β-Secretase 1’s Targeting Reduces Hyperphosphorilated Tau, Implying Autophagy Actors in 3xTg-AD Mice

β-site APP cleaving enzyme 1 (BACE1) initiates APP cleavage, which has been reported to be an inducer of tau pathology by altering proteasome functions in Alzheimer’s disease (AD). However, the exact relationship between BACE1 and PHF (Paired Helical Filaments) formation is not clear. In this study, we confirm that BACE1 and Hsc70 are upregulated in the brains of AD patients, and we demonstrate that both proteins show enhanced expression in lipid rafts from AD-affected triple transgenic mouse brains. BACE1 targeting increased Hsc70 levels in the membrane and cytoplasm fractions and downregulated Hsp90 and CHIP in the nucleus in the hippocampi of 3xTg-AD mice. However, these observations occurred in a proteasome-independent manner in vitro. The BACE1miR-induced reduction of soluble hyperphosphorylated tau was associated with a decrease in MAPK activity. However, the BACE1 RNAi-mediated reduction of hyperphosphorylated tau was only blocked by 3-MA (3-methyladenine) in vitro, and it resulted in the increase of Hsc70 and LAMP2 in lipid rafts from hippocampi of 3xTg-AD mice, and upregulation of survival and homeostasis signaling. In summary, our findings suggest that BACE1 silencing neuroprotects reducing soluble hyperphosphorylated tau, modulating certain autophagy-related proteins in aged 3xTg-AD mice.

AD patient brains contain significantly higher levels of Hsp70 and Hsc70 (Perez et al., 1991;Lee et al., 2008), and some studies have described the presence of heat shock proteins in lipid rafts (Triantafilou et al., 2002;Broquet et al., 2003). Previous studies have shown that β-amyloid induces tau pathology through direct alterations of proteasome functions (Oddo et al., 2008). Recent reports have strongly implicated Hsp70/Hsp90 in tauopathy, because these chaperones regulate stability and degradation of unfolding protein as, pair helical filaments (PHF); but its hyperphosphorylation state overload the efficiency of proteasome-dependent degradation (Bonini, 2002;Sakahira et al., 2002;Petrucelli et al., 2004;Shimura et al., 2004;Dickey et al., 2007;Luo et al., 2007;Oddo et al., 2008;Jinwal et al., 2011). So, when the refolding or degradation of abnormal tau protein is not executed by the proteasome, autophagy pathways enter in the scene. Also, depending on the phosphorylation state of tau can be degraded by the proteasome and by the autophagy-lysosome system (Ikeda et al., 1998;Murakami et al., 1998;Oyama et al., 1998;Hamano et al., 2008;Wang et al., 2009). In this way, autophagy represents a homeostatic regulatory mechanism to control metabolism and cellular stressinduced protein aggregation (Singh and Cuervo, 2011), by macroautophagy or CMA, which faults under neurodegeneration condition, allowing the hyperphosphorylation of tau and NTFs formation (Villamil-Ortiz and Cardona-Gomez, 2015).
Therefore, since Hsc70 has been previously linked to proteasome and CMA, and its dysfunction to tau pathogenesis. In addition, BACE1 and Hsc70 reside in lipid rafts. Then, in this study, we have focused in to solve the link between BACE1 and tau pathogenesis.

EXPERIMENTAL PROCEDURES Human Brains
Human brains from the Neuroscience Group of Antioquia's Neurobank (Universidad de Antioquia) were used. A total of ten brains from either sex were evaluated, five brains from patients diagnosed with AD and five control brains from adult patients without a clinical or family history of AD or any other neurodegenerative disease. The brains were optimally preserved by fixing one hemisphere in a buffered solution of 37% formaldehyde and freezing the other hemisphere at −80 • C. The inferior temporal gyrus region was dissected from the frozen hemisphere for biochemical analysis, and the fixed brains were sectioned at 1 cm 3 and sections of 50 µm were obtained for histological, immunohistochemical and immunofluorescence analyses.

