PS1-NTF PS1-CTF(20 kD)

Scheme 3.5.7. Presenilin architecture.

gel failed. This is partially due to the strong binding affinity of the III-31-C 12 core for the target protein complex and partially due to the deep and narrow binding site, which requires strong denaturing conditions to break up the binding interactions. Co-precipitation of the inhibited y-secretase with its substrates C83 and C99 gave rise to speculation about additional binding sites, where the substrate is recognized prior to transfer to the active site. These speculations are in accordance with the observed promiscuous nature of the cleavage, because they assign the specific recognition to other complex domains.

In comparison with ^-secretase, for which detailed structural information and the enzyme kinetics are available, very little is known about the more complex y-secretase. The diversity of selective, non-peptidic y-secretase inhibitors is, therefore, something of a surprise. Elan's development of DAPT (16, difluorophenylacetyl-aminopropionylaminophenylacetic acid tert-butyl ester; Scheme 3.5.8) originated from an N-dichlorophenylalanine ester 15, which had an IC50 > 30 mM in cellular screening. After several cycles of refinement and hundreds of compounds, activity peaked for the dipeptide mimetic DAPT 16 featuring the non-natural amino acid phenylglycine, which is crucial for activity (human embryonic kidney cells IC50 = 20 nM). Another key contribution stems from the difluorophenylacetic acid, resulting in a very steep structure-activity relationship; broad variation of the di-fluorophenyl moiety confirmed the demand for small electron-withdrawing sub-stituents. Branched esters had similar activity and, despite speculations about the labile nature of tertiary butyl ester, which might be cleaved at the low pH of the gut, DAPT is not a prodrug.

Subcutaneous application of 100 mg kg-1 to mice resulted in a 50% reduction of Ab brain levels after 3 h. The reduction was 40% 3 h after an oral dose of 100 mg kg-1. Further development of the compound by Eli Lilly included stereo-selective placement of the hydroxyl group and locking of the spatial arrangement of two phenyl rings in a seven-membered lactam to yield 17 (LY 411575, HEK IC50 < 1 nM).

After several failures with peptidic structures such as 2, all of which suffered from toxicity problems during development, Bristol Myers Squibb and Merck [32] published details of almost 1000 derivatives of 4-chloro-N-(2,5-difluorophenyl)-benzenesulfonamides. Five-hundred of these were reported to be very good in-

0 0

Post a comment