α-Tocopherol Modulates Non-Amyloidogenic Pathway and Autophagy in an In Vitro Model of Alzheimer’s Disease: A Transcriptional Study

Alzheimer’s disease (AD) is the most common form of dementia worldwide. The hallmarks of AD are the extracellular amyloid plaques, which are formed by amyloid β (Aβ) aggregates derived from the processing of the amyloid precursor protein (APP), and the intraneuronal neurofibrillary tangles, which are formed by the hyperphosphorylated tau protein. The aim of this work was to study the effects of α-tocopherol in retinoic acid differentiated SH-SY5Y neuroblastoma cells exposed to Aβ1-42 evaluating the transcriptional profile by next-generation sequencing. We observed that α-tocopherol was able to reduce the cytotoxicity induced by Aβ treatment, as demonstrated by Thiazolyl Blue Tetrazolium Bromide (MTT) assay. Moreover, the transcriptomic analysis evidenced that α-tocopherol treatment upregulated genes involved in the non-amyloidogenic processing of APP, while it downregulated the amyloidogenic pathway. Moreover, α-tocopherol modulated the expression of the genes involved in autophagy and the cell cycle, which are both known to be altered in AD. The treatment with α-tocopherol was also able to reduce oxidative stress, restoring nuclear factor erythroid-derived 2-like 2 (Nrf2) and decreasing inducible nitric oxide synthase (iNOS) levels, as demonstrated by immunocytochemistry.


Introduction
Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by cognitive and functional deficits, and in particular defects in short-term memory [1]. It represents the most common type of dementia. Electroencephalography (EEG) and different tests, such as the Mini Mental State Examination allow the detection of patients affected by AD and the differentiation between different types of dementia [2][3][4]. The number of patients affected by AD is increasing and until now, no efficacious cure that was able to slow disease progression has been available. For these reasons, AD represents one of the most problematic and costly illnesses.
The hallmarks of AD from a molecular point of view are extracellular amyloid plaques and intraneuronal neurofibrillary tangles (NFTs) [1]. Their accumulation leads to the degeneration and death of neurons in AD.
Amyloid plaques are composed of amyloid β (Aβ) aggregates that derived from the cleavage of the amyloid precursor protein (APP). APP can be processed through the amyloidogenic or the non-amyloidogenic pathway. Specifically, APP is mainly processed through the non-amyloidogenic pathway, where the α-secretase cleaved APP producing soluble APPα (sAPPα) and the 83-amino acid C-terminal fragment (C83). In turn, it can be cleaved by the γ-secretase to produce two small fragments, p3 and the APP intracellular domain (AICD). When APP follows the amyloidogenic pathway, APP

Cell Viability
Cell viability in the presence of different concentrations of α-tocopherol and Aβ 1-42 was evaluated using the Thiazolyl Blue Tetrazolium Bromide (MTT) assay. SH-SY5Y cells were cultured in 96-well plates, differentiated for 5 days with RA, and treated with α-tocopherol and Aβ  , as reported in the previous paragraph. At the end of the treatment, cells were incubated with medium containing MTT (0.5 mg/mL; Sigma-Aldrich) at 37 • C for 4 h. The formed formazan crystals were dissolved in acidic isopropanol at 37 • C for 1 h, and the optical density was evaluated by the spectrophotometric measurement of absorbance using the microplate reader Victor NIVO TM (PerkinElmer, Waltham, MA, USA). The assay was repeated in triplicate.

Extraction of Total RNA and cDNA Library Preparation
RNA was extracted using the Maxwell®RSC simplyRNA Cells Kit (Promega, Madison, Wisconsin, USA) following the manufacturer's instruction. The library preparation was carried out according to the TruSeq RNA Exome protocol (Illumina, San Diego, CA, USA) as previously described [28]. The experiment was repeated in triplicate.