Immunohistochemistry
Human brain sections (50 µm) were pre-treated with 10 Mm Tris, pH 6.0, at 85 • C for 5 min. The slices were treated with formic acid (20%) for 20 min to detect the β-amyloid protein. The mouse (50 µm coronal sections) and human brain sections were ''treated for 20 min in 0.1 M PB:methanol (1:1) with 1% hydrogen peroxide and then incubated for 1 h in 0.1 M PB with 1% BSA and 0.3% Triton X-100. Slices were then incubated with primary antibodies, mouse anti-human amyloid beta protein (1:250, SIGNET) and rabbit anti-BACE1 C-terminal (485-501; 1:100, Calbiochem), overnight at 4 • C in 0.1 M PB with 0.3% BSA and 0.3% Triton X-100. The slices were incubated with a biotinylated mouse secondary antibody and then incubated with ABC-HRP complex (Pierce Biotechnology) for 2 h. Diaminobenzidine (DAB) was used to develop the staining. The tissues were dehydrated, covered with mounting solution and observed on an Eclipse E200 optical microscope (Nikon)''.

Lipid Raft Isolation
The cerebral cortices and hippocampi were lysed in 1% Triton lysis buffer (25 mM HEPES, pH 6.5, 150 mM NaCl, 2% TX-100, 1 mM EDTA, and 1 mM PMSF) containing a protease inhibitor cocktail. An equal volume of 80% sucrose was transferred to an SW41Ti centrifuge tube and then overlaid with 6.5 ml 30% sucrose solution and 3.5 ml of 5% sucrose solution containing 25 mM HEPES and 150 mM NaCl at pH 6.5. The discontinuous sucrose gradients were ultra-centrifuged for 18 h at 4 • C with an SW41Ti rotor (Beckman Instruments, Palo Alto, CA, USA) at 200,000 g. The gradient was then fractionated into 12 fractions from the bottom to the top. Each fraction was then prepared for Western blotting and confirmed with flotillin and clathrin antibodies''.

RNAi Design
We designed shRNA miR sequences for silencing BACE1 (shRNAmiR-BACE1) using the BACE1 RNAi sequences (version 1.3.) previously published by Kao et al. (2004). Following the same methodologic strategy in Piedrahita et al., 2010, ''these sequences were cloned into human miR 30-base stem-loops by polymerase extension of overlapping DNA oligonucleotides. The following primers were used for polymerase extension to clone the RNAi into a lentiviral shuttle plasmid (pCMV-GIN-ZEO.GFP) for transfection in HEK-293T: shBACE1miR forward primer, 5 -CAGAAGGCTCGAGAAGGTATATGCTGTTGA CAGTGAGCGCGGACTGCAAGGAGTACAACTATAGTGAA GCCACAGATGTA-3 , and shBACE1miR reverse primer, 5 -CTAAAGTAGCCCCTTGAATTCCGAGGCAGTAGGCATGG ACTGCAAGGAGTACAACTATACATCTGTGGCTTCAC-3 . The extension products were digested with XhoI and EcoRI for directional cloning into the pCMV-GIN-ZEO.GFP vector (Open Biosystem). The following primers were used for polymerase extension to clone the RNAi vectors for adeno-associated virus (AAV) production: shBACE1miR forward primer, 5 -AAAACTCGAGGAGCTCGTGAGCG CTGGACTGCAAGGAGTACAACTCTGTGAAGCCACAGAT GGG-3 , and shBACE1miR reverse primer, 5 -TTTTGGATCCATTAATAGGCAATGGACTGCAAGGAGTA CAACTCCCATCTGTGGCTTCACAG-3 . These extension products were digested with XhoI and SpeI for directional cloning into a U6 expression plasmid that had been digested with XhoI and XbaI (Boudreau et al., 2009)

