quantitative proteomics reveals that pea15 regulates...
TRANSCRIPT
J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com
ScienceDirect
www.e l sev i e r . com/ loca te / j p ro t
Quantitative proteomics reveals that PEA15
regulates astroglial Aβ phagocytosis in anAlzheimer's disease mouse modelJunniao Lva, Shuaipeng Maa, Xuefei Zhanga, Liangjun Zhenga, Yuanhui Maa,Xuyang Zhaob, Wenjia Laic, Hongyan Shena, Qingsong Wanga,⁎, Jianguo Jia,⁎aState Key Laboratory of Protein and Plant Gene Research, College of Life Sciences, Peking University, Beijing 100871, ChinabInstitute of Systems Biomedicine, Peking University, Beijing 100191, ChinacNational Center for Nanoscience and Technology, Beijing 100190, China
A R T I C L E I N F O
Abbreviations: AD, Alzheimer's disease; AβSP, senile plaque; NFT, neurofibrillary tangleApolipoprotein E; SR-A, scavenger receptorglycosylation end products; IDE, Insulin-degsample preparation; TMT, tandem massphosphate-buffered saline; PI, protease inhibbovine serum.⁎ Corresponding authors at: College of Life SciE-mail addresses: [email protected] (J. Lv)
[email protected] (L. Zheng), mayua(W. Lai), [email protected] (H. Shen), wang
http://dx.doi.org/10.1016/j.jprot.2014.07.0281874-3919/© 2014 Elsevier B.V. All rights rese
A B S T R A C T
Article history:Received 9 March 2014Accepted 29 July 2014Available online 7 August 2014
Amyloid-beta (Aβ) deposition plays a crucial role in the progression of Alzheimer's disease(AD). The Aβ deposited extracellularly can be phagocytosed and degraded by surroundingactivated astrocytes, but the precise mechanisms underlying Aβ clearance mediated byastrocytes remain unclear. In this study, we performed tandem mass tag-based quantitativeproteomic analysis on the cerebral cortices of 5-month-old APP/PS1 double-transgenic mice.Among the 2668 proteins quantified, 35 proteins were upregulated and 12 were downregu-lated, with most of these proteins being shown here for the first time to be differentlyexpressed in the APP/PS1 mouse. The altered proteins were involved in molecular transport,lipid metabolism, autophagy, inflammation, and oxidative stress. One specific protein, PEA15(phosphoprotein enriched in astrocytes 15 kDa) upregulated in APP/PS1 mice, was verified toplay a critical role in astrocyte-mediated Aβ phagocytosis. Furthermore, PEA15 levels weredetermined to increase with age in APP/PS1 mice, indicating that Aβ stimulated theupregulation of PEA15 in the APP/PS1 mouse. These results highlight the function of PEA15in astrocyte-mediated Aβ phagocytosis, and thus provide novel insight into the molecularmechanism underlying Aβ clearance. The protein-expression profile revealed here shouldoffer new clues to understand the pathogenesis of ADand potential therapeutic targets for AD.
Biological significanceActivated astrocytes are known to clear the Aβ deposited in the extracellular milieu, whichis why they play a key role in regulating the progression of Alzheimer's disease (AD).
Keywords:Quantitative proteomicsPEA15AβPhagocytosisAlzheimer's diseaseAPP/PS1 mouse
, Amyloid-beta; APP, β-amyloid precursor protein; WT, wild type; APP/PS1, APPswe/PS1dE9;; ROS, reactive oxygen species; PEA15, 15 kDa phosphoprotein enriched in astrocytes; ApoE,A; CR3, complement receptor type 3; TLR, Toll-like receptor; RAGE, Receptor for advancedrading enzyme; NPC, Nuclear pore complex; DED, death effecter domain; FASP, filter aidedtag; IHC, immunohistochemistry; OCT, optimum cutting temperature compound; PBS,itor; ECL, electrochemiluminescence; DMEM, Dulbecco's Modified Eagle Medium; FBS, fetal
ences, Peking University, Beijing 100871, China. Tel.: +86 10 6275 5470; fax: +86 10 6275 1526., [email protected] (S. Ma), [email protected] (X. Zhang),[email protected] (Y. Ma), [email protected] (X. Zhao), [email protected]@pku.edu.cn (Q. Wang), [email protected] (J. Ji).
rved.
46 J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
However, the molecular mechanism underlying astrocyte-mediated Aβ phagocytosis anddegradation remains unclear. By performing tandem mass tag-based quantitativeproteomic analysis, we identified 47 proteins that were differentially expressed in APP/PS1 double-transgenic. To our knowledge, this is the first time most of these proteins havebeen reported to exhibit altered expression in the mouse model of AD. Furthermore, ourresults indicate that one of the proteins upregulated in the APP/PS1 mouse, PEA15(phosphoprotein enriched in astrocytes 15 kDa), regulates astroglial phagocytosis of Aβ.Our findings provide new insights into the molecular mechanism underlying Aβ clearancein AD. The altered profile of protein expression in APP/PS1 mice described here should offervaluable clues to understand the pathogenesis of AD and facilitate the identification ofpotential targets for the treatment of AD.
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
Alzheimer's disease (AD) is a neurodegenerative disease whoseclinical symptoms include learning andmemory impairment [1],and AD has attracted considerable attention because of the highmorbidity of the disease in aged population. AD is characterizedby two principal pathological features, the presence of senileplaques (SPs) composedmainly of the amyloid-beta (Aβ) peptide,and the occurrence of neurofibrillary tangles (NFTs) containinghyperphosphorylated microtubule-associated protein tau [2]. Aβpeptides are generated through the sequential enzymatic cleav-age of amyloid precursor protein (APP) by β- and γ-secretases.The Aβ deposited extracellularly can induce neuronal cell death,inflammation, oxidative stress, and mitochondrial dysfunction,and ultimately lead to cognitive and behavioral abnormalities[3–6]. Transgenicmicewithmutations inAPPproduceAβplaquesand present some of the symptoms of AD [7–10], and thereforethese mice have been widely used as animal models of AD.
The hypothesis has been proposed that accumulation of Aβin the brain is caused by an imbalance in Aβ generation andclearance. The Aβ plaques in the brain can activate bothmicroglia and astrocytes, and the reactive glial cells play animportant role in removing Aβ in the early stages of AD[3,11–15]. Contrary to microglia, astrocytes can phagocytoseand degrade Aβ in the absence of cytokine stimulation [16,17].Both postnatal and adult astrocytes internalized and degradedAβ present in culture medium, brain slices, and brain tissues[18–20], and transplanted astrocytes successfully cleared Aβdeposits in the AD mouse by upregulating the expression ofspecific proteases [21,22]. Several key proteins have beenidentified to be involved in the uptake and degradation of Aβ,including low-density lipoprotein receptor (LDLR), apolipopro-tein E (ApoE), CD36, CD47, advanced glycation end productsreceptor (RAGE), neprilysin (NEP), angiotensin-convertingenzyme-1 (ACE-1), and endothelin-converting enzyme-2 (ECE-2)[16,19,22–26]. Although astrocytes are widely accepted to play aprominent role in Aβ clearance, the precise mechanism under-lying the clearance process remains unclear.
Astrocytes are required for maintaining normal brainfunctions, and astroglial dysfunction has been implicated invarious neurodegenerative diseases [27–29]. Specifically, in AD,glutamate metabolism and glucose uptake in astrocytes havebeen reported to be impaired [30,31]. These findings haveengendered the view that the ability of reactive astrocytes toclear Aβ might be overwhelmed at the late stages of AD, andthis might contribute to Aβ deposition.
In this study, to elucidate the molecular mechanism under-lying Aβ clearance that is mediated by reactive astrocytes andthe alteration of Aβ clearance in AD, we used the APP/PS1(presenilin 1) double-transgenic mouse, a well-recognized ADmouse model. Using cerebral cortices isolated from APP/PS1mice at 5 months old, when Aβ deposition occurred andastrocytes near the deposits were activated, we performedtandem mass tag (TMT) labeling and LC-MS/MS analysis. Ourproteomic study showed a novel profile of altered proteins in theAPP/PS1 mice, and the results of bioinformatics analysessuggested that these differentially expressed proteins wereinvolved in diverse biological processes such as moleculartransport, lipid metabolism, autophagy, immune system pro-cesses, and response to oxidative stress. Furthermore, weverified that one specific protein, PEA15 (phosphoproteinenriched in astrocytes 15 kDa), regulated astrocyte-mediatedAβ phagocytosis and that the expression of PEA15 increasedwithage in theAPP/PS1mouse,which confirmed thatAβpeptidecan stimulate the expression of PEA15. Our results highlight thefunction of PEA15 as a modulator in Aβ clearance and providenew insights into the pathogenesis of AD.
