The substrate-targeting activity detailed in the Nature report was revealed in studies that extended prior work (Eriksen et al., 2003) showing that certain non-steroidal anti-inflammatory drugs (NSAIDs) could selectively lower levels of Aβ42—the form of Aβ believed to be most neurotoxic—and that they do so independent of their well-known effects on inflammation (Weggen et al., 2001; ARF related news story; ARF related conference story). Intrigue grew when researchers led by Sascha Weggen, then at the University of California, San Diego, reported that those NSAIDs lowered Aβ42 by modulating γ-secretase activity (Weggen et al., 2003), putting them into a class of drugs known as γ-secretase modulators (GSMs). Those findings had AD researchers wondering whether the Aβ-lowering NSAIDs could be safely harnessed to fight AD.

To get a better grip on how GSMs influence Aβ production, lead author Kukar and colleagues set out to identify their molecular targets. Collaborating with Abdul Fauq at Mayo’s chemical synthesis core and Boris Schmidt at Darmstadt Technical University, Germany, the researchers made biotinylated photoprobes from derivatives of two GSMs—one that raises Aβ42 levels—fenofibrate—and one that lowers Aβ42—tarenflurbil (inactive R-enantiomer of the NSAID flurbiprofen) which has been licensed to Myriad Genetics and is undergoing Phase 3 testing as a treatment for mild AD (see ARF related news story and drugs in clinical trials).

The Mayo team expected the GSM photoprobes to label presenilin-1 or some other core component of the γ-secretase complex. “We actually spent a lot of time looking for this interaction and came up with a lot of negative results,” Kukar told the Alzforum, recalling early attempts to detect presenilin-1 labeling in an APP-overexpressing human neuroglioma cell line. Blaming the negative results on limited sensitivity of their GSM photoprobes, Kukar and colleagues sought help from Harvard γ-secretase experts Michael Wolfe and then-postdoc Pat Fraering, who now heads an AD lab at the Swiss Federal Institute of Technology in Lausanne. After Wolfe and Fraering failed to detect labeling of any of the core proteins in a highly purified prep of active γ-secretase, “we went back to the drawing board and thought, what's left? That's when we thought about APP,” Kukar said.

Applying their photoprobes to a batch of recombinant APP, the researchers were delighted to see that APP lit up over a range of concentrations at which the GSMs typically modulate Aβ42 production. To establish the specificity of the GSM-APP interaction, the researchers threw Aβ42-lowering and -raising GSMs into the mix and found that each competed for labeling of APP, whereas a non-GSM NSAID did not. They showed that their photoprobes labeled APP more efficiently than Notch, another γ-secretase substrate.

Using a series of truncation mutants, the researchers mapped the GSM interaction domain to the Aβ region of APP and, furthermore, to an eight-amino-acid stretch of Aβ critical for aggregation. Co-authors Dominic Walsh and colleagues at University College Dublin, Ireland, then applied two Aβ42-raising GSMs and one Aβ42-lowering GSM to cultured CHO cells overexpressing APP and found that all three GSMs decreased Aβ oligomer formation.

“The idea that certain compounds will not only affect the toxic ratio of Aβ peptides generated, but could also lower toxicity by interfering with such oligomer formation, is very attractive indeed,” wrote Bart De Strooper of K.U. Leuven, Belgium, in an e-mail to ARF.

To further establish that binding to APP is, in fact, what enables some GSMs to decrease Aβ42 production, co-authors Edward Koo and colleagues at the University of California, San Diego, replaced a portion of the GSM binding site in APP with the analogous region of human Notch. Cleavage of this chimeric construct was not significantly affected by an Aβ42-lowering or -raising GSM but was inhibited by treatment with a γ-secretase inhibitor, the researchers found.

“This is an astonishing result,” writes Thomas Kodadek, University of Texas Southwestern Medical Center, Dallas, in a commentary accompanying the paper. “There are precious few examples of substrate-targeted enzyme inhibitors in the literature, all of which are peptides rather than drug-like small molecules. If Kukar and colleagues’ findings prove to be relevant to other proteases and their substrates, it would suggest that a completely new set of drug targets exists for the treatment of a variety of diseases.”

As a next step toward optimizing GSMs that show promise for AD drug development, Kukar and colleagues are taking a closer look not only at how the compounds reduce levels of Aβ42 but also at whether they raise levels of Aβ38 and shorter Aβ species. While studies have shown that increased levels of Aβ40 can prevent Aβ42 aggregation in animal models, the jury is still out as to whether shorter Aβ species confer similar protection. Kukar told ARF that his group has unpublished in vitro results suggesting that Aβ30, Aβ36 and Aβ37 might mimic Aβ40 in its ability to prevent Aβ42 aggregation.—Esther Landhuis.

References:
Kukar TL, Ladd TB, Bann MA, Fraering PC, Narlawar R, Maharvi GM, Healy B, Chapman R, Welzel AT, Price RW, Moore B, Rangachari V, Cusack B, Eriksen J, Jansen-West K, Verbeeck C, Yager D, Eckman C, Ye W, Sagi S, Cottrell BA, Torpey J, Rosenberry TL, Fauq A, Wolfe MS, Schmidt B, Walsh DM, Koo EH, Golde TE. Substrate-targeting gamma-secretase modulators. Nature. 2008 June 12;453:925-930. Abstract

Kodadek T. Molecular cloaking devices. Nature. 2008 June 12;453:861-2. Abstract