Trehalose Alters Subcellular Trafficking and the Metabolism of the Alzheimer-associated Amyloid Precursor Protein*

The disaccharide trehalose is commonly considered to stimulate autophagy. Cell treatment with trehalose could decrease cytosolic aggregates of potentially pathogenic proteins, including mutant huntingtin, α-synuclein, and phosphorylated tau that are associated with neurodegenerative diseases. Here, we demonstrate that trehalose also alters the metabolism of the Alzheimer disease-related amyloid precursor protein (APP). Cell treatment with trehalose decreased the degradation of full-length APP and its C-terminal fragments. Trehalose also reduced the secretion of the amyloid-β peptide. Biochemical and cell biological experiments revealed that trehalose alters the subcellular distribution and decreases the degradation of APP C-terminal fragments in endolysosomal compartments. Trehalose also led to strong accumulation of the autophagic marker proteins LC3-II and p62, and decreased the proteolytic activation of the lysosomal hydrolase cathepsin D. The combined data indicate that trehalose decreases the lysosomal metabolism of APP by altering its endocytic vesicular transport.

CTF␣. The subsequent cleavage of APP-CTF␣ by ␥-secretase results in secretion of a small peptide called p3 (2). It is important to note that APP is also metabolized by additional pathways, including proteasomal and lysosomal degradation (3)(4)(5).
The metabolism of APP is also regulated by macroautophagy (herein referred to as autophagy), a lysosome-dependent degradative pathway for long lived proteins, organelles, and nutrient recycling (6,7). Autophagy starts with the formation of double-membrane vesicles, the so-called autophagosomes, which further fuse with lysosomes to become autolysosomes for degradation of autophagosome contents by lysosomal hydrolases (8). Autophagy is an essential prosurvival pathway induced by a variety of stress factors, including nutrient deprivation, growth factor withdrawal, oxidative stress, infection, and hypoxia (9). These factors contribute to the etiology of multiple diseases such as cancer, stroke, heart disease, and infection (10). Eukaryotic cells have a basal autophagic activity under normal physiological conditions. Cells deficient in autophagy show diffuse abnormal protein accumulation and mitochondrial disorganization (11,12), suggesting that autophagy is important to maintain cellular homeostasis by eliminating protein aggregates and damaged organelles. Basal autophagy is particularly important and active in the liver and non-dividing cells such as neurons and myocytes (11,13,14). Dysfunction in autophagy is implicated in the pathogenesis of many human diseases, including different types of neurodegeneration (15,16). Incompletely degraded autophagic vacuoles contain both APP and ␥-secretase complex and could thus contribute to enhanced processing of APP into toxic A␤ (17)(18)(19).
Trehalose is a natural non-reducing disaccharide containing two D-glucose residues connected via an ␣,␣-1,1 linkage. It functions not only as an energy source but is also used to protect cells against heat, cold, oxidation, or dehydration (20). Trehalose is considered to enhance autophagic activity (21). It has been shown to promote the cellular clearance of pathogenic proteins like mutant huntingtin, ␣-synuclein (21,22), and phosphorylated tau (22)(23)(24) that are associated with Huntington, Parkinson, and Alzheimer diseases, respectively. However, the role of trehalose in the cellular metabolism of the Alzheimer disease-associated APP remains to be investigated.
In this study, we analyzed the role of trehalose in the cellular metabolism of APP. Interestingly, trehalose strongly decreased the degradation of APP and its CTFs. By biochemical and cell biological approaches, we demonstrate that trehalose interferes with endocytic vesicular trafficking as indicated by decreased processing of cathepsin D (Cat D), redistribution of lysosome-associated protein (LAMP)2, and accumulation of the autophagy-related proteins microtubule-associated protein 1 light chain 3 (LC3) and p62. The combined data demonstrate that trehalose exerts complex effects on vesicular trafficking, thereby decreasing the metabolism of APP and the secretion of A␤.
