Calcium homeostasis and amyloid production in Familial Alzheimer's Disease models based on mutant presenilin-2 and amyloid precursor protein: focus on the store-operated calcium channel subunit Orai2

Alzheimer's disease (AD) is the most common form of dementia among elderly population. More than twenty-five years ago, the so-called amyloid hypothesis was formulated based on one of the major histopathological hallmarks of AD, the senile plaques, which are formed by aggregates of beta-amyloid peptides (Aβ). This hypothesis was prompted by the discovery of three genes that, whereas mutated, are associated with the familial forms of the disease (FAD). One of these genes encodes for the amyloid precursor protein (APP), a single-pass type I transmembrane protein that, upon proteolytically processing by secretases (α, β, γ), leads to the formation of different intra- and extracellular peptides including Aβ, of which the most abundant are Aβ40 and Aβ42. This latter peptide forms toxic aggregates, initially as small oligomers and fibrils and, subsequently, as large extracellular amyloid plaques. The last and key cleavage leading to Aβ production is done by the γ-secretase. This enzyme is composed of four proteins, among which presenilin 1/2 (PS1/2) forms its catalytic core. Autosomic dominant mutations in APP, PSEN1 or PSEN2 cause accelerated Aβ deposition due to an increased Aβ42/Aβ40 ratio. The PSEN1 and PSEN2 genes, when mutated, are responsible for the majority of FAD cases, which however represent only a minor fraction (<5%) of AD forms. Many groups reported that PSs are capable of perturbing cellular Ca2+ homeostasis, and, particularly, our group demonstrated that PS2 mutants reduce endoplasmic-reticulum (ER) Ca2+ release and store-operated Ca2+ entry (SOCE), while increasing ER-mitochondria tethering. All these pleiotropic effects are independent of γ-secretase activity. Of note, modulation of SOCE has been linked to Aβ production/release, yet with contrasting results. The mechanism of SOCE, whose main molecular components are STIM1 and Orai1, has been known since many years. Recently, both PS2 and Orai2, an Orai1 homologue, were suggested to have a function in ER Ca2+ leak. Moreover, Orai2 was demonstrated to negatively modulate SOCE. Taken together, this body of information offers an interesting background to study the role of Orai2 in FAD models with a focus on SOCE modulation and Aβ accumulation at the brain level. What is more, this study also aims at verifying whether Orai2 is a suitable therapeutic target for AD. Taking advantage of the PS2-based AD mouse models available in our laboratory, namely the homozygous single transgenic (TG) line expressing the FAD-linked mutant PS2-N141I (line PS2.30H) and the homozygous double transgenic (2TG) line expressing the PS2-N141I together with the Swedish double mutant APP-K670M/N671L (line B6.152H), I investigated early markers of disease progression and the expression pattern of Orai2 and the other SOCE components in nervous cells. Subsequently, I tested the crosstalk between SOCE and Aβ42 accumulation in neuroglioma cells expressing the APP Swedish mutation. In these cells, upon modulation of Orai2 expression, I performed cytosolic Ca2+ measurements - by the aequorin approach - to evaluate ER Ca2+ release and SOCE, and ELISA measurements to estimate Aβ levels in the conditioned media. By these integrated approaches, I concluded that downregulation of Orai2 potentiates SOCE and reduces the Aβ42/Aβ40 ratio, supporting Orai2 as a possible target in AD therapy. For the purpose of evaluating the effects of Orai2 downregulation on Ca2+ dynamics and Aβ production in AD mouse models, I designed microRNAs (miRNAs) suitable to target mouse Orai2, selectively in neurons or astrocytes, upon insertion into lentiviral particles (LPVs) to reach high infection efficiency in both cultured cells and in vivo. One of these miRNAs has been validated in a cell line and in mouse astrocytes for efficient Orai2 downregulation, SOCE potentiation and reduction of Aβ levels. Studies are now in progress at in vivo level by infection with LVPs to allow either astrocytic or neuronal Orai2-miRNA expression in the somato-sensory cortex of control and AD mice. As an alternative strategy to study how Orai2 expression modulates Ca2+ dynamics at the brain level and Aβ accumulation in AD mouse models, I considered the chronic reduction of Orai2 level obtained by crossing homozygous 2TG (B6.152H) with Orai2-/- mice. The F1 hemizygous (He)-2TG-Orai2+/- mice were thus compared to F1 hemizygous 2TG (He2TG) mice, obtained by crossing B6.152H with WT mice. Preliminary data suggest that the two mouse groups have similar disease onset, as estimated by plaque density and gliosis, however He2TG-Orai2+/- are not impaired in hippocampal slow oscillation connectivity with respect to He-2TG mice. Considered that investigation into Orai2 functions in health and disease is in its infancy, here I also provided a first characterization of Ca2+ handling in primary cultures of cortical neurons or astrocytes from newborn Orai2-/- mice. Surprisingly, a chronic halved Orai2 level reduces, rather potentiating, SOCE in astrocytes while it increases depolarization-induced Ca2+ entry in neurons, highlighting a novel role for neuronal Orai2. In summary, I collected evidence on the feasibility of addressing SOCE modulation by Orai2 in AD mouse models. Orai2 downregulation can allow to rescue astrocytic Ca2+ defects and, possibly, also dampening Aβ accumulation.