Presenilins (PSs) were first identified in early 1990s as multipass transmembrane proteins those mutations causing familial early-onset forms of Alzheimer disease in which symptoms usually develop between a person’s early 40s and mid-50s. Alzheimer’s disease (AD), as the most common form of dementia, is a major public health problem in the world especially in developed country. Presenilins, the core units of the γ -secretase complex, participate in the process of amyloid beta protein (Aβ) that plays central role in the pathogenesis of AD. However, there are numerous pieces of evidence that PS mutations have several γ -secretase-independent effects.
Presenilins and γ-Secretase Assembly
PSs are highly conserved transmembrane proteins with aspartyl protease activity, characterized by nine helical transmembrane domains (TMD). In mammals, two homologs are present: PS1 and PS2. The homology between them is about 67%. PSs are mostly localized in the ER, Golgi, and in smaller quantity at the cell surface (Lai et al. 2003). PSs are synthesized as a 50 kDa holoprotein that undergo autoendoproteolysis in the ER, catalyzing the formation of 30 kDa N-terminal (NTF) and a 20 kDa C-terminal (CTF) fragments. These fragments together with nicastrin, PEN-2 (presenilin enhancer-2), and APH-1 (anterior pharynx defective1) in 1:1:1:1 stoichiometry build up the γ-secretase complex and are transported through the Golgi to the plasma membrane.
Pen-2 protein with two TMDs has important role in the endoproteolysis of PSs and also regulates the activity of γ-secretase in different way (Gertsik et al. 2014). It helps the stabilization of the complex through binding PSs’ fourth transmembrane domain (Li et al. 2014).
The Aph-1 protein with seven TMDs was discovered parallel with Pen-2 (Francis et al. 2002). In humans, three Aph-1 proteins are known: Aph-1a, -1b, and -1c and they differ in their production of shorter and longer Aβ peptides.
Nicastrin is a single-span membrane protein, and it has a large glycosylated extracellular domain (Li et al. 2014). It seems to regulate complex activity and to be involved in substrate selection.
Besides the four subunit of the complex, there are other proteins that have important role in γ-secretase regulation. During hypoxia, Hif1α, as the key regulator of the response, directly binds γ-secretase complex and thus stimulates its activity by increasing the ratio of active:inactive enzyme (Gertsik et al. 2014). Experimental data with a mouse model suggest that GSAP (γ-secretase activating protein) also has role in Aβ production but the whole mechanism is still mystery (Gertsik et al. 2014). Some suggests that a 16 kDa fragment of GSAP form a complex with γ-secretase and APP, but other data do not support this assumption. However, SNP of GSAP has found to be associated with AD (Gertsik et al. 2014). A proton myo-inositol cotransporter was identified as a novel γ-secretase associated protein that regulates Aβ production without affecting Notch cleavage in vitro (Teranishi et al. 2015).
It was Alois Alzheimer who first published the “presenile dementia” which was then named as Alzheimer’s disease by Emil Kraepelin in the eighth edition of his Handbook of Psychiatry in 1910 (Freeman 2004).
Less than 1% of Alzheimer’s diseases belong to the early onset familiar form that is caused by mutations of PSs, amyloid-beta precursor protein (APP), or apolipoprotein E. So far, approximately 250 AD causing mutations of PSs’ have been identified (http://www.molgen.vib-ua.be/ADMutations).
Neuronal loss, neurotic plaques, and neurofibrillary tangles (NTFs) belong to AD’s histopathology. NTFs are formed by aggregated, hyperphosphorylated tau proteins. The insoluble aggregation of tau proteins associates with other neurodegenerative diseases too, called taupathy (Ballatore et al. 2007).
APP proteins are abundant in synapses; the produced, extracellular-released, soluble Aβ, mostly as Aβ40 circulates in brain intestinal fluid, cerebrospinal fluid, and plasma. When PSs, the catalytic unit of γ-secretase are mutated (mostly loss-of-function mutations), that results incomplete cleavage of APP peptide, longer (42–43 residues) and more fibrillogenic Aβs are formed accordingly. They are hydrophobic and accumulate faster than the shorter ones. Not only the level of longer Aβs but also the ratio of longer to shorter version have important role in AD pathology.
Other γ-Secretase Substrates
ϒ-secretase preferentially cleaves type I transmembrane proteins at two positions: γ-site and downstream ε-site and thus mediates release of their intracellular domain translocated to the nucleus and functions as transcription regulator. By now, more than 90 proteins have been identified as γ-secretase substrates. Many of them functions in cell adhesion, migration, formation of synapses, neurite outgrowth, etc. (Stiller et al. 2014).
A study in C. elegans provided the first evidence that PS mutations cause loss of Notch signaling (Levitan et al. 1996). PS-deficient mice also showed Notch-knockout phenotype (Wong et al. 1997). In FAD-associated PS1 mutated cell culture, the Notch cleaving function of PS is injured (Song et al. 1999).