Cells were cultured in DMEM supplemented with 10% FCS and 1% penicillin/streptomycin. H4-mCherry-GFP-LC3 stable clones were generated from transfection of H4 cells with mCherry-GFP-LC3 constructs. The protocol was performed with Lipofectamine (Invitrogen) according to the manufacturer's instructions. Clones were selected with 25 g/ml puromycin.
Immunocytochemistry-Cells were processed for immunocytochemistry as described previously (28). Briefly, cells were cultured on poly-L-lysine-coated coverslips and fixed with 4% paraformaldehyde for 20 min. Subsequently, cells were kept in blocking solution (90% PBS, 10% FCS, 0.1% Triton X-100) at room temperature for 1 h, incubated with primary antibodies at 4°C overnight or at room temperature for 1-2 h, and then incubated with secondary antibodies for 1 h at room temperature. Cells were imaged with a Carl Zeiss Axio Imager 2 Apo-Tome fluorescence microscope for optical sectioning. Signal intensity and co-localization were analyzed with AxioVision software. For quantification, randomly selected images (n ϭ 10) were used. Signals were quantified in two randomly chosen boxes (150 m 2 ) per cell. Values represent means Ϯ standard deviation (S.D.).
To isolate cellular membranes, cells were incubated in hypotonic buffer D (10 mM Tris-HCl, pH 7.5, 10 mM NaCl, 0.1 mM EGTA, 25 mM glycerol 2-phosphate, 1 mM DTT) supplemented with 1% proteinase inhibitor on ice for 10 min. Cells were then homogenized using 21-gauge needles and centrifuged at 1000 rpm for 10 min to pellet nuclei. The supernatant was then centrifuged at 13,200 rpm for 30 min to obtain the membrane fraction as a pellet.
Membrane fractions or cellular lysates were separated by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes (Schleicher & Schüll). Membranes were blocked in 5% nonfat milk for 1 h and incubated with primary antibodies overnight and appropriate secondary antibodies for 1 h. Proteins were detected by enhanced chemiluminescence (ECL) using ECL reagent (GE Healthcare) and an ECL imaging station (ChemiDoc XRS, Bio-Rad). Quantification of signals was done by the Quantity One software package (Bio-Rad).
Subcellular Fractionation-For subcellular fractionation, cells were subjected to a hypotonic shock as described above. Postnuclear fractions (3800 rpm, 5 min) were loaded onto a discontinuous gradient (30, 20, 17.5, 15, 12.5, 10, 7.5, 5, and 2.5% OptiPrep TM ; Sigma Aldrich; 1.2 ml each) and separated at 100,000 ϫ g at 4°C for 8 h without using the brake. Single fractions were collected at a volume of 1 ml, and proteins were precipitated with trichloroacetic acid.
Measurement of A␤ Variants-Cells were grown on 6-well culture plates until 70% confluence in DMEM as described above. For collection of A␤, 750 l of fresh medium was added overnight. Conditioned media were cleared by centrifugation. Cells were briefly washed and lysed in radioimmune precipitation assay buffer. Both cell lysates and conditioned media were analyzed by electrochemiluminescence technology (MesoScale Discovery) for A␤38, A␤40, and A␤42 according to the manufacturer's protocol.
Metabolic Radiolabeling and Pulse-Chase Experiments-Cells were grown in Petri dishes until 80% confluence. Cells were washed with PBS and incubated in serum and methionine-free medium for 30 min. Cells were then incubated with 20 Ci of 35 S-radiolabeled methionine/cysteine for 30 min to pulse label newly synthesized proteins. After pulse labeling, cells were washed with label free-medium containing 10% serum and a 5-fold excess of unlabeled methionine and chased for the indicated time periods. Cells were then lysed, and proteins were separated by SDS-PAGE. After transfer to PVDF membrane, radiolabeled proteins were quantified by phosphorimaging.
In Vitro ␥-Secretase Assay-The in vitro ␥-secretase assay was carried out as described previously (18,29). Briefly, cells were lysed in hypotonic buffer D, and membranes were isolated as described above. The membrane pellet was then resuspended in citrate buffer (150 mM sodium citrate, distilled H 2 O, pH 6.4 adjusted with citric acid) supplemented with 1% protease/phosphatase inhibitors and incubated in the absence or presence of ␥-secretase modulators at 37°C for 2 h. Samples were centrifuged at 13,200 rpm for 1 h. The resulting pellets and supernatants were separated by SDS-PAGE, and proteins were detected by Western immunoblotting.