Viral Particle Production and Neuron Culture Transduction
The protocol for producing ''AAV particles for large-scale production of heterologous proteins used Sf9 insect cell culture with co-infection of recombinant baculovirus derived from the Autographa californica nuclear polyhedrosis virus (Urabe et al., 2002). The shRNAmir-BACE1 expression cassettes were driven by the mouse U6 promoter and were cloned into pAAV.CMV.hrGFP, which contained AAV serotype 2/5 inverted terminal repeats and a CMV-humanized Renilla GFP (hrGFP)simian virus 40 poly-(A) reporter cassette (Urabe et al., 2002;Boudreau et al., 2009). AAV titers were determined using quantitative PCR and/or DNA slot blot analysis. The AAV particles were dialyzed before use'' (Castro-Alvarez et al., 2014).
Neuronal primary cultures were prepared from 10 pregnant Wistar rats at E17-E18. Mice from either sex were used, for a total of 10 C57BL/6 wild-type mice (5-9 months old), 20 18-month-old triple transgenic Alzheimer's mice (3xTg-AD) that were treated for 6 months and 30 15-month-old 3xTg-AD mice that were treated for 3 weeks. How previously we have described ''The 3xTg-AD mice (Oddo et al., 2003) and wild-type mice were injected with 2 µL of AAV2/5-BACE1miR or AAV2/5-GFP (control) with a titer of 10 12 genomes per ml into the right hippocampus (Bregma coordinates: −1.7 antero-posterior, −0.7 lateral and −1.75 depth). The injections were performed with a 10 ml Hamilton syringe at a rate of 0.2 µl/min, and 5 min elapsed after the infusion before the syringe was withdrawn. The animals were transcardially perfused with 4% paraformaldehyde in PBS and processed for immunodetection assays. The brains were cryopreserved with 30% sucrose and stored at −20 • C. The hippocampi and cerebral cortices were dissected, immediately frozen, placed on dry ice and stored at −80 • C until use'' (Piedrahita et al., 2010;Gutiérrez-Vargas et al., 2015).

Soluble and Insoluble Tau Quantification
The hippocampi from 15-month-old 3xTg-AD mice that had been treated with AAV2/5-BACE1miR (BACE1miR) or AAV2/5-GFP (GFP) for three weeks were lysed in ''150 mM NaCl, 20 mM Tris, pH 7.4, 10% glycerol, 1 mM EDTA, 1% NP40, 100 µM phenylmethylsulfonyl fluoride, 1 µg/ml aprotinin and leupeptin (Sigma), 100 µM orthovanadate (Cardona-Gomez et al., 2004) and a protease inhibitor cocktail. The lysates were centrifuged at 13,000 rpm at 4 • C for 10 min. A fraction of the supernatant was stored as the soluble fraction. The remaining fraction was diluted in sarkosyl buffer (50 mM Tris HCl, pH 7.4; 0.15 M NaCl, 1% lauryl sarcosamine, and protease inhibitor cocktail) and centrifuged at 13,000 rpm at 4 • C for 10 min. The supernatant was incubated for 30 min at RT and centrifuged at 170,000 g for 2 h. The pellet was diluted in sarkosyl buffer and stored as the insoluble fraction. The soluble and insoluble fractions were analyzed by Western blotting, as described above''.

