The Journal of Neuroscience, March 1, 2017 • 37(9):2449 –2462 • 2449

Neurobiology of Disease

Selenomethionine Mitigates Cognitive Decline by Targeting

Both Tau Hyperphosphorylation and Autophagic Clearance
in an Alzheimer’s Disease Mouse Model
Zhong-Hao Zhang,1 Qiu-Yan Wu,1 Rui Zheng,1 Chen Chen,1 Yao Chen,1 X Qiong Liu,1 Peter R. Hoffmann,2 Jia-Zuan Ni,1
and X Guo-Li Song1
1Shenzhen Key Laboratory of Marine Bioresources and Ecology, College of Life Sciences and Oceanography, Shenzhen University, Shenzhen, 518060, China,
and 2Department of Cell and Molecular Biology, John A. Burns School of Medicine, University of Hawaii, Honolulu, Hawaii 96813

Tau pathology was recently identified as a key driver of disease progression and an attractive therapeutic target in Alzheimer’s disease
(AD). Selenomethionine (Se-Met), a major bioactive form of selenium (Se) in organisms with significant antioxidant capacity, reduced
the levels of total tau and hyperphosphorylated tau and ameliorated cognitive deficits in younger triple transgenic AD (3xTg-AD) mice.
Whether Se-Met has a similar effect on tau pathology and the specific mechanism of action in older 3xTg-AD mice remains unknown.
Autophagy is a major self-degradative process to maintain cellular homeostasis and function. Autophagic dysfunction has been impli-
cated in the pathogenesis of multiple age-dependent diseases, including AD. Modulation of autophagy has been shown to retard the
accumulation of misfolded and aggregated proteins and to delay the progression of AD. Here, we found that 3xTg-AD mice showed
significant improvement in cognitive ability after a 3-month treatment with Se-Met beginning at 8 months of age. In addition to attenu-
ating the hyperphosphorylation of tau by modulating the activity of Akt/glycogen synthase kinase-3␤ and protein phosphatase 2A,
Se-Met-induced reduction of tau was also mediated by an autophagy-based pathway. Specifically, Se-Met improved the initiation of
autophagy via the AMP-activated protein kinase–mTOR (mammalian target of rapamycin) signaling pathway and enhanced autophagic
flux to promote the clearance of tau in 3xTg-AD mice and primary 3xTg neurons. Thus, our results demonstrate for the first time that
Se-Met mitigates cognitive decline by targeting both the hyperphosphorylation of tau and the autophagic clearance of tau in AD mice.
These data strongly support Se-Met as a potent nutraceutical for AD therapy.
Key words: Alzheimer’s disease; autophagy; hyperphosphorylated tau; selenomethionine; tau; tauopathy

Significance Statement
Selenium has been widely recognized as a vital trace element abundant in the brain with effects of antioxidant, anticancer, and
anti-inflammation. In this study, we report that selenomethionine rescues spatial learning and memory impairments in aged
3xTg-AD mice via decreasing the level of tau protein and tau hyperphosphorylation. We find that selenomethionine promotes the
initiation of autophagy via the AMPK–mTOR pathway and enhances autophagic flux, thereby facilitating tau clearance in vivo and
in vitro. We have now identified an additional, novel mechanism by which selenomethionine improves the cognitive function of
AD mice. Specifically, our data suggest the effect of selenium/selenomethionine on an autophagic pathway in Alzheimer’s disease.

Introduction acterized by abnormal accumulations of ␤-Amyloid (A␤) oli-

Alzheimer’s disease (AD), the most common type of dementia in gomers and intracellular neurofibrillary tangles (NFTs) in the
aged people, is an untreatable neurodegenerative disorder char- brain attributable to hyperphosphorylated tau that results in pro-

Received Oct. 18, 2016; revised Dec. 29, 2016; accepted Jan. 23, 2017. Prof. Yunwu Zhang (Institute for Biomedical Research, Xiamen University, Xiamen, China) for providing the N2a-sw
Author contributions: Z.-H.Z., Q.L., J.-Z.N., and G.-L.S. designed research; Z.-H.Z., Q.-Y.W., R.Z., C.C., and Y.C. cells.
performed research; Z.-H.Z. contributed unpublished reagents/analytic tools; Z.-H.Z. and G.-L.S. analyzed data; The authors declare no competing financial interests.
Z.-H.Z., P.R.H., and G.-L.S. wrote the paper. Correspondence should be addressed to Prof. Guo-Li Song, Shenzhen Key Laboratory of Marine Biological Re-
This work was supported by the National Natural Sciences Foundation of China (Grants 81400847, 31470804, and sources, College of Life Sciences, Shenzhen University, Shenzhen, 518060, Guangdong Province, P.R. China. E-mail:
21271131) and the Shenzhen Bureau of Science, Technology, and Information (Grants JCYJ20150529164656093 [email protected].
and JSGG2014070316383 8793). We are extremely thankful to Dr. Cuihong Zhou (Beijing Nuclear Magnetic Reso- DOI:10.1523/JNEUROSCI.3229-16.2017
nance Center, Peking University, Beijing, China) for providing the mTagRFP-mWasabi-LC3 plasmid. We also thank Copyright © 2017 the authors 0270-6474/17/372449-14$15.00/0
2450 • J. Neurosci., March 1, 2017 • 37(9):2449 –2462 Zhang et al. • Selenomethionine Targets Tauopathy

gressive synaptic dysfunction and cognitive deficits (Balducci and and Harris-White, 2015). Based on our research and other previous
Forloni, 2014; Orr et al., 2014; Shibuya et al., 2014; Alzheimer’s studies, integrated pathological characteristics of tau occur at ⬃8 –12
Association, 2015; Sui et al., 2015). A␤ has been widely viewed as months of age in the 3xTg-AD mouse (Oddo et al., 2003a,b). Thus,
the principal therapeutic target of AD. However, increasing evi- in this study, the therapeutic effect and underlying mechanisms of
dence suggests that the generation of A␤ and tau dysfunction Se-Met on tau pathology were explored in 8-month-old AD mice.
have an intrinsic relationship and together may synergistically We found that Se-Met rescued cognitive impairments of 3xTg-AD
trigger disease pathogenesis (Eckert et al., 2010; Hurtado et al., mice, modulated the activity of Akt/GSK3␤ and PP2A to decrease
2010; Yu et al., 2014). Pathological alterations in tau mediate A␤- hyperphosphorylated tau, initiated autophagy via the AMP-acti-
and inflammation-induced neurotoxicity, and a reduction in tau vated protein kinase (AMPK)–mammalian target of rapamycin
can prevent A␤- and inflammation-induced neurodegeneration (mTOR) pathway, and enhanced autophagic flux to promote the
(Vossel et al., 2010, 2015; Maphis et al., 2015). Tau pathology autophagic clearance of tau in aged 3xTg-AD mice.
(tauopathy) better correlates with cognitive impairment in AD
patients than A␤ pathology. Thus, tau has emerged as a meaning- Materials and Methods
ful target for pathological research and pharmaceutical interven- Animals and treatment. 3xTg-AD and wild-type (WT) mice were pur-
tions for AD. chased from The Jackson Laboratory (hybrid 129/C57BL6 background,
Tau is a microtubule-associated protein encoded by the MAPT JAX order number 3591206). These mice express the mutant human
gene located on chromosome 17 that dynamically regulates the po- APPswe and tauP301L genes and the mutant mouse PS1M146V gene.
3xTg-AD mice (the fourth generation) at 8 months of age (n ⫽ 12; six
lymerization, stability, and assembly of axonal microtubules (Gend-
males and six females) were treated with 6 ␮g/ml Se-Met (Sigma-
ron and Petrucelli, 2009). An increasing body of evidence suggests Aldrich) in their drinking water for 12 weeks. The control mice and WT
that tauopathies caused by pathological changes in tau proteins (e.g., mice (n ⫽ 12 per group; six males and six females in each group) received
misfolding, mislocalization, and hyperphosphorylation) exist in normal drinking water. The body weight of each mouse was recorded
⬎30 neurodegenerative diseases, with hyperphosphorylated tau every 2 weeks. The mice were maintained under standard laboratory
presenting as a pathological feature of AD (Berger et al., 2007; conditions, including a 12 h light/dark cycle, a temperature of 22 ⫾ 2°C,
Kopeikina et al., 2012). Tau hyperphosphorylation is site specific and ad libitum access to water and food. Each cage housed three to four
and mainly occurs at serine/threonine residues, such as Thr231, mice with the same genotype.
Thr235, Ser396, Ser404, Ser262, and Ser202 (Ding et al., 2006; Wang After treatment with Se-Met for 12 weeks and the subsequent behav-
et al., 2014). Thus, serine/threonine protein kinases and phospha- ioral testing, mice were killed and their brains were rapidly removed. The
left hemisphere was immersion fixed with 4% paraformaldehyde for
tases that regulate tau phosphorylation directly, such as glycogen
24 h, dehydrated with a graded ethanol series, cleared with xylene, infil-
synthase kinase-3␤ (GSK3␤) and protein phosphatase 2A (PP2A), trated with wax, and cut into 5-␮m-thick sections. The right hemisphere
may play a crucial role in AD-related tau pathology (Hu et al., 2006). was dissected into hippocampal and cortical samples, snap frozen in
Inhibition of GSK3␤ or activation of PP2A ameliorated cognitive liquid nitrogen, and stored at ⫺80°C until analysis. The experiments and
impairment in AD mice by attenuating tau hyperphosphorylation procedures were performed in strict accordance with the institutional
(Y. Zhang et al., 2014; Castro-Alvarez et al., 2015). We previously guidelines regarding experimental animal use at Shenzhen University.
reported that selenomethionine (Se-Met), a major form of dietary The protocol was approved by the Animal Ethical and Welfare Commit-
selenium (Se) with powerful antioxidant capacity, reduced tau hy- tee of Shenzhen University (permit number AEWC-20140615-002). All
perphosphorylation by modulating the activity of GSK3␤ and PP2A efforts were made to minimize suffering.
in 4-month-old triple transgenic AD mice (3xTg-AD; Song et al., Morris water maze. WT mice (11 months old), 3xTg-AD mice, and
Se-Met-treated mice were trained in a round, water-filled tub (160 cm in
2014). Our study was the first to reveal that Se-Met has a positive
diameter and 50 cm in height) as described previously (Song et al., 2014).
effect on cognitive ability and tau pathology in AD. However, poten- Each mouse was given four trials per day for 5 consecutive days with a
tial mechanisms by which this occurs remain unclear. In 3xTg-AD 15 min intertrial interval. The first day was designated as the training day,
mice, human mutant P301L-tau is an abnormally overexpressed and the maximum trial length was 60 s. If mice did not reach the platform
protein that is particularly prone to misfolding and site-specific hy- in the allotted time, they were manually guided to it, where they re-
perphosphorylation (Santacruz et al., 2005; Terwel et al., 2005; mained for 15 s. After the 5 d of task acquisition, probe trials were per-
Hunsberger et al., 2014). Therefore, besides reducing hyperphos- formed 24 and 72 h later to assess short-term and long-term memory
phorylation, clearance of this mutant tau protein or abnormally hy- consolidation. The platform was removed, and the mice were placed into
perphosphorylated tau could represent another method to prevent the quadrant of the pool opposite to the original platform quadrant. In
tauopathies. each probe trial, mice were allowed to swim for 60 s. The time mice spent
in the original platform quadrant and the number of times the mice
Macroautophagy is a major cellular pathway for the degradation
crossed the platform position were recorded. All trials were recorded
and clearance of damaged organelles and/or denatured and aggre- using an HVS water maze program for subsequent analyses of escape
gated proteins/peptides. The autophagy process serves to maintain latency and swimming speed (WaterMaze 3; Actimetrics).
cellular homeostasis, which depends on the fusion of autophago- Step-down test. The mice were subjected to a step-down test in a pas-
somes and lysosomes (i.e., autophagolysosomal fusion; Levine and sive avoidance chamber. The floor of the chamber consisted of copper
Klionsky, 2004; Mizushima et al., 2008). Previous studies have rods and a well-insulated platform made of rubber in a corner of the
shown that autophagic dysfunction in the CNS causes neurodegen- chamber. Mice were placed in the chamber for a 3 min adaptation period
eration in mice and that defects in autophagosome formation and and then placed on the platform. The latency to step down on the grid
autophagosome–lysosome fusion occur early during AD pathogen- with all four paws was measured. After stepping down on the copper bars,
esis (Komatsu et al., 2006; Zare-Shahabadi et al., 2015). Promoting the mice received an immediate mild electrical shock (36 V). To avoid the
shock, mice showed an instinctive reaction to jump back onto the plat-
autophagy initiation, autolysosome formation, and lysosome bio-
form. Mice were tested in this manner for 5 min. The number of times
genesis reduced the expression and aggregation of abnormal tau mice stepped down from the platform within 5 min was defined as the
protein in AD models (Shibuya et al., 2014; Y. Kim et al., 2015; Siman acquisition error. After 24 h, this procedure was repeated, and the step-
et al., 2015). Oxidative stress can regulate autophagic pathways in down latency (SDL) was used as a measurement of memory retention.
neurodegeneration, and autophagy may functionally bridge oxida- The number of times that mice stepped down onto the platform within
tive stress and the development of AD (Zheng et al., 2006; Hensley the 3 min interval was recorded as the error time.
Zhang et al. • Selenomethionine Targets Tauopathy J. Neurosci., March 1, 2017 • 37(9):2449 –2462 • 2451

