Unlike other PI 5-phosphatase family members, INPP5A recognizes only soluble inositol 1,4,5-trisphosphate [Ins(1,4,5)P3] and inositol 1,3,4,5-tetrakisphosphate [Ins(1,3,4,5)P4] as substrates. INPP5A is ubiquitously expressed including in the hippocampus and prefrontal cortex, the brain regions highly affected in AD, and is abundantly detected in cerebellum (Liu et al., 2020). INPP5A negatively controls the mobilization of intracellular calcium by decreasing Ins(1,4,5)P3 levels (De Smedt et al., 1997). DNA methylation of the INPP5A gene is increased in association with aging in neurons (Gasparoni et al., 2018). Meta-analysis of blood-based DNA methylation has shown that the methylation of cg12507869 located in the INPP5A gene had a significant negative correlation with phonemic verbal fluency and was associated with logical memory and vocabulary (Marioni et al., 2018). Blood-based DNA methylation of cg12507869 in the INPP5A could be thus considered as a potential biomarker for aging and cognitive functions.

SHIP1 is a hematopoietic-specific PI 5-phosphatase activated downstream of a multitude of receptors for growth factors, cytokines, antigens, immunoglobulin and toll-like receptor agonists. Once activated and correctly localized, SHIP1 generally acts as a negative regulator of signaling processes in hematopoietic cells, for example on the B cell receptor activation signaling pathway (Ramos et al., 2019). SHIP1 is detected in the brain, primarily in microglia reflecting its myeloid origin. Genome-wide association studies (GWAS) have identified the risk variant rs35349669 in INPP5D, the gene encoding human SHIP1 for late-onset AD (Lambert et al., 2013). INPP5D mRNA is significantly upregulated in human AD brains and in transgenic mouse brains with knock-in mutations of APP^{NL−G−F/NL−G−F} (Castillo et al., 2017). INPP5D mRNA expression in peripheral leucocytes is elevated in early AD but is decreased with cognitive decline (Yoshino et al., 2017). Further long-time follow-up of the participants would be necessary to decipher the correlation between the level of INPP5D mRNA and cognitive decline. Since SHIP1 converts PI(3,4,5)P3 to PI(3,4)P2, the amounts of these PIs in the blood leucocytes may also be altered and needs to be further investigated (as discussed in section PI Metabolism and Autophagic-Endosomal-Lysosomal Abnormalities). Taken together, the level of INPP5D mRNA in leucocytes could be an interesting target to develop a blood-based biomarker in the early stages of AD.

SHIP2, encoded by INPPL1, is ubiquitously expressed including in the brain (Muraille et al., 1999). By using PI(3,4,5)P3 as substrate, SHIP2 controls PI(3,4)P2 content, a major SHIP2 product (Ghosh et al., 2018). SHIP2 can also dephosphorylate PI(4,5)P2, another albeit less potent substrate (Elong Edimo et al., 2016). PI(3,4)P2 is scarce under normal conditions but increases through signaling following PI 3-kinase activation. This lipid plays critical roles as a second messenger in cell migration, polarity, feedback control of PI(3,4,5)P3 generation, and basal mTORC1 activity (Ramos et al., 2019). SHIP2 is directly implicated in several human diseases: mutations in INPPL1 cause opsismodysplasia, a rare autosomal recessive disease characterized by delayed bone maturation (Fradet and Fitzgerald, 2017). SHIP2 is also upregulated in some cancer cells, particularly in aggressive human breast cancer cells (Ghosh et al., 2018). SHIP2 negatively regulates insulin/IGF-I actions and is implicated in type 2 diabetes and metabolic syndrome (Marion et al., 2002). Recent network-based approach has unraveled that SHIP2 is also linked to AD and cognitive decline: upregulation of INPPL1 transcript in the brain significantly correlates with cognitive decline in human AD patients (Mostafavi et al., 2018). The same study also reported that SHIP2 immunoreactivity was detected in astrocytes and neurons in the post-mortem human brain tissues of AD patients and that lentivirus-mediated down regulation of SHIP2 in cultured astrocytes significantly reduced Aß production (Mostafavi et al., 2018). Other independent studies have reported SHIP2 functions as a mediator of amyloid toxicity via tau hyperphosphorylation (Kam et al., 2016) and actin-cytoskeleton reorganization (Lee et al., 2019). Kam et al. reported that the interaction between Aß and the FcγRIIb immuno-receptor leads to a translocation of SHIP2 to the plasma membrane to form a protein complex in which SHIP2 dephosphorylates PI(3,4,5)P3 into PI(3,4)P2. Increased amounts of PI(3,4)P2 lead to decreased inhibitory phosphorylation of GSK3ß at Ser9 via endoplasmic reticulum (ER) stress in cultured neurons (Kam et al., 2016). Consequently, tau phosphorylation by GSK3ß is increased by Aß via FcγRIIb-SHIP2 complex (Kam et al., 2016). SHIP2 inhibitors are thus under active scrutiny as a novel therapeutic target for AD. Actually, SHIP2 inhibitors represent new treatments for several diseases: SHIP2 inhibition has been reported to partially rescue memory deficits in transgenic mouse models of diabetes and AD (Soeda et al., 2010; Kam et al., 2016) and to prevent metastasis in breast cancer cells (Ghosh et al., 2018). Since both SHIP1 and SHIP2 play critical roles in antagonizing microglial proliferation and phagocytosis, the use of both SHIP1 and SHIP2 inhibitors has been proposed in AD to enhance basal microglial homeostatic functions for therapeutic purposes (Pedicone et al., 2020). Although SHIP2 could be a potential biomarker and a valuable therapeutic target for AD, it remains largely elusive whether SHIP2 undergoes a significant alteration in subcellular localization and post-translational modifications during the progression of the disease. SHIP2 has more than 20 putative phosphorylation sites and its phosphatase activity and substrate recognition are, at least partially, regulated by phosphorylation, protein-protein interaction and subcellular localization (Elong Edimo et al., 2011). Given that SHIP2 is translocated to plasma membranes upon Aß-FcγRIIb interaction (Kam et al., 2016), subcellular localization of SHIP2 should be significantly altered in AD brains. Since FcγRIIb activation leads to tyrosine phosphorylation of SHIP2 (Muraille et al., 1999), the post-translational modifications of SHIP2 could be altered in the affected areas of AD brains. It remains to be carefully determined in post-mortem brain tissues of AD patients whether there are changes in SHIP2 subcellular localizations, post-translational modifications and the impact of SHIP2 upregulation in AD on PI amounts, particularly PI(3,4,5)P3 and PI(3,4)P2.

