Vitamin E: Curse or benefit in Alzheimer's disease? A systematic investigation of the impact of α-, γ- and δ-tocopherol on Aβ generation and degradation in neuroblastoma cells

Abstract

Objectives

The E vitamins are a class of lipophilic compounds including tocopherols, which have high antioxidative properties. Because of the elevated lipid peroxidation and increased reactive oxidative species in Alzheimer's disease (AD) many attempts have been made to slow down the progression of AD by utilizing the antioxidative action of vitamin E. Beside the mixed results of these studies nothing is known about the impact of vitamin E on the mechanisms leading to amyloid-β production and degradation being responsible for the plaque formation, one of the characteristic pathological hallmarks in AD. Here we systematically investigate the influence of different tocopherols on Aβ production and degradation in neuronal cell lines.

Measurements

Beside amyloid-β level the mechanisms leading to Aβ production and degradation are examined.

Results

Surprisingly, all tocopherols have shown to increase Aβ level by enhancing the Aβ production and decreasing the Aβ degradation. Aβ production is enhanced by an elevated activity of the involved enzymes, the β- and γ-secretase. These secretases are not directly affected, but tocopherols increase their protein level and expression. We could identify significant differences between the single tocopherols; whereas α-tocopherol had only minor effects on Aβ production, δ-tocopherol showed the highest potency to increase Aβ generation. Beside Aβ production, Aβ clearance was decreased by affecting IDE, one of the major Aβ degrading enzymes.

Conclusions

Our results suggest that beside the beneficial antioxidative effects of vitamin E, tocopherol has in respect to AD also a potency to increase the amyloid-β level, which differ for the analysed tocopherols. We therefore recommend that further studies are needed to clarify the potential role of these various vitamin E species in respect to AD and to identify the form which comprises an antioxidative property without having an amyloidogenic potential.

Introduction

Tocopherols are lipophilic compounds consisting of various methylated phenols. They are essential for humans and animals and are found in all cellular membranes (1). Vitamin E comprises three homologous series, tocopherol, tocotrienol and tocomonoenol, which differ in the saturation grade of the acylchain. Tocopherols are saturated, trienols are three times unsaturated having an acylchain being composed of three isoprene units, and the monounsaturated tocomonoenol. Additionally, the chroman ring of tocopherols and tocotrienols can be methylated at position C5 and C7 leading to α-, β-, γ-, δ- tocopherol or toctotrienol (figure 1A). The a-form has two methyl groups at position C5 and C7, the β- and y-form contain one methyl group whereas the δ-form is not methylated at position C5 or C7. Due to their three chiralic centers (at position C 2, 4’ and 8’) each tocopherol has 8 possible isomers. However, only the RRR isoform is produced by photosynthetically active organisms like plants or cyanobacteria whereas synthetically produced vitamin E often consists of racemats (2). The α-tocopherol was long time considered as the only significant member of the vitamin

Figure 1.

Effects of Tocopherols on Aβ production. Error bars represent standard deviation of the mean. Asterisks show the statistical significance calculated by unpaired Student’s t-test compared to solvent control. A: Skeletal formulas of α-, β-, γ- and 5-tocopherol. B: Cell viability of SH-SY5Ywt cells after incubation with tocopherols (n>4). C: Effect of α-, γ- and 5-tocopherol on Aβ-production in SH-SY5Y APP695 cells compared to control (n=5). A dashed line indicates that samples were not loaded next to each other

E group and other members were mainly ignored. Indeed a-tocopherol is preferentially absorbed and accumulated in humans (3). The “vitamin E activity” given in international units is therefore based on the fertility enhancement quantified by the prevention of miscarriages in pregnant rats compared to a-tocopherol. The a-tocopherol is mainly found in olive and sunflower oils and is the major form used in supplements (4). In contrast to the typical European diet, where a-tocopherol is the most present form, y-tocopherol is more pronounced in a more typical American diet due to higher consumption of corn and soybean oil (5). The U.S. dietary reference intake (DRI) recommended daily amount (RDA) is 15 mg per day of vitamin E for 25 year old males. The Scientific Comitee of food of the European Food Safety Authority (EFSA) has decided to set the tolerable upper intake level of vitamin E for adults of 300 mg a-tocopherol equivalents per day. Intake of 26.7 mg (40 IU) per day of tocopherol is considered as normal intake, up to 400 IU (267 mg) as supplemental and application rates higher than 400 IU as pharmacological dose sometimes also named megadose (6). Doses over 3000 IU are known to cause unwanted effects as e.g. abdominal cramps, dizziness, bleedings, worsening diabetes mellitus, coagulation defects and increased cholesterol levels (7). Normal plasma level of tocopherol are 11.6-30.8 µmol/l (8) and vitamin E is known to pass the blood-brain-barrier and long term treatment increases brain concentration of tocopherols (9).