RNA-Seq Data Analysis and Gene Evaluation
In order to perform the cDNA analysis, differentiated SH-SY5Y treated with Aβ (SH-SY5Y-Aβ) and cells treated with α-tocopherol and Aβ (SH-SY5Y-Aβ-αT) were sequenced by a MiSeq Next-Generation Sequencing instrument (Illumina, San Diego, California, USA). The MiSeq output files in "bcl" format were demultiplexed by CASAVA software (version 1.8) (Illumina) in "Fastq" format files. The software fastQC (Babraham Institute, Cambridge, UK) was used to perform the quality check of the files. Adapters and low-quality bases were removed by Trimmomatic (Usadel Lab, Aachen, Germany) [29] (version 0.38) (LEADING: 30 TRAILING: 28 SLIDINGWINDOW: 4:28 MINLEN: 75). The alignments of the reads were made against the "Homo Sapiens" reference genome available on the University of California Santa Cruz (UCSC) website [30] by Spliced Transcripts Alignment to a Reference (STAR) RNA-seq aligner [31]. Then, the alignment was sorted by the same program. The changes in the expression profile between SH-SY5Y-Aβ versus SH-SY5Y-Aβ-αT were studied with Cuffdiff software (Trapnell lab, Washington, USA) version 2.2.1 [32], taking advantage of the Homo Sapiens GTF file still available on the UCSC website. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database [33,34] was used to inspect the role of the genes involved in the Alzheimer's disease pathway (hsa05010), in the oxidative phosphorylation pathway (hsa00190), in the cell cycle pathway (hsa04110), and in the autophagy-animal pathway (hsa04140).

Immunocytochemistry
SH-SY5Y cells were grown on coverslips (12 mm; Thermo Fisher Scientific, Waltham, MA, USA) and treated with Aβ 1-42 and α-tocopherol, as reported in the previous paragraph. At the end of the treatment, cells were fixed with 4% paraformaldehyde at room temperature for 30 min. Afterwards, cells were washed with PBS (pH 7.5) and incubated with 3% H 2 O 2 at room temperature for 15 min, with the aim to suppress the endogenous peroxidase activity. SH-SY5Y cells were washed three times with PBS, and they were blocked with horse serum +0.1% Triton X-100 for 20 min. Afterwards, cells were incubated overnight at 4 • C with the primary antibodies anti-inducible nitric oxide synthase (iNOS) (1:50; Santa Cruz Biotechnology) and anti-nuclear factor erythroid-derived 2-like 2 (Nrf2) (1:50; Santa Cruz Biotechnology). Afterwards, cells were washed with PBS and incubated with biotinylated secondary antibody (1:200; Vector Laboratories, Inc., Burlingame, CA, USA) and streptavidin AB Complex-horseradish peroxidase (HRP) (ABC-kit from Dako, Glostrup, Denmark). The immunostaining was developed with the 3,3 -Diaminobenzidine (DAB) peroxidase substrate kit (Vector Laboratories, DBA Italia S.r.l., Milan, Italy; brown color; positive staining) and counterstaining with nuclear fast red (Vector Laboratories, DBA Italia S.r.l., Milan, Italy; pink background; negative staining). The immunocytochemical assays were repeated three times, and each experimental group was plated in duplicate, for a total of six coverslips for each antibody. The images were captured using a light microscopy (LEICA DM 2000 combined with LEICA ICC50 HD camera; Leica, Wetzlar, Germania).

Statistical Analysis
Statistical analysis of cell viability was carried out using GraphPad Prism version 7.0 software (GraphPad Software, La Jolla, CA, USA). The multiple comparison was performed using a one-way ANOVA test and the Bonferroni post hoc test. A p value less than or equal to 0.05 was considered statistically significant. The results are expressed by mean ± standard deviation (SD).

Cell Viability
The results of the MTT assay evidenced that α-tocopherol treatment did not cause cytotoxicity in the range 10 to 200 µM ( Figure 1A). On the contrary, the treatment of RA-differentiated SH-SY5Y with 10 µM of Aβ 1-42 for 24 h caused a reduction of cell viability. However, 100 µM of α-tocopherol was able to significantly increase cell viability compared to Aβ 1-42 treated cells ( Figure 1B). Lower doses of α-tocopherol, although increasing the percentage of cell viability, did not induce a significant increase. The highest dose of α-tocopherol tested did not increase the protective effect against Aβ 1-42 . On the basis of these results, we performed the transcriptomic analysis using the α-tocopherol concentration of 100 µM.

Statistical Analysis
Statistical analysis of cell viability was carried out using GraphPad Prism version 7.0 software (GraphPad Software, La Jolla, CA, USA). The multiple comparison was performed using a one-way ANOVA test and the Bonferroni post hoc test. A p value less than or equal to 0.05 was considered statistically significant. The results are expressed by mean ± standard deviation (SD).