Data Analysis and Statistics-Statistical analyses were carried out by calculation of S.D. and Student's t test. Significance is indicated by asterisks as follows: * for p Ͻ 0.05, ** for p Ͻ 0.01, *** for p Ͻ 0.001, and n.s. for not significant.

Results
Trehalose Impairs the Metabolism of APP and Decreases the Secretion of A␤-The degradation of APP and its CTFs involves autophagic and lysosomal pathways (17-19, 25, 26). To assess the effect of trehalose on APP metabolism, human neuroglioma H4 cells that endogenously express APP were incubated with trehalose for different time periods, and APP levels were analyzed by Western immunoblotting. Cell treatment with trehalose strongly increased levels of APP full length (FL) (Fig. 1, A and B) and, even more pronounced, of APP-CTFs ( Fig. 1, A, C, and D). Trehalose also significantly increased the secretion of soluble APP derived from ␣-secretory processing into conditioned media (Fig. 1, E and F), suggesting that the increased levels of cellular APP-FL were not caused by inhibition of APP secretion.
To test whether trehalose exerts similar effects in other cell types, we also used human neuroblastoma SH-SY5Y, liver Hep-G2, and embryonic kidney 293 cells. Trehalose induced accumulation of APP-FL and particularly of APP-CTFs in all tested cell lines (Fig. 1G), very similar to the effects observed in human H4 cells.
We next wanted to test whether the trehalose-dependent increase in APP-FL and APP-CTFs is due to altered degradation. To inhibit de novo protein biosynthesis, H4 cells were incubated with cycloheximide and chased in the absence or presence of trehalose. In the absence of trehalose, levels of APP-FL and its CTFs declined within 2 h after block of de novo protein synthesis (Fig. 1F). In the presence of trehalose, levels of APP-FL also declined. Notably, levels of APP-CTFs were stabilized and even increased after 1 and 2 h of cycloheximide treatment in the presence of trehalose (Fig. 1H). The increase in APP-CTFs after block of protein biosynthesis could be explained by the ongoing processing of APP-FL by ␣and ␤-secretases. These data indicate that trehalose inhibits the degradation of APP-CTFs without inhibition of secretory processing of APP-FL.
Levels of A␤ derived from endogenous APP in the different cell lines were below the detection limit (not shown). Thus, we used previously described fibroblasts that express either APP-WT or the familial Alzheimer disease-associated Swedish mutant variant of APP (APP-SW) (30). As expected, levels of A␤ in conditioned media from cells expressing the APP-SW variant were much higher than that in media from APP-WTexpressing cells (Fig. 2, A and B). Incubation of cells with trehalose reduced the secretion of the different length variants A␤38, A␤40, and A␤42 from both APP-WT-and APP-SWexpressing cells (Fig. 2, A and B). Levels of cell-associated or intracellular A␤ were very low. Only A␤40 could be detected in lysates of APP-SW-expressing cells. Here, levels were slightly increased upon cell treatment with trehalose (Fig. 2B). These data demonstrate that trehalose decreases the secretion of A␤ and impairs the degradation of APP-CTFs.
