Cleavage of APP within the Aβ domain by α-secretases is of great physiological interest, not only because it precludes the formation of Aβ, but also because it generates a soluble N-terminal fragment, sAPPα, that exhibits neuroprotective properties [26, 27]. Moreover, shedding of the ectodomain is a prerequisite for cleavage of the intracellular domain by γ-secretases; a process that liberates a C-terminal fragment with transcriptional activity [28–30]. Although the up-regulation of APP shedding by activation of PKC-dependent signaling pathways has been well-documented , the mechanism mediating this response is still obscure.
The present study was undertaken to determine if inhibitors of dynamin function would affect ectodomain shedding of APP. We first showed that APP internalization is dependent on the activity of dynamin, a large molecular weight GTPase that mediates both clathrin-dependent endocytosis, and internalization of caveolae, by promoting the separation of endocytic vesicles from the plasma membrane [22, 31]. In confirmation of a recent study , we found that overexpression of a dominant negative dynamin mutant protein in HEK cells increased surface expression of full-length APP, and release of sAPPα. Thus, although cleavage of APP by α-secretases occurs largely in an intracellular compartment in many cell types (reviewed in ), our results suggest that inhibition of dynamin function, by preventing internalization of APP, increases its dwell-time on the cell surface, and prolongs its interaction with α-secretases at the plasma membrane. Similar elevations in APP secretion are induced by mutations of the APP cytoplasmic domain that inhibit internalization [23–25]. Consistent with the observed increase in α-secretase mediated cleavage, expression of the dynamin mutant increased cellular levels of C83, the C-terminal stub remaining after α-secretase-mediated cleavage of APP (Fig. 2B and 2C).
The increase in sAPPα release in HEK cells overexpressing dyn I K44A was associated with a reduction in the release of Aβ1–40, (Fig. 4), a result in keeping with reports that Aβ is generated in an endocytic compartment [24, 25, 32–34]. Our results are also in agreement with a study by Ehehalt et al.  who found that overexpression of a dyn K44A mutant protein reduced formation of the Aβ peptide in mouse neuroblastoma N2a cells. In contrast, Chyung and Selkoe reported that Aβ generation was increased in HeLa cells following induction of dyn K44A expression . The increased Aβ formation observed in the latter study occurred in the absence of any alteration in the synthesis or maturation of APP, and suggested that, in HeLa cells, processing of APP by β- and γ-secretases occurs at the plasma membrane . Indeed, an active γ-secretase complex was subsequently isolated from the plasma membrane of HeLa cells . As a possible explanation for the reduction in Aβ observed by Ehehalt et al.  in cells overexpressing dyn K44A, Chyung and Selkoe pointed out that those workers measured formation of radiolabeled Aβ in cells labeled for 1 hour with [35S]methionine, and surmised that the mutant dynamin reduced generation of labeled Aβ by increasing the amount of unlabeled APP at the cell surface, and diluting the concentration of labeled precursor available for cleavage by β- and γ-secretases. In support of the notion that Aβ can be generated at the cell surface, Ehehalt et al  showed that when endocytosis was blocked by transfection with dyn K44A, the reduction in Aβ could be partially rescued by antibody cross-linking of APP and the β-secretase, β-site APP-cleaving enzyme (BACE). The decrease in total Aβ1–40 generation in HEK cells overexpressing dyn I K44A described in the present report might simply reflect reductions in the precursor pool due to increased cleavage of APP by α-secretase. This result is consistent with earlier studies showing that upregulation of α-secretase cleavage by PKC activation in HEK cells , or via mutations of the APP cytoplasmic domain in stably transfected HEK or Chinese hamster ovary (CHO) cells [23–25], is associated with decreased Aβ formation. The discrepancies among these studies might be due at least in part to cell-specific differences in the compartments where APP comes into contact with α- and β/γ-secretases, or in the relative capacities of the different secretases to cleave APP within a specific compartment.
Modulation of endocytosis might represent a mechanism for physiological regulation of APP processing by PKC-dependent signaling pathways. PKC phosphorylates dynamin, thereby activating its GTPase activity , and inhibiting its association with phospholipids in vitro . In nerve terminals, dynamin must be dephosphorylated in order to promote retrieval of synaptic vesicles following exocytosis, and re-phosphorylation is required for the next round of endocytosis that follows a second stimulus . Persistent phosphorylation of dynamin might therefore be predicted to interfere with endocytosis. Contrary to expectation, the PKC activator PMA did not affect the rate of APP internalization, as determined by reversible biotinylation in the presence of the α-secretase inhibitor TAPI-1 (Fig. 5). Thus, although PKC activation can modulate endocytosis of a variety of transmembrane proteins, either positively, in the case of β1 integrin, GABA receptors, and the dopamine transporter [38–41], or negatively, as is the case with μ-opioid receptors , we could not find evidence for a modulatory effect of phorbol esters on APP internalization. Others have shown that PKC activation increases APP ectodomain shedding in PC12 cells by stimulating trafficking of APP through the secretory pathway . In contrast, surface expression of APP was reduced in CHO cells that were surface biotinylated following treatment with PMA and TAPI, suggesting that in these cells, PMA did not increase trafficking of APP to the plasma membrane, but possibly stimulated α-secretase-mediated cleavage within an intracellular compartment that was partially resistant to TAPI . Interestingly, the motor neuron-derived trophic factor neuregulin-1, a ligand for the tyrosine kinase receptors ErbB3 and ErbB4, was found to increase the rate of internalization and degradation of APP in cultured myotubes, while decreasing release of the ectodomain . This report lends credence to the hypothesis that modulation of APP internalization may represent a physiological mechanism for regulation of sAPPα release.