Cell Viability
The results of the MTT assay evidenced that α-tocopherol treatment did not cause cytotoxicity in the range 10 to 200 µM ( Figure 1A). On the contrary, the treatment of RA-differentiated SH-SY5Y with 10 µM of Aβ1-42 for 24 h caused a reduction of cell viability. However, 100 µM of α-tocopherol was able to significantly increase cell viability compared to Aβ1-42 treated cells ( Figure 1B). Lower doses of α-tocopherol, although increasing the percentage of cell viability, did not induce a significant increase. The highest dose of α-tocopherol tested did not increase the protective effect against Aβ1-42. On the basis of these results, we performed the transcriptomic analysis using the αtocopherol concentration of 100 µM.

Gene Inspection
The RNA-Seq analysis was performed by Cuffdiff software. The comparison between SH-SY5Y-Aβ versus SH-SY5Y-Aβ-αT samples revealed 11,907 genes, of which 1966 genes were statistically different in their expression levels (q < 0.05). Regarding the significant genes, 971 genes were upregulated in SH-SY5Y-Aβ-αT, while 995 were downregulated. The Alzheimer's disease pathway revealed that 11 genes were upregulated and five genes were downregulated in SH-SY5Y-Aβ-αT compared to SH-SY5Y-Aβ (Table 1). Table 1. Genes significantly affected by amyloid β (Aβ) and α-tocopherol involved in the Alzheimer's disease pathway. Each gene in the first column was associated with the expression value in each group, the fold change was computed as the Log 2 (SH-SY5Y-Aβ versus SH-SY5Y-Aβ-αT) and the name of the related protein in KEGG. The values were rounded to the second decimal digit.
In the autophagy-animal pathway, 16 genes took part, of which nine genes were upregulated and seven genes were downregulated in SH-SY5Y-Aβ-αT (Table 4). The representation of the genes in the pathways is shown in Figure 2, while the summing of the changes in gene expression is represented in Figure 3. Table 2. Genes significantly affected by Aβ and α-tocopherol involved in the oxidative phosphorylation pathway. Each gene in the first column was associated with the expression value in each group, the fold change was computed as the Log 2 (SH-SY5Y-Aβ versus SH-SY5Y-Aβ-αT) and the name of the related protein in KEGG. The values were rounded to the second decimal digit.

Immunocytochemistry for iNOS and Nrf2
Given the antioxidant capacity of α-tocopherol, we also evaluated markers of oxidative stress by immunocytochemistry. We observed an increase of iNOS and a reduction of Nrf2 in Aβ-treated SH-SY5Y cells. However, the treatment with α-tocopherol was able to restore Nrf2 levels and

Immunocytochemistry for iNOS and Nrf2
Given the antioxidant capacity of α-tocopherol, we also evaluated markers of oxidative stress by immunocytochemistry. We observed an increase of iNOS and a reduction of Nrf2 in Aβ-treated SH-SY5Y cells. However, the treatment with α-tocopherol was able to restore Nrf2 levels and decrease iNOS protein levels ( Figure 4).

Immunocytochemistry for iNOS and Nrf2
Given the antioxidant capacity of α-tocopherol, we also evaluated markers of oxidative stress by immunocytochemistry. We observed an increase of iNOS and a reduction of Nrf2 in Aβ-treated SH-SY5Y cells. However, the treatment with α-tocopherol was able to restore Nrf2 levels and decrease iNOS protein levels ( Figure 4).

Discussion
Previous reports in vitro and in vivo indicated the ability of vitamin E to reduce the toxicity of Aβ peptides. In particular, vitamin E was able to reduce oxidative stress and cognitive deficits [21]. Given the promising results obtained in experimental models, clinical trials evaluated vitamin E supplementation in AD patients, but only some of them indicated improvements, showing that vitamin E was able to slow AD progression and cognitive decline [21,35,36].

