23 November 2004. Since the discovery that one peptide, amyloid-β, is the major constituent of amyloid plaques, there has been a tremendous focus on the N-terminal end of amyloid-β precursor protein (AβPP). But as emphasized at the 34th annual meeting of the Society for Neuroscience in San Diego last month, the C-terminal end of AβPP has its claim to fame, too.
Work from the labs of Dale Bredesen and Eddie Koo has previously shown that a single caspase-mediated cleavage releases the last 31 amino acids from the C-terminus of AβPP. This C31 peptide causes apoptosis in cultured cells and can be found in brain samples from AD patients but not in samples from control brains (see ARF related news story). But does it have a role in vivo?
Apparently, it does. Continuing the Californian collaboration, Veronica Galvan from the Bredesen lab at the Buck Institute for Age Research, Novato, and Brock Schroeder from Eddie Koo’s lab at University of California San Diego reported that in transgenic animal models of AD, lack of this cleavage site in the C-terminal of AβPP has profound effects.
To test how mutation of this cleavage site may affect the progression of AD-like pathology, Galvan substituted an alanine for asparagine at position 664 of the human AβPP gene in PDAPP mice. (Asparagine 664 is where the C31 cleavage occurs; see SfN abstract 488.4.) PDAPP mice express human AβPP harboring mutations that cause early-onset AD in humans (phenylalanine for a valine at position 617, asparagine for lysine at 670, and leucine for methionine at position 671). These mice develop Aβ plaques and suffer synaptic and neuronal losses, and develop memory deficits as they age. When Galvan examined mice expressing the alanine664 AβPP, she found that they produced about the same quantity of Aβ and about the same number of plaques as asparagine644 PDAPP animals. However, synaptic and hippocampal changes were a different story. Galvan found that hippocampal volume—normally reduced by about one-third in PDAPP transgenics—was maintained in mice devoid of C31. The loss of synaptic densities (up to 50 percent) that normally occurs as the PDAPP mice age was also prevented, as judged by immunoreactivity of the synaptic marker, synaptophysin. The results suggest that the neurotoxicity that accompanies accumulation of Aβ plaques in these mice can be prevented if the C31 cleavage is abolished.
Continuing with this theme, Schroeder showed that synapses are indeed protected in PDAPP mice that can’t produce C31 (see SfN abstract 488.5). In these animals, not only is loss of the synaptic protein synaptophysin attenuated, but also extracellular field potentials recorded from hippocampal neurons are maintained instead of reduced as they are in standard PDAPP transgenics (by about 50 percent).
All told, the physiological data suggest a major role for the C-terminal cleavage of APP in the pathology of AD. So might it contribute to the most worrisome syndrome of AD, cognitive impairment? To test this, Schroeder evaluated PDAPP mice in the Morris water maze, which measures spatial learning and memory. The animals lacking C31 failed to show the loss of spatial learning and memory that PDAPP mice normally exhibit as they grow older.
So how may the C-terminal cleavage of APP contribute to pathology in these animal models? Galvan suggested two possibilities: Cleavage at asp644 not only yields C31, which is toxic in vitro, but also lops off endocytic signals and binding sites for the transcriptional associated factors Fe65 and X11, leaving the AβPP intracellular domain (AICD) with no means to attract these partners. This C-terminal cleavage may, therefore, interrupt both endocytosis and AICD signal transduction to the nucleus. If this turns out to be true, then facilitating AICD signaling could be a novel therapeutic approach to treating AD.
So can we now forget about the head of AβPP, the Aβ? That is unlikely, given the amount of evidence linking it to neurotoxicity, and also recent observations showing that Aβ may bind to AβPP, induce its dimerization, and favor the C-terminal cleavage of the precursor (see Lu et al., 2003 and also related news), and that C31 alone is insufficient for toxicity—full-length AβPP is needed as a co-conspirator (see Lu et al., 2003). So while the observed lack of AD-like pathology in PDAPP-ala664 supports the involvement of both the C- and N-terminal ends of AβPP, perhaps the tail does wag the dog a little.—Tom Fagan.