Alzheimer’s disease (AD) is a devastating neurodegenerative disorder; the pathology is characterized by extracellular senile plaques, mainly consisting of aggregated β-amyloid (Aβ), a 40-42 amino acid long peptide (10). Aβ is generated by sequential proteolytic cleavage of the amyloid precursor protein (APP) involving β- and y-secretase (reviewed in (11)). Both, APP and the secretases are transmembrane proteins, the y-secretase cleavage even occurs within the membrane. Several lines of evidences therefore suggest that the lipid environment plays a crucial role in the processing of APP (1218). Plasmalogens and polyunsaturated fatty acids (PUFAs) like docosahexaenoic acid (DHA) are altered in AD brains and are able to decrease APP processing by several synergistic mechanisms leading to reduced Aβ (15, 19). Moreover, it has been reported that Aβ increases the production and formation of reactive oxidative species (ROS) and free radicals (20). Furthermore, prooxidant species such as reactive nitrogen intermediates are elevated by the Aβ mediated stimulation of inflammatory processes and microglia cells (21). PUFAs and plasmalogens are prime targets for ROS and free radicals or peroxide radicals (22). Because tocopherols are natural antioxidants acting as scavengers of radicals and peroxides and protecting lipids from lipid peroxidation in membranes (23), vitamin E has been proposed to be beneficial in AD (24). However, nothing is known about the effect of vitamin E as a highly lipophilic compound, being itself part of the membrane, on the impact on Aβ generation and degradation. In addition, evidences exist that beside the antioxidant effect of vitamin E, tocopherol influences the expression pattern of several proteins in the brain (25, 26). Interestingly this effect seems to be dependent on the individual tocopherol. In this study we aim to investigate the effect of different tocopherols on the Aβ level and to identify potential underlying mechanisms in neuroblastoma cell lines.

Materials and Methods

Chemicals

(+)-α-Tocopherol, (+)-γ-Tocopherol and (+)-δ-Tocopherol, further chemicals and cell culture media were purchased from Sigma (Taufkirchen, Germany) if not stated otherwise.

Cell Culture

SH-SY5Y wildtype (wt), SH-SY5Y APP (overexpressing the human APP695 isoform) and SH-SY5Y C99 cells (overexpressing the β-cleaved C-terminal fragment) [27] were cultivated in Dulbecco’s modified Eagle’s medium (DMEM) containing 10 % fetal calf serum (FCS) (PAN Biotech, Aidenbach, Germany). For Neuro 2a (N2a) cells 10 % FCS/DMEM was supplemented with 1 % Penicillin/Streptomycin solution, 2 mM L-Glutamine, 0.1 mM MEM and 1 mM

Sodium-Pyruvat. In addition, Hygromycin B (PAN Biotech, Aidenbach, Germany) was added in a final concentration of 0.3 mg/ml for cultivating SH-SY5Y APP, SH-SY5Y C99 and stably transfected N2a IDE-knock down cells.

Before incubation, cells were treated for 12 h with 0.1 % FCS/DMEM. Tocopherols (dissolved in ethanol) were incubated 10 in 0.1 % FCS/DMEM for 24 h (8+16h); controls were treated with ethanol in a final concentration of 1 %% corresponding to the concentration in the incubation media. After incubation, cells were lysed chemically in lysis buffer (0.1 % NP-40, 0.1 % Triton-X-100, 10 mM Tris, 2 mM EDTA) with protease inhibitor (Roche Diagnostics, Mannheim, Germany) or mechanically using a PotterS homogenizer (B. Braun, Melsungen, Germany) depending on further usage.

N2a IDE-knock down cells

N2a cells were stably transfected with sh-RNA KM-05515H (Qiagen, Hilden, Germany) according to manufacturer’s informations. In these cells, IDE protein level is reduced to 13.1 % (± 5.4 %; p<0.001) (figure 4C).

Figure 4.

Effects of tocopherols on Aß degradation. A: Increased protein level of non-degraded human Aß incubated in combination with α-, γ-, or δ-tocopherol on murine N2a cells (n≥4). B: Protein level of non-degraded human Aß incubated in combination with α-, γ-, or δ-tocopherol on murine N2a IDEknock down cells (n>4). C: Remaining IDE protein level in IDE-knock down N2a cells (n=3). Statistical significance as described for figure 1

Determination of protein concentration

Protein concentration was measured using bicinchoninic acid according to (28). For use in experiments, samples were adjusted to equal protein amount.