Discussion
Previous reports in vitro and in vivo indicated the ability of vitamin E to reduce the toxicity of Aβ peptides. In particular, vitamin E was able to reduce oxidative stress and cognitive deficits [21]. Given the promising results obtained in experimental models, clinical trials evaluated vitamin E supplementation in AD patients, but only some of them indicated improvements, showing that vitamin E was able to slow AD progression and cognitive decline [21,35,36].
In this study, we aimed to characterize the changes in the transcriptional profile induced by α-tocopherol in SH-SY5Y cells treated with Aβ using next generation sequencing (NGS) technology and focusing on pathways known to be altered in AD.
We observed that the treatment with α-tocopherol was able to increase the processing of APP through the non-amyloidogenic pathway. We observed the increased expression of APP and ADAM metallopeptidase domain 17 (ADAM17) genes. ADAM17 encodes for α-secretase, that as said in the introduction, processes APP leading to the production of sAPPα. This result is in line with a previous work reporting that carnosic acid-a compound derived from rosemary that is able to exert an antioxidant action activating Nrf2-increased the mRNA expression of ADAM17, but only marginally increased the mRNA expression of ADAM metallopeptidase domain 10 (ADAM10), suggesting that it reduces Aβ production by activating ADAM17 in human neuroblastoma cells [37]. On the contrary, the processing of APP through the amyloidogenic pathway seemed to be reduced. Indeed, we found the upregulation of Reticulon 4 (RTN4), while the expression of presenilin 2 (PSEN2), part of the γ-secretase, was reduced. RTN4 is known to inhibit the β-secretase action reducing the production of Aβ [38,39]. However, aph-1 homolog B, gamma-secretase subunit (APH1B) was increased, but it also has a role in inhibiting apoptosis [40,41]. These results may indicate the increased production of sAPPα, which showed neuroprotective and pro-survival effects, and has a role in promoting neurogenesis [42][43][44]. Different stress conditions induce APP upregulation, given its neuroprotective and neurotrophic roles. Indeed, APP maintains synapses and plays a role in neuronal survival [9]. The induction of α-secretase activity exerts a neuroprotective effect against AD, thanks to its ability to promote some neurotrophic signaling pathways; for this reason, it may be a new strategy to reduce Aβ [45].
AD is characterized by mitochondrial dysfunctions [46], and mitochondria represent a site where Aβ is accumulated in AD neurons [47]. It was reported that Aβ altered the function of some proteins of mitochondrial oxidative phosphorylation [48]. We found that α-tocopherol treatment was able to increase the expression of the genes encoding for proteins that take part in Complex I (NDUFB1, NDUFB3, NDUFC2, NDUFS4), Complex III (UQCRB, UQCRH), and Complex IV (COX7A2, COX10). Only the genes NDUFA1 and NDUFA6, which encoded for subunits of Complex I, and SDHA, encoding for Complex II, were downregulated. These results may indicate a better functioning of mitochondrial complexes, and that α-tocopherol may counteract the toxicity induced by Aβ in mitochondrial oxidative phosphorylation. Aβ-induced mitochondrial dysfunctions are linked to oxidative stress. We also investigated the levels of iNOS and Nrf2 to evaluate if α-tocopherol was able to counteract oxidative stress. We observed that Aβ treatment increased oxidative stress, as demonstrated by the increased iNOS levels, while Nrf2 was reduced. On the contrary, α-tocopherol restored Nrf2 levels and inhibited iNOS. Interestingly, the genes encoding for subunits of the V-type ATPase (ATP6V1C1 and ATP6V1H) were upregulated, with only ATP6V0A2 downregulated. V-type ATPase is a complex that acts as an ATP-dependent proton pump found in the endomembrane system, where it is involved in the acidification of the lumen of subcellular organelles such as lysosomes and endosomes. The V-type ATPase function is necessary for different actions other than pH, indeed it is even necessary for the formation of synaptic vesicles [49]. Data showed the involvement of V-ATPase deficiency in the pathogenesis of AD [50].
The apoptotic peptidase activating factor 1 (APAF1) has a main role in inducing apoptosis, taking part in apoptosome formation [51]. The treatment with α-tocopherol was able to reduce APAF1 expression, indicating a reduction of apoptosis.
