Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3

http://dx.doi.org/100006-8993/Published(http://creativecomm

E-mail addressemeghan.makley.ctrjeffery.gearhart.ctr@

Research Report

Quantitative proteomic analysis of the brainstemfollowing lethal sarin exposure

Mitchell L. Meadea,c, Andrea Hoffmanna, Meghan K. Makleya, ThomasH. Sniderb, John J. Schlagerc, Jeffery M. Gearharta,d

aHenry M. Jackson Foundation for the Advancement of Military Medicine, 2729 R Street, Wright Patterson AFB, Dayton,OH 45433, USAbBattelle Biomedical Research Center, 1425 Plain City Georgesville Road, West Jefferson, OH 43162, USAcMolecular Bioeffects Branch, Bioeffects Division, 711 Human Performance Wing, Human Effectiveness Directorate, AirForce Research Laboratory (711 HPW/RHDJ), WPAFB, Dayton, OH 45433, USAdBoonShoft School of Medicine, Wright State University, 3640 Col. Glenn Highway, Dayton, OH 45433, USA

a r t i c l e i n f o

Article history:

Accepted 24 March 2015

The brainstem represents a major tissue area affected by sarin organophosphate poisoning due

to its function in respiratory and cardiovascular control. While the acute toxic effects of sarin on

Available online 1 April 2015

Keywords:

Sarin

Brain effect

Quantitative proteomics

Cytoscape

Reactome functional interaction

network analysis

.1016/j.brainres.2015.03.04by Elsevier B.V. This isons.org/licenses/by-nc-

s: mitchell.meade.1@us.a@us.af.mil (M.K. Makley),us.af.mil (J.M. Gearhart)

a b s t r a c t

brainstem-related responses are relatively unknown, other brain areas e.g., cortex or cerebellum,

have been studied more extensively. The study objective was to analyze the guinea pig

brainstem toxicology response following sarin (2� LD50) exposure by proteome pathway

analysis to gain insight into the complex regulatory mechanisms that lead to impairment of

respiratory and cardiovascular control. Guinea pig exposure to sarin resulted in the typical acute

behavior/physiology outcomes with death between 15 and 25min. In addition, brain and blood

acetylcholinesterase activity was significantly reduced in the presence of sarin to 95%, and 89%,

respectively, of control values. Isobaric-tagged (iTRAQ) liquid chromatography tandem mass

spectrometry (LC–MS/MS) identified 198 total proteins of which 23% were upregulated, and 18%

were downregulated following sarin exposure. Direct gene ontology (GO) analysis revealed a

sarin-specific broad-spectrum proteomic profile including glutamate-mediated excitotoxicity,

calcium overload, energy depletion responses, and compensatory carbohydrate metabolism,

increases in ROS defense, DNA damage and chromatin remodeling, HSP response, targeted

protein degradation (ubiquitination) and cell death response. With regards to the sarin-

dependent effect on respiration, our study supports the potential interference of sarin with

CO2/Hþ sensitive chemoreceptor neurons of the brainstem retrotrapezoid nucleus (RTN) that

send excitatory glutamergic projections to the respiratory centers. In conclusion, this study gives

insight into the brainstem broad-spectrum proteome following acute sarin exposure and the

gained information will assist in the development of novel countermeasures.

Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1an open access article under the CC BY-NC-ND licensend/4.0/).

f.mil (M.L. Meade), drahoffmann1@gmail.com (A. Hoffmann),snidert@battelle.org (T.H. Snider), john.schlager@us.af.mil (J.J. Schlager),

.

