09 Feb 2012

A paper in the February 8 Journal of Neuroscience reports that γ-secretase inhibitors do nothing to improve memory in AD mice when given subchronically, and can even worsen cognition in normal mice. By contrast, under the same conditions a γ-secretase modulator, which shifts where γ-secretase cuts APP but does not inhibit the enzyme, sharpened memory in transgenic mice. First author Yasuyuki Mitani at Astellas Pharma in Tsukuba, Japan, previously presented some of these data in a poster session at the 2011 AD/PD Conference held 9-13 March in Barcelona, Spain (see ARF related news story). The findings add to the growing disenchantment with γ-secretase inhibition as an AD treatment strategy. In 2010, Eli Lilly and Company halted their clinical trial of the γ-secretase inhibitor semagacestat when they found that patients on the drug fared worse than those on placebo (see ARF related news story and ARF news story).

To compare the drugs, the authors fed semagacestat, a Notch-sparing γ-secretase inhibitor called avagacestat, or the γ-secretase modulator GSM-2, to five-month-old Tg2576 mice overexpressing human APP. After one day of dosing, all three drugs improved mouse performance in a Y-maze test of learning and memory. However, after eight days, mice that ate either semagacestat or avagacestat did no better on the task than untreated mice, while rodents on GSM-2 improved. The researchers also tested three-month-old Tg2576 animals, which do not yet show cognitive deficits. In these animals, eight days of γ-secretase inhibitor treatment worsened their performance. Likewise, semagacestat and avagacestat impaired performance in wild-type mice, although that took a 10-fold higher concentration of the drugs. GSM-2 treatment had no effect on wild-type mice.

What explains the negative effect of γ-secretase inhibitors on cognition? Mitani and colleagues lay the blame on β-CTF, which is produced when β-secretase (BACE1) cuts APP. Normally, γ-secretase snips β-CTF to produce Aβ and an intracellular domain, but in the absence of γ-secretase activity, β-CTF accumulates. Levels of β-CTF increased in Tg2576 mice treated with inhibitors, and correlated with poor performance, the authors note. Immunostaining revealed β-CTF building up at presynaptic terminals in the mossy fibers of the hippocampus. By contrast, GSM-2 did not affect levels of β-CTF. Moreover, the negative effects of γ-secretase inhibitors seem to depend upon high levels of APP, since the inhibitor dose had to be dramatically raised to affect wild-type mice, which have normal levels of the protein. This further implicates β-CTF as the culprit, the authors suggest. Several other studies have pointed a finger at β-CTF as a toxic entity, with some showing that BACE1 inhibition improves cognition while γ-secretase inhibition does not (see, e.g., Neve et al., 1992; ARF related news story and ARF news story.—Madolyn Bowman Rogers