Early-onset FAD (familial Alzheimer's disease) is caused by mutations of PS1 (presenilin 1), PS2 (presenilin 2) and APP (amyloid precursor protein). Beyond the effects of PS1 mutations on proteolytic functions of the γ-secretase complex, mutant or deficient PS1 disrupts lysosomal function and Ca2+ homoeostasis, both of which are considered strong pathogenic factors in FAD. Loss of PS1 function compromises assembly and proton-pumping activity of the vacuolar-ATPase on lysosomes, leading to defective lysosomal acidification and marked impairment of autophagy. Additional dysregulation of cellular Ca2+ by mutant PS1 in FAD has been ascribed to altered ion channels in the endoplasmic reticulum; however, rich stores of Ca2+ in lysosomes are also abnormally released in PS1-deficient cells secondary to the lysosomal acidification defect. The resultant rise in cytosolic Ca2+ activates Ca2+-dependent enzymes, contributing substantially to calpain overactivation that is a final common pathway leading to neurofibrillary degeneration in all forms of AD (Alzheimer's disease). In the present review, we discuss the close inter-relationships among deficits of lysosomal function, autophagy and Ca2+ homoeostasis as a pathogenic process in PS1-related FAD and their relevance to sporadic AD.
Autosomal dominant mutations of PS1 (presenilin 1), PS2 (presenilin 2) and APP (amyloid precursor protein) cause EOAD (early-onset Alzheimer's disease). Although EOAD accounts for fewer than 5% of all AD cases, investigation of the three genes responsible has so far provided many of the available clues to suspected pathogenic mechanisms in AD. Mutations of PS1, which are responsible for the vast majority of EOAD cases, can accelerate disease onset to ages as early as the late 20s. In most cases of EOAD, the defining lesions of AD, i.e. NFTs (neurofibrillary tangles) and neuritic plaques, as well as characteristic autophagic–lysosomal pathology, resemble the features of later-onset sAD (sporadic AD), although these abnormalities are usually more severe. Notable clinical and neuropathological heterogeneity, however, is sometimes seen among families carrying one of the >120 known PS1 mutations [1–3].
PS1, a ubiquitous protein with nine transmembrane domains, exists as a 65 kDa holoprotein in the ER (endoplasmic reticulum). The molecule is cleaved by a furin-like ‘presenilinase’ to generate a two-chain form [4,5], which constitutes the catalytic subunit of the γ-secretase enzyme complex composed of four additional subunits. Delivered from the ER to diverse vesicular destinations in the cell, γ-secretase mediates the intramembranous cleavage of well over 25 different substrates, which are mainly type 1 membrane proteins [6,7], including most notably APP. The γ-secretase generates aneurotoxic Aβ (amyloid β-peptide) from a membrane-bound C-terminal fragment of APP generated by BACE-1 (β-site APP-cleaving enzyme 1) . The pathogenic effects of PS1 mutations in AD are commonly ascribed to a small increase in the generation of a toxic 42-amino-acid peptide (Aβ42) relative to a less toxic 40-amino-acid form (Aβ40) [7,9,10]. Recent findings, however, indicate that AD (Alzheimer's disease)-causing PS1 mutations often confer loss of function of the secretase, so that, even if the Aβ42/Aβ40 ratio is modestly higher, as is often, although not invariably, the case , the absolute levels of these peptides may be markedly lowered [12,13] and not always a higher Aβ42/Aβ40 ratio . Thus additional or alternative explanations for pathogenicity of PS1 mutations have been sought [14–16].
PS1 may contribute to AD pathogenesis through loss of its other functions  while serving as a component of γ-secretase, which may include cell adhesion, neurite outgrowth and synaptic plasticity [18,19] or, alternatively, by acting via its secretase-independent roles as a PS1 holoprotein, which include lysosomal acidification essential for autophagic protein degradation , Wnt signalling  and cellular Ca2+ regulation [22–24]. In the present brief review, we focus specifically on the multiple roles of the PS1 holoprotein and how loss of its function in the ER may link ER and lysosome pathogenic mechanisms in EOAD, which accelerate neurodegeneration.