Table 1. Antibody information cells/well. For confocal microscopy, cells were seeded onto 13 mm glass
Antibody Host Application Source coverslips precoated with poly-D-lysine and incubated in 24-well cell culture
plates at a density of 1.5 ⫻ 10 5 cells/well. The medium was completely
Tau5 Mouse IMB/IMF Abcam replaced after 4 h, and half of the medium was subsequently replaced every
Tau-pS404 Rabbit IMB/IMF Abcam 3 d. To observe the effects of Se-Met on tau pathology, the neurons were
Akt Rabbit IMB Cell Signaling treated on day 8 with 1 ␮M Se-Met for 24 h.
p-Akt Rabbit IMB Cell Signaling Cell culture, transfection, and bafilomycin A1 treatment. Mouse neuro-
GSK3␤ Mouse IMB Cell Signaling blastoma Neuro-2a (N2a) cells stably expressing human APP695
p-GSK3␤ Rabbit IMB Cell Signaling (N2a-sw cells) were obtained from Prof. Yunwu Zhang (Institute for
PP2A Rabbit IMB Cell Signaling Biomedical Research, Xiamen University, Xiamen, China). Cells were
p-PP2A Rabbit IMB Epitomics cultured in DMEM/Opti-MEM (1:1, v/v) containing 200 ␮g/ml G418,
Map2 Chicken IMF Abcam 10% fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomy-
mTOR Rabbit IMB Cell Signaling cin at 37°C in humidified 5% CO2/95% air. The cells were passaged every
p-mTOR Rabbit IMB Cell Signaling 3 d and grown to 80% confluence. The mTagRFP-mWasabi-LC3 plas-
P70S6K Rabbit IMB Cell Signaling mid was a generous gift from Dr. Cuihong Zhou (Peking University,
p-P70S6K Rabbit IMB Cell Signaling Beijing, China). Approximately 1.4 ⫻ 10 5 cells were plated into each well
AMPK␣ Rabbit IMB Cell Signaling of a 24-well plate 24 h before transfection. Plasmid (500 ng/well) trans-
p-AMPK␣ Rabbit IMB Cell Signaling fections were performed with Lipofectamine 2000 (11668; Invitrogen)
SQSTM1/p62 Rabbit IMB Santa Cruz Biotechnology according to the manufacturer’s protocol. To study the effect of Se-Met
Cathepsin D Rabbit IMB/IMF Cell Signaling on autophagic flux in N2a-sw cells, cells were pretreated with bafilomy-
LC3 Rabbit IMB/IMF Sigma cin A1 (10, 50, 100, 200, or 400 nM) 4, 8, or 16 h before Se-Met treatment.
LC3B Rabbit IMB/IMF Genetex The cells were treated with Se-Met for 24 h.
GAPDH Mouse IMB Proteintech Isolation of autophagic vacuoles and lysosomes from brain tissue. For
␤-Actin Mouse IMB Proteintech each group, six different fresh hemibrains were pooled. Using a modified
IMB, Immunoblot; IMF, immunofluorescence. protocol from Marzella L (Marzella et al., 1982), the samples were
homogenized and subjected to differential centrifugation. The nuclear
Western blot analysis. The right hemibrain was dissected into hip- fraction and any unbroken cells were precipitated; then, the fraction
pocampal and cortical samples and homogenized in 9 vol of Tris- enriched in autophagic vacuoles, lysosomes, and mitochondria was col-
buffered saline (TBS) with a protease inhibitor mixture and phosphatase lected (Cuervo et al., 1995; Yang et al., 2014). The different organelles in
inhibitors (Roche). The samples were centrifuged at 13,000 ⫻ g for 1.5 h this fraction were isolated by flotation in a discontinuous gradient of
at 4°C. The TBS-soluble supernatants were collected, and the pellets were Nycodenz (50, 26, 24, 20, and 10%). A fraction containing mainly au-
resuspended in 2 vol of 5% SDS containing a protease inhibitor mixture tophagosomes was recovered from the 10 –20% interface, a fraction en-
and phosphatase inhibitors. The TBS-insoluble pellet mixtures were son- riched in autolysosomes was isolated from the 20 –24% interface, and a
icated for 1 min in an ice bath and centrifuged at 13,000 ⫻ g for 30 min at lysosome-enriched fraction was recovered from the 24 –26% interface.
4°C. The supernatants of TBS-insoluble homogenates were also col- These fractions were washed by centrifugation in 0.25 M sucrose. Next,
lected. The protein concentration of each sample was determined using the pellets were homogenized in 3 vol of TBS with a protease inhibitor
the bicinchoninic acid assay (Sigma-Aldrich). Proteins (20 ␮g) were mixture and phosphatase inhibitors (Roche), sonicated for 1 min in an
loaded into each lane of a 10 –15% SDS-polyacrylamide gel. After elec- ice bath, and centrifuged at 13,000 ⫻ g for 30 min at 4°C. The samples
trophoresis, proteins were transferred onto 0.45 nm polyvinylidene di- were then stored for subsequent Western blotting.
fluoride membranes (Millipore) at 100 mA for 1.5 h. The membrane was RNA isolation, cDNA synthesis, and RT-PCR for tauP301L mRNA. Total
then blocked with 5% fat-free milk in TBS for 2 h at 37°C, followed by RNA was extracted from the hippocampi of AD mouse brains using TRIzol
incubation with primary antibodies (Table 1) overnight at 4°C and reagent. Total RNA (2.5 ␮g) was reverse transcribed with an oligo dT primer
horseradish peroxidase-conjugated secondary antibodies (anti-mouse in a 10 ␮l volume using the PrimeScript first-strand cDNA synthesis kit
and anti-rabbit; NeoBioscience) for 1 h at 37°C. The immunoreactive (Takara Bio) according to the manufacturer’s instructions. PCR was per-
bands were visualized with an electroluminescence kit and scanned for formed with 25 cycles of 1 min at 94°C, 1 min at 48°C, and 2 min at 72°C in
densitometric analysis using an imaging system (Image Station 4000 M; a 25 ␮l reaction mixture containing 12.5 ␮l of GoTag Green Master Mix
Kodak) and Quantity One software (Bio-Rad). ␤-Actin and GAPDH (Promega), 0.5 ␮l of the forward and reverse Tau5 primers (1 ␮M each) or
were used as the loading control. ␤-actin primers (10 pM each), 0.5 ␮l of first-strand cDNA, and 11 ␮l of
Immunofluorescent staining. Sagittal paraffin sections (5 ␮m thick) of nuclease-free water. The primer sequences used were 5⬘-TCGCAGTCAC
mouse brains were mounted on glass slides. The sections were pretreated CGCCACCCAC-3⬘ (forward) and 5⬘-TGTCATCGCTTCCAGTCCCG
with 0.01 mol/L citrate buffer, pH 6.0, in hyperthermy for 5 min and TC-3⬘ (reverse) for tauP301L and 5⬘-GGAGAAGATCTGGCACCACA
blocked with 5% goat serum in PBS for 10 min. Next, the sections were CC-3⬘ (forward) and 5⬘-CCTGCTTGCTGATC CACATCTGCTGG-3⬘ (re-
incubated with primary antibodies overnight at 4°C, followed by incuba- verse) for mouse ␤-actin. The ␤-actin primer was used as an internal control.
tion with secondary antibodies (1:500 in PBS) for 1 h at 37°C. The pri- Each PCR product (10 ␮l) was run on a 2% agarose gel containing 0.1 ␮g/ml
mary antibodies used in this study were summarized in Table 1. The ethidium bromide, and the amplified products (750 bp for tauP301L and
primary antibodies were detected using Alexa Fluor fluorescent dye- 840 bp for ␤-actin) were visualized using an imaging system (G: Box;
conjugated secondary antibodies (anti-mouse and anti-rabbit; Alexa Syngene).
Fluor 488 and 695, Multi Sciences Biotech). Three equidistant sec- Transmission electron microscopy. Transmission electron microscopy
tions that included the whole hippocampal and frontal cortical areas was performed as described previously (Su et al., 2014; Jiang et al., 2015).
were evaluated per animal. For the analysis and quantification of Briefly, mouse hippocampi were fixed with 2.5% glutaraldehyde for 2 h at
immunoreactive areas, the sections were imaged by fluorescence mi- 4°C and 1% osmium tetroxide for 2 h at 4°C. After washing three times in
croscopy (Olympus). distilled water, samples were dehydrated in a graded ethanol series and
Primary neuron cultures. Primary cortical neurons were obtained from embedded in epoxy resin. Next, the samples were ultrathin sectioned
postnatal (P0 –P1) pups of WT and 3xTg-AD mice. Briefly, cortices were (DiATOME knife and Leica ultramicrotome), stained with 2% uranyl
dissected from the brains and digested with 2 mg/ml papain for 30 min at acetate and 2% lead citrate, and analyzed using an FEI Tecnai G2 Spirit
37°C. Digested tissue was triturated by pipetting in DMEM with 10% FBS. transmission electron microscope.
Dissociated cells were cultured in neurobasal medium supplemented with Statistical analysis. The data were analyzed using GraphPad Prism soft-
2% B27, 0.5 mM L-glutamine, and 50 U/ml penicillin–streptomycin using ware. All data are expressed as the mean ⫾ SEM and considered statisti-
poly-D-lysine-coated six-well cell culture plates at a density of 0.5 ⫻ 10 6 cally significant at a level of p ⬍ 0.05. A repeated-measures ANOVA was
2452 • J. Neurosci., March 1, 2017 • 37(9):2449 –2462 Zhang et al. • Selenomethionine Targets Tauopathy

