Response to Shilling et al. (10.1074/jbc.M111.300491)

We demonstrated previously that presenilins (PS) function as endoplasmic reticulum (ER) Ca²⁺ leak channels and are directly involved in controlling steady-state ER Ca²⁺ levels, independently of γ-secretase function. We also discovered that many familial Alzheimer disease (FAD) mutations in PS result in loss of Ca²⁺ leak function and elevated ER Ca²⁺ levels. Experimental support for these claims is presented in a series of five original publications our groups have published since 2006, using a wide variety of experimental systems and multiple methods. Our hypothesis has been directly challenged in a recent paper by Shilling et al. (1), who categorized our observations as a case of data misinterpretation and experimental artifact. We agree that our hypothesis remains controversial, but we respectfully disagree with 1) the interpretation of their data and 2) how our observations and conclusions are represented in their publication. Key points of discussion follow.

Shilling et al. (1) did not test PS double knock-out (DKO) mouse embryonic fibroblasts (MEFs) and MEFs harboring PS1-M146V mutations for an augmented bradykinin response, a well documented and consistent observation (2, 3). Verification of enhanced Ca²⁺ response to bradykinin is an essential control for these experiments, as these cells are subject to genetic drift.

In our studies and in the studies by others, application of ionomycin to PS DKO and PS1-M146V fibroblasts resulted in a cytosolic Ca²⁺ increase of enhanced amplitude (2, 3). This is in contrast to the data presented by Shilling et al. (1), where the main difference observed in ionomycin experiments with PS DKO and PS1-M146V was due to delayed cytosolic Ca²⁺ extrusion.

We directly measured ER Ca²⁺ levels in PS DKO and PS FAD mutant cells using Mag-Fura-2 (3) and D1ER (4) indicators and demonstrated increased steady-state ER Ca²⁺ levels. We did not rely exclusively on measurements of the ionomycin-released Ca²⁺ pool size, as implied by Shilling et al. (1).

Shilling et al. (1) did not demonstrate increased intracellular Ca²⁺ response to glutamate or caffeine in PS1-M146V neuronal culture, another important control experiment (4, 5). Moreover, using heterogeneous cortical cultures is problematic when assessing global Ca²⁺ leak function, which we demonstrated to be highly variable between different neuronal subtypes (4).

Shilling et al. (1) found that ryanodine receptor expression is down-regulated in PS1-M146V cortical cultures, which is contrary to the consistent observation of ryanodine receptor up-regulation in PS1 FAD neurons by multiple laboratories (4, 5).

In conclusion, we do not agree that conclusive evidence is presented by Shilling et al. (1) to refute the hypothesis that presenilins act as passive ER Ca²⁺ leak channels.