Implications of presenilin 1 mutations in Alzheimer's disease

Vesicular trafficking and secretion of matrix metalloproteinases-2, -9 and tissue inhibitor of metalloproteinases-1 in neuronal cells

Matrix metalloproteinases (MMPs) are endopeptidases that cleave matrix, soluble and membrane-bound proteins and are regulated by their endogenous inhibitors the tissue inhibitors of MMPs (TIMPs). Nothing is known about MMP/TIMP trafficking and secretion in neuronal cells. We focussed our attention on the gelatinases MMP-2 and MMP-9, and their inhibitor TIMP-1. MMPs and TIMP-1 fused to GFP were expressed in N2a neuroblastoma and primary neuronal cells to study trafficking and secretion using real time video-microscopy, imaging, electron microscopy and biochemical approaches. We show that MMPs and TIMP-1 are secreted in 160-200 nm vesicles in a Golgi-dependent pathway. These vesicles distribute along microtubules and microfilaments, co-localise differentially with the molecular motors kinesin and myosin Va and undergo both anterograde and retrograde trafficking. MMP-9 retrograde transport involves the dynein/dynactin molecular motor. In hippocampal neurons, MMP-2 and MMP-9 vesicles are preferentially distributed in the somato-dendritic compartment and are found in dendritic spines. Non-transfected hippocampal neurons also demonstrate vesicular secretion of MMP-2 in both its pro- and active forms and gelatinolytic activity localised within dendritic spines. Our results show differential trafficking of MMP and TIMP-1-containing vesicles in neuronal cells and suggest that these vesicles could play a role in neuronal and synaptic plasticity.

Intracellular site of gamma-secretase cleavage for Abeta42 generation in neuro 2a cells harbouring a presenilin 1 mutation

Previously, we reported that mutations in presenilin 1 (PS1) increased the intracellular levels of amyloid beta-protein (Abeta)42. However, it is still not known at which cellular site or how PS1 mutations exert their effect of enhancing Abeta42-gamma-secretase cleavage. In this study, to clarify the molecular mechanisms underlying this enhancement of Abeta42-gamma-secretase cleavage, we focused on determining the intracellular site of the cleavage. To address this issue, we used APP-C100 encoding the C-terminal beta-amyloid precursor protein (APP) fragment truncated at the N terminus of Abeta (C100); C100 requires only gamma-secretase cleavage to yield Abeta. Mutated PS1 (M146L)-induced Neuro 2a cells showed enhanced Abeta1-42 generation from transiently expressed C100 as well as from full-length APP, whereas the generation of Abeta1-40 was not increased. The intracellular generation of Abeta1-42 from transiently expressed C100 in both mutated PS1-induced and wild-type Neuro 2a cells was inhibited by brefeldin A. Moreover, the generation of Abeta1-42 and Abeta1-40 from a C100 mutant containing a di-lysine endoplasmic reticulum retention signal was greatly decreased, indicating that the major intracellular site of gamma-secretase cleavage is not the endoplasmic reticulum. The intracellular generation of Abeta1-42/40 from C100 was not influenced by monensin treatment, and the level of Abeta1-42/40 generated from C100 carrying a sorting signal for the trans-Golgi network was higher than that generated from wild-type C100. These results using PS1-mutation-harbouring and wild-type Neuro 2a cells suggest that Abeta42/40-gamma-secretase cleavages occur in the Golgi compartment and the trans-Golgi network, and that the PS1 mutation does not alter the intracelluar site of Abeta42-gamma-secretase cleavage in the normal APP proteolytic processing pathway.