Western blot analysis

Total protein concentration in samples was determined as mentioned above and samples were adjusted to an equal concentration afterwards. For sodium dodecyl sulfate Polyacrylamid gel electrophoresis (SDS-PAGE) 10 % - 20 % Tricine gradient gels (Anamed, Groß-Bieberau, Germany) were used and proteins were transferred onto nitrocellulose membranes (Whatman, Dassel, Germany). For Western blot (WB) analysis, following antibodies were employed: anti-BACE1 (B0806) (1:1000, Sigma, Taufkirchen, Germany), anti-Presenilin 1 (1:500, sc 7860) (Santa Cruz, Dallas, USA), anti-IDE (0.5 µg/ml, ST1120) (Calbiochem, Darmstadt, Germany), anti-sAPP beta (1:250, MBS492139) (MyBioSource, San Diego, USA) and W02 antibody (5 µg/ml) (Millipore, Billerica, MA, USA) as primary antibodies, anti-rabbit (W401B) (1:5000, Promega, Mannheim, Germany) and anti-mouse (P0260) (1:10000, Dako, Hamburg, Germany) as secondary antibodies. Aß analysis was performed as described earlier (29). To detect secreted Aß in WB analysis, immunoprecipitation with 20 µl protein G-Sepharose (Sigma, Taufkirchen, Germany) and W02 antibody (5 µg/ml) was conducted. Proteins were detected by ECL-method (Perkin Elmer, Rodgau-Jügesheim, Germany). Densitometric quantification was performed with Image Gauge V3.45 software.

Preparation of purified membranes

SH-SY5Ywt cells were washed three times with phosphate buffered saline (PBS, precooled to 4 °C) and scraped in sucrose buffer (10 mM Tris/HCl, pH 7.4, 1 mM EDTA, 200 mM sucrose). To homogenize cells Minilys (Peqlab, Erlangen, Germany) was used at maximum speed for 30 seconds. Protein concentration was determined and equaled. After centrifugation at 900 rcf for 10 min at 4 °C, supernatants were collected in a new tube and centrifuged (Optima MAX Ultracentrifuge, Beckman Coulter, Krefeld, Germany) for 75 min at 55,000 rpm and 4 °C. Pellets were resuspended using Minilys (Peqlab, Erlangen, Germany) at medium speed for 10 seconds in sucrose buffer (15).

Endosome isolation

Endosome isolation was performed as described in (17). Briefly, endosomes of mechanically lysed SH-SY5Y cells were isolated in compliance with the application sheet S20 from Optiprep (Axis-Shield, Oslo, Norway) by density gradient centrifugation (gradients adjusted to 5 mg protein) for 18 h at 16,800 rpm and 4 °C. Fractions were collected in 550 µl steps. EEAl-positive fractions were pooled and 100 µl of endosomal fraction was transferred to a 96-well plate for determination of ß-secretase activity.

Measurement of ß- and y-secretase activity

Determination of ß- and y-secretase activity has been described in (13). Briefly, fluorogenic ß-secretase substrate IV (Calbiochem, Darmstadt, Germany) or y-secretase substrate (Calbiochem, Darmstadt, Germany) was added in a final concentration of 20 or 10 to samples, respectively. Resulting fluorescence was determined continuously at an excitation wavelength of 345 ± 5 nm and an emission wavelength of 500 ± 2.5 nm for ß-secretase as well as an excitation wavelength of 355 ± 10 nm and an emission wavelength of 440 ± 10 nm for y-secretase under light preclusion with a Safire2 Fluorometer (Tecan, Crailsheim, Germany).

Measurement of secretase activity in living cells

After incubation with tocopherols, SH-SY5Y cells were washed with HEPES buffer. 30 fluorogenic ß-secretase substrate (Calbiochem, Darmstadt, Germany) or 12 y-secretase substrate (Calbiochem, Darmstadt, Germany) in 50 of cell imaging solution (140 mM NaCl, 5 mM KCl, 8 mM CaCl2, 1 mM MgCh, 20 mM HEPES, pH 7.4) was added after washing. Under light exclusion and at 37 °C, fluorescence was measured in a Safire2 Fluorometer (Tecan, Crailsheim, Germany). Excitation and emission wavelength for the substrates are described above.

Determination of total Aß degradation

N2a wt and N2a IDE-knock down cells were pre-incubated for 6 h in 0.1 % FCS/DMEM. After 18 h of incubation with tocopherols (10 µM), cells were additionally treated for 6 h with tocopherols combined with 1 µg/ml human Aß40. Cell culture supernatant was used in SDS-PAGE separation and remaining human Aß40 was detected in Western blot analysis using the W02 antibody as described above.