Figure 1. Se-Met administration ameliorates cognitive deficits in 3xTg-AD mice. A–F, The Morris water maze test was used to evaluate spatial memory in mice. A, B, During the 4 d of spatial
orientation trials, escape latencies and swimming speed were measured to assess the memory and the locomotor ability of the mice, respectively. C–F, For the 24 and 72 h probe trials, respectively,
the time spent in the target quadrant and the number of crossings over the original platform area were recorded. G, H, The latency period and number of errors during a step-down-type passive
avoidance test were detected. The values are shown as the mean ⫾ SEM, and 12 mice were tested per group. *p ⬍ 0.05; **p ⬍ 0.01.

used to analyze the behavioral test data. The one-way ANOVA and mul- trol mice ( p ⬍ 0.05; Fig. 1D). Similar results were obtained in the
tiple comparisons test were used to analyze the immunoblot and RT- 72 h memory trial. Se-Met-treated 3xTg-AD mice spent more
PCR data. time in the target quadrant ( p ⬍ 0.05; Fig. 1E). Se-Met-treated
mice showed a trend toward increased platform crossings, but
Results these data were not significantly different from those of the con-
Se-Met rescues spatial memory deficits in 8-month-old
trol mice (Fig. 1F ).
3xTg-AD mice
The step-down passive avoidance test was also used to evalu-
To examine the therapeutic efficacy of Se-Met, we used 8-month-
ate the effect of Se-Met on the learning and memory of the trans-
old 3xTg-AD mice, which express mutant human APPswe,
genic mice. Compared with WT mice, 3xTg-AD mice exhibited a
tauP301l, and mutant mouse PS1M146V. These mice along with
poor performance consisting of a significant increase in the num-
age-matched WT mice were administered drinking water with
ber of errors ( p ⬍ 0.05; Fig. 1H ) and a significant decrease in SDL
Se-Met (6 ␮g/ml, n ⫽ 12) or nontreated water (n ⫽ 12) for
( p ⬍ 0.05; Fig. 1G). Se-Met treatment did not significantly affect
3 months, followed by behavioral analyses.
the SDL of 3xTg-AD mice, but the number of errors was signifi-
Spatial learning was assessed using the Morris water maze test,
cantly decreased. These findings suggest that Se-Met treatment
which measures the time required for mice to find the hidden
effectively rescues the spatial memory deficits in 11-month-old
platform (escape latency). A shorter escape latency was consid-
AD mice.
ered better spatial learning. We found that the escape latencies of
the 3xTg-AD mice were markedly longer than those of WT
mice. However, the prolonged latency was mostly reversed with Se-Met reduces the accumulation of total tau and
Se-Met treatment (Fig. 1A). A t test revealed that Se-Met-treated hyperphosphorylated tau
3xTg-AD mice performed significantly better than those of the Our previous work found that Se-Met ameliorated tau pathology
control mice beginning on the second day (day 2 and day 4, in the brains of 4-month-old 3xTg-AD mice (Song et al., 2014).
p ⬍ 0.05). Moreover, Se-Met-treated mice exhibited significantly To determine whether Se-Met had a similar effect on older
shorter total escape latencies over 5 consecutive training days. No 3xTg-AD mice when the integrated pathological features of tau
significant differences were observed in the swimming speed have developed in the brain, the pathological changes of tau were
among the three groups (Fig. 1 A, B), indicating that the effect of investigated in the hippocampus and cortex, which represent the
Se-Met on learning and memory was independent of physical most sensitive regions affected in the AD brain (Bryan et al.,
performance. 2009). After 3 months of treatment with Se-Met, Western blot
Probe trials were conducted to evaluate the spatial reference analysis of total tau levels showed significant reductions in Tris-
memory of the mice 24 and 72 h after the 5 d training. The time buffered saline (TBS)-soluble and TBS-insoluble tau in the hip-
spent in the target quadrant and the number of times the mice pocampus ( p ⬍ 0.01 and p ⬍ 0.05, respectively; Fig. 2 A, B) and
crossed the previous platform location during a single 60 s trial cortex ( p ⬍ 0.05 and p ⬍ 0.001, respectively; Fig. 2C,D) in Se-
were measured. Twenty-four hours after training, 3xTg-AD mice Met-treated mice compared with untreated controls. Besides re-
spent significantly less time in the target quadrant than WT mice. ducing total tau protein levels (P301L-tau, human full-length tau
Treatment with Se-Met significantly reversed this effect in trans- protein, about 55 kDa; Fig. 2), Se-Met treatment also decreased
genic mice ( p ⬍ 0.05; Fig. 1C). In addition, Se-Met-treated mice the levels of 35– 45 kDa tau fragments, but the difference was not
swam across the platform location more frequently than the con- statistically significant (data not shown). The level of pS404-tau
Zhang et al. • Selenomethionine Targets Tauopathy J. Neurosci., March 1, 2017 • 37(9):2449 –2462 • 2453

Figure 2. Se-Met decreases the level of total tau and hyperphosphorylated tau. A–D, Se-Met significantly reduced the levels of total tau and pS404-tau protein in the TBS-soluble and
TBS-insoluble fractions of the hippocampus (A, B) and cortex (C, D) in 3xTg-AD mice. E, F, Immunofluorescence images of tau in the hippocampus (DG and CA1) and cortex and pS404-tau in the
hippocampus (CA3). G, H, Representative Western blot images and densitometric analysis of total tau and pS404-tau in primary cortical neurons. I, J, Immunofluorescence images of tau and
pS404-tau in primary cortical neurons. Quantitative results were normalized against the levels of GAPDH. The data were expressed as percentages of the control (set to 100%) and presented as the
group mean ⫾ SEM (n ⫽ 6). *p ⬍ 0.05; **p ⬍ 0.01; ***p ⬍ 0.001. Scale bars, 20 and 50 ␮m. Hipp, Hippocampus; DG, dentate gyrus; CA1, cornu ammonis 1; CA3, cornu ammonis 3.