Trehalose Does Not Inhibit ␥-Secretase Activity or Stimulate mTOR-dependent Autophagy-␥-Secretase is critically involved in the cleavage of APP-CTFs and A␤ generation. Thus, we next tested the potential effect of trehalose on ␥-secretase activity. Purified membranes of H4 cells were incubated in the presence or absence of trehalose, and ␥-secretase activity was monitored by the detection of the APP intracellular domain (AICD) released into the supernatant. In control samples, AICD was readily detected, indicating efficient cleavage of APP-CFTs by ␥-secretase (Fig. 3, A and B). As expected, pharmacological inhibition of ␥-secretase by the specific inhibitor N-[N- (3,5-difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester efficiently reduced the generation of AICD (Fig. 3, A and B). However, trehalose did not decrease the formation of AICD (Fig. 3, A and B), demonstrating that trehalose does not inhibit ␥-secretase activity. To further test whether trehalose induced accumulation of APP-CTFs independently of ␥-secretase inhibition, we next used fibroblasts from WT and PSdKO mice. PSdKO cells lack both catalytically active variants of presenilins and thus have no ␥-secretase activity. Accordingly, basal levels of APP-CTFs were higher in PSdKO than in WT cells (Fig. 3, C and D). Interestingly, trehalose further increased APP-CTF levels in PSdKO cells (Fig. 3, C and D), indicating that the observed accumulation of APP-CTFs upon cell treatment with trehalose is not caused by inhibition of ␥-secretase activity. The combined data demonstrate that trehalose decreases the degradation of APP-CTFs without inhibiting ␥-secretase.
We next analyzed the effects of trehalose on LC3, which is widely used as a marker for autophagosomes. Upon induction of autophagy, LC3 is converted from a cytosolic form (LC3-I) to a phosphatidylethanolamine-conjugated form (LC3-II) and thereby recruited to membranes of autophagosomes (31). Consistent with previous findings (21,22), trehalose induced a strong increase of LC3-II levels over time (Fig. 4A), resulting in elevated LC3-II/LC3-I ratios (Fig. 4B). However, this result would indicate that trehalose either increases autophagosome formation and/or decreases autophagic flux. To further verify the effects of trehalose on autophagy, we assessed p62 levels by immunoblotting. p62 binds directly to LC3 and is also degraded during autophagy (32). Consistent with increased LC3-II, we detected a significant increase of p62 upon treatment with trehalose (Fig. 4, A and C). Thus, the accumulation of p62 supports a decreased clearance capacity of trehalose-treated cells. Interestingly, immunocytochemistry revealed that accumulated LC3 partially co-localized with APP-positive vesicles upon tre-halose treatment, whereas both proteins were segregated in control cells (Fig. 4, D-F).
Although trehalose is commonly used to modulate autophagic activity, very little is known regarding how it regulates the autophagic machinery. We next analyzed proteins involved in the induction of autophagy, including the mTOR found in mTOR complexes and beclin-1 (BECN1), a major regulator protein of the class III phosphatidylinositol 3-kinase (PI3K) (8,33). Western blotting analyses revealed no significant changes in the levels of BECN1 (Fig. 5, A and B), total mTOR, and phosphorylated mTOR (Ser-2448) (Fig. 5, A, C, and D) upon cell treatment with trehalose, suggesting that trehalose did not affect these proteins involved in the induction of autophagy. The results are consistent with previous findings showing that effects of trehalose are independent of mTOR activity (21).
To further analyze the role of trehalose in the autophagic process and the metabolism of APP, we used Atg5KO cells. Atg5 plays a role at early steps of autophagosome formation and is required for the lipidation of LC3 (31). Accordingly, cells deficient for Atg5 are defective in the formation of autophagosomes (34). Fibroblasts of WT and Atg5-KO mice were treated with trehalose, and LC3 and APP were analyzed by Western immunoblotting. In WT cells, trehalose induced an increase in  LC3-II, resulting in an increased LC3-II/LC3-I ratio (Fig. 5, E  and F). In contrast, LC3-II was not detected in Atg5 KO cells. Due to defective conversion of LC3-I to LC3-II, LC3-I levels were strongly elevated in Atg5-KO cells (Fig. 5, E and F). Interestingly, the treatment of Atg5KO cells with trehalose still led to accumulation of APP-FL and particularly of its CTFs (Fig. 5,  E and G). The response of Atg5KO cells was very similar to that of WT cells, indicating that the effects of trehalose on APP metabolism are independent of Atg5-dependent induction of autophagy. Together with the unchanged levels of BECN1 and mTOR phosphorylation (Ser-2448) in trehalose-treated cells, increases in LC3-II, APP-CTFs, and p62 strongly suggest alterations of vesicular transport and/or autophagic flux upon cell treatment with trehalose. Trehalose also decreased the degra-dation of long lived proteins (Fig. 5H), further indicating an impairment of autophagic flux.
