Abstract Presenilins 1 and 2 (PS1 and PS2) are predominantly known as the catalytic subunits of the γ-secretase complex which generates the amyloid-β (Aβ) peptide, the major constituent of the senile plaques found in the brain of Alzheimer’s disease (AD) patients. Apart from their role in γ-secretase activity, a growing number of cellular functions have been recently attributed to PSs. They are involved in synaptic transmission, endo-lysosomal function and calcium homeostasis. PSs were also found to be enriched in mitochondria-associated membranes (MAMs) where mitochondria and endoplasmic reticulum (ER) interact. PS2 was more specifically reported to regulate calcium shuttling between the ER and mitochondria by controlling the formation of functional MAMs through its interaction with the Mitofusin2 protein. We have previously demonstrated that the absence of PS2 (PS2KO) alters mitochondrial morphology and function. Indeed, a PS2KO cell line showed reduced mitochondrial respiration along with disrupted mitochondrial cristae and increased glycolysis. This phenotype is restored by the stable re-expression of human PS2. Still, all these results were obtained in immortalized Mouse Embryonic Fibroblasts (MEF) and one bottom-line question is to know whether these observations hold true for the Central Nervous System (CNS) cells, and in particular neurons and astrocytes. To that end, we carried out primary PS1KO, PS2KO and PS1/PS2KO (PSdKO) neuronal and astrocyte cultures. All the conditions were obtained in the same litter by crossing PS2 heterozygous and PS1 floxed (PS2 +/− ; PS1 flox/flox ) animals. Indeed, contrary to PS2KO mice, PS1KO are not viable and therefore require the use of the Cre-LoxP system to achieve gene deletion in vitro . Strikingly, we did not observe any mitochondrial phenotype in PS1KO, PS2KO or PSdKO primary cultures. Mitochondrial respiration and membrane potential were similar in all models, as were the glycolytic flux and NAD + /NADH ratio. We further investigated the discrepancies between these results and the ones previously reported in the MEF PS2KO cell line by analyzing PS2KO primary fibroblasts. No mitochondrial dysfunction was observed in this model, in line with observations in PS2KO primary neurons and astrocytes. These results indicate that the mitochondrial phenotype observed in immortalized PS2-deficient cell lines cannot be extrapolated to primary neurons, astrocytes and even to primary fibroblasts. The PS-dependent mitochondrial phenotype reported so far might be the consequence of a cell immortalization process and, therefore, should be critically reconsidered regarding its relevance to AD.

Presenilin-deficient neurons and astrocytes display normal mitochondrial phenotypes