was also significantly decreased in the TBS-soluble and TBS- decreasing total tau and tau hyperphosphorylation both in vitro
insoluble fractions from the hippocampus ( p ⬍ 0.01 and p ⬍ and in vivo.
0.01, respectively; Fig. 2 A, B) and TBS-soluble fractions from the
cortex ( p ⬍ 0.05; Fig. 2C,D) of Se-Met-treated mice. A histo- Akt and GSK3␤ are involved in Se-Met-induced tau
pathological analysis of total tau in the dentate gyrus, CA1 dephosphorylation in aged 3xTg-AD mice
(cornu ammonis 1), and cortical regions of the brain revealed To investigate the mechanisms underlying the effect of Se-Met on
that Se-Met-treated mice showed markedly less tau immuno- tau phosphorylation, several enzymes that regulate this post-
staining than control mice (Fig. 2E). Similarly, neuronal stain- translational modification were evaluated. GSK3␤ is a major ki-
ing using a specific antibody against pS404-tau revealed lower nase that regulates tau hyperphosphorylation, and its activity is
fluorescence intensity in the hippocampus of Se-Met-treated inhibited by phosphorylation at residue Ser9 (Hernandez et al.,
mice than in control mice (Fig. 2F ). 2013). Thus, the levels of GSK3␤ and GSK3␤ phosphorylated at
To further ascertain the effect of Se-Met on tau pathology, Ser9 (p-GSK3␤) were assessed by Western blot. The ratio of
primary cortical neurons were isolated from newborn 3xTg-AD p-GSK3␤/GSK3␤ inversely reflects the activity of GSK3␤ in vivo.
and WT mice. The neurons were treated with Se-Met (1 ␮M) for Compared with WT mice, the ratio of p-GSK3␤/GSK3␤ signifi-
24 h, and the levels of total tau and pS404-tau proteins were cantly decreased in the hippocampus and cortex of 3xTg-AD
detected by Western blot. There was a significant increase in total mice ( p ⬍ 0.05 and p ⬍ 0.05, respectively; Fig. 3A–D). Se-Met
tau and pS404-tau in primary 3xTg neurons compared with significantly increased the ratio of p-GSK3␤/GSK3␤ in the hip-
WT neurons ( p ⬍ 0.001 and p ⬍ 0.001, respectively; Fig. 2G,H ). pocampus and cortex ( p ⬍ 0.001 and p ⬍ 0.05, respectively; Fig.
Treatment with Se-Met significantly decreased the levels of both 3A–D), suggesting that Se-Met inhibits GSK3␤ activity in
total tau and pS404-tau in 3xTg neurons ( p ⬍ 0.01 and p ⬍ 0.01, 3xTg-AD mice. PP2A is part of a large family of enzymes that
respectively; Fig. 2G,H ). Consistent with the immunoblot results, accounts for the majority of Ser/Thr phosphatase activity in the
decreased fluorescence staining of tau5 and pS404-tau was ob- brain and a major tau phosphatase in vivo (Sontag and Sontag,
served in Se-Met-treated 3xTg neurons compared with the con- 2014). PP2A is inactivated by phosphorylation at Tyr-307. The
trol group (Fig. 2 I, J ). Thus, Se-Met impedes tau pathology by ratio of p-PP2A/PP2A significantly decreased in the hippocam-
2454 • J. Neurosci., March 1, 2017 • 37(9):2449 –2462 Zhang et al. • Selenomethionine Targets Tauopathy

Figure 3. Se-Met decreases tau phosphorylation via activation of Akt and inhibition of GSK3␤. A, B, Representative Western blot images and quantification of GSK3␤, pGSK3␤ (Ser9), PP2A,
p-PP2A (Tyr307), Akt, and p-Akt (Ser473) in the hippocampus of WT, Se-Met-treated, and control mice. C, D, Representative Western blot images and analysis of these kinases in the cortex of WT,
Se-Met-treated, and control mice. E, F, Western blot images and analysis of these kinases in primary cortical neurons. Densitometry results were normalized against the levels of GAPDH. The values
are expressed as percentages of the control (set to 100%) and presented as the group mean ⫾ SEM (n ⫽ 6). *p ⬍ 0.05; **p ⬍ 0.01.

pus of 3xTg-AD mice after the treatment of Se-Met ( p ⬍ 0.05, activities of those enzymes in WT neurons. The ratio of p-GSK3␤/
Fig. 3 A, B). However, there was no difference in the ratio of GSK3␤ significantly decreased and the ratio of p-PP2A/PP2A signif-
p-PP2A/PP2A in the cortex among WT, Se-Met-treated, and icantly increased in 3xTg neurons compared with WT neurons (p ⬍
control groups. Akt is an important upstream protein kinase that 0.001 and p ⬍ 0.01, respectively; Fig. 3E,F). Similar to the in vivo
can promote the phosphorylation of GSK3␤ at Ser9 (Dolan and results, Se-Met treatment significantly improved the p-GSK3␤/
Johnson, 2010). Similarly, Akt can be activated by phosphoryla- GSK3␤ and p-Akt/Akt ratios (p ⬍ 0.05 and p ⬍ 0.05, respectively;
tion at Ser473, and the ratio of p-Akt/Akt reflects Akt activity. The Fig. 3E,F) and had no effect on the p-PP2A/PP2A ratio.
ratio of p-Akt/Akt significantly decreased in the hippocampus
and cortex of 3xTg-AD mice compared with WT mice ( p ⬍ 0.05 Se-Met promotes tau clearance by modulating the
and p ⬍ 0.05, respectively; Fig. 3A–D), and Se-Met treatment AMPK–mTOR-mediated autophagic pathway
significantly increased p-Akt/Akt in the hippocampus ( p ⬍ 0.01; The decrease in total tau could be the result of inhibition of the
Fig. 3 A, B). transcription and translation of P301L-tau gene or the elimina-
We also examined these enzyme levels in primary cortical neu- tion of abnormal tau proteins. We found no significant difference
rons using Western blotting. Se-Met had no significant effect on the in the level of tau mRNA between Se-Met-treated and control
Zhang et al. • Selenomethionine Targets Tauopathy J. Neurosci., March 1, 2017 • 37(9):2449 –2462 • 2455

hippocampus) and p-AMPK (in the cor-

tex) were significantly decreased in the
brain of 3xTg-AD mice compared with
WT mice ( p ⬍ 0.05 and p ⬍ 0.01, respec-
tively; Fig. 5A–D). Treatment with Se-Met
significantly increased AMPK (in the
hippocampus; p ⬍ 0.5; Fig. 5 A, B) and
p-AMPK (in the hippocampus and cor-
tex; p ⬍ 0.001 and p ⬍ 0.01, respectively;
Fig. 5A–D). To validate the relationship
between tau clearance and autophagy, the
localization of LC3-II and total tau in the
hippocampus was determined using immu-
nofluorescence (Fig. 5E). Se-Met reduced
the intensity of total tau fluorescence, in-
creased the fluorescence intensity of LC3-II,
and increased the colocalization of LC3-II
with tau.
A similar phenomenon was observed
in primary cortical neurons. The level of
Figure 4. Se-Met has no effect on tau P301L transcription and increases the formation of autophagosomes in the hippocampus LC3-II in 3xTg neurons was significantly
of 3xTg-AD mice. A, B, The level of tauP301L mRNA in the hippocampus was investigated by RT-PCR. Quantitative results were lower than in WT neurons ( p ⬍ 0.05; Fig.
normalized against the level of actin. Values are expressed as percentages of the control (set to 100%) and presented as the group 5 F, G), and Se-Met treatment signifi-
mean ⫾ SEM (n ⫽ 4). C, The double-membrane-bound autophagosomes in the hippocampus of 3xTg-AD mice were observed cantly restored the level of LC3-II in 3xTg
using transmission electron microscopy. Scale bars, 0.2 ␮m. neurons ( p ⬍ 0.05; Fig. 5 F, G). The ex-
pression of mTOR and p-mTOR in 3xTg
mice (Fig. 4 A, B), which implied that the reduction in tau was neurons was significantly higher than in WT neurons ( p ⬍ 0.01
attributable to a clearance mechanism. Thus, we next focused on and p ⬍ 0.05, respectively; Fig. 5 F, G). Treatment with Se-Met
the autophagic pathway. Autophagy, a vital catabolic process in significantly inhibited mTOR and p-mTOR expression ( p ⬍ 0.05
cells, is a primary driving force for degradation of abnormal and and p ⬍ 0.01, respectively; Fig. 5 F, G). Se-Met treatment also
aggregated proteins and an essential protective response to reversed the increased expression of p70S6K and p-p70S6K
pathologic stress. Light chain protein 3 (LC3) is a commonly used ( p ⬍ 0.01 and p ⬍ 0.01, respectively; Fig. 5 F, G) and improved
marker of autophagy. This protein is converted from its inactive the level of p-AMPK ( p ⬍ 0.05; Fig. 5 F, G) in 3xTg neurons.
LC3-I form to active LC3-II during formation of the autophago- These data suggest that Se-Met promotes the clearance of tau
some. To assess autophagy after Se-Met treatment, LC3-II in the protein by regulating the AMPK–mTOR-mediated autophagic
brain was detected by Western blotting. LC3-II in the hippocam- pathway.
pus and cortex of 3xTg-AD mice was significantly lower com-
pared with WT mice ( p ⬍ 0.05 and p ⬍ 0.05, respectively; Fig. Se-Met enhances the degradation of p62 and increases
5A–D). However, this effect was significantly attenuated in cathepsin D levels in the autophagy–lysosomal pathway
Se-Met-treated mice ( p ⬍ 0.05 and p ⬍ 0.05, respectively; Fig. Defects in autophagic maturation may be a general feature of AD
5A–D). Consistent with the immunoblot results, transmission pathology. p62/SQSTM1 is involved in the targeting of ubiquiti-
electron microscopy revealed more autophagosomes in the cyto- nated proteins to autophagic vacuoles and has been identified as
plasm of hippocampal neurons from Se-Met-treated mice than one of the specific substrates that are degraded through the au-
control mice (Fig. 4C). Collectively, these results suggest that tophagy–lysosomal pathway. Since the expression of p62 is in-
Se-Met promotes the initiation of autophagy to enhance tau versely correlated with autophagic activity (Komatsu et al., 2007;
clearance. Mizushima et al., 2010), we assessed p62 levels by Western blot-
mTOR, a serine/threonine protein kinase, is a major regulator ting. The 3xTg-AD mice showed significantly increased p62 ex-
of mammalian autophagy (Schmelzle and Hall, 2000). Both the pression compared with WT mice (hippocampus, p ⬍ 0.01;
hippocampus and cortex of 3xTg-AD mice showed significantly cortex, p ⬍ 0.01; Fig. 6A–D). Treatment with Se-Met significantly
increased mTOR ( p ⬍ 0.01 and p ⬍ 0.05, respectively; Fig. 5A–D) reduced p62 levels (hippocampus, p ⬍ 0.05; cortex, p ⬍ 0.01; Fig.
and p-mTOR (Ser2448; p ⬍ 0.05 and p ⬍ 0.05, respectively; Fig. 6A–D). Cathepsin D (CatD), a principal lysosomal aspartyl
5A–D) levels compared with WT mice, and treatment with Se- protease, is highly abundant in the brain and is activated by pro-
Met significantly reduced levels of mTOR ( p ⬍ 0.01 and p ⬍ 0.05, teolysis in the acidified lysosome to produce a mature proteolytic
respectively; Fig. 5A–D) and p-mTOR ( p ⬍ 0.001 and p ⬍ 0.05, product (mCatD) (Avrahami et al., 2013). Precursor CatD
respectively; Fig. 5A–D) in both hippocampus and cortex. mTOR (46 –52 kDa) and mCatD (32 kDa) were detected in the brains of
activity is routinely measured by the phosphorylation level of its WT, Se-Met-treated, and 3xTg-AD mice. The expression of
direct downstream target p70S6K at Thr389. After treatment mCatD was significantly reduced in 3xTg-AD mice compared
with Se-Met, the levels of p70S6K and p-p70S6K in the hip- with WT mice (hippocampus, p ⬍ 0.01; cortex, p ⬍ 0.01; Fig.
pocampus were significantly decreased compared with control 6A–D). Treatment with Se-Met significantly elevated mCatD
mice ( p ⬍ 0.01 and p ⬍ 0.05, respectively; Fig. 5A–D). In addition expression in the hippocampus and cortex of 3xTg-AD mice
to mTOR, AMPK plays a significant role in the promotion of (hippocampus, p ⬍ 0.05; cortex, p ⬍ 0.01; Fig. 6A–D). Immuno-
autophagy. When activated by phosphorylation at Thr172, fluorescence staining of CatD in the hippocampus and cortex of
AMPK inhibits mTOR activation. The levels of AMPK (in the the brain revealed that the Se-Met-treated group showed a
2456 • J. Neurosci., March 1, 2017 • 37(9):2449 –2462 Zhang et al. • Selenomethionine Targets Tauopathy