Mutations or deletion of PS1 cause autophagy defects by disrupting v-ATPase assembly and lysosome acidification
Autophagy is a lysosomal degradative pathway for recycling diverse cellular constituents [25,26], particularly under conditions of metabolic stress (Figure 1). Essential for survival of neurons, autophagy is solely responsible for the cellular turnover of damaged or obsolete organelles and is vital for eliminating misfolded and aggregated proteins, which are poorly degraded by the ubiquitin–proteasome system. Autophagy has been reported to be altered in different neurological disorders, and is considered to be a pathogenic factor in several of these diseases, particularly AD and Parkinson's disease . In AD, disruption of autophagy results in a particularly profuse accumulation of autophagic vacuoles within grossly swollen dystrophic neurites of affected neurons  which stems from defects in the lysosomal clearance of autophagy substrates by lysosomes [29,30]. Growing evidence has linked lysosome system failure during autophagy to multiple pathological outcomes in AD, including accelerated amyloidogenesis, neuritic dystrophy, apoptosis and, possibly, tauopathy [31–34].
An overview of macroautophagy (autophagy)
Mutations of PS1 considerably exacerbate autophagy pathology in EOAD and mouse models of AD , and similar pathology can be induced in neurons and non-neuronal cells by deleting PS1 [20,36,37]. Exploring the basis for these effects, Lee et al.  demonstrated that the PS1 holoprotein serves as a chaperone in the ER for the v-ATPase (vacuolar ATPase) V0a1 subunit, a six-pass transmembrane protein constituent of the proton pump responsible for full acidification of lysosomes. Binding to the PS1 holoprotein, as this subunit is being translated in the ER, stabilizes the correct folded state and facilitates its glycosylation, believed to be necessary for transit from the ER [20,38–40]. In cells lacking PS1, the V0a1 subunit of the v-ATPase is poorly glycosylated and unstable, which prevents adequate delivery to lysosomes and assembly of the multisubunit v-ATPase, leading to greatly diminished levels and activity of the proton pump  and resulting in failure to fully acidify lysosomes, which is a requirement for lysosomal protease activation and autophagy [20,30]. Similar lysosomal acidification defects have now been observed in PS1-knockout blastocysts, PS1-knockout and PS1/2-knockout MEFs (mouse embryonic fibroblasts), PS1-FAD (familial Alzheimer's disease) human fibroblast lines, PS1-knockout neurons and PS1/APP mice [30,41]. The roles of the V0a1 subunit and of PS1 in lysosome acidification have been confirmed in studies of PS1-ablated cells or mutant PS1 mouse models and other systems [30,38–40,42], although Coen et al.  has proposed that neither of these components is involved in lysosomal pH regulation. The importance of this PS1 chaperone function in AD pathogenesis is underscored by the ability of pharmacological inhibitors of v-ATPase to induce identical AD-related autophagy dysfunction in normal cells and by the complete reversal of this phenotype in PS1-deficient cells by pharmacologically normalizing lysosomal acidification . The connection of PS1 loss of function to Wnt signalling was demonstrated recently in which the loss of PS1 led to an increase in nuclear β-catenin and phosphorylated LRP6 (low-density lipoprotein receptor-related protein 6) due to increased Wnt activity, all of which was attributed to defective acidification of the multivesicular body .
