© 2016 - IOS Press and the authors. All rights reserved.The majority of mutations causing familial Alzheimer's disease (fAD) have been found in the gene PRESENILIN1 (PSEN1) with additional mutations in the related gene PRESENILIN2 (PSEN2). The best characterized function of PRESENILIN (PSEN) proteins is in γ-secretase enzyme activity. One substrate of γ-secretase is encoded by the gene AMYLOID BETA A4 PRECURSOR PROTEIN (AβPP/APP) that is a fAD mutation locus. AβPP is the source of the amyloid-β (Aβ) peptide enriched in the brains of people with fAD or the more common, late onset, sporadic form of AD, sAD. These observations have resulted in a focus on γ-secretase activity and Aβ as we attempt to understand the molecular basis of AD pathology. In this paper we briefly review some of the history of research on γ-secretase in AD. We then discuss the main ideas regarding the role of γ-secretase and the PSEN genes in this disease. We examine the significance of the "fAD mutation reading frame preservation rule" that applies to PSEN1 and PSEN2 (and AβPP) and look at alternative roles for AβPP and Aβ in fAD. We present a case for an alternative interpretation of published data on the role of γ-secretase activity and fAD-associated mutations in AD pathology. Evidence supports a "PSEN holoprotein multimer hypothesis" where PSEN fAD mutations generate mutant PSEN holoproteins that multimerize with wild type holoprotein and dominantly interfere with an AD-critical function(s) such as autophagy or secretion of Aβ. Holoprotein multimerization may be required for the endoproteolysis that activates PSENs' γ-secretase activity.