In Vitro CDK5 Kinase Assay
''Neuronal primary cultures were transduced for 7 days with AAV2/5-BACE1miR or AAV2/5-GFP (transduction control) and then placed in 1.5 ml microfuge tubes containing lysis buffer, rapidly frozen using liquid nitrogen immersion, and kept frozen until the assay was performed. Then, the sample was thawed on ice, homogenized, incubated for 15 min on ice, and centrifuged at 13,000 rpm at 4 • C. The supernatant was recovered in clean microfuge tubes, and the protein concentration was measured with the bicinchonic acid method (Thermo Fisher Scientific). CDK5 was immunoprecipitated from 250 µg of total protein using 1 µg of the rabbit polyclonal anti-CDK5 (C-8) antibody (Santa Cruz). The antibody was incubated with the protein extract overnight at 4 • C on a rotator. Protein G-Sepharose (Sigma-Aldrich) was added, and the samples were incubated for an additional 1 h at 23 • C. The Protein G-Sepharose beads were washed five times with immunoprecipitation (IP) buffer (Sigma-Aldrich), while maintaining the sample at 4 • C.
After the fifth wash, the Protein G-Sepharose beads were resuspended in 200 µl of kinase assay buffer (20 mM Tris-HCl, pH 7.5, 100 µM sodium orthovanadate, 10 mM MgCl 2 , 50 mM NaCl, 1 mM DTT, and 1 mM NaF), and ATP was added to the resuspended beads at a 10-fold excess (0.5 mM). Histones from calf thymus type III-S (Sigma-Aldrich) were added at a final concentration of 6 µM as a substrate for CDK5, and then the reaction was gently vortexed, aliquoted and incubated at 37 • C for 30 min. To stop the reaction, 5 µl of SDS-PAGE loading buffer (250 mM Tris-HCl, 10% SDS, 30% glycerol, 0.5 M DTT, 0.02% bromophenol blue) was added, and the samples were immediately incubated for 5 min at 95 • C. The samples were separated electrophoretically at 120 V for 2 h and transferred to a nitrocellulose membrane at 200 mA for 1.5 h. Ponceau Red in 5% acetic acid was used to stain the transferred proteins. The histones were clearly visible, and migrated at approximately 21 kDa on the gel. Western blots for CDK5 (C-8 antibody) and rabbit polyclonal antiphosphorylated histone H1 (Millipore; 06-597) were used. Goat anti-rabbit IRDye 800WE (LI-COR) was used as the secondary antibody and detected using an Odyssey Infrared Imaging System (LI-COR). The band intensities for the histones were measured with NIH ImageJ software and normalized to the IgG heavy chain intensity''.

In Vitro MAP Kinase Assay
We used the MAP Kinase/Erk IP Kinase Assay kit  according to the manufacturer's instructions. ''The assay kit is designed to measure phosphotransferase activity in an immunocomplex formed between the MAP Kinase R2 antibody and MAP Kinase (p44 mpk ). This precipitated enzyme is used to phosphorylate a specific substrate, myelin basic protein (MBP). The phosphorylated substrate is then analyzed by Western blotting using an antibody specific for phosphorylated MBP. The measurement of MAPK activity in most cell lysates is not accurate due to the phosphorylation of MBP by other kinases''.

Cellular Fractions
The cerebral cortices and hippocampi from wild-type and 3xTg-AD mice were ''homogenized in 100 mM Tris, pH 7.4, 3 mM MgCl 2 , 0.32 M sucrose, 0.1% Triton X-100 (Buffer A), and a protease inhibitor cocktail (Sigma-Aldrich). The soluble fraction was obtained after centrifuging at 2500 rpm for 15 min at 4 • C. The pellet was suspended in Buffer A and loaded on a two-layer sucrose cushion; the first one was Buffer B containing 1.9 M sucrose and the second, Buffer C, contained 2 M sucrose. The sample was then centrifuged for 1 h at 10,000 rpm at 4 • C. The membrane fraction was at the top of the 2 M sucrose layer, whereas the nuclear fraction was at the bottom''.

Statistical Analysis
''The n used for in vitro and in vivo experiments were 3-6. Parametric data were compared using multi-variable two-way analysis of variance (ANOVA) followed by Tukey's post hoc test for comparisons between several independent groups. A p < 0.05 confidence level using a two-tailed test was adopted as statistically significant. The Student's t-test was used to compare two groups. The data were expressed as the means ± SEM. The analyses were performed with SPSS (IBM) and GraphPad Prism version 4.00, 2003 (GraphPad Software Inc., San Diego, CA, USA)''.

BACE1 and Hsc70 were Upregulated in AD Brains, and Hsc70 is Retained in Lipid Rafts
The relationship between BACE1 and Hsc70 was evaluated in the inferior temporal gyrus of AD-affected human brains. The typical hallmarks of AD (β-amyloid and PHF-1) were detected, and BACE1 and Hsc70 immunoreactivity were increased ( Figure 1A). BACE1 and Hsc70 protein levels were upregulated, whereas a related protein, the CHIP carboxyl terminus, remained unchanged compared to the levels in control brains ( Figure 1B). It is widely accepted that BACE1 is enriched in lipid rafts in AD (Riddell et al., 2001;Cordy et al., 2003;Ehehalt et al., 2003). Lipid rafts were isolated from 15-month-old triple transgenic mouse brains (3xTg-AD) to determine whether BACE1 and Hsc70 were associated in these microdomains. BACE1 and Hsc70 proteins were significantly increased in isolated lipid rafts (flotillin-positive fractions, data not shown). CHIP was not changed and Hsp90 and PHF-1 proteins were not detected ( Figure 1C). Interestingly, Hsc70 levels were significantly reduced in the cytoplasmic fractions from the brains of the 3xTg-AD mice in comparison with the wild-type mice ( Figure 1D). These data suggest that the additional Hsc70 is retained in lipid rafts when BACE1 is increased in this micro-domain in AD brains.