Defective lysosome acidification: an emerging factor in pathogenesis of degenerative diseases
Interestingly, disruptions of normal v-ATPase activity appear to be important for the progression of other diseases. Mutations within the a3 subunit of the v-ATPase cause misfolding and ER retention of this subunit, leading to osteopetrosis, a bone disease that may be accompanied by neurodegeneration . Mice lacking the a3 subunit also have impaired acidification of lysosomes in macrophages . In a rare familial form of Parkinson's disease, the loss of function of ATPase13A2 (Park9), believed to be involved in lysosome acidification, leads to impaired autophagy and neurodegeneration . Mutations in LRRK2 (leucine-rich repeat kinase 2) increase the activity of the protein and lead to late-onset familial Parkinson's disease. Among other pathological effects that have been attributed to this protein kinase, increase in LRRK2 activity leads to pathologically increased autophagy induction and defective lysosomal acidification . Reduced v-ATPase activity and abnormally elevated lysosomal pH due to a deficiency of VMA21, an integral membrane protein required for the assembly of the v-ATPaseV0 complex in the ER, leads to X-linked myopathy, a form of muscle fibre degeneration .
Although lysosomal pH is predominantly maintained through the activity of the v-ATPase, other ion channels localized to the lysosomal membrane are suspected to be important in maintaining or regulating pH shifts during lysosomal proteolysis and may be particularly important when the intraluminal lysosomal environment is disturbed in disease states. The Cl− channel CLC7 has been implicated in fine-tuning lysosomal pH in conjunction with v-ATPase , and other evidence has suggested an involvement of Ca2+ channels, TRPML1 (transient receptor potential mucolipin 1) and TPC2 (two-pore channel 2), in lysosomal pH regulation [50–52]. Lysosomal storage disorders exhibit elevated lysosomal pH as a part of the progression of the disease. Elevated lysosomal pH has been detected in several forms of NCL (neuronal ceroid lipofuscinosis), including juvenile NCL (Batten disease) and infantile NCL, leading to altered lysosomal enzyme activity .
Ca2+ dysregulation in EOAD: the important contribution of defective lysosomal acidification
PS1 holoprotein has also been implicated in mediating the γ-secretase-independent effects of AD-causing PS1 mutations (or PS1 deletion) in dysregulating Ca2+ homoeostasis at several different subcellular levels, including the ER [54–57], the MAMs (mitochondria-associated membranes)  and lysosomes [43,59] (Figure 2). Ca2+ homoeostasis in the ER is normally maintained by regulating Ca2+ efflux through RyRs (ryanodine receptors) and IP3Rs [Ins(1,4,5)P3 receptors], and by regulating the import of Ca2+ through the SERCA (sarcoplasmic/endoplasmic reticulum Ca2+-ATPase) channel . Although the mechanism(s) is unresolved, loss of function of PS1 leads to ER Ca2+ dysregulation. Exaggerated Ca2+ release by IP3Rs in response to Ins(1,4,5)P3 was initially reported in 1994  and confirmed by other groups in different cell types [57,62,63], and more recent studies into the underlying mechanism of the IP3R-mediated Ca2+ release linked the exaggerated release to a gain-of-function mechanism in which mutations in PS1 stimulate IP3R gating . The expression of RyRs is also up-regulated in the brains of transgenic mice harbouring PS1 mutations  as well as in PC12 cells expressing PS1 mutations and in primary cells from PS1-knockin mice . This increase in RyR channels mediates Ca2+ release [63,65]. Additionally, PS1 may act as a passive Ca2+ leak channel within the ER: mutations within the PS1 holoprotein are proposed to disrupt the leak function, leading ER stores to overfill, thereby overactivating IP3Rs and increasing Ca2+ release [22,66]. Another group of investigators, however, was unable to observe changes in ER Ca2+ fill or leak rates in cells lacking presenilin or expressing FAD-linked presenilin mutations , and the basis for the observed differences is not yet resolved. The removal of Ca2+ from the cytosol back to the ER is controlled by the SERCA pump, but, in cells lacking PS1, PS2 or PS1/2, SERCA activity is diminished, leading to increased cytosolic Ca2+. A similar effect has been seen in cells expressing either PS2 or FAD-linked PS2 mutant , although others have observed increased SERCA function leading to the overfilling of ER Ca2+ stores in cells overexpressing PS2 or AD-linked mutant PS1 . Despite these uncertainties about the precise mechanisms, there is consensus in the field that ER Ca2+ is dysregulated in response to loss of PS1 function [56,60].