2. Materials and methods
2.1. Experimental animals
The APPswe/PS1dE9 double-transgenic mice were purchasedfrom the Institute of Laboratory Animal Science, ChineseAcademy of Medical Science (Beijing, China). The transgenicmice express APPswe (APP with Swedish mutations K670N/M671L) and PS1 with the exon-9 deletion under the control ofthemouse prion-protein promoter; thesemice develop amyloiddeposits in the brain and also exhibit cognitive impairment [7].The mice were crossed with C57/6J mice and the genotypes ofoffspring were identified by performing PCR tests on genomicDNA extracted from tail tissues. The primer sequences used forAPP were 5′-GACTGACCACTCGACCAGGTTCTG-3′ (sense) and5′-CTTGTAAGTTGGATTCTCATATCCG-3′ (antisense), and theproduct length was 344 bp; the PSI primer sequences were5′-AATAGAGAACGGCAGGAGCA-3′ (sense) and 5′-GCCATGAGGGCACTAATCAT-3′ (antisense), and the product length was608 bp. Female transgenic mice and matched non-transgeniccontrol (wild type; WT) littermates were used in this study. Allstudies on mice were conducted following the UniversityPolicies on the Use and Care of Animals and were approved by
47J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
the Institutional Animal Experiment Committee of PekingUniversity.
2.2. Brain tissue preparation
Animals were deeply anesthetized using chloral hydrate(400 mg/kg) and perfused transcardially with cold normalsaline. The brains were removed and carefully dissected, andthen snap-frozen in liquid nitrogen and stored at −80 °C foruse in experiments.
2.3. Histological analysis
Animals used in immunohistochemistry (IHC) studies wereanesthetized using chloral hydrate as mentioned above, andthen perfused transcardially with 4% paraformaldehyde pre-pared in 0.1 M phosphate-buffered saline (PBS, pH 7.4). Brainswere removed and immediately immersed in 4% paraformal-dehyde for post-fixation at 4 °C for 24 h, after which the brainsamples were dehydrated by treatment with 15% and 30%sucrose solutions. Next, the tissues were embedded in OCT(optimum cutting temperature compound) in a freezing micro-tome and 30-μm-thick sections were cut along the sagittalplane. Fresh sectionswere permeabilized using 1% Triton X-100in 0.1 M PBS and then blockedwith 10% fetal bovine serum (FBS;Hyclone, USA) for 1 h at 37 °C. The sectionswere incubatedwithprimary antibodies overnight at 4 °C and then with thesecondary antibodies, after which they weremounted on slidesfor observation. The following primary antibodies were used:anti-human Aβ, 6E10 (SIG-39320; Covance, Princeton, NJ),anti-GFAP (glial fibrillary acidic protein) (ab4674; Abcam,Cambridge, UK), and anti-PEA15 (AP8524b; ABGENT, CA).
2.4. Sample preparation and TMT labeling
For proteomic analysis, 10 female mice were killed fromeach strain at the age of 5 months. Cerebral tissue sampleswere homogenized in a lysis buffer containing 7 M urea,2 M thiourea, 4% 3-[(3-cholamidopropyl)-dimethylammonio]propanesulfonate, 1% dithiothreitol, 10% D/RNase, and aprotease-inhibitor cocktail. After centrifugation at 20,000 ×gfor 30 min at 4 °C, the proteins in the supernatant wereprecipitated using acetone and then re-dissolved in lysis buffer,followed by the determination of protein concentration using a2-D Quant Kit (GE Healthcare, Pittsburgh, PA, USA). The proteinsamples from 5 mice of each strain were equally mixed toreduce the impact of individual outliers [32], and digested withsequencing-grade trypsin (Promega, Madison, WI, USA) byusing the filter-aided sample preparation (FASP) method [33].After using C18 Stage Tips to desalt the samples, peptides weredissolved in tetraethylammonium bromide (TEAB, 100 mM,pH 8.5) and equally labeled with TMT reagents (Thermo FisherScientific, Rockford, IL, USA) according to manufacturer'sinstructions. The TMT reagents (0.8 mg) were dissolved in40 μL of anhydrous acetonitrile. Aliquots of the samples wereincubated with TMT reagents in separate tubes for 2 h at roomtemperature, and the reactions were quenched by adding 8 μLof 5% hydroxylamine solution and incubating for 20 min.Samples obtained fromWTmice 1–5 were labeledwith Reagent129 and transgenic mice 1–5 with Reagent 130 in the first
replicate while samples obtained from WT mice 6–10 werelabeled with Reagent 131 and transgenic mice 6–10 withReagent 128 in the second replicate (Fig. S5). All TMT-modifieddigests were then combined, desalted, and dissolved in anisoelectric focusing (IEF) buffer that contained 5% glycerol and2% IPG buffer (pH 3–10, GE Healthcare, USA). The peptidemixtures were loaded into 12 wells over a 13-cm ImmobilineDry Strip, pH 3–10 (GEHealthcare), and separated by performingIEF on a 3100 OFFGEL Fractionator (Agilent Technologies, SantaClara, CA, USA) according to themanufacturer's instructions. Intotal, 12 fractions were acidified using trifluoroacetic acid andthen desalted prior to LC-MS/MS analysis.
2.5. LC-MS/MS analysis
The digested samples were dissolved in 0.2% formic acid andwere separated using an Easy nLC 1000 system (ThermoScientific, USA) equipped with a C18 reverse-phase column.The column was eluted using linear gradients of 5–32%acetonitrile in 0.2% formic acid, at a constant flow rate of300 nL/min for 100 min. The HPLC system was coupled to anLTQ Orbitrap XL Mass Spectrometer (Thermo Fisher Scientific)that was coupled to a nano-ES ion source (Proxeon Biosystems,Denmark). The instrument was operated in the positive-ionmode with the ESI spray voltage set at 2.0 kV. Survey full-scanMS spectra (m/z 300–1600) were acquired in the Orbitrap at amass resolution of 30,000 after the accumulation of 1,000,000ions. Three peptide ions showing the most intense signal fromeach scan were selected for higher energy collisional dissocia-tion (HCD)-MS/MS analysis (normalized collision energy, 70%and activation time, 0.1 ms) in theOrbitrap at amass resolutionof 7500 and AGC value of 50,000. Maximal filling times were200 ms in the case of full scans and 200 ms (HCD) for theMS/MSscans. Ions with unassigned charge states and singly chargedspecieswere rejected. The dynamic exclusion listwas restrictedto amaximumof 500 entrieswith amaximal retention period of30 s and a relative mass window of 10 ppm. The data wereacquired using Xcalibur 2.2 (Thermo Fisher Scientific).
2.6. Protein identification and quantification
All MS raw data were processed using Proteome Discoverer(Version 1.4, Thermo Fisher Scientific). Data were searchedagainst themouseUniversal Protein Resource sequence database(UniProt, August, 2013). The search parameters were set asfollows: enzyme, trypsin; fixed modification, carbamidomethyl(Cys); variablemodification, oxidation (Met), TMT/+229D-K, TMT/+229D-N Terminal; maximum miss cleavage, 1; MS tolerance,5 ppm; MS/MS tolerance, 0.5 Da; false discovery rate (FDR) atpeptide and protein levels, <0.01; and required peptide length, >7amino acids. At least one unique peptide per protein group wasrequired for identifying proteins, and two quantified peptideswere required for quantifying protein. Differentially expressedproteins were selected with protein ratio ≥1.5 or ≤0.67 above the95% confidence level in each comparison (128/131, 130/129).
2.7. Bioinformatics
Gene Ontology (GO) enrichment of differentially expressedproteins was analyzed using the GeneCoDis web-based tool
48 J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
V.3 (http://genecodis.dacya.ucm.es/) [34]. The gene names ofthe proteins were input to analyze GO biological process(GOBP), GO molecular function (GOMF), and GO cellularcomponents (GOCCs). We determined whether the selectedGO categories were overrepresented in the input gene listsby using the hypergeometric test with simulation-basedcorrection.
2.8. RNA isolation and quantitative real-time PCR
The brains from transgenic and WT mice were divided intoright and left hemispheres for biochemical analysis. The totalRNA was extracted from the cerebral cortex of one hemi-sphere by using TRIzol (Invitrogen, NM, USA) according to themanufacturer's protocol, and 1 μg of the total RNA wasreverse-transcribed into cDNA by using an HIFI-MMLV cDNAfirst-strand synthesis kit (Cowin Biotech Co. Ltd, Beijing,China). Quantitative PCR was performed using a CFX96Real-Time system (BioRad, CA, USA) and GoTaq qPCR MasterMix (Promega, WI, USA). Data were analyzed by the 2−ΔΔCT
method [35,36]. Individual mouse was used in each replicateand 3 independent experiments were performed.
2.9. Western blotting
Proteins were extracted from the cerebral cortex of anotherhemisphere as described above (see Sample preparation andTMT labeling section). The protein concentration wasmeasuredusing a 2-D Quant Kit (GE Healthcare, Pittsburgh, PA, USA). Theprotein lysates weremixedwith gel-loading buffer, separated bypreforming sodium dodecyl sulfate-polyacrylamide gel electro-phoresis (SDS-PAGE), and then transferred onto polyvinylidenedifluoride (PVDF) membranes (BioRad). After blocking with 5%skimmed milk for 1 h, the membranes were incubated withprimary antibodies overnight at 4 °C and then with horseradishperoxidase (HRP)-conjugated secondary antibodies (JacksonImmunoResearch, West Grove, PA, USA) for 2 h at roomtemperature. Immunoreactive protein bands were detected byincubating blots in a chemiluminescence solution (Millipore, CA,USA) and then exposing them to X-ray films (Fuji medical X-rayfilm, FujiFilm). Individual mouse was used in each replicate and3 independent experiments were performed.