Specific Silencing of BACE 1 Reduces β-Amyloidosis in the Hippocampus of 3xTg-AD Mice
We generated a recombinant AAV (serotype 2/5), which expresses BACE1miR in addition to a GFP reporter. We evaluated the effect of the BACE1 shRNA-miR treatment on β-amyloidosis in 3xTg-AD mice. Initially, the AAV:BACE1miR vector was injected into the right hippocampus of 5-9-monthold wild-type mice to evaluate gene silencing at 3 weeks, 3 months and 6 months post-injection. We observed reduced BACE1 protein levels in the brains that were injected with the BACE1miR compared to the GFP control as detected by Western blotting (Figure 2A) and confocal immunofluorescence analysis; BACE2 expression was not affected (Figure 2B). In addition, there was a significant decrease in BACE1 as well as β-amyloid immunoreactivity in the 3xTg-AD mice at 6 months after injection with the BACE1miR compared to the GFPtreated animals ( Figure 2C). We also confirmed the reduced levels of BACE1 by Western blotting. Moreover, the level of the CTF-β fragment was significantly reduced; however, the levels of APP-CT, APP-NT, full-length PS1 and C-terminal PS1 proteins did not change ( Figure 2D). BACE1miR specifically reduced Aβ-42 levels, without changing the Aβ-40 levels ( Figure 2E). Therefore, our findings confirm that the BACE1miR reduced BACE1 levels and β-amyloidosis, as previously reported (Luo et al., 2001;Kao et al., 2004;Ohno et al., 2004;Laird et al., 2005).

BACE1 Targeting Decreases Soluble
Hyperphosphorylated Tau by Reducing MAP Kinase Activity in the Hippocampus of 3xTg-AD Mice Unexpectedly, a significant reduction in the number of PHFpositive cells was observed in the hippocampus ( Figure 3A). However, only the level of soluble tau was reduced by BACE1miR, and the level of insoluble tau was even increased ( Figure 3B). In addition, PHF-1 protein levels were reduced by BACEmiR, whereas the levels of AT-8, AT-100, AT-180 and TAU-5 were not changed ( Figure 3C). To understand the cellular mechanism of hyperphosphorylated tau reduction, we evaluated the various kinases involved, as well as protein phosphatase 2A (PP2A) and Bcl-2. The CDK5 protein levels (Figure 3D), CDK5 activity (Figure 3E), GSK-β, GSK-β (Figure 3F), Bcl-2 ( Figure 3G), ERK-1, ERK-2 ( Figure 3H) and PP2A activity ( Figure 3I) were not modified in BACE1miR-treated 3xTg-AD mice. Interestingly, our data showed that MAPK activity was significantly reduced by BACE miR (Figure 3H), while Bcl-2 was upregulated (Figure 3G), suggesting impacts on cell survival.

BACE1 Silencing Upregulated Hsc70 in the Cytoplasm and Reduced Hsp90 in the Nucleus
To examine the relationship between BACE1 and Hsc70, Hsc70 protein expression was evaluated in the brains of BACE1miRtreated 3xTg-AD mice. The brain slices showed a significant increase in Hsc70 fluorescence intensity in the CA1 area compared to the GFP control group (Figures 4A,B). The subcellular localization of Hsc70 is critical for its specific functions (D'Souza and Brown, 1998;Chen and Brown, 2007). The localization of Hsc70 and of other related proteins was determined by cellular fractionation and Western blotting from the hippocampi of BACE1miR-treated 3xTg-AD mice. Hsc70 was significantly increased in the cytoplasm and membrane fractions; however, its expression in the nucleus was not changed. Interestingly, the Hsp90 protein level was decreased in the cytoplasm and nuclear fractions, but it remained unchanged in the membrane fraction. Moreover, CHIP was significantly increased in the cytoplasm and decreased in the nucleus (Figures 4C-E). Also, HSF-1 was reduced by the BACEmiR treatment, meaning a non transcriptional upregulation of Hsc 70 ( Figure 4F).