Ca2+ dysregulation in wild-type PS1 (PS1wt; left) compared with mutant PS1-linked FAD (PS1-FAD; right) conditions
Although the ER contains the largest stores of Ca2+, with concentrations in the low millimolar range , lysosomes are not far behind with Ca2+ concentrations in the range 500–600 μM . Recently, efflux of Ca2+ from lysosomes in PS1/2-knockout MEFs was reported, although the underlying mechanism was not investigated . A subsequent study confirming this phenomenon also revealed reduced levels of the lysosome-associated NAADP-dependent Ca2+ channel TPC2 , although it remains unclear how diminished amounts of a Ca2+ efflux channel would explain elevated lysosomal Ca2+ efflux. An earlier study using the lysosomal pH modulators chloroquine and bafilomycin, however, provides an attractive possible basis for Ca2+ dysregulation in PS1 loss-of-function conditions by linking lysosomal Ca2+ efflux to changes in lysosomal pH . The possibility that decreased lysosomal Ca2+ is a direct consequence of elevated lysosomal pH is indeed consistent with our recent findings  that raising lysosomal pH in wild-type blastocysts not only reproduces the autophagy defects seen in PS1 mutant or deficient cells, but also induces the abnormal lysosomal Ca2+ efflux. Importantly, reversing lysosomal pH deficits in PS1-deleted cells rapidly rescues both the lysosomal Ca2+ abnormalities and the autophagy phenotype. By contrast, reversing the Ca2+ efflux abnormalities alone corrects neither the acidification defect nor the autophagy deficits in PS1-deleted cells, indicating that lysosomal Ca2+ dysregulation is a secondary consequence of pH elevation. These studies therefore link two γ-secretase-independent effects of PS1 loss of function. They imply that v-ATPase deficiency is the common basis for these deficits and suggest that Ca2+ efflux from lysosomes may be a significant contributor to the overall defect in cytosolic Ca2+ elevation seen in PS1-deficient cells.
Altered ER Ca2+ homoeostasis reflected in a depletion of ER stores leading to elevated cytosolic Ca2+ is common among several lysosomal storage disorders, and it is possible that the mechanisms are similar to those proposed in the case of PS1 loss of function. Enhanced release of Ca2+ by the RyRs has been observed in models of Gaucher's disease due to the accumulation of glycosylceramide [72–74]. The ER Ca2+ stores are depleted in GM1 gangliosidosis due to the interaction of GM1 with IP3Rs , and down-regulation of SERCA pumps has been observed in Sandhoff disease  and in Neimann–Pick type A disease  due to interactions with ganglioside GM2 resulting in elevation of cytosolic Ca2+ . Lowered amounts of lysosomal Ca2+ have also been observed in Neimann–Pick type C disease due to accumulation of sphingosine .