Inhibition of APP-CTF Degradation Is Selective for the Disaccharide Trehalose-The intracellular pathways by which trehalose is metabolized and regulates cell signaling are still unknown. Because trehalose is a disaccharide composed of two D-glucose residues, we tested whether supplementation of glucose induces similar effects on the metabolism of APP and LC3. H4 cells were incubated with trehalose or additional glucose in the culture media. As observed before, trehalose treatment led to strong accumulation of LC3-II, APP-FL, and APP-CTFs. In contrast, glucose treatment did not change the ratio of LC3-II/ LC3-I (Fig. 6, B and C). Similarly, glucose also did not alter the ratio of APP-CTFs/APP-FL (Fig. 6, B and D). Together, the data indicate that the effects of trehalose on autophagy-lysosomal function did not involve increased generation of glucose.
Maltose is a disaccharide similar to trehalose that is also composed of two glucose molecules. The two disaccharides only differ in the type of glycosidic linkage between the two glucose residues. Although the glucose residues in maltose are coupled by an ␣,␣-1,4 linkage, they are coupled by an ␣,␣-1,1 glycosidic linkage in trehalose (35) (Fig. 6A). Therefore, we tested whether trehalose-induced effects could be mimicked by the disaccharide maltose. However, in contrast to trehalose, maltose did not affect the levels of LC3 and APP (Fig. 6, E-G).
These findings indicate that the inhibition of LC3 and APP degradation is specific for the disaccharide trehalose.
Trehalose Alters the Subcellular Localization and Degradation of APP-CTFs-The present data strongly suggest an inhibitory effect of trehalose on APP metabolism in endolysosomal compartments. To test whether direct inhibition of lysosomal degradation could mimic the effects of trehalose on LC3 and APP, H4 cells were treated with trehalose or different lysosomal inhibitors including bafilomycin A1 and chloroquine. Bafilomycin A1 is a specific inhibitor of the vacuolar-type H ϩ -ATPase, which inhibits acidification and therefore protein degradation in lysosomes (36). Chloroquine is a lysosomotropic compound that also elevates/neutralizes the intraluminal pH of endolysosomal compartments (37). Both bafilomycin A1 and chloroquine led to accumulation of LC3-II and APP-CTFs (Fig.  7, A-C). However, the effects of bafilomycin A1 and chloroquine differed quantitatively from that of trehalose. Although LC3-II showed strongest accumulation upon direct inhibition of lysosomal activity with bafilomycin A1 or chloroquine, APP-CTF showed strongest accumulation upon cell treatment with trehalose (Fig. 7, A-C).
Cat D is an abundant aspartic endopeptidase in lysosomes (38,39). The precursor form of Cat D is proteolytically processed to an intermediate during its transport from the Golgi to acidic endolysosomal compartments. In the lysosomes, the intermediate Cat D is further processed to generate the active fragment as the C-terminal heavy chain (40,41). Thus, we assessed the effect of trehalose on the proteolytic processing of Cat D. Interestingly, cell treatment with trehalose strongly  (Fig. 7A). Accordingly, ratios of fully processed active Cat D to precursor and intermediate forms were strongly reduced upon cell incubation with trehalose (Fig. 7D). Consistent with previous reports showing that Cat D processing is affected by lysosomal pH (40), bafilomycin A1 and chloroquine also decreased the levels of the active forms of Cat D (Fig. 7, A and D).