2.10. Human Aβ1–42 preparation
Human Aβ1–42 powder (AnaSpec, San Jose, CA, USA) wasdissolved to a final concentration of 2 mM in PBS. Fibrils weregenerated by incubating the solutions at 37 °C for 1 weekbefore use. The ultrastructure of the peptide was examinedusing transmission electron microscopy (Tecnai G220; FEI,Hillsboro, OR, USA). HiLyte-488-conjugated Aβ1–42 (AnaSpec)was purchased for use in the fluorescent detection of Aβ inphagocytosis assays.
2.11. Primary astrocyte culture
Primary astrocytes were cultured from the cerebral cortices of1–2-day-old mice as described previously [37]. Briefly, themeninges and blood vessels were completely removed fromthe cortices, which were thenminced and digested with 0.25%
trypsin and D/RNase for 30 min at 37 °C. The digested tissueswere blown gently into single cells and plated in culturedishes (Corning, New York, NY, USA). Cells were cultured inDMEM supplemented with 10% FBS (Hyclone) and 1% antibi-otics (100 U/mL penicillin, 100 μg/mL streptomycin) for 7–9days at 37 °C in an incubator with 5% CO2. Microglial cellswere removed by shaking the culture plates for 16–24 h at260 rpm at 37 °C, and the cells that remained attached to thedishes were washed and digested for preparing continuouscultures. Based on immunoreactive detection of GFAP, thecultured astrocytes were identified to be >97% pure.
2.12. Uptake and degradation of Aβ by astrocytes
To measure the dynamics of Aβ clearance, astrocytes wereincubated with 0.2 μM fAβ1–42 (fibrillar Aβ) at 37 °C for 0, 0.5, 1,3, and 6 h. At each time point, the cells were harvested andlysed in a lysis buffer containing protease inhibitor cocktails,and the cell lysates were used in western blotting analyses todetermine intracellular Aβ levels.
To measure Aβ phagocytosis by astrocytes, cells wereincubated with 0.2 μM fAβ1–42 at 37 °C for 3 h, after which theAβ-containingmediumwas removed and intracellular Aβ levelswere determined by performing western blotting as describedabove. In flow-cytometry experiments, HiLyte-488-conjugatedfAβ1–42 was used, and fluorescence intensity was measuredusing a FACScan flow cytometer (BD Biosciences, Franklin Lakes,NJ, USA). At least 20,000 events were recorded for each sample.Data were analyzed using WinMDI 2.9 software.
To analyze intracellular Aβ degradation in astrocytes, cellswere incubated with 0.2 μM fAβ1–42 at 37 °C for 3 h. Subse-quently, the Aβ-containing medium was removed and freshmedium was added and cells were incubated for 0, 6, or 12 h;at each time point, cells were harvested and analyzed byperforming western blotting.
To determine the intracellular localization of Aβ in astro-cytes, cells were cultured on coverslip and incubated with0.2 μM HiLyte-488-conjugated Aβ and LysoTracker Red (C1046,Beyotime, China) for 3 h. Cells were fixed with 4% paraformal-dehyde and stained with DAPI (4,6-diamidino-2-phenylindole),after which the coverslips were mounted on slides for imagingthe labeled cells by using a Zeiss LSM 710 confocal microscope.
2.13. PEA15 RNA interference
Mouse primary astrocytes grown in 12-well plates for 24 h weretransfected with a PEA15-specific siRNA (sc-37486, Santa CruzBiotech, Europe) or a control siRNA (sc-37007, Santa CruzBiotechnology) by using Lipofectamine RNAiMAX (Invitrogen,Carlsbad, CA,USA); 12 h later, the siRNA-containingmediumwaschangedand the culturesweremaintained in regularmedium.At48 h after transfection, the cells were treated with 0.2 μM Aβ1–42
at 37 °C for 3 h and then the intracellular Aβ level wasmeasured.The interference efficiency of PEA15 was over 5 times.
2.14. PEA15 overexpression
A pEGFP-C1 plasmid carrying the PEA15 gene was constructedand then transfected into cells by using the Lipofectamine2000 (Invitrogen), with pEGFP-C1 serving as the control. At
49J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
48 h after transfection, cells were treated with 0.2 μMAβ1–42 at37 °C for 3 h, and then the intracellular Aβ level wasmeasured.
2.15. Statistical analysis
Except in the case of GO-enrichment analysis-based hierar-chical clustering, in which the hypergeometric test was used,all data are expressed asmeans ± SD of at least 2 independentexperiments; the data were analyzed using Student's t-testsor one-way ANOVA. The significance level was set at p < 0.05.
Fig. 1 – Distribution of identified protein ratios and scatterplot of altered protein ratios. A) APP/PS1 vs WT protein ratioswere presented on a log2 scale. The highest frequencyconcentrated around Zero (1-fold). B) APP/PS1 vs WT proteinratios of altered proteins from two replicates werenormalized by log2 and plotted on the x/y axis. Pearsoncorrelation was 0.9065 between the two rounds. The 95%prediction bond was displayed.
3. Results
3.1. Validation of the AD phenotype in APP/PS1 mice
Using IHC methods, we confirmed the presence of thepathological features of AD, including Aβ deposition and glialactivation, in brains of APP/PS1mice of various ages (3, 5, 9, and12 months). As described previously [38,39], APP/PS1 miceexhibited clear Aβ deposition in both the cerebral cortex andthe hippocampus at 5 months, and the number of Aβ plaquesincreased with age (Fig. S1A). Reactive astrocytes express highlevels of GFAP andmigrate to sources of chemotactic stimuli inthe brain [40–42]. Therefore, we used a GFAP-specific antibodyto identify astrocytes activated in response to Aβ plaques.GFAP-positive astrocytes were detected in the hippocampus ofboth APP/PS1 and WT mice (Fig. S1B). However, in the cerebralcortex area in 5-month-old APP/PS1 mice, GFAP-positiveastrocytes detected around Aβ plaques were observed to formglial scars (Fig. S1C), and the number of these GFAP-positiveastrocytes increased with age. These results indicated thatastrocytes were activated and recruited around Aβ plaques in5-month-old APP/PS1 mice [43].
3.2. TMT-based quantitative proteomics revealed that proteinexpression was altered in the APP/PS1 mouse brain
To determine how global protein expression was altered inAPP/PS1 mice, TMT-based quantitative proteomics was per-formed on cerebral cortices obtained from 5-month-old APP/PS1 and WT mice (Fig. S2). Cerebral protein mix from animals1–5 of each strain was used in the first replicate and fromanimals 6–10 in the second replicate. TMT labeling efficiencywas determined to be up to 99% (Table S1). After eliminatingcontaminants and proteins without unique peptides, 2672proteins were identified with a FDR of <1%, of which 2668proteins were quantified and then the protein ratios werenormalized. Detailed quantitative information at the proteinlevel is shown in Table S2, and that at the peptide level isshown in Table S3. The distribution of protein ratios is shownin Fig. 1A, and the highest frequency was concentratedaround zero (1-fold). By calculating the correlation betweenthe two replicates, we obtained a mean correlation of 0.9065for the differentially expressed proteins among measure-ments (Fig. 1B). By using the filter criteria of protein ratio ≥1.5or ≤0.67 at the 95% confidence level in both of the experimentsabove, 47 proteins were determined to be differentially
expressed, with 35 proteins being up-regulated and 12 beingdown-regulated in the APP/PS1 mice (Table 1).
To investigate the functional importance of the alteredproteins, we analyzed GO annotation and enrichment byusing GeneCoDis. Approximately 40% and 38% of the proteinswere identified to localize in the cytoplasm and in mem-branes, and 28% and 15% of the proteins were determined tolocalize in the nucleus and the mitochondrion, respectively,according to GO (Fig. S3A). The molecular functions of thealtered proteins were highly enriched in molecular interac-tion, transferase activity, catalytic activity, hydrolase activity,and oxidoreductase activity (Fig. S3B). In accordance with theenrichment of GO molecular function, the GO biologicalprocesses of the altered proteins were clustered in transport,the oxidation–reduction process, lipid metabolism, and theinflammatory process, according to the results obtained usingGeneCoDis (Fig. S3C). The functions of key proteins in AD arediscussed next.
Table 1 – List of differentially expressed proteins identified in the cerebral cortices of APP/PS1 mice.