Quantitative real-time experiments

Total RNA was isolated from SH-SY5Y cells using TRIzol Reagent (Invitrogen, Karlsruhe, Germany) and 2 of total RNA were reverse transcribed with High-Capacity cDNA Reverse Transcription Kit (Life Technologies, Darmstadt, Germany) according to manufacturers’ protocols. Quantitative real-time PCR analysis was performed using Fast SYBR Green Master Mix (Applied Biosystems, Darmstadt, Germany) on Piko Real Real-Time PCR System (Thermo Scientific, Waltham, USA). The obtained results were normalized to β-actin gene expression and expression changes were calculated using the 2^-(ΔΔCt) method [30]. The following primers (purchased from Eurofins MWG Operon, Eberberg, Germany) were used for qPCR: beta-site APP-cleaving enzyme 1: forward CTTCGTTTGCCCAAGAAAGT, reverse ACACCAGCT GCTCTCCTAGC; presenilin 1: forward CTCAATTCT GAATGCTGCCA, reverse GGCATGGATGACCTTATA GCA; presenilin 2: forward GATCAGCGTCATCGTGGTTA, reverse GGAACAGCAGCATCAGTGAA; presenilin enhancer: forward CATCTTCTGGTTCTTCCGAGAG, reverse AGAAGAGGAAGCCCACAGC; nicastrin: forward CTGTACGGAACCAGGTGGAG, reverse GAGAGGCTG GGACTGATTTG; anterior pharynx defective 1a: forward GCCTCTGTGGTCTGGTTCAT, reverse TCTGCCTTCT TAAGCAGCTTGT; anterior pharynx defective 1b: forward ATCGCCGGAGCTTTCTTC, reverse TTTCTGTGTTG GTCCATCTTTG; β-actin: forward CTTCCTGGGCA TGGAGTC, reverse AGCACTGTGTTGGCGTACAG.

Cell Viability

Cell viability after incubation with tocopherols was determined by measuring lactate dehydrogenase activity using the Cytotoxicity Detection KitPLUS (Roche Diagnostics, Mannheim, Germany) according to manufacturer’s protocol.

Statistical analysis

Data represent an average of at least three independent experiments. Error bars represent standard deviation of the mean. Statistical significance of differences between tocopherol treated samples to control samples was calculated using two-tailed Student’s t-test. Significance was set at *p < 0.05, **p < 0.01 and ***p < 0.001. For multiple comparison analysis of the different tocopherol species post hoc ANOVA was conducted. The normality of the data distribution was tested with Kolmogorov-Smirnov test (p>0.19).

Results

Effect of Tocopherols on Aβ level in SH-SY5Y APP695 cells

To analyse the effect of tocopherols on the Aβ level, we incubated stably transfected SH-SY5Y APP695 cells with 10 α-, γ- or δ-tocopherol. APP695 is an isoform of APP predominantly produced in neurons (31). All used tocopherols were the naturally occurring RRR isomers, in order to investigate the effect of the different tocopherol species independently from their chiralic structure. As ß-tocopherol was only available as racemat we excluded this tocopherol from this study. 10 of tocopherol is the lower limit of the physiological plasma concentration (8). At these concentrations all tocopherols showed no significant changes in cell viability compared to solvent treated control cells (figure 1B). The cell viability of all analysed tocopherols and the solvent control was >97% and the differences between the control and the tocopherols was <0.2% (figure 1B). Additionally, the total protein amount of control cells compared to tocopherol treated cells was unchanged (control: 5.1 ± 0.20 mg/ml; a-tocopherol: 5.7 ± 0.75 mg/ml, p=0.429; γ-tocopherol: 5.4 ± 0.50 mg/ml, p=0.542; ò-tocopherol: 5.6 ± 0.66 mg/ml, p=0.496) indicating that cell growth was not affected by tocopherols. Unexpectedly, all tocopherols showed a highly significant increase in the Aß level (a-tocopherol: 132.25 ± 4.45 %, p=0.0034, y-tocopherol: 134.22 ± 4.66 %, p=0.0026, ò-tocopherol: 136.86 ± 3.74 %, p=0.0011, mean: 134.44 ± 2.35 %, p=0.0038) (figure 1C). Our results suggest that, at least in cell culture, conditions exist where vitamin E, supplemented in physiological concentrations, has not only beneficial effects due to its antioxidative properties (32) but is also able to directly increase processes leading to AD, challenging the question whether vitamin E should be - independent of its isoform and species -unguarded advised to be ingested.

