Alzheimer's disease (AD) is the most common form of dementia and the most common neurodegenerative disorder [1, 2]. So far, no disease modifying therapies against AD are available. Only symptomatic treatment options are approved. Currently, about 30 million patients are suffering from AD worldwide. Due to an ageing population, it is estimated that by the year 2050 numbers will increase to 106 million AD patients [1, 2]. These numbers raise the concern of AD becoming an epidemic problem. The β-secretase, known as β-site amyloid precursor protein cleaving enzyme 1 (BACE1), plays a central role in AD pathogenesis as it initiates the production of (toxic) amyloid-β peptides (Aβ) that accumulate in AD brains [3]. BACE1 has become a prime therapeutic target for lowering Aβ and thus, treating AD, and clinical development of BACE1 inhibitors is being intensely pursued [4].

It has been shown that AD shares several key features with the rare, inherited, still untreatable, neurodegenerative disorder Niemann-Pick type C (NPC) [5, 6]. Common features include accumulation of intracellular Aβ and APP C-terminal fragments (APP-CTFs), accumulation of hyperphosphorylated tau, neurodegeneration, activation of astrocytes and microglia, and altered cholesterol metabolism. In contrast to AD, which is a complex disorder involving both genetic and environmental components, NPC is a monogenic autosomal recessive disorder caused by mutations in either NPC1 or NPC2 gene [7, 8, 9]. Both code for proteins involved in cholesterol transport. NPC can represent an innovative model in which a defect of a single gene could be used to dissect certain molecular aspects of complex AD. Due to similarities between NPC and AD, NPC disease has also been termed “juvenile Alzheimer’s disease”. However, these two neurodegenerative disorders also show clear distinctions. In contrast to AD, where neurodegeneration primarily occurs in hippocampus and cortex, while cerebellum seems to be protected, in NPC disease Purkinje cells in the cerebellum are the most affected cell types in the brain, while the hippocampus seems to be spared. Different vulnerabilities of brain regions (hippocampus vs. cerebellum) between these two neurodegenerative disorders may give us important clues about protective and/or vulnerable mechanism(s) of neurodegeneration and may also identify critical components in this process. Therefore, further studies of similarities and differences between AD and NPC are encouraged to increase our understanding of the molecular events taking place during the pathogenesis of these conditions, as well as our chances of finding disease-modifying therapies. In AD, for example, the pharmacological inhibition of β-secretase is one of the most promising therapeutic approaches which has entered clinical trial phase II/III [4]. However, recent discovery of multiple BACE1 substrates has raised the concern on potential side-effects that may occur upon BACE1-inhibition [10]. Indeed, prolonged inhibition of BACE1 has shown to impair synaptic plasticity [11, 12]. Thus, further studies are needed to elucidate the function and the effect of cleavage of BACE1 substrates to better evaluate therapeutic potential of BACE1 and foresee the side-effects of BACE1-inhibition.

We have previously reported that increased levels of Aβ and soluble APPβ (sAPPβ) fragments in NPC1-null cells are due to enhanced BACE1 processing of APP [13] which most likely involves cholesterol-mediated sequestration of both APP and BACE1 within the endocytic compartments [14] and/or within lipid rafts [15]. In line with the recent identification of additional BACE1 substrates [10, 16, 17], in this study, we tested whether enhanced BACE1-dependent proteolysis occurs in NPC1^-/- (NPC1) mouse brains and whether it is specific for APP or could also be observed for other BACE1 substrates that could be of relevance for disease pathology. Namely, we analysed processing and localization of seizure protein 6 (Sez6) and seizure 6-like protein (Sez6L) in NPC1 vs. wt mouse brains (cortex, hippocampus and cerebellum) at 4-weeks of age, to detect the earliest BACE1-mediated changes associated with the disease, and at 10-weeks of age to identify changes at the end-stage of disease. Sez6 and Sez6L were selected as being top BACE1 substrates [18] and nearly exclusively cleaved by BACE1 in neurons [10, 19] and not—like APP—additionally by an α-secretase [18, 20]. Although the function of Sez6/Sez6L proteins is unknown, both proteins may be involved in synaptic (dys)function [21, 22]–characteristic features of both AD and NPC diseases. In addition, these proteins were analysed due to the fact that patients with NPC disease exhibit seizures together with the loss of motor control, dementia and other neuropathological symptoms [9, 23]. We hypothesized that, upon NPC1-dysfunction, cholesterol-mediated sequestration of BACE1 within the endolysosomal system (where BACE1 is most active) causes enhanced cleavage of BACE1-substrates, and that increased BACE1-mediated proteolysis may also contribute to the pathological features of NPC disease.