Accession Protein name Genename
Score Coverage Uniquepeptidesa
Ratiob MWc[kDa]
6P8U6 Pancreatic triacylglycerol lipase Pnlip 15.14 3.01 1 3.568 51.4Q9DBB4 N-alpha-acetyltransferase 16, NatA auxiliary subunit Naa16 3.35 1.04 1 2.861 101.2Q9D0J1 Syntaxin 8 Stx8 2.56 5.38 1 2.313 14.9Q91VM5 RNA binding motif protein Rbmxl1 49.38 21.13 1 1.851 42.1Q8VEJ9 Vacuolarprotein sorting-associated protein 4A Vps4a 3.58 2.75 1 2.022 48.9F6TN80 Protein Srsf11 Srsf11 2.97 7.92 1 2.041 11.0Q4FJZ2 Importin subunit alpha Kpna6 6.10 2.06 1 1.843 59.6P06880 Somatotropin Gh1 1.61 3.24 1 2.079 24.7G3UWG1 MCG115977 Gm10108 117.93 49.52 6 1.757 11.7F6T1Y2 Rho GTPase-activating protein 44 Arhgap44 4.37 2.02 1 2.019 69.9Q922D4 Isoform4 of Serine/threonine-protein phosphatase
6 regulatory subunit 3Ppp6r3 1.85 1.41 1 2.108 56.7
O09131 Glutathione S-transferase omega-1 Gsto1 1.91 4.17 1 2.072 27.5P97478 Ubiquinone biosynthesis protein COQ7 homolog Coq7 2.62 5.07 1 2.266 24.0Q9CPV9 P2Y purinoceptor 12 P2ry12 2.09 2.31 1 1.786 39.4Q9D8B6 Protein FAM210B Fam210b 0.00 6.32 1 1.630 20.3Q80UK0 SEC14 domain and spectrin repeat-containing protein 1 Sestd1 2.58 1.29 1 2.126 79.3Q9CQA1 Trafficking protein particle complex subunit 5 Trappc5 4.31 4.79 1 3.121 20.8O35682 Myeloid-associated differentiation marker Myadm 2.08 3.13 1 2.098 35.3Q4QQM5 Isoform 2 of protein FAM73A Fam73a 0.00 1.58 1 2.782 56.9P24668 Cation-dependent mannose-6-phosphate receptor M6pr 3.90 3.96 1 2.778 31.2P19324 Serpin H1 Serpinh1 2.73 2.16 1 1.685 46.5Q8BR63 Protein FAM177A1 Fam177a1 3.11 5.31 1 1.946 23.6Q99JI4 26S proteasome non-ATPase regulatory subunit 6 Psmd6 32.14 12.60 4 1.753 45.5E9QKK8 Probable phospholipid-transporting ATPase 11C Atp11c 0.00 2.51 1 1.704 127.7D3Z375 Astrocytic phosphoprotein PEA-15 Pea15 12.91 43.48 3 1.548 10.7F6RAZ3 Short-chain-specific acyl-CoA dehydrogenase Acads 2.15 5.96 1 2.104 32.6P53395 Lipoamide acyltransferase component of branched-chain
alpha-keto acid dehydrogenase complexDbt 2.63 3.11 1 1.582 53.2
P34884 Macrophage migration inhibitory factor Mif 49.77 13.91 2 2.675 12.5Q5M956 BTB/POZ domain-containing protein KCTD1 Kctd1 1.91 2.72 1 1.657 29.4Q8R5F7 Isoform 2 of Interferon-induced helicase C domain-containing
protein 1Ifih1 2.76 0.82 1 1.670 110.8
Q3U186 Probable arginine–tRNA ligase Rars2 2.51 2.08 1 1.698 65.3P97326 Cadherin-6 Cdh6 1.91 1.01 1 1.650 88.3Q8BNL5 BTB/POZ domain-containing protein KCTD6 Kctd6 2.62 2.95 1 1.600 27.6Q91YS8 Calcium/calmodulin-dependent protein kinase type 1 Camk1 19.74 4.55 2 1.528 41.6Q9R1R8 Retinol dehydrogenase 11 Rdh11 1.98 2.67 1 1.680 33.2Q8CES0 N-alpha-acetyltransferase 30 Naa30 3.54 4.95 1 0.596 39.4P04370-5 Isoform 5 of myelin basic protein Mbp 579.12 52.66 1 0.522 18.5P98086 Complement C1q subcomponent subunit A C1qa 6.17 5.31 1 0.539 26.0Q91V35 Receptor-type tyrosine-protein phosphatase Ptpra 3.31 1.13 1 0.659 89.8E9Q0F0 Protein Krt78 Krt78 21.01 1.78 1 0.627 112.2H9KV03 Nephrocystin-3 Nphp3 2.99 0.58 1 0.637 137.5F6X3X7 Protein Uggt2 Uggt2 2.34 13.21 1 0.381 11.9E9Q1I5 Multidrug resistance-associated protein 1 Abcc1 2.37 1.56 1 0.379 73.5O35280 Isoform 2 of erine/threonine-protein kinase Chk1 Chek1 2.56 3.93 1 0.437 43.6P59017 Bcl-2-like protein 13 Bcl2l13 2.98 2.30 1 0.493 46.7J3QP68 Uncharacterized protein Gm4204 6.20 5.93 1 0.330 41.1Q2UY11 Collagen alpha-1 Col28a1 3.03 1.49 1 0.331 118.7
a Number of unique high-confidence peptide that matches the protein sequence.b Ratio represents the average fold-change of APP/PS1 versus WT mice.c Theoretical molecular weight (MW) of unprocessed precursor protein.
50 J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
ATP-binding cassette (ABC) transporters play a crucial rolein Aβ clearance [44]. The deficiency of ABCC1 increased Aβlevels without enhancing the activity of APP-processingenzymes, while the activation of ABCC1 substantially reducedthe Aβ load in an AD mouse model [45,46], implying a criticalrole of ABCC1 in Aβ metabolism.
Lipid homoeostasis has been suggested to be closelyassociated with AD pathology [47–49]. Pancreatic triacylglyc-erol lipase (PNLIP), Acyl-CoA dehydrogenase (ACADS) anddihydrolipoamide branched-chain transacylase (Dbt) participatein lipidmetabolism and they can generate acetyl CoA, the directmaterial of cholesterol, which promotes Aβ deposition [48].
Fig. 2 – Expression levels and location of altered proteins.A) The gene expression levels of 6 proteins were analyzed byreal-time PCR test. Values shown are mean ± SD for N = 3.B) PEA15 expression change was verified by western blot.PEA15 was significantly reduced in the cerebral cortex of5-month-old APP/PS1 transgenic mouse. Values shown aremean ± SD for N = 3. C) PEA15 location was detected in thebrain of 5-month-old mice. Micrographs showed theexpression of PEA15 in astrocytes. Scale bar: 20 μm. Datawere analyzed by Student's t test; *p < 0.05; **p < 0.01;***p < 0.001. WT: wild type mice; Tg: APP/PS1 mice.
51J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
Macrophage migration-inhibitory factor (MIF) was physi-cally associated and co-localized with Aβ peptide [50]; MIFwas substantially up-regulated in AD patients and relatedwith Aβ-induced neurotoxicity [51], implying that MIF is apotential therapeutic target for AD.
MBP can bind Aβ peptides and inhibit the assembly offibrillar Aβ [52]; it also can prevent Aβ-mediated cytotoxicity[53], and degrade fibrillar Aβ in vitro and vivo [54], indicatingthat MBP possesses Aβ-degrading activity.
PEA15 is a multifunctional protein that has been suggestedto participate in various biological processes including cellproliferation and migration, glucose metabolism, apoptosis,and autophagy [55–57]. PEA15 inhibits TNF-α intracellularsignaling and regulates the MAPK cascade by directly bindingto ERK or JNK and anchoring them in the cytoplasm [37,58,59].PEA15 has been implicated in diverse diseases including type IIdiabetes [60], glioma [61,62], breast cancer [63], and neurode-generative disorders [64]. Deletion of PEA15 in mice impairedspatial-learning abilities in the animals [56], and PEA15 wasreported to be downregulated in Huntington's disease (HD) [64]but upregulated in the AD mouse model [65]. However, theprecise molecular function of PEA15 in AD remains poorlystudied.
3.3. Verification of differential protein expression in APP/PSIand WT mice
To confirm the differential expression of proteins identified tobe altered in the APP/PSI mice by proteomics, we performedquantitative real-time PCR analysis. As shown in Fig. 2A, themRNA expression levels of five proteins, ACADS, PNLIP, MIF,MBP, and PEA15, were up-regulated, whereas the expressionof ABCC1 mRNA was down-regulated, which agreed with theproteomics results.
Next, using western blotting analysis, we confirmed thechange in PEA15 protein expression. PEA15 was expressed athigher levels in the cerebral cortices of 5-month-old APP/PS1mice than in the cerebral cortices of age-matched control mice(Fig. 2B), which is consistent with the proteomics and real-timePCR results. Lastly,weused IHC to identify the cell type inwhichPEA15 was expressed in vivo, and we determined that PEA15was co-localized with GFAP (Fig. 2C); this result indicated thatPEA15 is expressed in astrocytes, which agrees with theprevious findings [66].
