The news article states that "γ-secretase appears primarily in endosomes," but the results of Area-Gomez et al. (Area-Gomez et al., 2009) do not support this. Most of the γ-secretase activity (at least in neurons) appears to be in the mitochondrial associated membranes (aka "MAM"). In a second paper (Schon and Area-Gomez, 2010), they describe why this has not previously been observed.