AD is characterized by an abnormal cell cycle re-entry of neurons, which may have a causal role in neurodegeneration. AD neurons showed high levels of the cell cycle markers, such as cyclins and cyclin-dependent kinases (CDK), and markers of DNA replication [52]. In particular, it was reported that Aβ was able to activate the cell cycle in SH-SY5Y [53]. The cell cycle re-entry seems to lead to apoptosis in neurological diseases. In particular, oxidative stress may be responsible for cell cycle re-entry. Indeed, oxidative stress may induce DNA damage with the consequent activation of enzymes involved in DNA repair. In turn, they can modulate downstream targets that regulate the proteins involved in the cell cycle checkpoint, leading at the end to cell cycle re-entry [54]. On the contrary, α-tocopherol treatment was able to reduce the expression of cyclins D, A, and B, which are encoded by CCND1, CCNA2, CCNB1 and CCNB2 respectively, together with BUB1 mitotic checkpoint serine/threonine kinase B (BUB1B), BUB3 mitotic checkpoint protein (BUB3), minichromosome maintenance complex component 4 (MCM4), minichromosome maintenance complex component 6 (MCM6), and Polo like kinase 1 (PLK1), compared to cells treated with Aβ. PLK1 acts during mitosis from G2 until the last phase of cytokinesis [55]. The proteins MCM4 and MCM6 are part of a heterohexameric complex formed by MCM2-MCM7, which plays a role as a replicative DNA helicase during DNA replication [56]. BubR1, together with Bub3, is part of the spindle assembly checkpoint that acts in preventing the loss of sister chromatid cohesion and premature chromosome segregation in the presence of unattached or incorrectly attached chromosomes [57]. In association, also mitotic arrest deficient 2 like 1 (MAD2L1), encoding for MAD2-which is part of the spindle assembly checkpoint together with BubR1 and Bub3-was significantly reduced in cells treated with α-tocopherol and Aβ. CDK1, encoding for CDK1, was also significantly reduced in cells treated with α-tocopherol and Aβ. CDKs are protein kinases that are associated with cyclins' control cell cycle [58]. Anaphase-promoting complex/cyclosome (APC/C) is responsible for the destruction of cyclins and other cell cycle proteins, and promotes anaphase [59]. We observed the upregulation of the genes encoded for APC/C (CDC27, CDC16), while only ANAPC1 was downregulated. APC/C was reported to be downregulated and inactivated in AD, probably causing the aberrant cycle re-entry [60]. Also, the genes encoding for securin and structural maintenance of chromosomes 1A, PTTG1 and SMC1A respectively, were downregulated in SH-SY5Y cells treated with α-tocopherol and Aβ. Cell Division Cycle-25 (CDC25) acts through dephosphorylating and activating CDKs, and for this reason, it has a main role in the transition between the phases of the cell cycle. We found the downregulation of CDC25B, which is involved in G2/M progression [61]. In parallel, we observed the upregulation of the genes encoding for 14-3-3 proteins (YWHAE, YWHAQ, YWHAZ, YWHAG) in cells treated with α-tocopherol and Aβ. 14-3-3 proteins inhibited CDC25 activity [62]. Transcriptional analysis revealed the upregulation of WEE1 in SH-SY5Y cells treated with α-tocopherol and Aβ compared to those exposed to Aβ. The Wee protein phosphorylated, causing the inhibition of CDK1 [63]. ABL1, which plays a role in spindle orientation control, was reduced, while SKP1, which is part of the SKP1-Cullin 1-F-box protein (SCF) complex [64], was increased.
It was demonstrated that the abnormal re-entry of neurons into the cell cycle is linked to changes of both mitochondria and autophagy, and then to neuronal degeneration. Neuroblastoma cells treated with an inhibitor of autophagy showed that the proportion of cells in G2/M was significantly increased, while the number of cells in G0/G1 was reduced. In parallel, autophagic proteins were accumulated in association with the changes in proteins responsible for the mitochondrial cycle [65].
Dysfunction of the autophagy is involved in different neurodegenerative disorders, including AD [12,13]. The autophagosomes in AD brains may be a reservoir of Aβ; then, enhancing the formation of new autophagosomes without the associated increase in their degradation may cause an increase in Aβ production, accumulation, and as a consequence, its toxicity [13].
The process of autophagy is composed by different phases: the initiation and nucleation of the phagophore, its expansion and autophagophore formation, the fusion of the autophagosome with lysosomes, and the final degradation in autolysosomes [66]. In healthy neurons, autophagosomes are scarce thanks to the efficiency of the autophagic flux, while in AD, the accumulation of autophagic vacuoles was observed [13]. Different approaches that induce the upregulation of autophagy were reported to be potential therapeutic strategies against AD [12,67]. Lipinski et al. found the transcriptional upregulation of autophagy in the brains of AD patients, speculating about a possible compensatory regulation of autophagy [68]. Moreover, they found an increased initiation of autophagy in response to Aβ and a suppression of lysosomal proteolysis [68].