Effect of Tocopherols on y-secretase activity

The y-secretase is a multienzyme complex, consisting of at least four proteins, nicastrin, aphla or aphlb (anterior pharynx defective 1a or b), pen-2 (presenilin enhancer 2), and the presenilin with its proteolytically active aspartates (33., 34., 35.). The y-secretase cleaves C99, the C-terminal membrane-bound APP protein fragment which is already cleaved by P-secretase, and is therefore the last step in the Aβ generation. To analyse whether the elevated Aβ level in presence of tocopherols are caused by increased y-secretase processing, we incubated SH-SY5Y cells stably transfected with C99. As the resulting Aβ production in these cells is only dependent on y-secretase activity, the Aβ levels reflect potential alterations caused by y-secretase processing. γ- and δ- tocopherols increased the Aβ level compared to control, and statistical analysis of a-tocopherol did not show a significant effect (a-tocopherol: 109.35 ± 6.43 %, p=0.2428, y-tocopherol: 139.34 ± 7.13 %, p=0.0012, 5-tocopherol: 144.64 ± 8.35 %, p=0.0012, mean: 131.11 ± 5.71 %, p=0.0003) (figure 2A). Interestingly, using multicomparison post hoc analysis also revealed significant changes between a-tocopherol and the other tocopherols (supplement table S1), suggesting that a-tocopherol has the weakest effect on y-secretase dependent Aβ generation. Similar results were obtained by directly analysing y-secretase activity in living SH-SY5Y wt cells (a-tocopherol: 114.97 ± 3.1 %, p=0.0004, y-tocopherol: 116.22 ± 3.40 %, p=0.0004, 5-tocopherol: 135.21 ± 2.94 %, p=6.3x10^-12, mean: 122.13 ± 2.11 %, p=5.1x10^-9) (figure 2B). In line with the Aβ level obtained by C99 transfected cells, a-tocopherol showed the lowest and 5-tocopherol the most pronounced effect on y-secretase activity. The difference between a- and 5-tocopherol was significant as shown in supplemental table S1. Contrary, incubating purified membranes containing the y-secretase complex with tocopherols, revealed no effect (a-tocopherol: 98.2 ± 2.0 %, p=0.698, y-tocopherol: 102.2 ± 2.4 %, p=0.643, 5-tocopherol: 97.8 ± 1.2 %, p=0.619, mean: 99.4 ± 1.1 %, p=0.8963) (figure 2C). Our results suggest that tocopherols have no direct effect on the enzyme activity caused e.g. by direct interaction, but that the tocopherol-induced observed alterations are rather mediated via cellular processes depending on metabolically active cells. We therefore tested whether the protein level of presenilin 1 (PS1), the protein including the active centre of the y-secretase (36), is changed after tocopherol incubation. Indeed in presence of tocopherols the PS1 protein level was increased. Statistical analysis revealed significance for all analysed tocopherols in average and for a-tocopherol, whereas y- and 5- tocopherol showed a non-significant increase of PS1 protein level (a-tocopherol: 128.35 ± 8.14 %, p=0.029, y-tocopherol: 118.03 ± 10.3 %, p=0.1641, 5-tocopherol: 117.97 ± 8.56 %, p=0.1138, mean: 121.45 ± 4.84 %, p=0.0027) (figure 2D). Increased PS1 protein level was accompanied by an elevated gene expression in SH-SY5Y cells supplemented with vitamin E, in line with the westernblot results. However, only the combined results of the different tocopherols reached significance for PS1. Because the composition of the y-secretase is not completely known and varies (37, 38), the average of the gene expression of all proteins of the y-secretase complex is presented in figure 2E. The increased PS1 expression was accompanied by an increased expression of the average of all proteins of the y-secretase complex. (a-tocopherol: 131.14 ± 22.85 %, p=0.0197, y-tocopherol: 119.21 ± 20.77 %, p=0.0269, 5-tocopherol: 107.54 ± 18.81 %, p= 0.1666, mean: 119.30 ± 12.01 %, p=0.0454) (figure 2E). A summary table of all single proteins of the multiprotein complex is shown in supplementary table S2. Our results suggest that the increased Ap level found by supplementing SH-SY5Y APP695 cells with tocopherols is due to an increased y-secretase activity. The effect of 5-tocopherol on y-secretase was most pronounced and a-tocopherol showed the smallest effect. More mechanistically, the observed effect of tocopherols on y-secretase processing is not direct, but mediated by a change in PS1 protein level and increased expression of the y-secretase complex.

Figure 2.