BACE1miR Reduced Hyperphosphorylated Tau in a Proteasome-Independent Manner In Vitro
As in the in vivo model, using protein extracts from transduced cultured cortical neurons, we confirmed that AAV2/5-BACE1miR downregulated BACE1, PHF-1, and Hsp90 and upregulated Hsc 70, while CHIP remained unchanged ( Figure 5A). BACE1miR did not affect the typical enzymes involved in tau hyperphosphorylation, such as GSK-3 and CDK5; however, pSer473 Akt showed a modest decrease ( Figure 5A). We used these primary neuronal cultures and the proteasome inhibitor lactacystin to assess whether PHF clearance by BACE1 silencing was proteasome-dependent.
Neuronal cultures transduced with AAV2/5-BACE1miR maintained a significant reduction in PHF-1 protein levels despite treatment with lactacystin (proteasome inhibitor) and KNK437 (heat shock protein inhibitor; Figure 5B). As expected, lactacystin increased the levels of Hsp90 and Hsc70, whereas cells treated with BACE1miR did not show increases in Hsp90 (B) BACE1, Hsc70 and CHIP Western blotting from the temporal gyrus of AD brains (AD); normal human brains were used as controls. A representative plot is shown. βIII-tubulin was used as a loading control. Densitometric quantification was performed; RU = relative units. n = 4, * p < 0.05. (C) Western blotting of BACE1, Hsc70, Hsp90, CHIP and PHF-1 proteins in lipid rafts isolated from the cerebral cortex of 15-month-old wild-type and 3xTg-AD mice. Flotillin-1 was used as a loading control. Densitometric quantification was performed; RU = relative units. n = 3, * p < 0.05. (D) Western blotting of Hsc70 in the cytoplasmic fraction from the cerebral cortex of wild-type and 3xTg-AD mice. RU = relative units. n = 3, * p < 0.05. Representative blots are shown. The arrow shows the band corresponding to the expected molecular weight of the BACE1 protein.
and Hsc70. KNK437 produced similar effects. CHIP was not modified by any treatment (Figures 5C-E).

The BACE1miR-Mediated Reduction in Hyperphosphorylated Tau was Blocked by 3-MA, an Autophagy Inhibitor
In addition to the proteasome pathway, macroautophagy is another important mechanism that is involved in the unfolded protein response and is specifically related to aggregated tau (Ikeda et al., 1998;Murakami et al., 1998;Wang et al., 2006aWang et al., , 2009Hamano et al., 2008). Dysfunction of this pathway has been reported in AD (Cataldo et al., 2004a,b;Yu et al., 2004Yu et al., , 2005Nixon et al., 2005;Boland et al., 2008). A macroautophagy marker, LC3B, was evaluated in the in vitro experiment. We found that LC3B was significantly increased in BACE1miRtreated neuronal primary cultures, and this increase was not reversed by lactacystin or KNK437 ( Figure 5F). Together, these data suggest a proteasome-independent mechanism for hyperphosphorylation tau loss induced by BACE1miR that involves the up-regulation of LC3B.

BACE1 Targeting Induces Hsc70/LAMP2 Upregulation in Lipid Rafts from Hippocampi of 3xTg-AD Mice
The LC3B protein was significantly upregulated in the cytoplasm and decreased in membrane fractions ( Figure 7A) from the hippocampi of 3xTg-AD mice treated with BACE1miR compared to the untreated mice.