Figure 5. Se-Met enhances tau clearance by promoting the initiation of autophagy via the AMPK–mTOR pathway. A, B, Representative Western blot images and densitometric analysis
of LC3-II, mTOR, p-mTOR (Ser2448), p70S6K, p-p70S6K (Thr389), AMPK␣, and pAMPK␣ (Thr172) expression in the hippocampus of WT, Se-Met-treated, and control mice. C, D,
Representative Western blot images and densitometric analysis of these kinases in the cortex of WT, Se-Met-treated, and control mice. E, Immunofluorescence staining was performed
to detect the expression and localization of LC3 and total tau in the hippocampus (DG, CA1, and CA3). F, G, Representative Western blot images and densitometric analysis of LC3-II, mTOR,
p-mTOR (Ser2448), p70S6K, p-p70S6K (Thr389), AMPK␣, and pAMPK␣ (Thr172) in primary cortical neurons. Densitometry results were normalized against the levels of GAPDH. Values
are expressed as percentages of the control (set to 100%) and presented as the group mean ⫾ SEM (n ⫽ 6). *p ⬍ 0.05; **p ⬍ 0.01; ***p ⬍ 0.001. Scale bar, 50 ␮m. DG, Dentate gyrus;
CA1, cornu ammonis 1; CA3, cornu ammonis 3.

marked increase in CatD-positive area compared with the con- in the autolysosome fraction in 3xTg-AD mice. After treatment
trol group (Fig. 6E). Similar phenomena were observed in the with Se-Met, the levels of p62 in both fractions were restored to
primary cortical neurons. Se-Met treatment significantly inhib- WT levels (Fig. 7A).
ited the p62 level ( p ⬍ 0.001; Fig. 6 F, G) and increased mCatD Additional immunoblot analyses of these subcellular frac-
expression ( p ⬍ 0.05; Fig. 6 F, G) in 3xTg neurons. tions using anti-tau/pS404-tau antibodies revealed increased
tau protein accumulation within the lysosomal compartments
Se-Met enhances autophagic proteolysis and autophagic flux of WT and Se-Met-treated mice compared with 3xTg-AD mice
LC3-II first appears on newly formed autophagosomes and is (Fig. 7A,B). Interestingly, pS404-tau levels were increased in the
then quickly degraded by lysosomal enzymes after autophagoly- autophagosome and decreased in the autolysosome in 3xTg-AD
sosomal fusion. To investigate whether autophagic proteolysis mice. Se-Met treatment potently restored pS404-tau levels in
and protein turnover can be restored by Se-Met treatment, lyso- these fractions to WT levels (Fig. 7 A, C).
somes and autophagic vacuoles of relatively lower (autophago- Based on these data, a N2a cell line overexpressing mutant
some) and higher (autolysosome) density were isolated from the human swedish-APP695 was treated with bafilomycin A1. Bafilo-
brains of three groups of mice by subcellular fractionation using mycin A1 is an autophagic flux inhibitor that blocks the fusion of
a Nycodenz density gradient. The identities of these organelles autophagosomes with the lysosome, thereby inhibiting the au-
were verified by the presence of LC3-II and/or LAMP1 (a marker tophagic degradation of proteins, including LC3-II and p62. The
for lysosome) immunoreactivity using immunoblot analyses optimal concentration and treatment time with bafilomycin A1
(Fig. 7A). The lysosome fractions of 3xTg-AD mice contained were determined to be 50 nM and 4 h, respectively (data not
abnormally low levels of LC3-II compared with WT mice. Levels shown). The ratio of LC3-II/LC3-I was significantly increased by
of p62 were higher in the autophagosome fraction and decreased blocking the transformation of the autophagosome to the autoly-
Zhang et al. • Selenomethionine Targets Tauopathy J. Neurosci., March 1, 2017 • 37(9):2449 –2462 • 2457

Figure 6. Se-Met decreases p62 protein expression and increases cathepsin D protein expression. A–D, Representative Western blot images of p62 and CatD expression in the
hippocampus (A, B) and cortex (C, D) of WT, Se-Met-treated, and control mice. E, Immunofluorescence staining of CatD in the hippocampus (DG, CA1, and CA3) and cortex. F, G,
Representative Western blot images of p62 and CatD expression in primary cortical neurons. Densitometry results were normalized against the levels of GAPDH. Values are expressed as
percentages of the control (set to 100%) and presented as the group mean ⫾ SEM (n ⫽ 6). *p ⬍ 0.05; **p ⬍ 0.01. Scale bar, 50 ␮m. DG, Dentate gyrus; CA1, cornu ammonis 1; CA3,
cornu ammonis 3.

sosome with bafilomycin A1 ( p ⬍ 0.001; Fig. 8 A, B). Cotreatment lysosome quenches the fluorescent signal of mWasabi but ex-
of bafilomycin A1 with Se-Met significantly inhibited the bafilo- erts a weak effect on mRFP. In green/red merged images, yel-
mycin A1-induced increase in LC3-II/LC3-I ( p ⬍ 0.05; Fig. low puncta (i.e., RFP ⫹mWasabi ⫹) indicate autophagosomes,
8 A, B). Immunofluorescence staining of LC3 (Fig. 8C) confirmed whereas red puncta (i.e., RFP ⫹mWasabi ⫺) indicate autolyso-
that Se-Met restored autophagic degradation. somes (Zhou et al., 2012). Thus, autophagic flux is increased
We also used an improved tandem fluorescent-tagged LC3 when the presence of both yellow and red puncta is increased
(mTagRFP-mWasabi-LC3) to monitor autophagic flux based in the cells, and autophagic flux is blocked when only the
on the different pH stabilities of green and red fluorescent presence of yellow puncta is increased in cells (Zhou et al.,
proteins. The acidic environment (less than pH 5) inside the 2012). The mTagRFP-mWasabi-LC3 plasmid was transfected
2458 • J. Neurosci., March 1, 2017 • 37(9):2449 –2462 Zhang et al. • Selenomethionine Targets Tauopathy

Figure 7. Se-Met influences the levels of total tau and phosphorylated tau in the autolysosome, autolysosome, and lysosome fractions isolated from the brain of 3xTg-AD mice. A, Autophago-
some, autolysosome, and lysosome fractions were isolated from the brain of WT, Se-Met-treated, and control mice. The relative enrichment of autophagic and lysosomal protein markers was
detected by Western blot with anti-LC3 and anti-Lamp1 antibodies (top two rows), respectively. The expression of p62, total tau, and phosphorylated tau in these fractions was detected (bottom
three rows). B, C, Densitometry analysis of total tau (B) and phosphorylated tau (C) expression in different fractions. CBB, Coomassie Brilliant Blue.

into N2a-sw cells, followed by treatment with bafilomycin A1 strated that long-term treatment with Se-Met reduced tau hyper-
and/or Se-Met. Bafilomycin A1 treatment markedly increased phosphorylation by regulation of the Akt/GSK3␤ pathway and
the number of yellow puncta in cells, whereas the cotreatment cleared tau protein by the AMPK–mTOR-mediated autophagic
with bafilomycin A1 and Se-Met notably improved the num- pathway. Furthermore, Se-Met ameliorated the cognitive impair-
ber of yellow and red puncta in the cells compared with bafi- ment in older 3xTg-AD mice. Our results provide the first evi-
lomycin A1 treatment only (Fig. 8D). These data indicate that dence for the effect of Se on autophagy in neurodegenerative
Se-Met modulates autophagic flux to promote autophagic disease.
proteolysis. Tau is mainly expressed in neurons and is abundant in neuronal
axons (Z. Zhang et al., 2014). Normal post-translational phosphor-
Discussion ylation of tau controls microtubule dynamics and establishes neuro-
Se is widely recognized as an essential trace element and an abun- nal polarity, axonal outgrowth, and axonal transport. Pathological
dant micronutrient in the brain. Evidence suggests that Se levels
hyperphosphorylated tau, the main component of NFTs in AD, se-
in the brain decrease with age and are associated with a decline of
verely interferes with tau’s ability to regulate microtubules dynamics
cognitive function in AD patients (Gao et al., 2007, 2012). Thus,
(Caceres and Kosik, 1990; Harada et al., 1994), displays an increased
Se is vital for central neuron system function. Se supplementation
propensity to form paired helical filaments in vitro, and sequesters
significantly enhanced mitochondrial function and protected
cells from apoptosis in primary neurons; in addition, the treat- full-length tau and other microtubule-associated proteins (Alonso et
ment reduced A␤ production, mitigated tau pathology, and ame- al., 2001). As expected, Se-Met reduced the hyperphosphorylation of
liorated cognitive deficits and oxidative damage in AD animal tau in 11-month-old 3xTg-AD mice (after a 3-month treatment).
models (Gwon et al., 2010; van Eersel et al., 2010; Mendelev et al., These data suggest that Se-Met has a significant therapeutic effect on
2012). Se-Met, the predominant chemical form of Se ingested by tauopathy. An imbalanced regulation in protein kinases and protein
humans, delayed the decline in cognitive function, reduced tau phosphatases is the direct cause for tau hyperphosphorylation in
hyperphosphorylation, and ameliorated synaptic deficits in AD. As the kinase and phosphatase playing key roles in the regula-
4-month-old 3xTg-AD mice (Song et al., 2014). The principal tion of tau phosphorylation, GSK3␤ phosphorylates tau at multiple
pathological hallmarks (e.g., hyperphosphorylation of tau, A␤ AD-related sites in a site-specific manner (Hanger et al., 2009) and
deposition, and inflammation) were significantly increased in the PP2A accounts for ⬃70% of the total tau phosphatase activity in the
hippocampus and cortex of 8-month-old 3xTg-AD mice. Thus, brain and plays a dominant role in the regulation of tau dephosphor-
determining the effect of Se-Met in aged AD mice is important ylation. Se-Met significantly inhibited the activity of GSK3␤ and
because AD is not diagnosed in patients until advanced stages. In increased that of PP2A in aged 3xTg-AD mice. This result can be
our present study, 8-month-old 3xTg-AD mice were treated with explained by the dynamic cross talk between GSK3␤ and PP2A at
6 ␮g/ml Se-Met in their drinking water for 12 weeks to explore its different stages of AD. Indeed, these proteins are regulated by each
effect on the phenotype, especially tau pathology. We demon- other and control tau phosphorylation both directly and indirectly
Zhang et al. • Selenomethionine Targets Tauopathy J. Neurosci., March 1, 2017 • 37(9):2449 –2462 • 2459

