Role of BACE1 in Alzheimer’s synaptic function

BACE1 deficiency alters synaptic plasticity in relation to APP cleavage: Long-term changes in the strength of synaptic transmission are the basis of memory formation. Any correlated activity in the pre- and postsynaptic compartments of a synapse in a repeated pattern either strengthens synaptic connections (long term potentiation; LTP) or conversely weakens them (long term depression; LTD) [55]. High levels of Aβ disturbs the balance of reactive oxygen species (ROS) in synaptic boutons and can interfere with pre- and postsynaptic function, presumably by affecting NMDARs, presynaptic P/Q type Ca²⁺ channels, and/or α7-nAChRs, and thus interrupting subsequent Ca²⁺ signaling and leading to altered synaptic function [56].

Mice lacking the BACE1 gene show no β-secretase activity and thus have nearly abolished Aβ (Aβ40 and Aβ42) production in the brain compared to wild-type controls. A deletion of the BACE1 gene in mouse models of AD was able to rescue hippocampal-dependent memory deficits resulting from Aβ accumulation [49] and to ameliorate impaired hippocampal cholinergic regulation of neuronal excitability [57]. Alternatively, BACE1 cleavage of APP will also produce a 99-amino acid C-terminal fragment, referred to as βAPPc or APP-C99. This βAPPc has been shown to impair synaptic functions [58]. BACE1 deficiency benefits AD patients likely through reducing this toxic fragment. These findings implicate that BACE1 may be a good therapeutic target for treating AD [59, 60].

However, recent research progress may suggest otherwise. Since BACE1 has normal physiological functions in synaptic transmission and plasticity in CA1 region of hippocampus, BACE1-null mice displays deficits in both synaptic transmission and plasticity at the hippocampal Schaffer collateral to CA1 synapses [49, 61]. There is a significant increase in the pair pulse ratio (PPF) in BACE1-null mice when compared to wild-type [49]. Because changes in PPF ratio have been attributed to alterations in presynaptic release probability [62], the increased PPF ratio seen in BACE1-null mice may indicate a deficit in presynaptic release [49]. Consistently, BACE1-null mice display altered synaptic plasticity in CA1 and CA3 regions [49, 63]. In addition to presynaptic alterations, changes in PPF ratio can also be attributed to postsynaptic modifications, such as in the subunit composition of AMPA receptors (AMPARs) [64]. Physiological concentrations of Aβ (in pM range) have been shown to facilitate synaptic plasticity [65] and BACE1 deficiency will cause a remarkable reduction in Aβ. Such a loss of physiological levels of Aβ may also lead to synaptic deficits.

BACE1 deficiency alters synaptic plasticity in relation to neuregulin-1 cleavage: An alternative possibility is that the synaptic dysfunctions in BACE1-null mice may arise from abnormal processing of substrates other than APP, i.e., neuregulin-1 (Nrg1) [66‐68]. Nrg1 has a plethora of functions in the central and peripheral nervous systems, which include regulation of myelination, radial and tangential neuronal migration of glutamatergic and GABAergic neurons, and synaptic plasticity [69]. To exert these functions, Nrg1 is required to be cleaved by membrane-anchored proteases, and BACE1 is one such protease. BACE1 deficiency reduces Nrg1 signaling activity and causes defects in these functions as manifested in BACE1-null mice [70, 71].

Many behaviors in animals with Nrg1 mutations exhibit a close resemblance to putative characteristics of schizophrenia, such as impaired pre-pulse inhibition, and spontaneous hyperactivity, which can be reversed by clozapine [72]. Nrg1 and its signaling receptor, the ErbB4 receptor, have been identified as leading candidates for schizophrenia susceptibility genes [73]. It has also been shown that Nrg1-ErbB4 signaling enhances excitatory synapse formation on interneurons and inhibitory synapse formation on pyramidal neurons [74, 75]. Specifically, deletion of ErbB4 from fast-spiking interneurons, such as chandelier and basket cells, has been shown to cause relatively subtle but consistent synaptic defects [74]. Deletion of ErbB4 in interneurons increases miniature excitatory postsynaptic current (mEPSC) frequency and amplitude, but increases miniature inhibitory postsynaptic current frequency in pyramidal neurons [75, 76]. In addition, Nrg1 increases both the number and size of PSD-95 puncta, indicating that Nrg1 stimulates the formation of new synapses and strengthens existing synapses. Nrg1 could also stimulate the stability of PSD-95 in a manner that requires tyrosine kinase activity of ErbB4 [77]. Together, these results suggest that Nrg1 plays a significant role in excitatory synapse development, possibly via stabilizing PSD-95 [76]. By abolishing Nrg1 cleavage to reduce Nrg1-ErbB4 signaling in synapses, BACE1 deficiency likely contributes to synaptic dysfunctions as reported in BACE1-null mice discussed above.