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3102

1. Introduction

Sarin (O-isopropyl methylphosphonofluoridate, also known asGB) is a toxic organophosphate (OP) nerve agent developed inGermany just preceding World War II. Together with soman(GD), tabun (GA) and cyclosarin (GF), the chemical is classifiedamong the group of “G” agents based upon its syntheticorigination and its ability to be applied as a chemical warfareweapon. Sarin has been discharged in both military and civilianattacks, among the most recent include the Iran-Iraq war (1980–1988), 1994 in open crowds in Matsumoto, and 1995 in theTokyo subway (Yanagisawa et al., 2006; Yokoyama et al., 1996).Reports also suggest that this agent was released in Syria duringthe current civil war (http://www.bbc.co.uk/news/world-middle-east-22773268), and in the most recent attack on the Damascussuburb of Ghouta in August 2013 (http://s3.documentcloud.org/documents/787427/u-n-syria-chemical-report.pdf) that incurredover 1700 fatalities. Sarin is an effective lethal toxic chemical inboth vapor and liquid form, due to its relatively high volatilityand its rapid uptake by inhalation and absorption through theskin or eyes (Hulet et al., 2014; Eason, 2013). Following uptake,acetylcholinesterase (AChE) is themainmolecular protein targetfor sarin and most other OPs. Irreversible inhibition of AChEresults in accumulation of the neurotransmitter acetylcholine atcholinergic synapses, which can produce a lethal cholinergiccrisis due to overstimulation of muscarinic and nicotinic recep-tors in the central and peripheral nervous system, including theneuromuscular junction (Bajgar, 2005). Systemic AChE inhibi-tion triggers seizures, lacrimation, defecation and asphyxiation(Klein-Rodewald et al., 2011; Solberg and Belkin, 1997). The time-course of nerve agent poisoning is very rapid, and typical deathby asphyxiation may occur within minutes, depending on thedose and route of exposure (Bajgar, 2005; Carey et al., 2013).Lethality is caused by a central depressant effect on therespiratory center in brainstem, constriction of airways, signifi-cant production of lung fluids, and paralysis of respiratorymusculature (Rickett et al., 1986).

Neuroproteomics represents an emerging approachtoward improvement in understanding a portion of thecomplex molecular response mechanisms that underliechanges including brain dysfunction during lethality. Massspectrometry (MS), while uniquely suited for identificationand quantitation of large numbers of proteins in a singleanalysis, has evolved into an indispensable tool for proteomicresearch. While previous acute and chronic sarin exposurestudies have focused on analyzing changes in mRNA expres-sion, and selective protein analysis, no study has analyzedthe broad-spectrum proteomic profile following OP exposure(Bhardwaj et al., 2012; Bhardwaj et al., 2012; Torres-Altoroet al., 2011). Accordingly, this study used proteomic profiling

Table 1 – Clinical observations exhibited by guinea pigs time (

Animal ID Fasciculations Ataxia Chewin

176195196 1–5 1–5197 1–5

of guinea pig brainstem following acute sarin poisoning toidentify pathways and compensatory regulatory mechanismsthat respond during acute OP-dependent toxicity within thebrainstem. This study profiles a significant number of sarin-modulated proteins simultaneously across multiple samples byisobaric tag (iTRAQ) quantitation using reverse phase liquidchromatography and tandem mass spectrometry (LC–MS/MS)(Tsuchida et al., 2013). Selected proteins that demonstratedchanges in expression were confirmed by quantitative immu-noblotting. To gain further insight into the detailed proteininteraction network following sarin exposure, the MS data werefurther analyzed by direct gene ontology (GO) analysis, andproteome pathway analysis using the Reactome FI applicationof Cytoscape 3.0.2.

2. Results

2.1. Clinical results

All guinea pigs in the sarin exposure group died within 15 to30 min of exposure. Three of four exposed animals experi-enced tremors within 5 to 15 min following exposure, whilesome exhibited fasciculations, salivation and prostration.Table 1 summarizes the clinical observations for each animal.

2.2. AChE activity in blood and brain following exposure

As expected, in line with previously published studies, bloodand brain AChE activities following sarin exposure wereseverely decreased (Young et al., 2001; Khan et al., 2000).Total brain AChE activity was significantly reduced to 95%(po0.001, t-test) with a mean value of 0.6370.03 U/mg in thesarin exposure group, while the non-exposed naïve controlgroup of equally processed brains from a companion studyhad an activity of 14.4071.20 U/mg tissue (Fig. 1). Similarly,total blood AChE activity demonstrated an initial reductionfrom 1.1270.04 U/ml pre-exposure to 0.9970.12 U/ml at 45 sinterim exposure to a final significant reduction (po0.0001,One-way ANOVA) of 0.1070.01 U/ml post exposure, averagingan approximate 89% loss in AChE activity.

2.3. Mass spectrometry and Western blot analysis

MS profiling detected 198 proteins quantitated through iTRAQanalysis. Of these, 46 proteins (23%) demonstrated anincreased expression, with an average iTRAQ ratio higherthan 1.20, and 36 proteins (18%) were decreased, with anaverage iTRAQ ratio of less than 0.83. The identified modu-lated proteins belonged to all subcellular regions and haddiverse functions throughout the cell (Fig. 2a). A complete list

min) post-challenge.

g Salivation Tremors Time of death

151–5 15

1–15 1–15 251–5 15

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3 103

of modulated proteins can be found in the Supplementarymaterials. The highest number of upregulated proteins wasin the cytosolic fraction with increased expression in 17proteins followed by increased expression in 9 nuclear and8 mitochondrial proteins. In addition, expression of 6 cytos-keletal, 4 membrane, and 1 extracellular matrix (ECM) pro-teins were upregulated. The downregulated protein panelincluded 13 cytosolic, 10 cytoskeletal, 5 membrane, 1 mito-chondrial, 1 ECM protein and 4 nuclear proteins.

