17 Jul 2000

Almost 12% (152) of the 1,307 abstracts presented at the 7th International Conference on Alzheimer’s Disease and related disorders made reference to presenilin (PS). For an overview of PS biology check out the webcast of Bruce Yankner’s plenary session lecture at: www.alzheimer2000.org/news/webcast.htm. This news report will limit itself to a summary of data presented which suggest a role for PS as γ-secretase (Abstracts: 311, 312, 313, 524, 603, 614, 635, 820, 821, 828, 837, 1267, 1270, 1274).

1. Cells lacking both PS-1 and PS-2 do not secrete detectable levels of Aβ40 or Aβ42 and show a marked increase in APP C-terminal fragments (CTFs).
2. Mutation of Asp257 or Asp385 of PS-1 prevents PS endoproteolysis, causes accumulation of APP CTFs and drastically reduces Aβ secretion. Mutation of homologous aspartate residues in PS-2 have a similar effect.
3. APP CTFs (the direct substrates for γ-secretase) can be co-immunoprecipitated with both PS-1 and PS-2. These complexes accumulate when PS is inactivated (either pharmacologically or by mutation) and are readily detectable in the principal intracellular sites of Aβ generation.
4. PS-1 co-elutes on size exclusion with a partially purified γ-secretase activity (obtained from solubilized membranes) as part of a protein complex.
5. SDS-PAGE and subsequent silver staining of immunoaffinity purified γ-secretase activity (derived from solubilized membranes) identified hetrodimeric (i.e., NTF + CTF) PS and six other protein bands.
6. "Naked PS-1" does not possess γ-secretase activity, indicating that it requires a number of critical cofactors for enzymatic activity.
7. Potent γ-secretase inhibitors designed to function as transition state analogue inhibitors of aspartyl proteases (produced and characterized by two different groups) affinity label heterodimeric forms of PS-1 and PS-2.
8. A biotinylated analogue of the general aspartyl protease inhibitor, Pepstatin A, also labels both PS-1 and PS-2.
9. Small amounts of PS-1 are detected on the cell surface of neurons at tight junctions.
10. PS-1 shares a similar sequence motif around Asp 385 as that seen in a family of bacterial proteases.

Although, as Colin Masters rightly said, "The pendulum has now swung in favor of the hypothesis that PS is γ-secretase," there still remain unanswered questions about the role of PS in amyloidogenesis. For instance, Robert Doms (Abstract 313) showed that cortical neurons from PS-1 KO mice display a dramatic decrease in secreted Aβ, but little or no change in intracellular (formic acid extracted) Aβ. Does this lack of change in intracellular Aβ in PS-1 KO mice mean that PS-2 is normally responsible for production of this pool of Aβ, or is there another γ-secretase activity responsible for the production of the insoluble intracellular pool of Aβ? This question will be readily addressed by comparing intracelluar pools of Aβ in PS-1 Vs PS-2 KOs and in PS-1 and PS-2 double KOs. Interestingly, Katja Fletcher (Abstract 1274) indicated that certain of her γ-secretase inhibitors could alter secreted and intracellular Aβ differentially. Given the recent recognition of the importance of intracellular Aβ any anti-secretase-based therapy should inhibit production of intracellular and secreted pools of Aβ. Note: Bristol-Myers Squibb is currently in phase I trials with a potent γ-secretase inhibitor which has no cross-reactivity with any other known protease and seems to be well-tolerated in humans. Importantly this γ-secretase inhibitor is more effective (by three orders of magnitude) at inhibiting cleavage of APP than inhibiting cleavage of Notch.—Dominic Walsh

(Editor’s note: Dr. Walsh is an author on Abstract 312, reported on in this news summary.)