The protease BACE1 is currently the most promising drug target for developing treatments against AD [18, 31]. The recently identified APP variant Ala673Thr, which is less efficiently cleaved by BACE1 [32], together with studies on AD-transgenic mice crossed with genetically deleted BACE1 mice, showing a decrease in Aβ levels and protection against AD and cognitive decline [33, 34, 35, 36, 37], strongly suggest that BACE1-inhibition should prove effective for AD treatment. However, recent studies on BACE1^-/- mice have shown complex neurological phenotypes, raising a concern that a complete BACE1-inhibition might not be as safe as initially expected [38]. Indeed, recently several clinical trials were stopped because of unacceptable side-effects [39]. Such side-effects may result from a reduced cleavage of other physiologically relevant BACE1 substrates, as BACE1 has additional functions in the brain besides cleaving APP [40, 41]. Thus, it is important to fully understand the basic biology of BACE1 and the effects of the cleavage of its substrates to better evaluate the therapeutic potential of BACE1 and make clinical trials safer.

In this work, we employed a single gene lysosomal disorder of cholesterol accumulation—NPC, which was previously shown to share several similarities with complex AD, to further characterize proteolysis carried out by BACE1. NPC disease is also referred to as “juvenile Alzheimer’s disease” and is characterized by progressive neurodegeneration causing symptoms which include ataxia, dystonia, seizures and dementia. Here, we show for the first time that BACE1-dependent proteolytic cleavage is enhanced in a mouse model of NPC disease suggesting that upregulation of BACE1-dependent cleavage could represent an additional common feature between NPC disease and AD. To characterize BACE1 proteolysis in NPC disease, we tested Sez6 and Sez6L, which were shown to be predominantly cleaved by BACE1 in primary cultured neurons [10, 19]. In parallel, we analysed proteolysis of APP which is, besides BACE1, also cleaved by another protease—ADAM10 [18, 20]. In agreement with our previous result of increased amyloidogenic APP processing by BACE1 in the NPC1-null cellular model [30], here, we describe increased proteolysis of the BACE1 substrates Sez6 and its homolog Sez6L, as well as APP, in the cortex and the hippocampus of NPC1 mice at 4-weeks of age while, in the cerebellum, increased BACE1 proteolysis was prominent only for Sez6L protein. The detected upregulation of Sez6, Sez6L and APP cleavage by BACE1 in NPC1 mouse brains, at 4-weeks of age, i.e. early stages in NPC disease pathogenesis [25, 26] is of particular relevance as it implies that enhanced BACE1-dependent cleavage is an early pathological event that precedes neurodegeneration and fully blown disease pathology. Increased BACE1 proteolysis could potentially be used as an early biomarker of NPC disease. It is worth to mention that the observed elevated BACE1 proteolysis of all three BACE1 substrates in nearly all three brain regions analysed (cortex, hippocampus and cerebellum) indicates that their enhanced BACE1 cleavage at this age is likely to be a general effect of NPC1 loss-of-function. However, upon neurodegeneration at 10-weeks of age [25, 26] increased BACE1-dependent proteolysis in NPC1 mice persisted only for Sez6L protein in all three brain regions analysed (it was even 2-fold increased in the cortex and hippocampus). Interestingly, Sez6 cleavage by BACE1 seemed to be strongly upregulated only in the cerebellum at this age, while it was substantially decreased in the hippocampus compared to the levels of cleaved Sez6 at 4-weeks of age. Although Sez6 and Sez6L belong to the family of Sez6 proteins, these results indicate that the efficacy of their cleavage by BACE1 is spatiotemporally differently regulated, particularly in the hippocampus and cerebellum (between 10- and 4-weeks of age), while it was similar in the cortex. At 10-weeks of age we also observed loss of immunoreactivity of APP and Sez6L in Purkinje neurons upon their degeneration in NPC1 vs. wt mice.