ErbB4 is a single-pass type I transmembrane protein with a receptor tyrosine-protein kinase activity. Its proteolysis by γ-secretase results the liberation of ErbB4 intracellular domain that has important role in cell-fate determination, e.g., in brain (Haapasalo and Kovacs 2011). Sardi et al. presented that γ-secretase cleaved ErbB4 and regulated timing of astrogenesis in the developing brain of mouse model (Sardi et al. 2006).
The controlling role of γ-secretase in neurite outgrowth, axonal migration, formation of synapses, etc., is important not only during development but also in the adult brain, e.g., during learning. Tomita et al. demonstrated the regulated intramembrane cleavage of ephrin-B1 by γ-secretase and thus the release of its intracellular domain that prevents cellular protrusions in COS cell line (Tomita et al. 2006). A very similar function has been reported to be connected to DCC (deleted in colorectal cancer) receptor where the PS truncated DCC resulted in inhibited neurite outgrowth in mouse neuroblastoma cells (Parent et al. 2005).
p75 neurotrophin receptor is also a γ-secretase substrate and has been implicated in death signaling of neurons in vitro and in vivo. An early significant feature in AD is the loss of basal cholinergic neurons and this is associated with memory and learning deficits (Duan et al. 2014). A high level of p75 neurotrophin receptors expression and induced p75 neurotrophin-mediated cell death have also been observed in these cells (Volosin et al. 2006).
It seems that the more than 90 substrates of γ-secretase ensure its contribution to a wide variety of cellular processes.
γ-Secretase Independent Functions
Presenilin Role in Ca2+ Homeostasis
Calcium is a key signaling molecule, a second messenger in the nervous system involved in a variety of diverse processes, including synaptic plasticity, neuronal excitability, and apoptosis. Neurons are very sensitive to any changes in intracellular Ca2+ concentration, so proper control of its signaling is vital for cell survival. Impaired Ca2+ signaling is an important factor in the pathogenesis of AD and other neurodegenerative diseases too (Brini et al. 2014). However, it is still a question whether it is a contribution to or results of AD. Several studies purpose that PSs have role in the regulation of calcium homeostasis and this function is independent of γ-secretase activity.
The interaction of PSs with inositol trisphosphate receptor (IP3R) is also assumed (Cheung et al. 2008, 2010; Mak et al. 2015). In the presence of FAD-linked mutant PS1 and PS2 in different cells, connection between mutant PSs and IP3 receptor and enhanced gating activity of the IP3R were showed.
Interestingly, the results of Cheung et al. (2008) indicate that FAD-mutant PS1 interaction with IP3R has strong effect on Aβ processing either by Ca2+ release or the interaction.
In the maintenance of cytosolic Ca2+, low level has crucial role in the SR/ER Ca2+-ATPase (SERCA) pumps. In immortalized mouse embryonic fibroblast (MEF) cells from presenilin double-knockout (PSDKO) mice, SERCA activity was reduced. During comparison of PS1 and PS2 effect separately, it appears that PS2 plays larger role in supporting of SERCA activity and thus maintenance of Ca2+ level. Interestingly, in the same PSDKO cells the steady state expression of SERCA2b (the brain isoform) showed elevation, hypothesizing that cells try to compensate the reduced SERCA activity. Physical interaction between PSs and Ca2+ pump was also demonstrated. The SERCA activity correlated with Aβ level; increased SERCA function caused increased Aβ level (Green et al. 2008).
Wild-type PS itself is able to form a passive Ca2+ ion channel, and in PSDKO fibroblast cells it was concluded that PS-mediated transport covers ~80% of Ca2+ leak from ER to cytoplasm (Duggan and McCarthy 2016). FAD-linked mutant PSs alter channel conductance.
PSs are able to modulate cytosolic Ca2+ concentration not only through interaction with Ca2+ channels but with proteins functionally involved in Ca2+ signaling, like sorcin, calmodulin, calpain, and calsenilin (Duggan and McCarthy 2016).
Presenilin Role in the Modulation of ER Stress, Autophagy, and Apoptosis
Ca2+ storage is one of the main functions attributed to the ER. The alteration in ER Ca2+ concentration triggers ER stress that contributes to the induction of unfolded protein response (UPR) and ER-associated degradation (ERAD). These molecular mechanisms try to cope with the stress; however, when it is too extreme or sustained, death of cells will be occurred (Holczer et al. 2015; Banhegyi et al. 2007).
Taken into consideration that PSs seem to have fundamental role in Ca2+ signaling it is not a surprise that several studies cover their contribution to ER stress, UPR, and apoptosis (Katayama et al. 1999; Jin et al. 2010); however, results are controversial.
