Abstract. The subcellular localization of presenilin-1 (PS1) and presenilin-2 (PS2), two proteins that, when mutated, cause familial Alzheimer's disease (AD), is controversial. We have discovered that mitochondria-associated membranes (MAM) -a specialized subcompartment of the endoplasmic reticulum (ER) involved in lipid metabolism and calcium homeostasis that physically connects ER to mitochondria -is the predominant subcellular location for PS1 and PS2, and for γ-secretase activity. We hypothesize that presenilins play a role in maintaining MAM function, and that not only altered amyloid-β levels and hyperphosphorylated tau, but also many other features of AD (e.g., altered phospholipid and cholesterol metabolism, aberrant calcium homeostasis, and abnormal mitochondrial dynamics) result from compromised MAM function. The localization of presenilins and γ-secretase in MAM may help reconcile disparate ideas regarding the pathogenesis of AD, under a unifying hypothesis that could explain many features of both sporadic and familial AD, thereby taking AD research in a new and fruitful direction.Keywords: Alzheimer's disease, calcium, cholesterol, endoplasmic reticulum (ER), mitochondria, mitochondria-associated membranes (MAM), phospholipids
ALZHEIMER'S DISEASEAlzheimer's disease (AD), the most common late onset neurodegenerative dementing disorder, is characterized by progressive neuronal loss, especially in the hippocampus and cortex [1]. The two main histopathological hallmarks of AD are the accumulation of extracellular neuritic plaques, containing amyloid-β (Aβ), and of neurofibrillary tangles, consisting mainly of hyperphosphorylated forms of the microtubule-associated protein tau [1]. Most AD patients are sporadic (SAD), but three genes have been associated with the familial form (FAD): amyloid-β protein precursor (AβPP), presenilin-1 (PS1), and presenilin-2 (PS2). Clinically, FAD is similar to SAD but has earlier onset. Presenilins are components of the γ-secretase complex (also containing APH1, nicastrin, and PEN2) that, together with β-secretase, processes AβPP to produce Aβ [1]. In the mainstream view, both SAD and FAD arise when AβPP is processed to Aβ, which accumulates in extracellular plaques. Aβ is toxic to cells and the resulting stress promotes tau hyperphosphorylation, leading to the tangles. The overall process has been called the "amyloid cascade" hypothesis [2,3]. This hypothesis reconciles findings from different approaches to the disease and has served as the basis for many key experiments in vivo and in vitro. However, certain questions that are central to understanding the pathogenesis of AD and the processing of Aβ remain unsolved.The first question concerns features of AD that are not obviously linked to plaque or tangle formation. While plaques and tangles are hallmarks of the disease, other apparently unrelated laboratory abnormali-