3.4. PEA15 is involved in astrocyte-mediated Aβ phagocytosis
Because our aforementioned results confirmed the upregula-tion of PEA15 in the brain of the APP/PS1 mouse, we sought todetermine whether extracellular Aβ can stimulate the increasein PEA15 expression. Primary astrocytes isolated fromWTmicewere left untreated or were treated with 0.2 μM fAβ for 3 h, andthen PEA15 levels were measured using western blottinganalysis. As shown in Fig. 3A, after Aβ treatment, the PEA15level in the astrocytes of WT mice was markedly up-regulated,indicating that Aβ can induce PEA15 expression.
One intriguing question was whether the Aβ-stimulatedincrease in PEA15 expression can affect Aβ clearance duringthe AD process. PEA15 has been suggested to modulateautophagy through the JNK/MAPK signaling pathway [67,68],
and, conversely, autophagy can regulate phagocytosis [69]and the MAPK signaling pathway is involved in phagocytosis[70]. Therefore, we tested the role of PEA15 in Aβ clearance.
To elucidate the dynamics of Aβ clearance in primaryastrocytes, we incubated the cells with 0.2 μM fAβ, harvestedthe cells at 0, 0.5, 1, 3, and 6 h, and then performed westernblotting to determine the levels of intracellular Aβ. Further-more, to investigate the function of PEA15 in the Aβ-clearanceprocess, before treating cells with Aβ, we used RNA interfer-ence to inhibit PEA15 expression and the protein level ofPEA15 was decreased to 20%. We also transfected the cellswith the pCMV-Tag2B-PEA15 plasmid to overexpress PEA15(Fig. 3B). In all groups of cells, the levels of intracellular Aβpeaked at 3 h (Fig. 3B), consistent with previous resultsobtained using microglia [71,72]. By comparison, at 6 h, theintracellular Aβ levels were substantially lower, indicatingthat the rate of protein degradation far exceeded the rate of
53J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
protein uptake. Furthermore, at both 1 and 3 h, the levels ofintracellular Aβ were diminished when PEA15 expression wasinhibited, whereas the Aβ levels were elevated when PEA15was over-expressed (Fig. 3C). These results suggested thatPEA15 promoted Aβ uptake by astrocytes.
In our experiments, PEA15's effect on Aβ uptake byastrocytes could have resulted from either an enhancement ofAβ phagocytosis or a dysfunction of Aβ degradation. Todetermine which of these processes was affected by PEA15, weanalyzed both Aβ phagocytosis and degradation. To analyze Aβphagocytosis, astrocytes were incubated with 0.2-μM fAβ for3 h and then the intracellular Aβ level was measured bymeans of western blotting; moreover, we performed flowcytometry by using HiLyte-488-conjugated fAβ. The results ofboth flow-cytometry andwestern blotting experiments showedthat the ability of astrocytes to phagocytose Aβwas diminishedwhen PEA15 expression was suppressed, and was enhancedwhen PEA15 was overexpressed (Fig. 3D and E). To analyze Aβdegradation, astrocytes were treated with 0.2 μM fAβ at 37 °Cfor 3 h, after which the Aβ-containing medium was replacedwith fresh Aβ-free medium and the cultures were maintainedfor various periods. Cells were harvested at 0, 6, and 12 h andthe levels of intracellular Aβwere determined through westernblotting. Although the initial amounts of intracellular Aβ weredistinct in the cultures under various conditions of PEAexpression, the progressive decline of Aβ beyond 0 h showedno significant differences (Fig. 3F), indicating that the rate of Aβdegradation was not affected by PEA15. Collectively, theseresults indicated that PEA15 regulates Aβ phagocytosis but hasno impact on Aβ degradation.
WeconfirmedAβphagocytosis anddegradation in astrocytesby performing confocal microscopy. Hilyte-488-conjugated fAβwas clearly detected inside cells and was colocalized withLysoTracker Red in primary astrocytes that had been exposedto the Aβ probe for 3 h (Fig. 3G). This result suggested that Aβ inthe medium was internalized and trafficked to endosomal/lysosomal compartments for degradation.
3.5. Dysfunctional Aβ phagocytosis and age-dependent alterationin PEA15 expression in APP/PS1 mouse
Astroglial dysfunction has been proposed to underlie variousneurodegenerative diseases [27–29]. To investigate whetherAβ phagocytosis is altered in the APP/PS1 mouse brain,primary astrocytes were isolated from APP/PS1 mice. Inthese cells, PEA15 protein expression after Aβ treatment andthe phagocytosis of Aβ were analyzed as described in thepreceding section. As compared with its levels in WT mice,PEA15 was downregulated in primary astrocytes isolated from
Fig. 3 – PEA15 modulates Aβ phagocytosis in primary astrocytesastrocytes was detected by western blot. Aβ can stimulate the upControl: treated with FBS free medium; Aβ: treated with 0.2 μMPEA15. C–D and F) Primary astrocytes were transfected with PEAcells were treated with 0.2 μM fAβ and the ability of Aβ clearancewestern blotting. E) Results of flow cytometer analysis. 0.2 μM Hwere analyzed by Student's t test; *p < 0.05; **p < 0.01; ***p < 0.00Hilyte-488-conjugated fAβ and lysotracker for 3 h. Aβ (green flulysotracker (red fluorescence). Scale bars: 5 μm.
APP/PS1 mice under Aβ stimulation (Fig. 4A), although theexpression of PEA15 increased in APP/PS1 mice in the absenceof Aβ treatment (Fig. 3A). Consistent with this change inPEA15 expression, the amount of Aβ engulfed by astrocyteswas substantially lower in APP/PS1 mice than in WT mice,indicating that Aβ phagocytosis was weaker in APP/PS1 micethan in control mice. These results further confirmed the roleof PEA15 in Aβ phagocytosis and revealed that the phagocy-tosis in response to Aβ stimulation was overwhelmed in theAPP/PS1 mouse. Lastly, age-dependent changes in PEA15expression in the cerebral cortices of WT and APP/PS1 micewere explored using western blotting: in bothWT and APP/PS1mice, PEA15 exhibited a progressive increase in expressionfrom the age of 1 month to 10 months (Fig. 4B).
4. Discussion
4.1. Molecular function of PEA15
PEA15 has been suggested to be involved in various biologicalprocesses including cell proliferation and migration, glucosemetabolism, apoptosis, and autophagy [55–57]. These func-tions of PEA15 depend mainly on its protein structure: PEA15features a leucine-rich nuclear-export sequence (NSE), whichhelps retain PEA15 and its binding proteins in the cytoplasm[37]. PEA15 inhibits TNF-α intracellular signaling and regu-lates the MAPK cascade by binding and anchoring ERK or JNKin the cytoplasm [37,58,59]. Because PEA15 plays key roles indiverse biological processes, abnormal PEA15 expression hasbeen associated with various disorders: overexpression of thePED/PEA15 gene may contribute to insulin resistance in type IIdiabetes [60]; aberrant expression of MAT1/PEA15 isoformswas detected during neoplastic changes in breast cancer [63];deletion of PEA15 in mice impaired spatial-learning abilities[56]; and PEA15 was downregulated in HD [64] but upregulatedin AD [65].
Although PEA15 has been demonstrated to regulate theMAPK signaling pathway, which plays a critical role in AD, thefunction of PEA15 in AD has not been studied in detail.Recently, PEA overexpression was reported to increase au-tophagy through JNK activation, and the downregulation ofendogenous PEA15 was reported to abrogate JNK-mediatedautophagy [67], suggesting that PEA15 regulated autophagy ina JNK-dependent manner [67,68]. Autophagy and phagocyto-sis are related cellular processes [73–75] that involve similarconstitutive processes [76] and share identical moleculesin their respective signaling pathway. For example, LC3, amolecule previously thought to function specifically in
. A) The protein levels of PEA15 under Aβ stimulation in-regulation of PEA15. Values shown are mean ± SD for N = 3.fAβ. B) The efficiency of inhibition and over-expression of15 specific siRNA and pEGFP-C1-PEA15 plasmid. 48 h later,was measured by detecting the intracellular Aβ levels usingilyte-488-conjugated fAβ was used in this experiment. Data1. G) Astrocytes were incubated with 0.2 μMorescence) was detectable intracellular and co-localized with
Fig. 4 – Astroglial dysfunctional of Aβ phagocytosis in APP/PS1 mice and the progressive up-regulation of PEA15. A) Primaryastrocytes isolated from both APP/PS1 and WT mice were incubated with 0.2 μM fAβ for 3 h. Under the condition of Aβstimulation, PEA15 was down-regulated in astrocytes from APP/PS1 mice, compared to astrocytes from WT mice.Correspondingly, the ability of Aβ phagocytosis was inhibited in astrocytes isolated from APP/PS1 mice. Transgenic −: WT;transgenic +: APP/PS1. B) The expression levels of PEA15 in the cerebral cortices of mice at different ages were detected usingwestern blotting. ACT, astrocytes; transgenic, APP/PS1. Data were analyzed by Student's t test; *p < 0.05; **p < 0.01;***p < 0.001.
