Efforts to derail the production of Aβ42 by targeting the enzyme that produces it, γ-secretase, have been complicated by the secretase’s complexity and promiscuity for substrates. In a paper published in Cell Chemical Biology on January 5, researchers addressed a relatively simple question: What part of γ-secretase binds the inhibitor avagacestat? Using a modified version of the inhibitor as a photoactive probe, researchers led by Douglas Johnson of Pfizer in Cambridge, Massachusetts, and Yueming Li at Memorial Sloan Kettering Cancer Center in New York found that avagacestat latches onto presenilin1 (PS1) next to its endoproteolytic region—where the enzyme must undergo self-proteolysis before it can cleave substrates. The findings paint the drug as a pan inhibitor of the secretase, which spares no substrate cleavage, contrary to previous suggestions. The modus operandi dovetails with disappointing side effects that brought avagacestat’s clinical development to an end.

The γ-secretase complex consists of four subunits—PS1, nicastrin, Aph-1, and Pen2—that all together traverse the cellular membrane 20 times. Researchers recently resolved the structure of the complex at atomic resolution, revealing a horseshoe-shaped structure lying flat in the cell membrane, with the two catalytic aspartate residues juxtaposed in transmembrane domains six and seven of PS1 (see May 2015 newsAug 2015 news). Despite this high-resolution picture of the enzyme, it is still unclear how it contorts to process different substrates, and how different inhibitors engage to affect its activity. Therapeutic attempts to block the secretase—first with Eli Lilly’s semagacestat and then with Bristol Myers Squibb's avagacestat—failed due to side effects and worsening cognition, likely due to the inhibitors blocking Notch1 processing (see Nov 2012 news). Although BMS had initially touted avagacestat as a “Notch-sparing” inhibitor, subsequent studies by Johnson, Li, and others challenged that notion (see Crump et al., 2012). 

Catalytic Clincher?

The avagacestat-based photoprobe (yellow) interacts with hydrophobic residues on PS1 (blue), near an intracellular loop (orange) adjacent the endoproteolytic site. These interactions could mess with the alignment of catalytic aspartate residues (green). [Image courtesy of Gertsik et al., Cell Chemical Biology 2017.]

With the aim of pinpointing exactly where avagacestat engages the enzyme, first author Natalya Gertsik of Memorial Sloan Kettering and colleagues developed a collection of cleavable photoprobes. The probes serve a dual purpose—they bind to the enzyme like the parent inhibitor does, and also facilitate purification of labeled proteins because they bind covalently after activation. Mass spectrometry analysis of these purified targets then allow researchers to home in on the exact residue to which the probes are attached.

After incubating the best probe with membranes from cells overexpressing all four subunits of γ-secretase, the researchers discovered that the inhibitor touched down on leucine 282 in PS1, which is smack dab in the middle of a region essential for catalytic activity. As a requisite for activation PS1 cleaves itself into two fragments—an N- and a C-terminal. L282 resides within a loop adjacent to this endoproteolytic cleavage site on the N-terminal side, in a region that becomes essential for catalytic function following endoproteolysis. The researchers ran a molecular dynamics simulation based on this binding site. It suggested that following endoproteolytic cleavage, the inhibitor latches onto PS1-NTF, and then prevents substrate binding and/or blocks catalysis. In support of this mechanism, the simulation revealed that the inhibitor interacts with glycine-382, which is near the catalytic aspartate-385. The findings suggest that avagacestat likely inhibits γ-secretase regardless of the substrate, Johnson told Alzforum.

Johnson said that he and colleagues are designing similar photoprobes to pinpoint binding sites of γ-secretase modulators currently in development. These modulators, including Pfizer’s PF-06648671, are designed to specifically affect amyloidogenic processing of APP, while sparing cleavage of essential proteins such as Notch1 (see Dec 2016 conference news).—Jessica Shugart 


  1. High-resolution structures of γ-secretase have revealed the architecture of one of the most intriguing proteases. Moreover, co-structures with the allosteric inhibitor DAPT and a short co-purifying peptide uncovered the pocket and interaction mode for DAPT and possibly the substrate binding site in presenilin. But most importantly, the elucidation of the γ-secretase structure opened a new era for experimentalists working on the protease complex family.

    In this paper, Li, Johnson, and colleagues present a beautiful exercise in which elegant experimental approaches, assisted with theoretical structural predictions, provide complementary insights into how avagacestat, an inhibitor tested in clinical trials a few years ago, binds and functionally silences γ-secretase. The authors tested alternative chemistries to label and identify the residues in γ-secretase that are involved in the binding of avagacestat. The experimental information guided theoretical docking studies of the inhibitor in the high-resolution structure and provided a model for its interaction with the catalytic presenilin subunit. Intriguingly, the binding sites of the allosteric DAPT and avagacestat inhibitors partially overlap.

    Available models for different presenilin-inhibitor interactions offer the opportunity to investigate, in depth, the molecular bases of γ-secretase inhibition. The outcome may have significance for the development of “selective” probes that target the processing of particular substrates, or specific types of proteases complexes. Of relevance, similar exercises using allosteric modulators may reveal the bases of allosterism in γ-secretases and add motion and functional significance to the current static views of the protease.

Make a Comment

To make a comment you must login or register.


Therapeutics Citations

  1. Avagacestat
  2. Semagacestat
  3. PF-06648671

News Citations

  1. Gamma Secretase: Intramembrane Liaisons Revealed
  2. γ-Secretase Revealed in Atomic Glory
  3. Déjà Vu? AD Patients Again Look Worse on γ-Secretase Inhibitor
  4. Emerging Alzheimer’s Therapies Test the Waters at CTAD

Paper Citations

  1. . BMS-708,163 targets presenilin and lacks notch-sparing activity. Biochemistry. 2012 Sep 18;51(37):7209-11. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Mapping the Binding Site of BMS-708163 on γ-Secretase with Cleavable Photoprobes. Cell Chem Biol. 2017 Jan 19;24(1):3-8. Epub 2017 Jan 5 PubMed.