These results suggest that upregulation of soluble ectodomains of BACE1 substrates, such as Sez6 and Sez6L, could have important implications in NPC disease phenotype. Indeed, Gunnersen et al. [21] showed that secreted and membrane-bound isoforms of Sez6 exert opposing actions on neurite branching, indicating that secreted ectodomain of Sez6 (Sez6 isoform III) may act to enhance neurite number (to positively control dendritic branching) and, thus, influence synaptic connectivity. As synaptic dysfunction is an early pathological feature of both AD and NPC diseases, it remains to be determined whether BACE1 substrates, Sez6 and Sez6L, may play a role in NPC/AD disease pathogenesis. The common clinical features between these two disorders, such as dementia, presentation of seizures, dysarthria and dysphagia, indicate that both AD and NPC may share common molecular and/or cellular pathway(s) in the disease pathogenesis. A new data has recently been reported for the function of Sez6 protein in maintaining normal dendritic spine dynamics and synaptic structure and function, indicating that prolonged BACE1-inhibition impairs synaptic plasticity via Sez6 protein [12]. Since enhanced BACE1-proteolysis of Sez6L persisted in all three brain regions analysed at 4- and at 10-weeks of age of NPC1 mice, we reasoned that BACE1-inhibition may affect Sez6L function, in addition to Sez6 [12]. Further studies are, however, needed to clarify this in order to comprehend the consequence of prolonged BACE1-inhibition and its therapeutic potential against AD and, possibly, NPC disease.

This study is the first to provide data on spatiotemporal cleavage and distribution/localization of Sez6 and its homologue Sez6L in both wt and NPC1 mouse brains and primary cortical neurons. Distribution of Sez6L in the brains of wt and NPC1 mice was similar to that of Sez6 (mostly stained hippocampus and lower layers in the cortex), except for cerebellum which showed a strong Purkinje staining of Sez6L in wt-mice. Osaki et al. [22] have reported that immunoreactivity of Sez6 gradually decreased by age in the early postnatal period and in the adult forebrain (from P0-P14, and 10-months old mice), however, the strongest levels of Sez6 were detected in CA1 hippocampi, which is in accordance to our findings.

Due to neuropathological similarities between AD and NPC diseases, Kodam and colleagues previously examined protein levels and distribution of APP and its processing enzymes, β- and γ-secretase, in the same mouse model of NPC disease, as used in our study, and detected a statistically significant increase in BACE1 protein levels in NPC1 vs. wt mice in cerebellum at 7- and 10-weeks of age and in hippocampus at 10-weeks of age [27]. Although, in our study, we did not detect an increase in BACE1 protein levels in NPC1 vs. wt mice, our results demonstrate significantly increased processing of the exclusive BACE1 substrates Sez6 and Sez6L, as well as APP, in NPC1 mice, suggesting that BACE1-dependent cleavage in this mouse model of NPC disease is enhanced. Indeed, Kodam et al. [27] showed that activity of β-secretase (BACE1), but not that of α-secretase, was significantly increased both in the hippocampus and cerebellum of NPC1 mouse brains, supporting our findings.