Effects of tocopherols on y-secretase. A: Increased Aβ production in SH-SY5Y C99 cells after incubation with α-, γ-, or δ-tocopherol (n=5). B: Enhanced y-secretase activity in SH-SY5Ywt cells treated with α-, γ-, or δ-tocopherol (n=24). C: No significant changes in y-secretase activity in purified membranes of SH-SY5Ywt cells in presence of α-, γ-, or 5-tocopherol (n=6). D: Altered presenilin 1 protein level (n=3) and E: Average gene expression of proteins of the y-secretase complex in SH-SY5Ywt cells after incubation with ↑-, γ-, or 5-tocopherol (n≥5). Statistical significance as described for figure 1

Effect of Tocopherols on P-secretase

C99 is generated by P-secretase cleavage of APP, the initial step in the generation of Aβ, releasing also the soluble N-terminal fragment sAPPp (39, 40). sAPPp level are therefore dependent on β-secretase activity and reflect an alteration in P-secretase activity. Incubation of APP695 overexpressing SH-SY5Y cells with tocopherols revealed increased sAPPβ level for y- and 5-tocopherol and the average of tocopherols (a-tocopherol: 104.33 ± 5.28 %, p=0.5884, y-tocopherol: 125.44 ± 5.1 %, p=0.0045, 5-tocopherol: 131.39 ± 8.57 %, p=0.0079, mean: 120.39 ± 4.37 %, p=0.0173) (figure 3A). In line with the effects of tocopherols on y-secretase activity, a-tocopherol showed no significant and 5-tocopherol the highest potency to increase sAPPp. The differences in the effect strength between a-tocopherol and the other tocopherols were significant (supplementary table S1). Consistent with sAPPp level, the P-secretase activity showed comparable results (a-tocopherol: 103.12 ± 3.26 %, p=0.598, y-tocopherol: 125.76 ± 4.64 %, p=0.0005, 5-tocopherol: 125.68 ± 4.91 %, p=0.0009, mean: 118.19 ± 2.94 %, p=0.0019) (figure 3B and table S1) whereas incubation of tocopherols on purified membranes has, as already observed for y-secretase, no effect (a-tocopherol: 98.6 ± 1.1 %, p=0.614, y-tocopherol: 100.0 ± 1.8 %, p=0.993, 5-tocopherol: 99.1 ± 0.8 %, p=0.722, mean: 99.2 ± 0.7 %, p=0.775 (figure 3C). BACE1, the protein being identified as the P-secretase in brain [41], showed an increased gene expression. Although the effects between the individual tocopherols were not significant (see table S1), it is remarkable that the effect strength of the BACE1 gene expression seems to be inversely to the observed effect strength in the alterations observed for sAPPp or BACE1 activity. We assumed that additional effects beside a change in gene expression may contribute to the alterations in BACE activity. We therefore analysed the BACE1 protein levels in endosomes, the active cellular compartment of BACE1 [42, 43]. Indeed the BACE1 protein level in endosomes showed a more pronounced effect (ò-tocopherol: 147.4 ± 2.7 %, p=0.0002) (figure 3E) compared to the alterations measured by gene expression suggesting that protein stability or protein sorting might also be effected by tocopherols. The enhanced BACE1 protein level is accompanied by an increased ß-secretase activity in endosomes (figure 3F) further underlying the impact of endosomes in the ò-tocopherol mediated change in ß-secretase activity. In summary, tocopherols induced similar changes on ß-secretase compared to the effect on γ-secretase. α-Tocopherol revealed the lowest, δ-tocopherol the highest increase in ß-secretase activity or sAPPß level. The effect was not due to direct interaction as no effect was observed in purified membranes containing BACE1. Instead the expression level of BACE1 were slightly increased and the BACE1 protein level in endosomes of δ-tocopherol treated cells were elevated resulting in an increased ß-secretase activity suggesting a crucial role of tocopherol dependent endosomal BACE1 trafficking or degradation of BACE1.

Figure 3.

Effects of tocopherols on ß-secretase. A: Enhanced secretion of sAPPß by SH-SY5Y APP695 cells after incubation with α-, γ-, or δ-tocopherol (n=9). B: ß-secretase activity in SH-SY5Y living cells treated with a-, y-, or ò-tocopherol (n>15). C: ß-secretase activity in purified membranes of SH-SY5Y cells with addition of α-, γ-, or δ-tocopherol (n=3). D: BACE1 gene expression in SH-SY5Ywt cells incubated with α-, γ-, or δ-tocopherol (n>4). E: Increased BACE1 protein level in endosomes of SH-SY5Ywt cells after treatment with δ-tocopherol (n=3). F: Average kinetic of ß-secretase activity in endosomes of SH-SY5Y cells after δ-tocopherol incubation (n=3). Statistical significance is described in figure 1

