The function of APP remains unknown, but it has long been assumed that APP is a receptor. Indeed, the title of one of the original papers reporting the cloning of APP was “The precursor of Alzheimer’s disease amyloid A4 protein resembles a cell-surface receptor” (Kang et al., 1987). Interest in a ligand for APP has been fueled recently by emerging details of Notch signaling. Upon binding its cognate ligand, Delta-1, the Notch receptor is cleaved by ADAM 10, generating a membrane-bound C-terminal fragment (CTF). This CTF is subsequently cleaved by γ-secretase, releasing the Notch intracellular domain (NICD), which translocates to the nucleus where it binds to a highly conserved DNA-binding transcription factor called CSL (also known as RBP-Jκ, CBF1, Suppressor of Hairless, and Lag-1) and activates the transcription of target genes. Analogously, γ-secretase cleaves APP to release APP intracellular domain (AICD), which can bind Fe65 and translocate to the nucleus (  Read more

The function of APP remains unknown, but it has long been assumed that APP is a receptor. Indeed, the title of one of the original papers reporting the cloning of APP was “The precursor of Alzheimer’s disease amyloid A4 protein resembles a cell-surface receptor” (Kang et al., 1987). Interest in a ligand for APP has been fueled recently by emerging details of Notch signaling. Upon binding its cognate ligand, Delta-1, the Notch receptor is cleaved by ADAM 10, generating a membrane-bound C-terminal fragment (CTF). This CTF is subsequently cleaved by γ-secretase, releasing the Notch intracellular domain (NICD), which translocates to the nucleus where it binds to a highly conserved DNA-binding transcription factor called CSL (also known as RBP-Jκ, CBF1, Suppressor of Hairless, and Lag-1) and activates the transcription of target genes. Analogously, γ-secretase cleaves APP to release APP intracellular domain (AICD), which can bind Fe65 and translocate to the nucleus (Cao and Sudhof, 2001; Kimberly et al., 2001).

But does APP have a ligand equivalent to Delta-1? This is the question Ho and Sudhof address. Using rat membrane extracts as a source of ligand and the immobilized central domain of APP (CAPPD) encompassing residues 286-557 of human APP695 as an affinity reagent, they identified F-spondin as the major protein bound by APP. To confirm the specificity of this interaction, the authors transfected Cos cells with constructs encoding the Fc-region of Ig fused to various APP regions, and they confirmed that both full-length APP and CAPPD bound F-spondin. In reciprocal studies a series of F-spondin Ig fusion proteins were generated, and it was found that the 807 amino acid full-length F-spondin bound APP most avidly and that N- and C-terminal truncations of spondin decreased binding. Further, CAPPD and APLP1 and 2, the central domains of which are highly similar to APP, also bound to F-spondin, whereas they did not bind to the F-spondin-related protein, Mindin.

Most intriguingly, Ho and Sudhof report that cotransfection of HEK cells with F-spondin, APP, and BACE dramatically reduced the BACE-processing of APP compared to cells transfected with either APP and BACE or APP and BACE and an unrelated protein. The authors further demonstrate that the inhibitory effect of F-spondin was concentration-dependent and that F-spondin was capable of preventing AICD release. Under the conditions employed to investigate AICD release, α-secretase is predominantly responsible for ectodomain shedding, thus providing evidence that F-spondin can inhibit both α- and β-processing of APP.

While these data are of high quality and demonstrate a specific and avid interaction (most experiments were performed in the presence of one percent Triton X-100), the physiologic relevance of APP-F-spondin binding is unclear. As the authors acknowledge in their discussion, further experiments designed to address this issue will be required, and should include immunolocalization, crosslinking, and the use of genetically modified animals. Moreover, the data presented do not exclude the potential for other APP ligands. In this regard, it is noteworthy that rat brain membrane extract was the source from which F-spondin was identified, a source that would not facilitate detection of a freely soluble ligand. Similarly, the authors used the central domain of APP for their initial screen, and thus could have missed putative ligands directed to the structurally conserved N-terminus (Rossjohn et al., 1999).

Whether or not F-spondin is physiologically relevant or is the sole ligand for APP, the demonstration that F-spondin can inhibit ectodomain shedding of APP provides a novel therapeutic route for reducing Aβ production.

View all comments by Dominic Walsh