Neuroblastoma cells transfected with FAD-linked mutant PS1 showed decreased Grp78 protein level, as an ER stress marker, and increased susceptibility to provoked ER stress (Katayama et al. 1999). In primary cultured neurons from mutant mouse model that mimics human FAD also, significant reduction of Grp78 was demonstrated. Moreover, PS1 interaction with IRE1 protein, a stress sensor of ER, and the induction of IRE1 autophosphorylation by presenilin 1 were also presented. The expression of Grp78 and Grp94 is also reduced in FAD patients (Katayama et al. 1999).
On the other hand, in HepG2 cells, the silencing of PS1 resulted in increased Grp78 and PDI (protein disulfide isomerase) expression (Szaraz et al. 2013). The contradictory outcome of studies may be due to the different models or to the length of stress. The functions of PSs are different in cell types. Varied PS-mediated UPR responses were presented in three different cell lines by Jin et al. (2010). Moreover, tunicamycin-induced ER stress resulted in different level of increase in PS1 expression, depending on the length of treatment (Jin et al. 2010).
In the study of Szaraz et al., the cell viability of HepG2 cells was reduced after PS1 but not PS2 silencing. First, the elevated expression of Grp78 and PDI were showed, then other proteins (CHOP) further in the UPR/apoptosis pathway were induced (Szaraz et al. 2013). The increased expression of CHOP (CCAAT-enhancer-binding protein homologous protein), a protein that mediates apoptosis (Nishitoh 2012), has also supported that finding.
Under stress conditions, autophagy has crucial survival function in the turnover of cellular compartment, misfolded- or aggregate-prone proteins. During autophagy, autophagosomes containing dysfunctional cellular organelles are generated first, followed by their fusion with lysosomes. Decreased autophagy is shown in various human diseases including neurodegenerative disorders as well as in cancers (Ghavami et al. 2014).
More hypotheses exist for the PS’s role in the regulation of autophagy. First, the failure of autophagosome and lysosome fusion was connected to PSs. That hypothesis was supported by the findings in mutant PS1 transgenic mouse as well as in AD brain where immature autophagic vacuoles were accumulated (Nixon and Yang 2011). It has also been shown that the turnover of long-lived proteins, e.g., tencephalin, are altered in PS1−/− hippocampal neurons and that phenomenon was able to be rescued by wild-type PS1 (Esselens et al. 2004). The same phenomenon was detected in nonneuronal, in PS1 silenced HepG2 cells (Szaraz et al. 2013).
Recently, Lee et al. (2010) published that the autophagosomal dysfunction upon PS1 deficiency is the consequence of failed lysosomal acidification. They concluded that PS1 is essential to the ER-lysosomes delivery of V0a1 subunit of vacuolar (H+)-ATPase. It is a proton pump assembled on lysosomal membrane and acidifies the newly created autolysosome thus activates cathepsins and effects proteolysis (Yoshimori et al. 1991). In mouse brain, coexpressing mutated APP and PS1, impaired lysosomal acidification, and reduced N-glycosylation of V0a1 subunit were observed (Avrahami et al. 2013).
Lysosome alkalization is also prominent in FAD patients and fibroblasts from AD (Cataldo et al. 2004; Wolfe et al. 2013). Other study concluded that not the alkalization of lysosomes but the altered Ca2+ homeostasis and storage accounts for dysfunction in autophagy. It was observed that in PS1−/− cells, lysosomal acidification was not affected and mutation in PS1 has not influenced V0a1 trafficking either. However, lysosomal Ca2+ homeostasis was compromised elucidating impaired lysosomal fusion capacity (Coen et al. 2012).
Since the discovery of PS mutations in FAD, hundreds of publications have presented new information connected to this molecule.
PSs, as the catalytic unit of γ-secretase, control the proteolysis of over 90 substrates, thus play significant roles in wide variety of cellular processes. In addition to its γ-secretase-related roles, PSs have a number of independent functions. The further exploration of the structural biology of the PS complex is an essential step towards the molecular understanding of PS complex function. Despite the numerous studies on PS since its identification, more work is needed to define the molecular and cellular mechanisms by which PS controls brain functions and the connections between disease-causing mutations and consequent dysfunctions.
- Avrahami L, Farfara D, Shaham-Kol M, Vassar R, Frenkel D, Eldar-Finkelman H. Inhibition of glycogen synthase kinase-3 ameliorates beta-amyloid pathology and restores lysosomal acidification and mammalian target of rapamycin activity in the Alzheimer disease mouse model. J Biol Chem. 2013;288(2):1295–306.PubMedCrossRefGoogle Scholar
- Tomita T, Tanaka S, Morohashi Y, Iwatsubo T. Presenilin-dependent intramembrane cleavage of ephrin-B1. Mol Neurodegener. 2006;1.Google Scholar