PS1-mediated Ca2+ dysregulation: a pathway to accelerated calpain-dependent neurofibrillary degeneration in FAD
Altered ER and lysosomal Ca2+ homoeostasis leads to an increase in cytosolic Ca2+ levels . One consequence of this is the activation of calpain 1 (μ-calpain) and calpain 2 (m-calpain), proteases that are highly expressed in the central nervous system  (Figure 2). Targets of calpain cleavage include cytoskeletal proteins, signal transduction proteins, synaptic proteins and transcription factors . Abnormally increased activities of both calpain 1 and calpain 2 have been linked to AD [81–83], with calpain 1 actions possibly concentrated at synapses  and calpain 2 bound to NFTs and more activated within the neurites of neurons at risk of developing NFTs . Calpain overactivation is well established to have many deleterious effects, among them being the pathological activation of CDK5 (cyclin-dependent kinase 5) (through the cleavage of p35 to p25) , ERK1/2 (extracellular-signal-regulated kinase 1/2) , and GSK3β (glycogen synthase kinase 3β)  and consequent hyperphosphorylation of tau leading to NFT formation. Calpains can also cleave IP3Rs making them constituently active which may further disrupt ER Ca2+ homoeostasis in PS1-related EOAD . The pathogenic role of calpains in AD is indicated further by the marked depletion in AD patients' brains of calpastatin , the specific endogenous inhibitor of calpains [89,90]. Moreover, overexpressing calpastatin reduced calpain activity, tau phosphorylation, ERK activation and amyloid plaque load in either of two different mouse models of AD [91,92]. A synthetic calpain inhibitor, BDA-410, improved spatial memory, fear conditioning and synaptic plasticity associated with increased CREB (cAMP-response-element-binding protein) phosphorylation in APP/PS1 transgenic mice . Finally, we have recently observed that calpastatin overexpression in the JNPL3 tauopathy mouse model leads to decreased tau phosphorylation, increased lifespan and loss of motor neurons (MK. McBrayer, unpublished work). Because ER and lysosomal Ca2+ dysregulation is altered in mutant PS1-linked FAD and PS1-knockout cell lines [43,59], it is not surprising that we have also found calpains to be robustly activated in PS1-knockout blastocysts and mutant PS1-linked FAD fibroblasts as reflected in marked increase in calpain-cleaved spectrin and significantly decreased calpastatin levels (MK. McBrayer, unpublished work).
Conclusions: lysosomal and Ca2+ dysfunction in sAD
Although Ca2+ dysregulation may be initially driven by presenilin mutations in FAD, additional cellular Ca2+ derangements are also seen in the more prevalent late-onset forms of AD  and probably have a multifactorial origin, including contributions from excitotoxicity, ischaemia and Aβ neurotoxicity [95,96]. It is tempting to speculate that the notable disruption of lysosomal function in AD [97,98] may give rise to additional lysosome-related Ca2+ dysregulation. Interestingly, a further connection has been suggested by Yamashima  between calpain activation and lysosomal dysfunction leading to lysosomal membrane rupture, which involves increased calpain-mediated cleavage of carbonylated HSP70.1 (heat-shock protein 70.1). HSP70.1 has roles as both a protein chaperone and stabilizer of lysosomal membrane integrity. Calpain-mediated loss of HSP70.1, coupled with other factors impairing autophagic clearance , are proposed to lead to lysosomal membrane disruption and are likely to diverse downstream pathological events . Calpain activation is observed in patients with sAD, and research indicates that Ca2+ homoeostasis is also disrupted in sAD [99–101]. EOAD and sAD probably have different initiation points leading to dysregulated lysosome dysfunction and Ca2+ dysregulation, but ultimately share pathological outcomes, including calpain overactivation, that should prove to be useful drug targets for AD therapy.
5th Conference on Advances in Molecular Mechanisms Underlying Neurological Disorders: A joint Biochemical Society/European Society for Neurochemistry Focused Meeting held at the University of Bath, U.K., 23–26 June 2013. Organized and Edited by Marcus Rattray (University of Bradford, U.K.) and Rob Williams (University of Bath, U.K.).
amyloid precursor protein
heat-shock protein 70.1
leucine-rich repeat kinase 2
mouse embryonic fibroblast
neuronal ceroid lipofuscinosis
sarcoplasmic/endoplasmic reticulum Ca2+-ATPase
two-pore channel 2
We gratefully acknowledge the expert assistance of Nicole Piorkowski in preparation of the paper and Corrinne Peterhoff in Figure preparation, and thank Devin Wolfe for helpful discussions.
Studies from this laboratory are supported by the National Institute on Aging [grant numbers P01AG17617 and R01AG5604].