Figure 8. Se-Met enhances autophagic flux by promoting autophagolysosomal fusion. N2a-sw cells were treated with bafilomycin A1 (50 nM, 4 h) to block the fusion of autophagosomes with
lysosomes, and the effect of Se-Met (1 ␮M) on autophagic flux was assessed. A, B, The expression of p62 and LC3II was detected by Western blot. Quantitative results were normalized against the
levels of ␤-actin. Values are expressed as percentages of the control (set to 100%) and presented as the group mean ⫾ SEM (n ⫽ 3). *p ⬍ 0.05; **p ⬍ 0.01; ***p ⬍ 0.001. C, LC3 expression was
detected by immunofluorescence staining in N2a-sw cells. D, Representative fluorescent images of N2a-sw cells transiently transfected with mTagRFP-mWasabi-LC3, followed by bafilomycin A1
and/or Se-Met treatment. Scale bars, 10 and 20 ␮m.

(Wang et al., 2015). The activity of Akt, the most important up- esis of AD (Boland et al., 2008). The reduction of autophagic
stream kinase that phosphorylates GSK3␤ at Ser9, was significantly activity that occurs in tauopathies results in the accumulation of
increased after Se-Met treatment. This suggests Se-Met reduces tau abnormal and hyperphosphorylated tau (Inoue et al., 2012; Z.
hyperphosphorylation by promoting Akt activity, which subse- Zhang et al., 2014). Thus, modulation of the autophagic pathway
quently inhibits GSK3␤ activity. is a promising approach for the clearance of tau and hyperphos-
Although hyperphosphorylation is a potent inducer of tau phorylated tau. The protein kinase mTOR is important in the
pathology, tau-mediated neurodegeneration in AD may result autophagy pathway and plays a key role in controlling cellular
from the combination of toxic gain of function in pathological homeostasis via activation of p70S6 kinase and inhibition of au-
tau (abnormal or aggregated forms) and loss of function in nor- tophagy (Klionsky and Emr, 2000). Overwhelming evidence has
mal tau (Ballatore et al., 2007). Tau hyperphosphorylation at shown that pharmacological or genetic reduction in mTOR
specific sites reduces its ability to bind microtubules and mark- signaling increases the formation and activity of autophagy
edly promotes self-aggregation (Liu et al., 2007). Mutant human (Heras-Sandoval et al., 2014). Upregulated mTOR and p70S6K
P301L-tau is overexpressed in 3xTg-AD mice. Previous studies are associated with increased tau accumulation in AD brains, and
on these mice have demonstrated that the conformational change inhibition of mTOR decreases the level of total tau in vitro and in
in P301L-tau is a necessary intermediate for further phosphory- vivo (Pei et al., 2006; Caccamo et al., 2010). mTOR also colocal-
lation and aggregation (Shibuya et al., 2015). Thus, neurons in izes with NFTs, mediates tau phosphorylation, and enhances tau
the mouse brain are fully capable of forming tau aggregates and aggregation (Tang et al., 2013). Consistent with these findings,
tangles (Terwel et al., 2005). A reduction in tau has a protective we discovered that Se-Met significantly decreased mTOR and
effect in cultured neurons and animal models of AD (Roberson et p70S6K expression both in vitro and in vivo. AMPK inhibits
al., 2007). In this manner, the inhibition or clearance of P301L- mTORC1 activity through activation of the tuberous sclerosis
tau may be an effective therapeutic strategy for AD. 1/tuberous sclerosis 2 complex (Maiese, 2016). Thus, Se-Met
Se-Met has no influence on the expression of the TauP301L initiates autophagy by up-regulating AMPK and inhibiting the
gene, and the reduction of tau by Se-Met can be achieved only by activation of the mTOR pathway during the early stages of
enhancing tau clearance. Autophagy is an essential process that autophagy.
initiates the cellular clearance of misfolded and aggregated pro- Enhanced lysosomal activity by Se-Met may also play a key
teins via autophagolysosomal fusion (Mizushima et al., 2010). role in the clearance of aggregated tau by effectively degrading
Autophagic dysfunction has been directly linked to the pathogen- LC3-II-positive autophagosomes. The p62 protein is selectively
2460 • J. Neurosci., March 1, 2017 • 37(9):2449 –2462 Zhang et al. • Selenomethionine Targets Tauopathy