BACE 1 Knockdown Induces Survival Signaling in the Hippocampus of 3xTg-AD Mice
We evaluated two of the most important mechanisms of autophagic regulation pathways, Akt/mTOR and Bcl2/Beclin-1 pathways (Figures 8A-D). Although, we did not detect changes of pSer473 AKT in hippocampi total lysates (Figure 8B), we detected high protein levels of p2448 mTOR and increased mTOR activity (Figure 8E) in the hippocampi of 3xTg-AD treated, without changes in pThr389 p70S6K with BACE1miR compared with the control values, (Figures 8C,D).
Complementarily, the BACEmiR treatment produced increased protein levels of BCL2 and a reduction of Beclin 1 (Figure 8A), accompained by an increase of FoXO3, Hsc70, without changes in total lysates of Hsp90 and CHIP. These findings together, maybe suggest that the silencing of BACE1 induces survival and cellular homeostasis.
Cholesterol homeostasis is impaired in AD patients' brains, thereby enhancing ''β-and γ-secretase activities and Aβ production in human brains'' (Xiong et al., 2008). Cellular cholesterol levels can modulate APP metabolism, and ''cholesterol depletion reduces the association between APP and lipid rafts and disrupts the APP-PS1 interaction'' (Guardia-Laguarta et al., 2009). ''Aberrant cholesterol trafficking is associated with the potentiation of APP processing by BACE1, leading to an overall increase in Aβ levels'' (Burns et al., 2003). In an APP-overexpressing mouse model, a cholesterol-rich diet increases Aβ accumulation, accelerates AD-related pathology (Refolo et al., 2000), and correlates with tau hyper-phosphorylation (Ghribi et al., 2006). In the Tg2576 mouse model and in ADaffected human brains, Aβ dimers appear in lipid rafts and ApoE progressively accumulates with aging, thereby facilitating Aβ fibril formation (Kawarabayashi et al., 2004).
The ''two major protein degradation systems are the proteasome pathway and the autophagy-lysosome pathway''; damage to lysosomal function is a well-recognized event in AD neurodegeneration. Lysosome pathology in ADaffected brains involves changes in macroautophagy and increased APP proteolysis, and it contributes to Aβ deposition (Grbovic et al., 2003;Cataldo et al., 2004a,b;Yu et al., 2004;Nixon et al., 2005;Boland et al., 2008). Our results suggest that BACE1miR modified some autophagy mediators because we found that the protein levels of LC3B and LAMP2 were upregulated. The transcription of autophagy-related genes, such as LC3B and Bnip3, occurs during fasting, and Akt/PKB activation blocks autophagy (Mammucari et al., 2007). Therefore, pSer473 Akt downregulation by BACE1miR could upregulate autophagyrelated genes, thereby increasing LC3B. By contrast, we found that LC3B was reduced in the membrane fraction of BACE1miR-treated 3xTg-AD mouse hippocampi. However, a significant increase in Hsc70/LAMP2 in lipid rafts was accompanied by a reduction in hyperphosphorylated tau in vivo, which could suggest some repercussion of BACE1miR on CMA.
Similarly, two major CMA proteins, Hsc70 and Hsp90, play critical roles in LAMP-2A complex dynamics on the lysosome membrane (Bandyopadhyay et al., 2008). LAMP-2 is a lysosome protein in mammalian cells, and it is necessary for the fusion of lysosomes with autophagosomes (Tanaka et al., 2000;Bampton et al., 2005). The ''lysosomal levels of both Hsc70 and LAMP-2A increase when CMA is activated'' (Agarraberes et al., 1997;Cuervo and Dice, 2000b). Changes in the dynamic distribution of LAMP-2A into and out of discrete lysosome membrane microdomains contribute to the regulation of CMA, and the number of Hsc70-containing lysosomes increases in conditions that produce CMA activation . CMA declines with age because of decreased levels of the lysosome-associated membrane protein LAMP-2 and decreased substrate binding to the lysosomes (Cuervo and Dice, 2000a,b;Massey et al., 2006;Kiffin et al., 2007). Interestingly, our FIGURE 8 | Silencing of BACE1 upregulates the survival signaling in 3xTg-AD mice. Survival and quality control regulation pathways were analyzed. (A) Bcl2/Beclin-1, (B) Akt/FoxO3, (C) Hsc70/Hsp90/CHIP protein levels, (D) mTOR pathway and its activity (E) from the hippocampi of 3xTg-AD mice 24-month-old after 6 months of injection with AAV2/5-BACE1miR and AAV2/5-GFP version as control. Representative blots are shown. RU = relative unit. n = 4, * p < 0.05. findings showed that Hsc70 and LAMP2-A were increased by BACE1miR in lipid rafts and in the cytoplasm of 3xTg-AD mouse brains.