Effect of Tocopherols on Aß degradation

Both, ß- and γ-secretases being responsible for the Aß production, showed significant changes between the individual tocopherols. However, these differences were not observed in the Aß level obtained by analysing SH-SY5Y APP695 cells, leading to the assumption that additional mechanisms contribute to the observed effect of the Aß level in these cells. Aß level are not only dependent on APP cleavage, generating Aß, but also on Aß degradation (44). To evaluate whether Aß degradation is affected by tocopherols, we incubated the murine neuronal cell line N2a with tocopherol and additionally supplemented with human Aß peptides. As the antibody used to detect Aß only recognizes human Aß, the detected Aß after incubation is independent of the endogenous produced Aß level but dependent on degradation. For that reason the remaining Aß is inversely correlated with the activity of the Aß degrading enzymes. After tocopherol treatment a-, ò-tocopherol and the average of all tocopherols showed a significant increase in the remaining non-degraded Aß. The increase after y-tocopherol incubation failed to reach significance (a-tocopherol: 134.93 ± 8.04 %, p=0.0171, y-tocopherol: 113.39 ± 7.01 %, p=0.1006, ò-tocopherol: 136.60 ± 8.00 %, p=0.0145, mean: 125.32 ± 5.11 %, p=0.0003) (figure 4A). In SH-SY5Y APP695 cells a-tocopherol showed a significant increase on Aß level. However, this increase could not be explained by Aß production suggesting that a-tocopherol acts dominantly via another mechanism. Indeed, a-tocopherol showed a more pronounced effect on Aß degradation. To identify the Aß degrading enzyme being responsible for the tocopherol mediated effect, knock down cells of insulin-degrading enzyme (IDE) were generated. These cells have a highly decreased IDE protein level (13.1 ± 5.4 % remaining, p<0.001, figure 4C). IDE (45, 46) is beside Neprilysin (47., 48., 49.) one of the best characterized Aß degrading enzymes. In these cell lines the effects of the tocopherols were highly diminished or not present (a-tocopherol: 113.67 ± 1.62 %, p=0.0027, y-tocopherol: 102.37 ± 2.86 %, p=0.5591, ò-tocopherol: 99.57 ± 1.81 %, p=0.8971, mean: 104.80 ± 2.07 %, p=0.1837) (figure 4B). These results suggest that beside Aß production also Aß degradation, mediated by IDE, is affected by tocopherols.

Discussion

Vitamin E is widely used as an antioxidant and supplemented in food, where mostly racemats, mixtures of

different tocopherol species, or α-tocopherol are used. Here we show that beside the protective antioxidative properties of vitamin E, especially of a-tocopherol, tocopherols are able to increase Aβ level in neuronal cell lines utilizing physiological plasma concentrations. Aβ level were increased by an elevated β- and y-secretase activity resulting in an increased Aβ production. The detected elevated secretase activities were not observed when tocopherols were incubated on purified membranes containing the secretases, but accompanied by an increase in the protein level and gene expression of these proteins. Therefore our results suggest that the mechanism of action by which the tocopherols increase secretase activity is not direct, e.g. by a direct interaction of the tocopherols and the proteins. Instead only in metabolically active cells tocopherols increase secretase activity. In principal, the observed effect on β- and y-secretase could be explained by a changed gene expression, protein translation, protein stability or protein sorting. Indeed altered PS1 and BACE protein level and expression was found after tocopherol incubation and the effect varies between the different tocopherols. Interestingly it has been reported that beside its antioxidative properties vitamin E is able to alter the expression level of several proteins in brain (25, 50) mediated by several transcriptional pathways such as PPARy (51) and NF-KB (52, 53). It might be hypothesized that these or similar pathways are differently affected by the individual tocopherols. A similar result has been found for β-secretase, at least for δ-tocopherol, additional contributing effects may exist leading to a strong increase of BACE protein levels and β-secretase activity in endosomes, the cellular compartment where β-secretase is mainly active (42, 43). It might be speculated that in respect to β-secretase BACE1 transport is affected by the different tocopherols. At least for a-tocopherol a potential crosstalk has been shown in Niemann-Pick-C disease mouse models and the endosomal/lysosomal pathway (54). However, additional experiments are necessary to clarify the exact role of endosomes in tocopherol mediated effects on β-secretase activity. Importantly, we found significant different effects of tocopherols on Aβ production in C99 overexpressing cells. a-Tocopherol has no or minor effects whereas o-tocopherol has the strongest effect on both, β- and y-secretase. The effect of y-tocopherol was mainly between α-and o-tocopherol. Taking into consideration that a-tocopherol is methylated at position C5 and C7, y-tocopherol at one position and δ-tocopherol is neither methylated at C5 nor C7, our results indicate that the effect strength on Aβ production correlates with the grade of methylation at the chroman ring. A different situation was found for Aβ degradation, where α- and δ-tocopherol have comparable effects leading to a decreased IDE mediated Aβ degradation. Besides IDE we cannot rule out that other Aβ degrading enzymes also subsidise to the observed effect. This might especially be the case for α-tocopherol because IDE-knock down cells show a reduced but still significant α-tocopherol dependent effect strength on Aβ degradation. The combined effect of increased Aß production and decreased Aß degradation results for all tocopherols in highly significant elevated Aß levels in SH-SY5Y cells. Nevertheless it has to be pointed out that this study has - because the experiments are performed in cell culture - its clear limitations and the results have to be confirmed in vivo. However, here we show for the first time indications that beside the protective antioxidative properties of tocopherols (32), additional undesirable effects in respect to AD exist. Moreover, our study emphasizes that it is necessary to differentiate between different species of vitamin E and that additional studies are necessary to investigate the effect of the racemats, the other isoforms or the effect of the degree of unsaturation in vitamin E. This might also be of special interest taking into consideration that vitamin E content in food as supplement is commonly declared in IU compared to its biological activity in fertilisation of a-tocopherol making it not possible to notice the exact composition of the vitamin E species.