incorporated into autophagosomes via direct binding to LC3, Alzheimer’s Association (2015) 2015 Alzheimer’s disease facts and figures.
and p62 is efficiently degraded by autophagy. Thus, p62 levels are Alzheimers Dement 11:332–384. CrossRef Medline
Avrahami L, Farfara D, Shaham-Kol M, Vassar R, Frenkel D, Eldar-
inversely correlated with autophagic flux. The lysosomal proteo-
Finkelman H (2013) Inhibition of glycogen synthase kinase-3 amelio-
lytic enzyme cathepsin D is the only aspartic-type protease ubiq- rates beta-amyloid pathology and restores lysosomal acidification and
uitously expressed in all cells of the human body and abundantly mammalian target of rapamycin activity in the Alzheimer disease mouse
expressed in the brain (Shacka and Roth, 2007). Generally, ca- model: in vivo and in vitro studies. J Biol Chem 288:1295–1306. CrossRef
thepsin D-mediated proteolysis is essential to neuronal cell ho- Medline
meostasis and involves the degradation of unfolded or oxidized Balducci C, Forloni G (2014) In vivo application of beta amyloid oligomers:
protein aggregates delivered to lysosomes via autophagy. Neurons a simple tool to evaluate mechanisms of action and new therapeutic ap-
proaches. Curr Pharm Des 20:2491–2505. CrossRef Medline
are particularly vulnerable to the loss of cathepsin D function
Ballatore C, Lee VM, Trojanowski JQ (2007) Tau-mediated neurodegenera-
(Vidoni et al., 2016). A number of studies have demonstrated that tion in Alzheimer’s disease and related disorders. Nat Rev Neurosci
cathepsin D expression is decreased in AD (Urbanelli et al., 2008), 8:663– 672. CrossRef Medline
and enhancing cathepsin D expression and activity restored the Berger Z, Roder H, Hanna A, Carlson A, Rangachari V, Yue M, Wszolek Z,
A␤1– 42-disturbed acid environment of lysosomes and promoted Ashe K, Knight J, Dickson D, Andorfer C, Rosenberry TL, Lewis J, Hutton
the fusion and clearance of autophagosomes with lysosomes M, Janus C (2007) Accumulation of pathological tau species and mem-
(Zhang and Zhao, 2015). Se-Met decreased p62 expression and ory loss in a conditional model of tauopathy. J Neurosci 27:3650 –3662.
CrossRef Medline
increased cathepsin D expression in primary cortical neurons and
Boland B, Kumar A, Lee S, Platt FM, Wegiel J, Yu WH, Nixon RA (2008)
the hippocampus and cortex of 3xTg-AD mice. Thus, Se-Met Autophagy induction and autophagosome clearance in neurons: relation-
promoted autophagolysosomal fusion and improved autophagy ship to autophagic pathology in Alzheimer’s disease. J Neurosci 28:6926 –
flux, which enhanced the clearance of abnormal tau proteins. The 6937. CrossRef Medline
N2a-sw cell line is a well known cellular model of AD. These cells Bryan KJ, Lee H, Perry G, Smith MA, Casadesus G (2009) Transgenic mouse
were treated with bafilomycin A1 in the absence or presence of models of Alzheimer’s disease: behavioral testing and considerations. In:
Se-Met to verify the effect of Se-Met on autophagic flux in AD. In Methods of behavior analysis in neuroscience, Ed 2 (Buccafusco JJ, ed),
Chapter 1. Boca Raton, FL: CRC Press/Taylor & Francis.
agreement with our in vivo results, Se-Met reversed the accumu-
Caccamo A, Majumder S, Richardson A, Strong R, Oddo S (2010) Mo-
lation of LC-II and p62 and increased the presence of yellow lecular interplay between mammalian target of rapamycin (mTOR),
and red puncta in mRFP-mWasabi-tandem fluorescent LC3- amyloid-beta, and Tau: effects on cognitive impairments. J Biol Chem
transfected cells. These data show that Se-Met effectively restores 285:13107–13120. CrossRef Medline
bafilomycin-blocked autophagic flux. Caceres A, Kosik KS (1990) Inhibition of neurite polarity by tau antisense
Why does Se-Met influence the autophagy pathway in AD? oligonucleotides in primary cerebellar neurons. Nature 343:461– 463.
Accumulating evidence suggests that oxidative stress plays a cen- CrossRef Medline
Castro-Alvarez JF, Uribe-Arias A, Cardona-GómezGP (2015) Cyclin-
tral role in the common pathophysiology of neurodegenerative
dependent kinase 5 targeting prevents beta-amyloid aggregation involv-
diseases, including AD (G. H. Kim et al., 2015). The AD brain ing glycogen synthase kinase 3beta and phosphatases. J Neurosci Res 93:
exhibits a significant increase in oxidative damage to lipids, pro- 1258 –1266. CrossRef Medline
tein, DNA, and RNA (Venkateshappa et al., 2012). Oxidative Cuervo AM, Palmer A, Rivett AJ, Knecht E (1995) Degradation of protea-
stress, which is present early in AD development, is influenced by somes by lysosomes in rat liver. Eur J Biochem 227:792– 800. CrossRef
autophagy. Thus, autophagy may be the link between oxidative Medline
stress and AD (Zheng et al., 2006). Oxidative stress and free rad- Ding H, Matthews TA, Johnson GV (2006) Site-specific phosphorylation
and caspase cleavage differentially impact tau-microtubule interactions
ical damage play principal roles in cell death via the autophagic
and tau aggregation. J Biol Chem 281:19107–19114. CrossRef Medline
pathway and lysosome dysfunction (inhibition of lysosome en- Dolan PJ, Johnson GV (2010) The role of tau kinases in Alzheimer’s disease.
zyme function and lysosome membrane damage; Pivtoraiko et Curr Opin Drug Discov Dev 13:595– 603. Medline
al., 2009). Oxidative stress induces lysosomal membrane permea- Eckert A, Schulz KL, Rhein V, GötzJ (2010) Convergence of amyloid-beta
bilization, improves lysosomal pH, and inhibits cathepsin D ac- and tau pathologies on mitochondria in vivo. Mol Neurobiol 41:107–114.
tivity (Hensley and Harris-White, 2015). Conversely, autophagy CrossRef Medline
dysfunction can increase oxidative stress in AD (Zheng et al., Gao S, Jin Y, Hall KS, Liang C, Unverzagt FW, Ji R, Murrell JR, Cao J, Shen J,
Ma F, Matesan J, Ying B, Cheng Y, Bian J, Li P, Hendrie HC (2007)
2006). Well characterized are the major biochemical and phar-
Selenium level and cognitive function in rural elderly Chinese. Am J Epi-
macological properties of Se and Se-Met in contributing to oxi- demiol 165:955–965. CrossRef Medline
dation resistance exerted by related selenoproteins and Gao S, Jin Y, Unverzagt FW, Liang C, Hall KS, Cao J, Ma F, Murrell JR, Cheng
selenoenzymes. Therefore, there may be a close relationship be- Y, Li P, Bian J, Hendrie HC (2012) Selenium level and depressive symp-
tween the effect of Se-Met on autophagy activation and its anti- toms in a rural elderly Chinese cohort. BMC Psychiatry 12:72. CrossRef
oxidative activity. Additional studies are required to test this Medline
hypothesis. Gendron TF, Petrucelli L (2009) The role of tau in neurodegeneration. Mol
Neurodegener 4:13. CrossRef Medline
In conclusion, our present study is the first to demonstrate that
Gwon AR, Park JS, Park JH, Baik SH, Jeong HY, Hyun DH, Park KW, Jo DG
Se-Met reduces tau hyperphosphorylation via modulation of the (2010) Selenium attenuates A beta production and A beta-induced neu-
Akt/GSK3␤ pathway, enhances autophagic flux and increases tau ronal death. Neurosci Lett 469:391–395. CrossRef Medline
clearance through the autophagy–lysosome pathway, and rescues Hanger DP, Anderton BH, Noble W (2009) Tau phosphorylation: the ther-
spatial learning and memory impairments in aged 3xTg-AD mice. apeutic challenge for neurodegenerative disease. Trends Mol Med 15:
These data shed light on a novel mechanism behind the therapeutic 112–119. CrossRef Medline
effect of Se-Met in AD. Harada A, Oguchi K, Okabe S, Kuno J, Terada S, Ohshima T, Sato-Yoshitake
R, Takei Y, Noda T, Hirokawa N (1994) Altered microtubule organiza-
References tion in small-calibre axons of mice lacking tau protein. Nature 369:488 –
Alonso A, Zaidi T, Novak M, Grundke-Iqbal I, Iqbal K (2001) Hyperphos- 491. CrossRef Medline
phorylation induces self-assembly of tau into tangles of paired helical Hensley K, Harris-White ME (2015) Redox regulation of autophagy in
filaments/straight filaments. Proc Natl Acad Sci U S A 98:6923– 6928. healthy brain and neurodegeneration. Neurobiol Dis 84:50 –59. CrossRef
CrossRef Medline Medline
Zhang et al. • Selenomethionine Targets Tauopathy J. Neurosci., March 1, 2017 • 37(9):2449 –2462 • 2461