On the other hand, our findings correlate with BACE1miRinduced neuroprotection, which generated a reduction of sustained tau phosphorylation, and involved heat shock proteins, phosphatidylethanolamine and MAPK in a similar manner that to control of cell stress (Dodson et al., 2013) and cancer inhibition. Phosphatidylethanolamine binding proteins block MAPK activation and thereby have potential therapeutic implications in Alzheimer's disease and cancer (Al-Mulla et al., 2013;Ling et al., 2014). Additionally, Raf, a component of the MAPK cascade, interacts with Hsc70 in the mouse hippocampus in basal conditions (Bonfiglio et al., 2011;Al-Mulla et al., 2013). However, the exact mechanisms by which BACE1miR could modulate those targets remain unknown and require additional studies for understanding the molecular convergence of these actors and their concomitant actions.
Finally, autophagy inhibitors reduced the BACE1miRinduced LC3B levels in vitro, and the BACE1miR-mediated soluble hyperphosphorylated tau downregulation was only blocked by 3-MA (a class III PI3K inhibitor), but not by bafilomycin or NH 4 Cl in vitro. Together, these data maybe suggest a role for BACE1miR in the double-membrane vesicles and in membrane fusion to form autophagosomes, which requires phosphatidylethanolamine, but they do not imply a typical autophagic flow because insoluble tau was not reduced, neither by bafilomycin nor NH 4 Cl blocked the reduction of hyperphosphorylated tau. Recently, our unpublished data show the repercussion of shBACE1miR on the cellular homeostasis induced by a favorable change of the fatty acid composition of phospholipids, mainly on LPE and ePE, by reduction of arachidonic acid (20:4), increase of DHA (22:6), reduction of cPLA2 activity, reduction of COX2 levels and improving of the cognitive function after 6 and 12 months post-injection (Villamil-Ortiz and Cardona-Gomez, 2015). Concomitantly, it could be reflected in the reduction of proinflammatory signaling as MAPK activity obtained in this study under the same experimental condition, which could be to favor of the modulation of CMA and macroautophagy proteins, and clearance of hyperphosphorylated tau. Because, MAPK regulates downstream cPLA2 under lipid peroxidation (Shibata et al., 2011); being the cPLA2 encharged of the formation of derivated plasmanogens of PE (Makide et al., 2009), and this enzyme activation is close related to hyperphosphorylated tau, inflammation and neurodegeneration (Sundaram et al., 2012). Therefore, the inhibition of PE lipidation of LC3B by 3MA, could be affecting the action of BACE1 silencing on PE fatty acid composition. However, future studies must be developed to determine the specific mechanism of that regulation by BACE1miR. Also, it has recently been shown that phosphatidylethanolamine regulates autophagy and longevity (Rockenfeller et al., 2015), and LC3 lipidation are involved in autophagy and lipid metabolism turnover (Singh et al., 2009), which could be related with tissue homeostasis and improvement of cognitive function of 3xTg-AD mice treated with BACE1miR during 6 and 12 months (Villamil-Ortiz and Cardona-Gomez, 2015).
In summary, our study has extended the role of BACE1 beyond its role in APP cleavage by linking it to Hsc70 and other mediators of autophagy. BACE1 in lipid rafts is associated with tau aggregation, but its silencing induces cellular homeostasis in PE-dependent mode, produce soluble PHF clearance, involving MAPK inhibition, Hsc70 and LAMP2 in lipid rafts maybe favoring CMA. These results emphasize that BACE1 targeting is a promising neuroprotective therapy for Alzheimer's disease.