. Modulation of gamma-secretase specificity using small molecule allosteric inhibitors. Proc Natl Acad Sci U S A. 2009 Dec 1;106(48):20228-33. PubMed.


Please login to recommend the paper.


  1. This is an interesting paper. The lead structure, hydroxycoumarin, is rapidly accessible and thus quite well known; SciFinder lists 560 close analogues covering 193 publications. The compound class is a frequent hitter in high throughput screening assays. A fraction of these hits may be due to its intrinsic UV absorption and metal chelation properties. Several close analogues inhibit urease (WO 2008033466), tyrosinase (Casañola-Martín et al., 2008), HIV-1 integrase (Saíz-Urra et al., 2007) and other enzymes.

    Many of these hydroxycoumarins strongly complex iron(II/III) or copper(II). Some metal complexes are cytotoxic: Cerium(III) and neodymium(III) complexes act as scavengers of X/XO-derived superoxide radical (Medicinal Chemistry 2006, 463-470). The lipophilicity of these compounds is within druglikeness (CS-1 clog P = 4.1), however the topological polar surface area is close to the limits of CNS applications (CS-1 tPSA = 93). This, in combination with the iron complexation, will complicate proof of concept studies in animals.

    The monosubstituted compound CS-4 strongly resembles known rodenticides (Whittle, Alan John; Swanborough, Joseph John; Parry, David Rees; Sunley, Raymond Leo. (Syngenta Limited, UK). Brit. UK Pat. Appl. 2003, 38 pp. GB 2388595), ruling out rodent studies altogether. Fortunately, it was found to be inactive in the GSI assay!

    The authors may have good reason to publish only these inferior hits and not more druglike other findings. We will know more when we see their patent applications.

    Until then, the merit of the publication resides with the biochemistry, enzymology, and selectivity of APP cleavage over Notch. On this count, the observed selectivity of 25 of Abeta secretion inhibition over Notch cleavage is close to proclaimed safety limits.

    The mapping of the active site by different photoreactive derivatives of L-685,458 is a smart approach (although the legend for figure 2e makes it hard to understand).

    I can’t, however, find the control experiments necessary to rule out FRET-derived artifacts. The hydroxycoumarins are fluorescent on their own and may interfere with photoactivation of the benzophenones (Rahman et al., 1992). This interaction should be distance-dependent, thus different benzophenone positioning may interfere with such a process. The photocrosslink efficiency may be wavelength-dependent. The authors applied the crosslink protocol from reference 3, where the samples were irradiated on ice for 45 minutes using a Rayonet RPR-200 photochemical reactor with PRP-3500A lamps. This protocol does not list a filter or specify a particular wavelength.

    In summary, the results do indicate an interesting mode of action. Additional experiments are required to confirm it.


    . Atom- and bond-based 2D TOMOCOMD-CARDD approach and ligand-based virtual screening for the drug discovery of new tyrosinase inhibitors. J Biomol Screen. 2008 Dec;13(10):1014-24. PubMed.

    . Quantitative structure-activity relationship studies of HIV-1 integrase inhibition. 1. GETAWAY descriptors. Eur J Med Chem. 2007 Jan;42(1):64-70. PubMed.

    . Interaction mode of dicumarol and its derivatives with human serum albumin, alpha 1-acid glycoprotein and asialo alpha 1-acid glycoprotein. J Pharmacobiodyn. 1992 Jan;15(1):7-16. PubMed.

Make a Comment

To make a comment you must login or register.

This paper appears in the following:


  1. Top AD Trends of 2009
  2. Yueming Li and Lennart Mucke Win MetLife Foundation Award