54 J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
autophagy, can play a similar role in phagocytosis as it does inautophagy [77]. High-mobility group box-1 (HMGB1), a criticalregulator of autophagy, also modulates phagocytosis [78,79].Recent evidence has also revealed that the regulation ofautophagy can affect phagocytosis [69,80]. Moreover, theMAPK signaling pathway was determined to participate inboth autophagy and phagocytosis [70]. Because PEA15 wasreported to modulate autophagy by activating the MAPKpathway [67] and increased expression of PEA15 was reportedin reactive astrocytes of AD brains [65], we sought toinvestigate whether PEA15 can also regulate the phagocytosisof Aβ in APP/PS1 mice.
To elucidate the role of PEA15 in Aβ phagocytosis, weinhibited PEA15 expression in primary astrocytes by using aPEA15-specific siRNA and overexpressed PEA15 in the astrocytesby transfecting themwith the pCMV-Tag2B-PEA15 plasmid, andthen measured the ability of the cells to phagocytose Aβ. Theamount of Aβ phagocytosed by astrocytes was diminishedsubstantially after the suppression of PEA15 expression, whichsuggested that PEA15 is required for Aβ phagocytosis. Converse-ly, Aβ phagocytosis was clearly enhanced after PEA15 wasoverexpressed in astrocytes, which indicated that PEA15 canenhance Aβ phagocytosis. These results demonstrated thatPEA15 regulated the process of Aβ phagocytosis. Comparedwith the expression in controlmice, the expression of PEA15wasincreased in APP/PS1 mice at the age of 5 months, a period atwhich Aβ deposition and astrocyte activation occurred.
Furthermore, our results demonstrated that Aβ stimulated theexpression of PEA15 (Fig. 3A). Therefore, the upregulation ofPEA15 could represent an organism's response to Aβ plaqueformation and might promote astroglial cell-mediated Aβphagocytosis, and consequently affect Aβ clearance in AD.
We also investigated the temporal dynamics of PEA15expression in WT and APP/PS1 mice. PEA15 protein levelsincreased progressively with age in APP/PS1 mice, indicatingthat Aβ and aging can stimulate the expression of PEA15.In 1-month-old APP/PS1 mice, although no Aβ depositionoccurred, APP protein levels were clearly elevated. The up-regulation of APP could result in an overproduction of Aβpeptides and ultimately lead to the formation of Aβ deposits.Therefore, the increase in PEA15 expression in 1-month-oldAPP/PS1micemight enable enhancedAβ clearance and preventAβ deposition. However, the increase of PEA15 in APP/PS1 micedoes not appear to be sufficient for compensating for the speedof Aβ production: excess Aβ was only partially removed fromthe brain, and therefore, Aβ depositions were detected in5-month-old APP/PS1 mice. The WT mice examined here alsoshoweda progressive increase in PEA15 expression, indicating arole of PEA15 in the aging process.
4.2. Astrocyte-mediated phagocytosis in AD
Astrogliosis is a ubiquitous but poorly understood patholo-gical feature in neurodegenerative diseases [28,29,40,43,81].
55J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
Reactive astrocytes undergo morphological and molecularchanges, and this transformation of astrocytes can trigger achange in their metabolic phenotype [29]. This alteration inmetabolism appears to underlie the metabolic perturbationsin AD [82]. Accumulating evidence suggests that reactiveastrocytes can exert both beneficial and detrimental effects byregulating specific molecular-signaling cascades [11,28,29]. Inthe early stages of AD, reactive astrocytes perform beneficialfunctions by forming a boundary around the regions occupiedby Aβ plaques and playing a crucial role in Aβ clearance[20,21]. However, persistent activation of astrocytes becomesdetrimental because these cells then overproduce proinflam-matory cytokines and free radicals, which induce neuronaldeath [40]. In this study, we observed that astrocytes wereactivated around Aβ plaques and participated in clearingextracellular Aβ in 5-month-old APP/PS1 mice. However,reactive astrocytes appeared to be unable to eliminate thedeposited Aβ, which could be because of the insoluble natureand the substantial quantities of the amyloid present. Thus,Aβ depositions increased with age and continued to activateneighboring astrocytes over prolonged periods.
Astrocytes play a crucial role in maintaining brain homeo-stasis [83]. However, after transforming from a quiescent stateto a reactive state, astrocytes lose their neurosupportiveproperty and render neurons vulnerable to neurotoxicity. Weobserved that in 12-month-old APP/PS1 mice, reactive astro-cytes occupied almost the entire brain. During this period,astrocytes are considered neurotoxic rather than homeostaticbecause they release inflammatory factors and free radicals,and we observed this in our study. Reactive astrocytesgenerally help defend the brain against acute neurotoxicityoccurring locally at an early stage of development, but thiscould become counterproductive if the reactive status of thecells is widespread and sustained.
Astroglial dysfunction has been documented in variousneurodegenerative diseases [27–29]. Metabolic dysregulation,particularly the impairment of lactate efflux in astrocytes, islikely to play a crucial role in superoxide dismutase 1-relatedamyotrophic lateral sclerosis [84]. A series of cellular dysfunc-tions including abnormal tau phosphorylation, increasedradical superoxide generation, reduced mitochondrial mem-brane potential, and diminished glutamate uptake wereobserved in the astrocytes of senescence-accelerated mice(SAMP8) [28]. Abnormal glutamate metabolism and reducedglucose uptake in astrocytes have also been detected in AD[30,31]. These findings have given rise to the view that theability of astrocytes to clear Aβ might also becomeoverwhelmed during the AD process, which would contributeto Aβ deposition. In our study, primary astrocytes isolatedfrom APP/PS1 mice exhibited a substantially lower ability tophagocytose Aβ than did the astrocytes isolated from WTmice. Moreover, under Aβ stimulation, PEA15 expressiondecreased in primary astrocytes isolated from APP/PS1mouse. This positive correlation between PEA15 expressionand Aβ phagocytosis further confirmed the role of PEA15 inthe process of Aβ phagocytosis. Aβ could stimulate theup-regulation of PEA15 (Fig. 3A), while the expression ofPEA15 in 10-month-old APP/PS1 mice showed no increasecompared to WT mice (Fig. 4B). Moreover, the increase rate ofPEA15 expression from 5 months to 10 months was obviously
slower in APP/PS1 mice than in WT mice. These resultsindicated that the up-regulation of PEA15 responding to Aβover-production was gradually overwhelmed in transgenicmice as the increased deposition of extracellular Aβ, andthus may contribute to the astroglial dysfunction in Aβphagocytosis. These results strongly support the view thatAβ phagocytosis might be impaired in AD, which in turncontribute to Aβ deposits during the AD process.
5. Conclusions
In this study, we used the cerebral cortices of 5-month-old APP/PS1 mice and WT mice and TMT-based quantitative proteomicanalysis to obtain a profile of proteins that exhibit alteredexpression in the AD mouse model; this is the first time thatmost of these proteins have been reported to be differentiallyexpressed in this model. Our bioinformatics analysis suggestedthat the altered proteins participate in AD progression becauseof its role in signal transduction, molecular transport, lipidmetabolism, inflammation, and response to oxidative stress.Moreover, we comprehensively investigated the function ofPEA15 in astrocyte-mediated Aβ clearance. The ability ofprimary astrocytes to phagocytose Aβwas markedly decreasedafter the inhibition of PEA15 expression, whereas this abilitywas enhanced following the overexpression of PEA15 in thecells, which indicated a regulatory role of PEA15 in the processof Aβ phagocytosis. Furthermore, by examining PEA15 expres-sion in the brains of 1-, 2-, 5-, and 10-month-old APP/PS1 mice,we determined that PEA15 levels increased progressively withage in the APP/PS1 mice, which suggested that Aβ can inducePEA15 expression. Interestingly, primary astrocytes isolatedfromAPP/PS1mice exhibitedaweaker ability to phagocytoseAβthan did the astrocytes isolated fromWTmice, which indicatedan astroglial dysfunction in Aβ clearance in the APP/PS1 mice.These results highlight the role of PEA15 in astrocyte-mediatedAβ clearance in the APP/PS1 mouse AD model, and thusimprove our understanding of the molecular mechanismsunderlying Aβ clearance and deposition. In future studies, themechanisms by which protein expression is altered in ADshould be elucidated, which will provide further insights intothe pathogenesis of AD and help identify therapeutic targets forthe disease.
Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.jprot.2014.07.028.
Transparency document
The Transparency document associated with this article canbe found, in the online version.
Acknowledgments
This work was supported by grants from the National NaturalScience Foundation of China (Nos. 31270872, 31470807,31200610) and the National Key Basic Research Program ofChina (Nos. 2010CB912203, 2011CB915504).
56 J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
R E F E R E N C E S
[1] Longridge RG. Alzheimer's disease. Proc R Soc Med 1939;32:222–4.
[2] Cummings JL. Alzheimer's disease. N Engl J Med 2004;351:56–67.
[3] Querfurth HW, LaFerla FM. Alzheimer's disease. N Engl J Med2010;362:329–44.
[4] Lee Y-J, Han SB, Nam S-Y, Oh K-W, Hong JT. Inflammationand Alzheimer's disease. Arch Pharm Res 2010;33:1539–56.
