Parker reported on experiments with a potent γ-secretase inhibitor that also inhibits Notch cleavage (LY-411,575), and a less potent stereoisomer to serve as a control (133.11). The work was conducted in large part by Gwen Wong, who is now at ALS-TDF. The researchers fed LY-411,575 for 15 days to wild-type mice and CRND8 APP transgenics, then they measured plasma and brain Aβ levels, studied effects on the immune system by flow cytometry, and examined tissues histologically.

As expected, LY-411,575 decreased plasma and brain Aβ levels. However, at the concentrations where it did so, it also increased the size of certain immature thymocyte and B lymphocyte populations while reducing the number of mature T cells in the thymus and mature B cells in spleen and blood. Specifically, LY-411,575 blocked the physiological transition of certain double-negative (CD4-CD8-) populations of immature lymphocytes (e.g., the Cd44+/25+ set) to their single-positive, differentiated state (e.g., the CD44+/25- set). What’s more, the normal cellular architecture of the mice’s intestinal villi looked abnormal, and the mucosa showed goblet cell hyperplasia. At the higher inhibitor doses tested, the mice lost weight and died, probably due to the intestinal side effects, Parker said. The histology of the brain, liver, kidney, lung, heart, adrenal gland, bone marrow and stomach appeared normal. The T cell effects predictably resulted from inhibition of Notch cleavage, as Notch is known to function in thymocyte development, but the B cell and intestinal effects were unexpected, Parker added.

These results still leave open the possibility of inhibiting γ-secretase with other inhibitors that distinguish between APP and Notch. Alas, recent reports suggest that’s easier said than done, at least with classic strategies of fitting competitive inhibitors into the enzyme’s catalytic pocket. Researchers at Merck, Sharpe and Dohm in Harlow, United Kingdom, reported this summer that a range of such inhibitors drawn from six different compound classes were all unable to distinguish clearly between APP and Notch in vitro (Lewis et al., 2003; see also Shearman section in ARF related news story). Others have reported that γ-secretase inhibitors cause developmental defects consistent with Notch inhibition in fruit flies (Micchelli et al., 2002) and zebra fish (Geling et al., 2002). For a review, see Tsai et al., 2002).

At the same time, however, a flurry of meeting presentations on NSAIDs is pointing to a new way of achieving this goal. These studies explore in more detail the mechanism by which certain NSAIDs inhibit γ-secretase in non-competitive ways. At this point, results vary depending on the assay, dose, and compounds used, and some data contradict each other. Overall, however, the new hope is that an existing NSAID can be found—or, more likely perhaps, a new one designed—that tweaks γ-secretase allosterically, i.e., outside of the active site. It would have to do so in such a way that APP cleavage shifts away from generating Aβ42 and toward Aβ38 (which by all accounts so far is safe to have in increased amounts), while leaving alone proteolysis of Notch and other targets, such as ErbB-4. For more detail, view abstracts 295.2, 295.7, 295.21, 295.22, 336.8, 336.9, 523.3, 549.4, 729.1, and 876.14 at the SfN/ScholarOne website.—Gabrielle Strobel.