Potential differences in the biological activity may also explain the mixed outcome of vitamin E studies used not only to treat AD (55), but also Parkinson Disease (reviewed in (56)), cancer including lung cancer (reviewed in (57)), prostate

cancer (reviewed in (58)), colon cancer (reviewed in (57, 59)), atheriosclerosis (reviewed in (60)), cardiovascular disease (reviewed in (61)), brain white matter lesions (62) and diabetes mellitus (reviewed in (63)).

In respect to AD meta-analysis revealed that the vitamin E status of AD patients is together with other nutrient levels significantly decreased (64). Recently a Finish cohort study showed that increased tocopherols and tocotrienols are associated with a reduced risk of cognitive impairment also indicating that the evaluation of the various vitamin E species might be important to understand the underlying correlations (65). Supplementation with antioxidative vitamin E and C for one year showed an increase of the antioxidants in CSF and a reduced lipid peroxidation. However, the clinical course of AD did not significantly differ between the vitamin and the control group (66). This is in line with the results found by Sung et al., reporting that Vitamin E supplementation in young but not aged transgenic AD mice reduces Aβ levels and amyloid deposition. Again, as a possible mechanism, the authors suggest, that vitamin E is mainly beneficial by reducing the oxidative stress, which is an important early event in AD pathogenesis (67). Moreover this might explain the differences in our results utilizing neuroblastoma cell lines in cell culture where the effect of reactive oxidative species on Aβ production is less pronounced or absent. Similar results were obtained by many different other studies (reviewed in (68)). However, another meta-analysis of the intake of antioxidants vitamin C, vitamin E and β-carotene revealed a protective effect in AD being most pronounced for vitamin E (69). Interestingly, a recent randomized clinical trial with vitamin E (800 IU/d of vitamin E (α-tocopherol) plus 500 mg/d of vitamin C plus 900 mg/d of a-lipoic acid (E/C/ALA); 400 mg of coenzyme Q 3 times/d) showed no influence on CSF biomarkers related to amyloid or tau pathology (70). However, some indications were also obtained that the oxidative stress in the brain was reduced but also that this treatment even raised the caution of faster cognitive decline (70).

In conclusion vitamin E has an obvious antioxidative effect which may also be beneficial under certain conditions in brain, where its antioxidative action was several times reported in literature. Nevertheless, here we could show that beside these beneficial properties vitamin E has also an Aß increasing potential caused by a decreased Aß degradation and an increased Aß production (figure 5), which differs in the single vitamin E species, where α-tocopherol has the weakest amyloidogenic potency. We therefore recommend that further studies are needed to clarify the potential role of these various vitamin E species in respect to AD and to identify the form which comprises an antioxidative property without having an amyloidogenic potential.

Figure 5.

Summary of effects of tocopherols on Aβ-production and clearance. Tocopherols enhance β-cleavage of APP via increased BACE expression and BACE1 level in endosomes. As a result, the secretion of sAPPβ is raised. The remaining β-CTF is further processed by the γ-secretase complex which leads to the production of Aβ peptides. By increasing γ-secretase expression, tocopherols enhance y-cleavage and, in combination with the aforementioned rise in β-cleavage, lead to elevated Aβ production. In addition, tocopherols lower IDE mediated degradation of Aβ peptides

Conflict of interest

The authors declare no conflict of interest.

Funding

According to the author guidelines, funding for the research leading to these results were received from: the EU FP7 project LipiDiDiet, Grant Agreement No. 211696 (TH), the DFG (HA2985/6-2) (TH), the Bundesministerium für Bildung, Forschung, Wissenschaft und Technologie via NGFNplus and KNDD (TH), and the HOMFOR (MG), the HOMFORexzellent 2011,2012 (MG) (Saarland University research grants).

Electronic supplementary material

Supplementary material is available for this article at 10.1007/s12603-015-0506-z and is accessible for authorized users.

Supplementary material, approximately 278 KB.