[5] Luo X, Weber GA, Zheng J, Gendelman HE, Ikezu T.C1q-calreticulin induced oxidative neurotoxicity: relevancefor the neuropathogenesis of Alzheimer's disease. JNeuroimmunol 2003;135:62–71.
[6] Voloboueva LA, Giffard RG. Inflammation, mitochondria, andthe inhibition of adult neurogenesis. J Neurosci Res 2011;89:1989–96.
[7] Crews L, Rockenstein E, Masliah E. APP transgenic modeling ofAlzheimer's disease: mechanisms of neurodegeneration andaberrant neurogenesis. Brain Struct Funct 2009;214:111–26.
[8] Giménez-Llort L, Blázquez G, Cañete T, Johansson B, Oddo S,Tobeña A, et al. Modeling behavioral and neuronalsymptoms of Alzheimer's disease in mice: a role forintraneuronal amyloid. Neurosci Biobehav Rev 2007;31:125–47.
[9] Balducci C, Forloni G. APP transgenic mice: their use andlimitations. Neuromolecular Med 2010;13:117–37.
[10] Duyckaerts C, Potier M-C, Delatour B. Alzheimer diseasemodels and human neuropathology: similarities anddifferences. Acta Neuropathol 2007;115:5–38.
[11] Mrak RE, Griffinbc WS. The role of activated astrocytes and ofthe neurotrophic cytokine S100B in the pathogenesis ofAlzheimer's disease. Neurobiol Aging 2001;22:915–22.
[12] Rogers J, Strohmeyer R, Kovelowski CJ, Li R. Microglia andinflammatory mechanisms in the clearance of amyloid betapeptide. Glia 2002;40:260–9.
[13] Thal DR, Sassin I, Schultz C, Haass C, Braak E, Braak H. Fleecyamyloid deposits in the internal layers of the humanentorhinal cortex are comprised of N-terminaltruncated fragments of Abeta. J Neuropathol Exp Neurol 1999;58:210–6.
[14] Funato H, Yoshimura M, Yamazaki T, Saido TC, Ito Y,Yokofujita J, et al. Astrocytes containing amyloid beta-protein(Abeta)-positive granules are associated withAbeta40-positive diffuse plaques in the aged human brain.Am J Pathol 1998;152:983–92.
[15] Ma Y, Bao J, Zhao X, Shen H, Lv J, Ma S, et al. Activatedcyclin-dependent kinase 5 promotes microglial phagocytosisof fibrillar beta-amyloid by up-regulating lipoprotein lipaseexpression. Mol Cell Proteomics 2013;12:2833–44.
[16] Tanzi RE, Moir RD, Wagner SL. Clearance of Alzheimer'sAbeta peptide: the many roads to perdition. Neuron 2004;43:605–8.
[17] Guenette SY. Astrocytes: a cellular player in Abeta clearanceand degradation. Trends Mol Med 2003;9:279–80.
[18] Shaffer LM, Dority MD, Gupta-Bansal R, Frederickson RC,Younkin SG, Brunden KR. Amyloid beta protein (A beta)removal by neuroglial cells in culture. Neurobiol Aging 1995;16:737–45.
[19] Guenette SY. Mechanisms of Abeta clearance and catabolism.Neuromolecular Med 2003;4:147–60.
[20] Wyss-Coray T, Loike JD, Brionne TC, Lu E, Anankov R, Yan F,et al. Adult mouse astrocytes degrade amyloid-β in vitro andin situ. Nat Med 2003;9:453–7.
[21] Pihlaja R, Koistinaho J, Malm T, Sikkila H, Vainio S, KoistinahoM. Transplanted astrocytes internalize depositedbeta-amyloid peptides in a transgenic mouse model ofAlzheimer's disease. Glia 2008;56:154–63.
[22] Pihlaja R, Koistinaho J, Kauppinen R, Sandholm J, Tanila H,Koistinaho M. Multiple cellular and molecular mechanismsare involved in human Abeta clearance by transplanted adultastrocytes. Glia 2011;59:1643–57.
[23] Basak JM, Verghese PB, Yoon H, Kim J, Holtzman DM.Low-density lipoprotein receptor represents an apolipoproteinE-independent pathway of Abeta uptake and degradation byastrocytes. J Biol Chem 2012;287:13959–71.
[24] Jones RS, Minogue AM, Connor TJ, Lynch MA.Amyloid-beta-induced astrocytic phagocytosis is mediatedby CD36, CD47 and RAGE. J Neuroimmune Pharmacol 2013;8:301–11.
[25] Rezai-Zadeh K, Gate D, Gowing G, Town T. How to get fromhere to there: macrophage recruitment in Alzheimer'sdisease. Curr Alzheimer Res 2011;8:156–63.
[26] Verghese PB, Castellano JM, Garai K, Wang Y, Jiang H, Shah A,et al. ApoE influences amyloid-beta (Abeta) clearance despiteminimal apoE/Abeta association in physiological conditions.Proc Natl Acad Sci U S A 2013;110:E1807–16.
[27] Seifert G, Schilling K, Steinhäuser C. Astrocyte dysfunction inneurological disorders: a molecular perspective. Nat RevNeurosci 2006;7:194–206.
[28] Garcia-Matas S, Gutierrez-Cuesta J, Coto-Montes A,Rubio-Acero R, Diez-Vives C, Camins A, et al. Dysfunction ofastrocytes in senescence-accelerated mice SAMP8 reducestheir neuroprotective capacity. Aging Cell 2008;7:630–40.
[29] Steele ML, Robinson SR. Reactive astrocytes give neurons lesssupport: implications for Alzheimer's disease. NeurobiolAging 2012;33:423 [e1-13].
[30] Freemantle E, Vandal M, Tremblay-Mercier J, Tremblay S,Blachere JC, Begin ME, et al. Omega-3 fatty acids, energysubstrates, and brain function during aging. ProstaglandinsLeukot Essent Fatty Acids 2006;75:213–20.
[31] Zou J, Wang YX, Dou FF, Lu HZ, Ma ZW, Lu PH, et al.Glutamine synthetase down-regulation reduces astrocyteprotection against glutamate excitotoxicity to neurons.Neurochem Int 2010;56:577–84.
[32] Rhein V, Song X, Wiesner A, Ittner LM, Baysang G, Meier F,et al. Amyloid-beta and tau synergistically impair theoxidative phosphorylation system in triple transgenicAlzheimer's disease mice. Proc Natl Acad Sci U S A 2009;106:20057–62.
[33] Wisniewski JR, Zougman A, Nagaraj N, Mann M. Universalsample preparation method for proteome analysis. NatMethods 2009;6:359–62.
[34] Tabas-Madrid D, Nogales-Cadenas R, Pascual-Montano A.GeneCodis3: a non-redundant and modular enrichmentanalysis tool for functional genomics. Nucleic Acids Res 2012;40:W478–83.
[35] Lai W, Wu J, Zou X, Xie J, Zhang L, Zhao X, et al. Secretomeanalyses of Abeta(1–42) stimulated hippocampal astrocytesreveal that CXCL10 is involved in astrocyte migration. JProteome Res 2013;12:832–43.
[36] Ermolinsky B, Pacheco Otalora LF, Arshadmansab MF, ZareiMM, Garrido-Sanabria ER. Differential changes in mGlu2 andmGlu3 gene expression following pilocarpine-induced statusepilepticus: a comparative real-time PCR analysis. Brain Res2008;1226:173–80.
[37] Fiory F, Formisano P, Perruolo G, Beguinot F. Frontiers: PED/PEA-15, a multifunctional protein controlling cell survivaland glucosemetabolism. Am J Physiol Endocrinol Metab 2009;297:E592–601.
[38] Horti AG, Gao Y, Ravert HT, Finley P, Valentine H, Wong DF,et al. Synthesis and biodistribution of [11C]A-836339, a newpotential radioligand for PET imaging of cannabinoid type 2receptors (CB2). Bioorg Med Chem 2010;18:5202–7.
[39] Ruan L, Kang Z, Pei G, Le Y. Amyloid deposition andinflammation in APPswe/PS1dE9 mouse model ofAlzheimer's disease. Curr Alzheimer Res 2009;6:531–40.
57J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
[40] Sofroniew MV. Molecular dissection of reactiveastrogliosis and glial scar formation. Trends Neurosci 2009;32:638–47.
[41] Weber M, Scherf N, Kahl T, Braumann UD, Scheibe P, KuskaJP, et al. Quantitative analysis of astrogliosis indrug-dependent humans. Brain Res 2013;1500:72–87.
[42] O'Callaghan JP. Quantitative features of reactive gliosisfollowing toxicant-induced damage of the CNS. Ann N Y AcadSci 1993;679:195–210.
[43] Eng LF, Ghirnikar RS. GFAP and astrogliosis. Brain Pathol1994;4:229–37.
[44] Pahnke J, Frohlich C, Krohn M, Schumacher T, Paarmann K.Impaired mitochondrial energy production and ABCtransporter function—a crucial interconnection in dementingproteopathies of the brain. Mech Ageing Dev 2013;134:506–15.
[45] Krohn M, Lange C, Hofrichter J, Scheffler K, Stenzel J, Steffen J,et al. Cerebral amyloid-beta proteostasis is regulated by themembrane transport protein ABCC1 in mice. J Clin Invest2011;121:3924–31.
