X11α, ShcC, Fe65 and ARH can also affect APP processing and Aβ levels [
16,
17,
36]. However, effects of RNAi knock-down of other APP adaptor proteins, including Dab and Numb, on APP processing and Aβ production have not been previously reported. Here, we show for the first time that RNAi knock-down of Dab and Numb significantly affects APP processing and Aβ levels.
RNAi knock-down of Dab and Numb caused increases in levels of APP-C99 and APP-C83 in absence of alterations in FL-APP levels, and RNAi for Dab and Numb decreased secreted Aβ levels in H4-FL-APP cells. As described in prior section, β-secretase and α-secretase cleaves FL-APP to produce APP-C99 and APP-C83 respectively, then γ-secretase cleaves APP-C99 and APP-C83 to generate Aβ and p3 respectively, and AICD. Therefore, the observed increases in APP-C99 and APP-C83 levels following Dab or Numb siRNA treatment could be due either to increases in the activities of β-secretase and/or α-secretase, or to decreases in γ-secretase cleavage of APP-C99 and C83. To distinguish these two possibilities, we repeated RNAi knock-down of Dab and Numb experiment in H4 cells over-expressing APP-C99 (H4-APP-C99 cells), the β-secretase cleavage product of APP, and found that Dab or Numb siRNA treatment still increased protein levels of APP-C99 and APP-C83 and decreased levels of secreted Aβ. These data indicate that the observed changes in APP processing and Aβ levels following RNAi knock-down of Dab and Numb are independent of β-secretase-, and most likely are due to inhibition of γ-secretase-mediated cleavage of APP.
One possible explanation for our current results is that Dab and Numb, APP adaptor proteins, are essential for γ-secretase cleavage of APP-C99 to generate Aβ, perhaps via facilitating APP trafficking to the sites where γ-secretase can cleave it. Reductions in levels of Dab and Numb by RNAi knock-down for their encoded genes, DAB and NUMB, will cause a "break" in APP trafficking to the sites where γ-secretase is located, thereby leading to the accumulation of APP-CTFs (APP-C99 and APP-C83), the substrates of γ-secretase, and reduction in levels of Aβ, the products of γ-secretase. In the future studies, we will assess effects of over-expression and reduction of Dab and Numb on APP trafficking to further test this hypothesis.
To date, many other substrates have been reported to undergo γ-secretase cleavage in addition to APP: Notch, ErbB-4, E-cadherin, the LDL receptor-related protein (LRP), CD44 and nectin-1-α [
37‐
43]. All of these substrates are type I membrane proteins residing at or near cell surface, which undergo ectodomain shedding prior to γ-secretase-like cleavage and release ICDs (
i ntra
c ellular
d omains) following proteolysis. In our future research, we will determine whether or not RNAi knock-down of Dab and Numb can affect processing of these other γ-secretase substrates. It is also important to determine the potential interactions between Notch signaling with APP adaptor proteins including Numb, Dab, X11α, ShcC and Fe65 in our established cellular model.
Taking together, our findings demonstrate, for the first time, that RNAi-mediated knock-down of Dab and Numb decrease secretion of Aβ, conceivably via inhibiting γ-secretase cleavage of APP. These data, together with those of previous studies, imply that blocking interaction of APP with either Dab or Numb, and perhaps other APP adaptor proteins (e.g., X11α, ShcC and Fe65), could be a novel therapeutical strategy for treating and/or preventing AD by lowering Aβ accumulation.
Increasing evidence suggest a role of Notch signaling pathway in neurodegeneration of adult vertebrate nervous system and in AD neuropathogenesis [
44,
45]. A recent study showed that overexpressions of APP, APLP1 and APLP2 induce Notch gain-of-function phenotypes in
Drosophila, suggesting a cross-talk between APP family and Notch [
46]. Numb and Dab have been suggested to be mediators of such APP and Notch interactions [
46]. Moreover, it has been suggested that pharmacological processing of APP in AD treatment may cause alterations in Notch phenotype through Dab and Numb, thereby contributing to the side effects of the treatment or even to increasing risk of AD [
46,
47]. Therefore, Dab and Numb could be important targets in the development of therapeutical strategies for AD. Studies to further identify the roles of Dab and Numb in AD neuropathogenesis, including assessment of effects of Dab and Numb on APP processing, Aβ accumulation, synaptic function and apoptosis, are warranted in the future.
A recent study [
24] showed that at low levels of Notch signaling, Numb and Numblike can negatively regulate Notch, however, high levels Notch can reduce protein levels of Numb and Numblike. These findings suggest that a reciprocal negative regulation between Notch and Numb/Numblike. In the present study, we have illustrated for the first time that reductions in APP levels can lead to reductions in Numb levels, and reductions in Numb levels can lead to reductions in Aβ levels. Collectively, these findings suggest that APP (and Aβ) and Numb can also have a reciprocal negative regulation, leading to reductions in Aβ levels.
In conclusion, RNAi-mediated knock-down of APP adaptor proteins, Dab and Numb, attenuate γ-secretase-mediated cleavage of APP, leading to decreased Aβ levels. Our findings, together with those of previous studies, suggest that pharmaceutical modulation of APP adaptor proteins, might potentially serve as a novel therapeutic approach to treating and preventing AD. The notion of attenuating γ-secretase cleavage of APP via APP adaptor proteins, such as X11α, ShcC, Fe65, Dab and Numb, is particularly attractive, because interactions with these APP adaptor proteins may bypass unwanted effects of γ-secretase inhibition owing to impaired proteolysis of other γ-secretase substrates. More studies are needed to further investigate these findings and to assess feasibility of such a therapeutic strategy aimed at lowering Aβ levels.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ZX carried out RNAi interference (RNAi) studies, Western blot analysis, experimental design, data analysis and wrote the manuscript. YD carried out RNAi studies and Western blot analysis. UM carried out cell culture and Western blot studies. WX carried out ELISA measurement of Aβ levels. RT carried out the data analysis and drafted the manuscript. All authors read and approved the final manuscript.