In collaboration with Eric Siemers and Robert Dean at Eli Lilly, Indianapolis, Portelius and colleagues report in the March 29 Alzheimer’s Research & Therapy that Aβ1-16 is elevated in CSF of patients enrolled in Lilly’s Phase 2 clinical trial of LY450139 (semagacestat), a γ-secretase inhibitor. Siemers said he found the data compelling. “When Kaj asked for our spinal fluid samples, we thought it would be interesting to look at, although we weren’t convinced that we would be able to see an effect,” he told ARF. “But there was this very nice, dose-dependent, and statistically significant increase.”

The findings highlight these β/α fragments as potentially valuable biomarkers for following γ-secretase inhibition in vivo. Portelius and colleagues looked at CSF of 35 patients with mild to moderate AD who were randomized to placebo, 100 mg, or 140 mg of LY450139. The patients had a lumbar puncture at baseline and again 14 weeks after treatment began. Between sampling, CSF levels of the β/α fragments—Aβ1-14, 1-15, and 1-16—rose 57, 21, and 30 percent, respectively, in the 100 mg group and 74, 35, and 67 percent in the 140 mg group. Aβ42 levels were unchanged.

The field of clinical AD research is abuzz about biomarkers of late, though sorting out which biomarker best serves what purpose is an ongoing process (see ARF related news story). Evidence suggests CSF Aβ42 reduction is a sign of impending dementia (see Hansson et al., 2006), but that marker does not turn out to be a good reporter of γ-secretase inhibition. “The problem with Aβ42 is that the timing of the lumbar puncture has to be precise because its turnover is so rapid,” Siemers told ARF. “You get a reduction [in Aβ], but as soon as you remove the inhibitor it goes back to baseline again. Siemers speculates that Aβ1-16 might be a better marker for γ-secretase inhibition because its turnover is much slower; hence, the timing of the lumbar puncture would not have to be as precise.

In contrast, while Aβ1-16 might provide a good readout of γ-secretase inhibition, it seems a poor marker for impending dementia according to Zetterberg. This Swedish group found that the peptide is elevated about 15-30 percent in both familial and sporadic AD samples but that the elevation is not diagnostically useful. It is also not clear why Aβ1-16 is elevated in AD, given that it may be a sign of reduced γ-secretase activity. Zetterberg speculated that product inhibition due to the elevated levels of Aβ42 might suppress γ-secretase, leaving the β-secretase-derived C99 ripe for α-secretase to pick off Aβ1-16. “More research is needed on what these isoforms can tell us about the pathogenesis of AD,” he told ARF.

As for testing Aβ1-16 in their ongoing Phase 3 trial of semagacestat, Siemers said that was not built into the study, though he thought if he had to do the development all over again, the peptide might be a great biomarker. But he suggested this touches on a larger question Alzforum readers might debate. “When you get into Phase 3, your real mission is to show you’ve had an effect on cognitive function. How important is it to people to demonstrate biomarker changes, especially if they were already shown in Phase 2?” he asked. We welcome thoughts or suggestions.—Tom Fagan.

Reference:
Portelius E, Dean RA, Gustavsson MK, Andreasson U, Zetterberg H, Siemers E, Blennow K. A novel Abeta isoform pattern in CSF reflects gamma-secretase inhibition in Alzheimer's disease. Alzheimers Res Ther. 2010 Mar 29;2(2):7. Abstract