SYNJ1 and SYNJ2 are both highly conserved and their genetic variants are associated with cognitive abilities in a cohort with a mean age of 70 (Lopez et al., 2012). SYNJ1 is a brain-enriched presynaptic phosphatase involved in synaptic vesicle recycling, clathrin-coated vesicle uncoating at synapse (Cremona et al., 1999) and autophagosomal maturation within presynaptic terminals (Vanhauwaert et al., 2017). SYNJ1, whose gene is located in chromosome 21, is linked to endolysosomal abnormalities in Down syndrome (Cossec et al., 2012). Several mutations in SYNJ1 gene are associated with early-onset Parkinsonism (Tran et al., 2020). Some of the polymorphisms in SYNJ1 are also linked with age of onset in familial AD, late-onset AD and Down syndrome with AD (Miranda et al., 2018). SYNJ1 is expressed in neurons and is implicated in Aß toxicity (Berman et al., 2008), synaptic toxicity (McIntire et al., 2012) and Aß clearance (Zhu et al., 2013). The mRNA level of SYNJ1 is significantly upregulated in post-mortem AD brains in association with APOE genotype (Zhu et al., 2015; Ando et al., 2020). SYNJ1 protein undergoes a significant solubility change and is co-enriched with PHF-tau in the sarkosyl-insoluble fraction (Ando et al., 2020). SYNJ1 immunoreactivity is detected in actin-positive Hirano bodies, some NFTs and plaque-associated dystrophic neurites in post-mortem human AD brains (Ando et al., 2020). Such aberrant alteration of mRNA levels, protein localization, and protein solubility of SYNJ1 could be applied to establish a valid biomarker for AD. While SYNJ1 is brain specific, its paralog SYNJ2 is ubiquitously expressed, but is also abundantly expressed in the synapse. In the temporal cortex from patients with depressive disorder, SYNJ2 transcript expression is significantly decreased (Aston et al., 2005). Furthermore, differential methylation in the gene of SYNJ2 has been also reported in association with aging in neuronal cells (Gasparoni et al., 2018).

The implication of the other members of PI 5-phosphatase family in AD remains largely unknown. Given that AD is associated with autophagic-endosomal-lysosomal dysfunction (Nixon et al., 2008), we speculate that INPP5E and OCRL, highly expressed in the brain and critical in autophagosome-lysosome fusion (De Leo et al., 2016; Hasegawa et al., 2016), might be involved in dysregulation of autophagy in AD brains.

Consistent with alterations of some PI 5-phosphatases observed in AD brains, there are substantial findings suggesting that PIs undergo dysregulation during the disease progression in AD brains (Stokes and Hawthorne, 1987) and in the AD blood plasma (Mapstone et al., 2014). In the AD prefrontal cortex where both amyloid and tau pathologies are abundant, the amounts of PI 3-phosphate (PI3P) and PI(4,5)P2 are significantly decreased (Morel et al., 2013). Deficiency of PIs in AD brains may be linked to autophagic-endosomal-lysosomal abnormalities observed in neurons of the AD patients even at an early stage (Nixon et al., 2008). Considering that PIs regulate membrane dynamics, we hypothesize that autophagic-endosomal-lysosomal abnormalities could be a potential target for developing AD biomarkers. For instance, endosomal morphology alteration has been observed in iPSC-neurons derived from AD fibroblasts (Israel et al., 2012) and AD blood monocytes (Corlier et al., 2015). Whereas the precise mechanisms underlying endosomal abnormalities remain to be determined, such endosomal alterations in peripheral cells could be considered as a novel potential approach to develop AD biomarkers.