GFP-LC3 is being widely used as a fluorescent marker of autophagosomes. However, GFP is acid-sensitive, and thus its fluorescence decreases in acidic compartments, i.e. when GFP-LC3-containing autophagosomes fuse with endosomes and lysosomes (26). In contrast, the monomeric red fluorescent protein mCherry is acid-stable. Therefore, a tandem fusion of the acid-insensitive red fluorescent protein (mCherry) and the acid-sensitive GFP at the N terminus of LC3 is a useful tool to assess fusion of LC3-positive autophagic vesicles with lysosomes by microscopy (37). H4 cells stably expressing mCherry-GFP-LC3 were cultured in the absence or presence of trehalose. As additional controls, cells were also incubated with chloroquine or starved in EBSS medium. Signals for mCherry and GFP were analyzed by fluorescence microscopy. Control cells showed predominantly mCherry-positive puncta but very few GFP-positive puncta (Fig. 8A). Quantification revealed that the number of mCherry-positive puncta was approximately 10 times higher than that of GFP-positive puncta (Fig. 8B). This is consistent with efficient quenching of the GFP fluorescence in acidic autolysosomes, whereas the mCherry fluorescence is not affected in these compartments. Interestingly, cell treatment with trehalose not only increased the number of GFP-positive puncta as compared with control cells but also led to a large overlap with mCherry-positive structures (Fig. 8, A and B), indicating inefficient quenching of the GFP signal. Cell incubation with chloroquine had very similar effects as trehalose (Fig. 8, A and B). Cell incubation with EBSS increased both mCherryand GFP-positive puncta compared with control cells (Fig. 8,  A and B). However, the number of mCherry-positive puncta was approximately 3 times higher than the number of GFP-LC3-positive puncta (Fig. 8B). As the incubation with EBSS enhances the induction of autophagy, LC3-I is efficiently converted to LC3-II and associates with autophagosomes. However, the efficient acidification of lysosomes and autolysosomes could lead to quenching of the GFP signal, thereby explaining the higher number of mCherry-positive vesicles as compared with that of GFP-positive puncta. The combined data strongly indicate that trehalose impairs the delivery of LC3 to and its degradation in lysosomal compartments.
To further test alterations in the subcellular distribution of APP-CTFs upon cell treatment with trehalose, we performed fractionation of vesicular compartments. In control cells, mature full-length APP and CTFs were detected in LC3-, Rab7-, Rab9-, and LAMP2-positive fractions (Fig. 9). The treatment with trehalose resulted in redistribution of APP-CTFs, Rab7, Rab9, and LAMP2. Cell incubation with bafilomycin A1 also led to accumulation of APP-CTFs in fractions that clearly segregated from APP-CTF-positive fractions from trehalosetreated cells (Fig. 9). The combined treatment of cells with trehalose and Baf A1 resulted in an APP-CTF distribution similar to that from cells treated with Baf A1 alone. Fractionation of marker proteins for the trans-Golgi network (TGN46) and endoplasmic reticulum (calnexin) was not affected by the treatment with trehalose or Baf A1. These data suggest that cell treatment with trehalose affects the subcellular distribution and/or trafficking of APP in late endosomal and lysosomal compartments.

Discussion
The present data demonstrate that trehalose decreases clearance of APP-CTFs and the secretion of A␤. Cell biological experiments showed redistribution of APP-CTFs and other endosomal and lysosomal proteins upon trehalose treatment. Trehalose also led to accumulation of LC3-II and decreased the processing of Cat D. These data are consistent with impaired vesicular transport and fusion of endosomal and lysosomal compartments. The cleavage of APP by ␣and ␤-secretases can occur at distinct cellular locations. Although ␣-secretase could predominantly cleave APP in secretory vesicles and at the cell surface, ␤-secretase has an acidic pH optimum and shows preferential cleavage in acidic compartments, including late endosomes and lysosomes. The C-terminal fragments of APP generated by ␣and ␤-secretase cleavage represent substrates for ␥-secretase that resides predominantly in endolysosomal compartments or at the cell surface (42)(43)(44)(45)(46). Although trehalose led to strong accumulation of APP-CTFs, we observed a rather decreased secretion of A␤. Because trehalose did not inhibit ␥-secretase, the findings suggest that trehalose rather decreases the delivery of the APP-CTFs to vesicular compartments containing ␥-secretase activity. We also observed decreased proteolytic activation of cathepsin D that also occurs during endosomal transport and delivery to lysosomes upon cell treatment with trehalose. Subcellular fractionation indeed showed redistribution of APP-CTFs together with marker proteins of late endosomal (Rab7 and Rab9) and lysosomal (LAMP2) compartments. The molecular mechanisms underlying these effects remain to be determined in more detail. However, the effects of trehalose on subcellular distribution differ from those induced by bafilomycin A1, suggesting that trehalose does not or at least does not only alter lysosomal pH regulation.