BACE1 deficiency alters synaptic plasticity in relation to Sez6 cleavage: Another family of proteins, the seizure-related gene 6 (Sez6) and its family member Sez6L, were identified as BACE1 substrates through an unbiased proteomic approach and were recently validated as strong substrates of BACE1 [78]. Sez6 and Sez6-like (Sez6L) are nearly exclusively cleaved by BACE1 and not by other proteases in the brain and are guided by their sub-cellular location and their function. They share an NPxY motif and a phosphotyrosine-binding domain (PTB) with another BACE1 substrate, amyloid precursor protein (APP). In BACE1-null and BACE1/2-double-null mice, a marked reduction in the shedding of Sez6 and Sez6L proteins has been confirmed. Their levels in BACE1-null cerebrospinal fluid (CSF) are significantly reduced to ~10% of the wild-type condition. Although the exact molecular functions of Sez6 and Sez6L are not yet fully understood, homology in their protein-binding domains to other cell surface receptors suggests that they may act as receptors at the cell surface [79], as they were originally identified as membrane proteins with five copies of short consensus repeat with a complement C3b/C4b binding site and were seen to be elevated after bursts of neuronal activity [80]. The interaction domains suggest adhesive and/or receptor trafficking functions of these proteins; however, their binding partners are not yet known. Sez-6 is required for normal dendritic arborization of cortical neurons, which is critical for neuronal transfer of information. Its localization along developing and mature dendritic branches and in dendritic spines modulates branch stability. In the absence of Sez-6, mice exhibit short dendrites while cultured cortical neurons display excessive neurite branching. Despite the noticeable effect on branching of dendrites, no obvious effect on an overall growth of the dendritic arbor is reported [81]. Excessive dendritic branching does not always mean a better condition for synapse formation, as studies have found that postsynaptic specializations on these branches (labeled with PSD-95) were dramatically reduced [82]. In the absence of Sez-6, spine numbers are reduced, with reduced excitatory synaptic connectivity between layers II/III and layer V pyramidal neurons in Sez-6-null mice. As spontaneous miniature EPSCs (mEPSCs) or EPSCs with minimal stimulation were not altered, the reduction might be because of uncoupling of pre- and postsynaptic ends of synapses due to altered branching patterns. There is also evidence for reduced synaptic density, punctate staining of PSD-95, and LTP in the frontal cortex of Sez6-null mice. Sez-6 proteins are therefore important for specifying proper dendritic arborization and for development of excitatory synapses on cortical neurons [81]. There is an activity-driven up-regulation of Sez-6 expression after 2 h post-high frequency stimulation [83]. Sez-6 expression levels are highly enriched in brain regions associated with ongoing morphological plasticity, such as the hippocampus and cerebellum in postnatal brain. In Sez-6 deficiency, animals exhibit poor motor coordination and balance, suppressed activity in the open field, reduced anxiety, as well as cognitive deficits. Thus Sez6 protein signaling is critical for excitatory synapse development and function [81] and synaptic circuit refinement [84]. Besides synapse formation and maintenance, Sez6 family members are also expressed and cleaved in lungs and pancreas [79, 85]. Since Sez6 and Sez6L are exclusive substrates of BACE1, they can be used as a direct readout for BACE1 activity in CSF and as a control condition where BACE1 inhibitors can be developed in a substrate-specific manner (for APP) without hampering the physiological actions of BACE1 on other essential proteins like Sez6 that are critical for proper synchronous synaptic transmission.

BACE1 deficiency alters synaptic plasticity in relation to jagged cleavage: Jagged-1 (Jag1) has been identified as a BACE1 substrate [86] and is known to play important roles in neural development and synaptic functions. Jag1 regulates astrogenesis/neurogenesis via the Notch signaling pathway [87‐89]. Because of abrogated Jag1 cleavage, BACE1-null mice exhibit increased astrogenesis and reduced neurogenesis due to increased Jag1-Notch interactions [87]. This is consistent with a prior report that astrocytes negatively regulate neurogenesis through the Notch pathway [90]. Although Notch and its ligands are expressed at low levels in the adult brain [91, 92], they are needed for long-term memory, which is dependent on ultra-structural remodeling of synapses. Hence Notch has an important role in the neural plasticity underlying consolidated memory. Loss of Notch function produces memory deficits in Drosophila melanogaster [93] and impairs proper morphology of dendritic spines [91] in the mouse hippocampus. Thus Jag, as a Notch regulator, is important for synaptic plasticity that contributes to memory formation.

On the other hand, a shift in the balance between neurogenesis and astrogenesis in BACE1-null mice likely contributes to aberrant synaptic transmission. Astrocytes regulate synaptic function and plasticity in close association with synapses [94]. They are involved in synaptogenesis as well as synapse function and elimination. This tight structural and functional partnership between the perisynaptic astrocytic process and the neuronal pre- and postsynaptic structures constitutes the “tripartite synapse” [95]. Astrocyte processes enclose synapses and define functional domains by ensheathing neuronal somas, axons, dendrites, and synapses occupying non overlapping territories, and thus establish gradually independent domains which are also developmentally regulated [96, 97]. This process of segregation, also known as astrocyte tiling, is thought to be regulated by “contact inhibition” between neighboring astrocytes and is crucial for normal functions of the nervous system because, in disease and post-injury conditions, astrocytes lose their tiling ability and display intermingled process morphology [98]. Astrocytes have also been known to regulate glutamatergic postsynaptic strength by increasing the number and stabilizing of AMPAR and NMDAR at the postsynaptic end of synapses [99]. Hence, BACE1 inhibition may impact synaptic functions due to an imbalance in total astrocytes and neurons.