[46] Hofrichter J, Krohn M, Schumacher T, Lange C, Feistel B,Walbroel B, et al. Reduced Alzheimer's disease pathology bySt. John's Wort treatment is independent of hyperforin andfacilitated by ABCC1 and microglia activation in mice. CurrAlzheimer Res 2013;10:1057–69.
[47] Bales KR. Brain lipid metabolism, apolipoprotein E and thepathophysiology of Alzheimer's disease. Neuropharmacology2010;59:295–302.
[48] Reiss AB, Voloshyna I. Regulation of cerebral cholesterolmetabolism in Alzheimer disease. J Invest Med 2012;60:576–82.
[49] Mascitelli L, Seneff S, Goldstein MR. Association ofAlzheimer disease pathology with abnormal lipidmetabolism: the Hisayama study. Neurology 2012;78:151[author reply −2].
[50] Oyama R, Yamamoto H, Titani K. Glutamine synthetase,hemoglobin alpha-chain, and macrophage migrationinhibitory factor binding to amyloid beta-protein: theiridentification in rat brain by a novel affinity chromatographyand in Alzheimer's disease brain by immunoprecipitation.Biochim Biophys Acta 2000;1479:91–102.
[51] Bacher M, Deuster O, Aljabari B, Egensperger R, Neff F, JessenF, et al. The role of macrophage migration inhibitory factor inAlzheimer's disease. Mol Med 2010;16:116–21.
[52] Hoos MD, Ahmed M, Smith SO, Van Nostrand WE. Myelinbasic protein binds to and inhibits the fibrillar assembly ofAbeta42 in vitro. Biochemistry 2009;48:4720–7.
[53] Liao MC, Hoos MD, Aucoin D, Ahmed M, Davis J, Smith SO,et al. N-terminal domain of myelin basic protein inhibitsamyloid beta-protein fibril assembly. J Biol Chem 2010;285:35590–8.
[54] Liao MC, AhmedM, Smith SO, Van NostrandWE. Degradationof amyloid beta protein by purified myelin basic protein. J BiolChem 2009;284:28917–25.
[55] Renault-Mihara F, Beuvon F, Iturrioz X, Canton B, De BouardS, Leonard N, et al. Phosphoprotein enriched inastrocytes-15 kDa expression inhibits astrocyte migration bya protein kinase C delta-dependent mechanism. Mol Biol Cell2006;17:5141–52.
[56] Ramos JW, Townsend DA, Piarulli D, Kolata S, Light K, Hale G,et al. Deletion of PEA-15 in mice is associated with specificimpairments of spatial learning abilities. BMC Neurosci 2009;10:134.
[57] Peacock JW, Palmer J, Fink D, Ip S, Pietras EM, Mui AL, et al.PTEN loss promotes mitochondrially dependent type IIFas-induced apoptosis via PEA-15. Mol Cell Biol 2009;29:1222–34.
[58] Formstecher E, Ramos JW, Fauquet M, Calderwood DA, HsiehJC, Canton B, et al. PEA-15 mediates cytoplasmicsequestration of ERK MAP kinase. Dev Cell 2001;1:239–50.
[59] Sharif A, Canton B, Junier MP, Chneiweiss H. PEA-15modulates TNFalpha intracellular signaling in astrocytes.Ann N Y Acad Sci 2003;1010:43–50.
[60] Condorelli G, Vigliotta G, Iavarone C, Caruso M, Tocchetti CG,Andreozzi F, et al. PED/PEA-15 gene controls glucosetransport and is overexpressed in type 2 diabetes mellitus.EMBO J 1998;17:3858–66.
[61] Hao C, Beguinot F, Condorelli G, Trencia A, Van Meir EG, YongVW, et al. Induction and intracellular regulation of tumornecrosis factor-related apoptosis-inducing ligand (TRAIL)mediated apotosis in human malignant glioma cells. CancerRes 2001;61:1162–70.
[62] Xiao C, Yang BF, Asadi N, Beguinot F, Hao C. Tumor necrosisfactor-related apoptosis-inducing ligand-induceddeath-inducing signaling complex and its modulation byc-FLIP and PED/PEA-15 in glioma cells. J Biol Chem 2002;277:25020–5.
[63] Tsukamoto T, Yoo J, Hwang SI, Guzman RC, Hirokawa Y,Chou YC, et al. Expression of MAT1/PEA-15 mRNA isoformsduring physiological and neoplastic changes in the mousemammary gland. Cancer Lett 2000;149:105–13.
[64] Zabel C, Sagi D, Kaindl AM, Steireif N, Klare Y, Mao L, et al.Comparative proteomics in neurodegenerative andnon-neurodegenerative diseases suggest nodal pointproteins in regulatory networking. J Proteome Res 2006;5:1948–58.
[65] Thomason LA, Smithson LJ, Hazrati LN, McLaurin J, KawajaMD. Reactive astrocytes associated with plaques in TgCRND8mouse brain and in human Alzheimer brain expressphosphoprotein enriched in astrocytes (PEA-15). FEBS Lett2013;587:2448–54.
[66] Danziger N, Yokoyama M, Jay T, Cordier J, Glowinski J,Chneiweiss H. Cellular expression, developmental regulation,and phylogenic conservation of PEA-15, the astrocytic majorphosphoprotein and protein kinase C substrate. J Neurochem1995;64:1016–25.
[67] Bock BC, Tagscherer KE, Fassl A, Kramer A, Oehme I, ZentgrafHW, et al. The PEA-15 protein regulates autophagy viaactivation of JNK. J Biol Chem 2010;285:21644–54.
[68] Iovino S, Oriente F, Botta G, Cabaro S, Iovane V, Paciello O,et al. PED/PEA-15 induces autophagy andmediates TGF-beta1effect onmuscle cell differentiation. Cell Death Differ 2012;19:1127–38.
[69] Bonilla DL, Bhattacharya A, Sha Y, Xu Y, Xiang Q, Kan A, et al.Autophagy regulates phagocytosis by modulating theexpression of scavenger receptors. Immunity 2013;39:537–47.
[70] Franchi N, Schiavon F, Betti M, Canesi L, Ballarin L. Insight onsignal transduction pathways involved in phagocytosis in thecolonial ascidian Botryllus schlosseri. J Invertebr Pathol 2013;112:260–6.
[71] Fleisher-Berkovich S, Filipovich-Rimon T, Ben-Shmuel S,Hulsmann C, Kummer MP, Heneka MT. Distinctmodulation of microglial amyloid beta phagocytosisand migration by neuropeptides (i). J Neuroinflammation2010;7:61.
[72] Floden AM, Combs CK. Beta-amyloid stimulates murinepostnatal and adult microglia cultures in a unique manner. JNeurosci 2006;26:4644–8.
[73] Aderem A, Underhill DM. Mechanisms of phagocytosis inmacrophages. Annu Rev Immunol 1999;17:593–623.
[74] Flannagan RS, Jaumouille V, Grinstein S. The cell biology ofphagocytosis. Annu Rev Pathol 2012;7:61–98.
[75] Deretic V, Levine B. Autophagy, immunity, and microbialadaptations. Cell Host Microbe 2009;5:527–49.
[76] Longatti A, Tooze SA. Vesicular trafficking andautophagosome formation. Cell Death Differ 2009;16:956–65.
[77] Martinez J, Almendinger J, Oberst A, Ness R, Dillon CP,Fitzgerald P, et al. Microtubule-associated protein 1 lightchain 3 alpha (LC3)-associated phagocytosis is required for
58 J O U R N A L O F P R O T E O M I C S 1 1 0 ( 2 0 1 4 ) 4 5 – 5 8
the efficient clearance of dead cells. Proc Natl Acad Sci U S A2011;108:17396–401.
[78] Friggeri A, Banerjee S, Biswas S, de Freitas A, Liu G, BierhausA, et al. Participation of the receptor for advancedglycation end products in efferocytosis. J Immunol 2011;186:6191–8.
[79] Tang D, Kang R, Coyne CB, Zeh HJ, Lotze MT. PAMPs andDAMPs: signal 0 s that spur autophagy and immunity.Immunol Rev 2012;249:158–75.
[80] Cadwell K, Philips JA. Autophagy meets phagocytosis.Immunity 2013;39:425–7.
[81] Little AR, O'Callagha JP. Astrogliosis in the adult anddeveloping CNS: is there a role for proinflammatorycytokines? Neurotoxicology 2001;22:607–18.
[82] Henry W FM. <Alzheimer's Disease.pdf>. 2010.[83] Verkhratsky A, Olabarria M, Noristani HN, Yeh CY, Rodriguez
JJ. Astrocytes in Alzheimer's disease. Neurotherapeutics 2010;7:399–412.
[84] Ferraiuolo L, Higginbottom A, Heath PR, Barber S, Greenald D,Kirby J, et al. Dysregulation of astrocyte-motoneuroncross-talk in mutant superoxide dismutase 1-relatedamyotrophic lateral sclerosis. Brain 2011;134:2627–41.