Heras-Sandoval D, Pérez-Rojas JM, Hernández-DamiánJ, Pedraza-Chaverri ate R, Mattson MP, Akbari Y, LaFerla FM (2003a) Triple-transgenic
J (2014) The role of PI3K/AKT/mTOR pathway in the modulation of model of Alzheimer’s disease with plaques and tangles: intracellular Abeta
autophagy and the clearance of protein aggregates in neurodegeneration. and synaptic dysfunction. Neuron 39:409 – 421. CrossRef Medline
Cell Signal 26:2694 –2701. CrossRef Medline Oddo S, Caccamo A, Kitazawa M, Tseng BP, LaFerla FM (2003b) Amyloid
Hernandez F, Lucas JJ, Avila J (2013) GSK3 and tau: two convergence points deposition precedes tangle formation in a triple transgenic model of Alz-
in Alzheimer’s disease. J Alzheimers Dis 33 [Suppl 1]:S141–S144. heimer’s disease. Neurobiol Aging 24:1063–1070. CrossRef Medline
Hu ZH, Tian Q, Wang JZ (2006) [Role of serine/threonine protein phos- Orr ME, Salinas A, Buffenstein R, Oddo S (2014) Mammalian target of rapa-
phatase in tau phosphorylation]. Sheng Li Ke Xue Jin Zhan 37:173–174. mycin hyperactivity mediates the detrimental effects of a high sucrose diet
Medline on Alzheimer’s disease pathology. Neurobiol Aging 35:1233–1242.
Hunsberger HC, Rudy CC, Weitzner DS, Zhang C, Tosto DE, Knowlan K, Xu CrossRef Medline
Y, Reed MN (2014) Effect size of memory deficits in mice with adult- Pei JJ, An WL, Zhou XW, Nishimura T, Norberg J, Benedikz E, Götz J, Win-
onset P301L tau expression. Behav Brain Res 272:181–195. CrossRef blad B (2006) P70 S6 kinase mediates tau phosphorylation and synthe-
Medline sis. FEBS Lett 580:107–114. CrossRef Medline
Hurtado DE, Molina-Porcel L, Iba M, Aboagye AK, Paul SM, Trojanowski JQ, Pivtoraiko VN, Stone SL, Roth KA, Shacka JJ (2009) Oxidative stress and
Lee VM (2010) A{beta} accelerates the spatiotemporal progression of autophagy in the regulation of lysosome-dependent neuron death. Anti-
tau pathology and augments tau amyloidosis in an Alzheimer mouse oxid Redox Signal 11:481– 496. CrossRef Medline
model. Am J Pathol 177:1977–1988. CrossRef Medline Roberson ED, Scearce-Levie K, Palop JJ, Yan F, Cheng IH, Wu T, Gerstein H,
Inoue K, Rispoli J, Kaphzan H, Klann E, Chen EI, Kim J, Komatsu M, Abe- Yu GQ, Mucke L (2007) Reducing endogenous tau ameliorates amyloid
liovich A (2012) Macroautophagy deficiency mediates age-dependent beta-induced deficits in an Alzheimer’s disease mouse model. Science
neurodegeneration through a phospho-tau pathway. Mol Neurodegener 316:750 –754. CrossRef Medline
7:48. CrossRef Medline Santacruz K, Lewis J, Spires T, Paulson J, Kotilinek L, Ingelsson M, Guimaraes
Jiang T, Yu JT, Zhu XC, Zhang QQ, Tan MS, Cao L, Wang HF, Shi JQ, Gao L, A, DeTure M, Ramsden M, McGowan E, Forster C, Yue M, Orne J, Janus
Qin H, Zhang YD, Tan L (2015) Ischemic preconditioning provides C, Mariash A, Kuskowski M, Hyman B, Hutton M, Ashe KH (2005) Tau
neuroprotection by induction of AMP-activated protein kinase- suppression in a neurodegenerative mouse model improves memory
dependent autophagy in a rat model of ischemic stroke. Mol Neurobiol function. Science 309:476 – 481. CrossRef Medline
51:220 –229. CrossRef Medline Schmelzle T, Hall MN (2000) TOR, a central controller of cell growth. Cell
Kim GH, Kim JE, Rhie SJ, Yoon S (2015) The role of oxidative stress in 103:253–262. CrossRef Medline
neurodegenerative diseases. Exp Neurobiol 24:325–340. CrossRef Shacka JJ, Roth KA (2007) Cathepsin D deficiency and NCL/Batten disease:
Medline there’s more to death than apoptosis. Autophagy 3:474 – 476. CrossRef
Kim Y, Park H, Nah J, Moon S, Lee W, Hong SH, Kam TI, Jung YK (2015) Medline
Essential role of POLDIP2 in Tau aggregation and neurotoxicity via au- Shibuya Y, Chang CC, Huang LH, Bryleva EY, Chang TY (2014) Inhibiting
tophagy/proteasome inhibition. Biochem Biophys Res Commun 462: ACAT1/SOAT1 in microglia stimulates autophagy-mediated lysosomal
112–118. CrossRef Medline proteolysis and increases Abeta1– 42 clearance. J Neurosci 34:14484 –
Klionsky DJ, Emr SD (2000) Autophagy as a regulated pathway of cellular 14501. CrossRef Medline
degradation. Science 290:1717–1721. CrossRef Medline Shibuya Y, Niu Z, Bryleva EY, Harris BT, Murphy SR, Kheirollah A, Bowen
Komatsu M, Waguri S, Chiba T, Murata S, Iwata J, Tanida I, Ueno T, Koike ZD, Chang CC, Chang TY (2015) Acyl-coenzyme A:cholesterol acyl-
M, Uchiyama Y, Kominami E, Tanaka K (2006) Loss of autophagy in the transferase 1 blockage enhances autophagy in the neurons of triple trans-
central nervous system causes neurodegeneration in mice. Nature 441: genic Alzheimer’s disease mouse and reduces human P301L-tau content
880 – 884. CrossRef Medline at the presymptomatic stage. Neurobiol Aging 36:2248 –2259. CrossRef
Komatsu M, Waguri S, Koike M, Sou YS, Ueno T, Hara T, Mizushima N, Medline
Iwata J, Ezaki J, Murata S, Hamazaki J, Nishito Y, Iemura S, Natsume T, Siman R, Cocca R, Dong Y (2015) The mTOR inhibitor rapamycin miti-
Yanagawa T, Uwayama J, Warabi E, Yoshida H, Ishii T, Kobayashi A, et al. gates perforant pathway neurodegeneration and synapse loss in a mouse
(2007) Homeostatic levels of p62 control cytoplasmic inclusion body model of early-stage Alzheimer-type tauopathy. PLoS One 10:e0142340.
formation in autophagy-deficient mice. Cell 131:1149 –1163. CrossRef CrossRef Medline
Medline Song G, Zhang Z, Wen L, Chen C, Shi Q, Zhang Y, Ni J, Liu Q (2014)
Kopeikina KJ, Hyman BT, Spires-Jones TL (2012) Soluble forms of tau are Selenomethionine ameliorates cognitive decline, reduces tau hyperphos-
toxic in Alzheimer’s disease. Transl Neurosci 3:223–233. Medline phorylation, and reverses synaptic deficit in the triple transgenic mouse
Levine B, Klionsky DJ (2004) Development by self-digestion: molecular model of Alzheimer’s disease. J Alzheimers Dis 41:85–99. Medline
mechanisms and biological functions of autophagy. Dev Cell 6:463– 477. Sontag JM, Sontag E (2014) Protein phosphatase 2A dysfunction in Alzhei-
CrossRef Medline mer’s disease. Front Mol Neurosci 7:16. Medline
Liu F, Li B, Tung EJ, Grundke-Iqbal I, Iqbal K, Gong CX (2007) Site-specific Su J, Zhang T, Wang K, Zhu T, Li X (2014) Autophagy activation contrib-
effects of tau phosphorylation on its microtubule assembly activity and utes to the neuroprotection of remote ischemic perconditioning against
self-aggregation. Eur J Neurosci 26:3429 –3436. CrossRef Medline focal cerebral ischemia in rats. Neurochem Res 39:2068 –2077. CrossRef
Maiese K (2016) Targeting molecules to medicine with mTOR, autophagy, Medline
and neurodegenerative disorders. Br J Clin Pharmacol 82:1245–1266. Sui HJ, Zhang LL, Liu Z, Jin Y (2015) Atorvastatin prevents Abeta oligomer-
CrossRef Medline induced neurotoxicity in cultured rat hippocampal neurons by inhibiting
Maphis N, Xu G, Kokiko-Cochran ON, Cardona AE, Ransohoff RM, Lamb Tau cleavage. Acta Pharmacol Sin 36:553–564. CrossRef Medline
BT, Bhaskar K (2015) Loss of tau rescues inflammation-mediated neu- Tang Z, Bereczki E, Zhang H, Wang S, Li C, Ji X, Branca RM, LehtiöJ, Guan Z,
rodegeneration. Front Neurosci 9:196. Medline Filipcik P, Xu S, Winblad B, Pei JJ (2013) Mammalian target of rapamy-
Marzella L, Ahlberg J, Glaumann H (1982) Isolation of autophagic vacuoles cin (mTor) mediates tau protein dyshomeostasis: implication for Alzhei-
from rat liver: morphological and biochemical characterization. J Cell mer disease. J Biol Chem 288:15556 –15570. CrossRef Medline
Biol 93:144 –154. CrossRef Medline Terwel D, Lasrado R, Snauwaert J, Vandeweert E, Van Haesendonck C,
Mendelev N, Mehta SL, Idris H, Kumari S, Li PA (2012) Selenite stimulates Borghgraef P, Van Leuven F (2005) Changed conformation of mutant
mitochondrial biogenesis signaling and enhances mitochondrial func- Tau-P301L underlies the moribund tauopathy, absent in progressive,
tional performance in murine hippocampal neuronal cells. PLoS One nonlethal axonopathy of Tau-4R/2N transgenic mice. J Biol Chem 280:
7:e47910. CrossRef Medline 3963–3973. CrossRef Medline
Mizushima N, Levine B, Cuervo AM, Klionsky DJ (2008) Autophagy fights Urbanelli L, Emiliani C, Massini C, Persichetti E, Orlacchio A, Pelicci G, Sorbi
disease through cellular self-digestion. Nature 451:1069 –1075. CrossRef S, Hasilik A, Bernardi G, Orlacchio A (2008) Cathepsin D expression is
Medline decreased in Alzheimer’s disease fibroblasts. Neurobiol Aging 29:12–22.
Mizushima N, Yoshimori T, Levine B (2010) Methods in mammalian au- CrossRef Medline
tophagy research. Cell 140:313–326. CrossRef Medline van Eersel J, Ke YD, Liu X, Delerue F, Kril JJ, Götz J, Ittner LM (2010)
Oddo S, Caccamo A, Shepherd JD, Murphy MP, Golde TE, Kayed R, Mether- Sodium selenate mitigates tau pathology, neurodegeneration, and func-
2462 • J. Neurosci., March 1, 2017 • 37(9):2449 –2462 Zhang et al. • Selenomethionine Targets Tauopathy

tional deficits in Alzheimer’s disease models. Proc Natl Acad Sci U S A mouse model: prevention by correcting lysosomal proteolytic deficits.
107:13888 –13893. CrossRef Medline Brain 137:3300 –3318. CrossRef Medline
Venkateshappa C, Harish G, Mahadevan A, Srinivas Bharath MM, Shankar Yu CJ, Liu W, Chen HY, Wang L, Zhang ZR (2014) BACE1 RNA interfer-
SK (2012) Elevated oxidative stress and decreased antioxidant function ence improves spatial memory and attenuates Abeta burden in a
in the human hippocampus and frontal cortex with increasing age: impli- streptozotocin-induced tau hyperphosphorylated rat model. Cell
cations for neurodegeneration in Alzheimer’s disease. Neurochem Res Biochem Funct 32:590 –596. CrossRef Medline
37:1601–1614. CrossRef Medline Zare-Shahabadi A, Masliah E, Johnson GV, Rezaei N (2015) Autophagy in
Vidoni C, Follo C, Savino M, Melone MA, Isidoro C (2016) The role of Alzheimer’s disease. Rev Neurosci 26:385–395. Medline
cathepsin D in the pathogenesis of human neurodegenerative disorders. Zhang Y, Ma RH, Li XC, Zhang JY, Shi HR, Wei W, Luo DJ, Wang Q, Wang
Med Res Rev 36:845– 870. CrossRef Medline JZ, Liu GP (2014) Silencing [Formula: see text] rescues tau pathologies
Vossel KA, Zhang K, Brodbeck J, Daub AC, Sharma P, Finkbeiner S, Cui B, and memory deficits through rescuing PP2A and inhibiting GSK-3beta
signaling in human tau transgenic mice. Front Aging Neurosci 6:123.
Mucke L (2010) Tau reduction prevents Abeta-induced defects in ax-
Medline
onal transport. Science 330:198. CrossRef Medline
Zhang YD, Zhao JJ (2015) TFEB participates in the Abeta-induced patho-
Vossel KA, Xu JC, Fomenko V, Miyamoto T, Suberbielle E, Knox JA, Ho K,
genesis of Alzheimer’s disease by regulating the autophagy-lysosome
Kim DH, Yu GQ, Mucke L (2015) Tau reduction prevents Abeta-
pathway. DNA Cell Biol 34:661– 668. CrossRef Medline
induced axonal transport deficits by blocking activation of GSK3beta.
Zhang Z, Song M, Liu X, Kang SS, Kwon IS, Duong DM, Seyfried NT, Hu WT,
J Cell Biol 209:419 – 433. CrossRef Medline Liu Z, Wang JZ, Cheng L, Sun YE, Yu SP, Levey AI, Ye K (2014) Cleavage
Wang JZ, Gao X, Wang ZH (2014) The physiology and pathology of of tau by asparagine endopeptidase mediates the neurofibrillary pathol-
microtubule-associated protein tau. Essays Biochem 56:111–123. CrossRef ogy in Alzheimer’s disease. Nat Med 20:1254 –1262. CrossRef Medline
Medline Zheng L, Marcusson J, Terman A (2006) Oxidative stress and Alzheimer
Wang Y, Yang R, Gu J, Yin X, Jin N, Xie S, Wang Y, Chang H, Qian W, Shi J, disease: the autophagy connection? Autophagy 2:143–145. CrossRef
Iqbal K, Gong CX, Cheng C, Liu F (2015) Cross talk between PI3K- Medline
AKT-GSK-3beta and PP2A pathways determines tau hyperphosphoryla- Zhou C, Zhong W, Zhou J, Sheng F, Fang Z, Wei Y, Chen Y, Deng X, Xia B,
tion. Neurobiol Aging 36:188 –200. CrossRef Medline Lin J (2012) Monitoring autophagic flux by an improved tandem
Yang DS, Stavrides P, Saito M, Kumar A, Rodriguez-Navarro JA, Pawlik M, fluorescent-tagged LC3 (mTagRFP-mWasabi-LC3) reveals that high-
Huo C, Walkley SU, Saito M, Cuervo AM, Nixon RA (2014) Defective dose rapamycin impairs autophagic flux in cancer cells. Autophagy
macroautophagic turnover of brain lipids in the TgCRND8 Alzheimer 8:1215–1226. CrossRef Medline