Additional experiments using a tandem fusion mCherry-GFP-tagged LC3 also demonstrated significant changes in the transport of LC3 to lysosomes upon trehalose treatment. Although control cells showed predominantly mCherry puncta due to quenching of GFP in efficiently acidified autolysosomes, trehalose treatment increased the co-localization of mCherry and GFP. These data indicate inefficient delivery of LC3 to lysosomes.
Trehalose is a disaccharide composed two D-glucose molecules linked by an ␣,␣-1,1 glycosidic linkage (35). Therefore, we tested whether trehalose-regulated effects might be caused by potential metabolites of trehalose or could be mimicked by other disaccharides. However, neither addition of glucose itself nor addition of the disaccharide maltose had overt effects on APP and LC3 levels, indicating that the observed effects are specific for trehalose and not due to increased supply with glucose. Our data also indicate that the effects of trehalose are independent of mTOR, BECN1, and Atg5. The data thus confirm and extend that trehalose-induced effects do not involve changes in mTOR complex 1 activity (21).
The present findings are surprising because trehalose has been shown to decrease cellular levels of the neurodegenerative disease-associated proteins huntingtin, ␣-synuclein, and phosphorylated tau in cellular and animal models (22)(23)(24), suggesting that trehalose promoted autophagy and lysosomal clearance of these proteins. A possible explanation for this contradiction might emerge from a recent study demonstrating the involvement of unconventional secretion of aggregated ␣-synuclein via exophagy in cases of impaired autophagosome-lysosome fusion or lysosomal function. Importantly, trehalose was shown to enhance ␣-synuclein secretion in PC12 catecholaminergic nerve cells (47). In addition, increased secretion of tau was also observed upon starvation and lysosomal inhibition in primary neurons, likely caused by release of autophagosomal contents into extracellular fluids (48). In our experiments, cellular levels of endogenously expressed tau and ␣-synuclein were not affected by trehalose (data not shown). Both proteins were not detected in conditioned media. However, we observed elevated levels of an overexpressed reporter protein for poly(Gln)-containing aggregates upon cell treatment with trehalose (data not shown), suggesting that trehalose might increase the release of protein aggregates. Secretion of pathogenic poly(Gln) protein from cultured cells has been reported previously (49).
Decreased levels of neurodegeneration-associated protein aggregates upon trehalose treatment might also result from upregulation of the ubiquitin-proteasome system. In contrast to APP and its CTFs, tau, huntingtin, and ␣-synuclein are cytosolic proteins. It has been shown that inhibition of the proteasome by epoxomicin increased the levels of ␣-synuclein and phosphorylated tau and correlated with enhanced cell death (22). Thus, although autophagy is considered as the major pathway to degrade protein aggregates, a contribution of proteasomal degradation to the reduction of aggregate-prone proteins could not be excluded. APP as an integral membrane protein is mainly targeted to endosomal-lysosomal compartments via endocytosis (42) and eventually degraded by lysosomal hydrolases (17-19, 44, 50, 51).
Decreased formation of cytosolic aggregates and lower toxicity were also observed with a mouse model for Huntington disease upon treatment with trehalose (21). However, this effect was at least partially attributed to the function of trehalose as a chemical chaperone that could bind to expanded poly(Gln) stretches and thereby stabilize the partially unfolded protein (52,53).
Trehalose appears to exert complex functions in cellular protein homeostasis. The effects of trehalose on vesicular trafficking and protein metabolism shown in this study should be considered not only for the interpretation of experimental results obtained with this disaccharide but also for its application in future therapeutic approaches. Author Contributions-N. T. T. and I. K. contributed equally to this study, performed and designed the experiments, and analyzed data. I. Y. T. also designed and analyzed the experiments. J. W. conceived the study, designed experiments, interpreted data, and wrote the manuscript. All authors read and edited the manuscript.