Since it was recently reported that a number of antibodies used against APP are non-specific (as they showed immunoreactivity in APP KO brains), we re-examined APP distribution in NPC1 mice using the previously reported specific APP antibody Y188 (Epitomics, raised against the YENPTY motif at the C-terminal fragment of APP). Kodam et al. [27] also used APP antibody against C-terminal APP fragment (raised against 15 amino acids at the C-terminus followed by the c-Myc epitope) [42]. Our results of APP immunostaining in NPC1 vs. wt mouse brains using the Y188 APP antibody are similar to findings by Kodam et al. [27] showing a profound APP loss in Purkinje neurons at 10-weeks of age. APP was mainly detected in neurons, in accord with the earlier study by Guo et al. [28]. In contrast to Kodam et al. [27], we did not observe the increased levels of fl-APP in the hippocampus and cerebellum of 10-weeks old NPC1 vs. wt mice.

Furthermore, our results on NPC1 mouse primary cortical neurons show that a trafficking defect of BACE1 and its substrates Sez6 and Sez6L, within the endosomal compartments may explain the mechanistic details of enhanced cleavage by BACE1 in a mouse model of NPC disease. In line with our previous results on APP and BACE1 [14] we show that cholesterol accumulation in NPC1 vs. wt mouse primary cortical neurons causes sequestration of the enzyme BACE1 and its substrates Sez6 and Sez6L within the endolysosomal compartments. As these acidic organelles present an optimal environment for the enzyme BACE1 to cleave its substrates, we assume that, in addition to APP, Sez6 and Sez6L, and any other BACE1 substrate that is functionally linked to endolysosomal routing would display its enhanced proteolysis by BACE1 in NPC disease. Our immunocytochemistry results revealed somewhat different subcellular localization of Sez6 and its homologue Sez6L. While Sez6 was localized in the somatodendritic region in wt cortical neurons (in accord to the findings described by Gunnersen et al. [21] and by Osaki et al. [22]), Sez6L was mainly found in the soma and axons in wt cortical neurons and was distributed in larger puncta throughout the neuronal processes in NPC1 cortical neurons. The described different spatiotemporal BACE1-cleavage of Sez6 and Sez6L in NPC1-mouse brains in this work, implies that these BACE1 substrates, although being members of the same Sez6 protein family, may exert somewhat different but complementary functions.

In NPC disease, disturbed cholesterol trafficking and its accumulation within the late endosomes/lysosomes due to NPC1/NPC2 dysfunction is the primary cause of the endolysosomal defect. Indeed, cholesterol-depletion by hydroxypropyl-β-cyclodextrin (HP-β-CD) showed beneficial effects in NPC1 mice [43, 44] and in a single NPC patient [45]. We propose that altered trafficking within the endolysosomal system caused by cholesterol accumulation may lead to some neuropathological features of NPC disease, including those mediated by altered BACE1-proteolysis of Sez6 and Sez6L proteins. Future studies are needed to better evaluate the potential role of Sez6 and Sez6L in NPC disease pathogenesis as well as to investigate the use of their enhanced BACE1-proteolysis as diagnostic and/or therapeutic biomarkers of NPC disease.

In conclusion, we show here for the first time that NPC1 loss-of-function in a mouse model of NPC disease leads to enhanced cleavage of BACE1 substrates Sez6 and Sez6L, in addition to APP, and that this effect most likely involves altered trafficking of BACE1 and of its substrates within the endolysosomal system. In this work we also describe the spatiotemporal distribution and processing of Sez6 and Sez6L, members of Sez6 family, in wt and NPC1-mouse brains. We show that the efficacy of Sez6 cleavage by BACE1 may be spatiotemporally differently regulated, particularly in the hippocampus and cerebellum, in contrast to the cleavage of Sez6L. This study suggests that enhanced BACE1-regulated proteolysis may play a role in NPC disease pathogenesis. Both AD and NPC diseases are still untreatable disorders. Thus, understanding the basic biology of BACE1 and the effects of the cleavage of its substrates is important to better evaluate the therapeutic potential of BACE1.