The combination of increased autophagy induction and damaged clearance of autophagic vacuoles represents a favorable condition to cause Aβ accumulation in AD [69]. The excessive induction of autophagy beyond the clearance capability of the lysosome led to the accumulation of autophagosomes in the neurons, leading to toxicity [12,70].
α-tocopherol treatment downregulated unc-51 like autophagy activating kinase 2 (ULK2), whose protein seems to have a redundant function with Unc-51 Like Autophagy Activating Kinase 1 (ULK1), which is part of a complex that has a role in autophagy initiation [71]. Moreover, we found the modulation of different genes whose encoded proteins act as regulators of the phosphoinositide-3-kinase (PI3K) complex of autophagy. In particular, vacuole membrane protein 1 (VMP1) and syntaxin 17 (STX17) were downregulated, while UV radiation resistance associated (UVRAG) and autophagy and beclin 1 regulator 1 (AMBRA1) were upregulated. The protein VMP1 is needed for autophagosome formation and autophagy induction [72]. STX17 is required for the autophagosome maturation phase [73]. AMBRA1 is needed for Beclin1 translocation to the ER and the subsequent induction of autophagy [72]. UVRAG has a role in autophagosome formation and maturation [72]. Myotubularin related protein 3 (MTMR3) and WD repeat domain, phosphoinositide interacting 1 (WIPI1) were upregulated, while WD repeat domain, phosphoinositide interacting 2 (WIPI2) was downregulated. The role of the protein encoded by MTMR3 was to inhibit autophagy through the dephosphorylation of phosphatidylinositol 3-phosphate (PI3P), which has a role in autophagosome formation [74]. The WIPI proteins interact with PI3P and play a role as PI3P effector proteins, recruiting downstream autophagy-related (ATG) proteins [75]. This data may indicate the induction of autophagy, even if the downregulation of some genes may limit the excessive autophagic flux. Also, GABARAP and GABARAPL1, encoding for members of the microtubule-associated protein 1A/1B-light chain 3 (LC3) family [76], and ATG3 were upregulated. Cleaved LC3 is conjugated to phosphatidylethanolamine (PE) by the sequential activation of Atg7 (E1-like enzyme), Atg3 (E2-like enzyme), and the Atg12 complex, to generate LC3-PE (a membrane-bound form of LC3), the level of which is correlated with the number of autophagosomes. In addition, during selective autophagy, LC3 acts as an adaptor protein to recruit selective cargo to the autophagosome via interaction with cargo receptors [77]. ATG4A and ATG4B were downregulated. Processing and delipidation reactions of LC3 by Atg4 are required for autophagosome formation [78]. Moreover, the genes CTSB and CTSD encoding for cathepsins were upregulated. On the contrary, the downregulation of CTSL was found, which encoded for cathepsin L, and is involved in the degradation of the autophagosomal membrane markers, such as GABARAP and LC3-in contrast to the other cathepsins, which are involved in the degradation of the content of autophagosomes [79]. The increased activity of cathepsins is in line with the results regarding the V-type ATPases, which is required to maintain the acidic pH in the lysosome that is needed for the functioning of lysosomal enzymes. These results may indicate a better degradation of autophagosome cargo, but the proteins involved in the autophagic flux are not degraded. Given that the literature data indicated the increase of autophagy initiation and the accumulation of autophagosomes in AD, our results may indicate that α-tocopherol is able to increase the degradation of the formed autophagosomes, leading to a clearance of Aβ.

Conclusions
In conclusion, our results suggested that α-tocopherol was able to counteract Aβ toxicity, inducing the non-amyloidogenic processing pathway of APP and modulating the autophagic flux in order to induce the degradation of the formed autophagosomes. Moreover, α-tocopherol inhibited the cell cycle that is known to be abnormally reactivated in AD. This data indicated that α-tocopherol was able to modulate pathways that are altered in AD, suggesting that it may be useful in AD treatment. Funding: This work was supported by Ricerca Finalizzata RF-2013-02359594 entitled "Vitamin E, miRNA and inflammation: A tunable network in Alzheimer's disease". Moreover, this work was supported by SensApp project that has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 829104.

Conflicts of Interest:
The authors declare no conflict of interest.