2.3.1. Upregulation of protein expression following sarinexposureThe highest percentage of MS-detected upregulated proteinsin response to sarin belonged to the categories of cellularmetabolism (4.04%), apoptosis/cell cycle (4.04%), and chroma-tin remodeling (4.04%) followed by the categories of structuralproteins important for neuronal degeneration/regeneration(3.53%), aspartate/glutamate transport/metabolism (1.51%),transcription/protein synthesis (1.51%), ubiquitin pathway

Fig. 1 – AChE activity in brain and blood following sarinexposure. Graph represents Ellman assay-determined AChEactivity levels from (a) whole brain; animal number (n¼32)non-exposed naïve control group from a companion study,(n¼4) in GB experimental, statistical analysis student's t-test, standard deviation displayed. (b) Blood; animal number(n¼4) in each group, statistical analysis one-way ANOVA,standard error displayed. po0.05 was used as a criterion forstatistic significance.

(1.51%), calcium binding (1.51%), and proteins for regulationof signaling/vesicle transport/ion transport (1.01%) (Fig. 2b).

2.3.2. Downregulation of protein expression following sarinexposureThe highest percentage of MS-detected downregulated pro-teins in response to sarin belonged to the category ofstructural proteins involved in neuronal degeneration/regen-eration (4.54%), followed by the categories of cellular meta-bolism (3.03%), transcription/protein synthesis (2.02%), apo-ptosis/cell cycle (1.51%), heat shock proteins (1.51%), andsignaling/vesicle transport/ion transport (1.51%) (Fig. 2b).Reactive oxygen species (ROS) defense/pH regulation andcalcium binding proteins were represented with 0.5%.

2.3.3. Effect of sarin on glutamate responseGlutamate-mediated excitotoxicity and the resulting delayedneuronal calcium overload represent major cytotoxic eventsduring OP-induced brain injury that also occurs within thebrainstem disrupting autonomic function involving respira-tory and cardiovascular control (Lallement et al., 1991;

Fig. 2 – Cellular location and biological function ofmodulated protein expression following sarin exposure.(a) Legend indicates color-coding for partitions of MSidentified modulated and non-modulated proteinscharacterized for “GO cell components” by UniprotKnowledge Base together with SEQUEST algorithm inProteome Discoverer 1.3. (b) Legend indicates color-codingfor partitions of MS identified modulated and non-modulated proteins characterized for “GO biologicalfunctions” by Uniprot Knowledge Base together withSEQUEST algorithm in Proteome Discoverer 1.3.

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3104

Solberg and Belkin, 1997; O’Donnell et al., 2011). Extracellularglutamate removal occurs through reuptake from thesynapse that is principally controlled by the glial andendothelial cell transporters EAAT-1 and EAAT-2 (Sheldonand Robinson, 2007) and alterations in these transporters hasbeen associated with different types of brain injury (Gilleyand Kernie, 2011; Yi and Hazell, 2006; Solberg and Belkin,1997). In accordance with the expected sarin-induced gluta-mate excitotoxicity, MS profiling demonstrated upregulationof EAAT-1 (GLAST-1/GLUT-1; 1.9770.50), and immunoblottingdemonstrated a significant upregulation (1.6170.12, po0.05)of EAAT-1 within the cytosolic fraction (Fig. 3a). No changesin other glutamate receptors including EAAT2, EAAT-3 andVGLUT-2 could be detected by immunoblotting with regardsto cytosolic fractions and whole cell lysates (not shown).While glial and endothelial cell EAAT-1 regulates uptake ofglutamate from synapses, the protein is also localized in theinner mitochondrial membrane as part of the aspartate/malate shuttle (Ralphe et al., 2004). Correspondingly, MS alsodetected upregulation of cytosolic aspartate aminotransfer-ase (1.2070.08) that plays a role in glutamate conversion bythe aspartate/malate shuttle, suggesting intracellular gluta-mate imbalance. Unchanged levels of glutamate dehydrogen-ase and glutamine synthetase detected in MS further supportthe lack of glutamate enzymatic partitioning and detoxifica-tion. Sarin also caused upregulation of glutamate vesicledocking protein syntaxin binding protein-1, and vesicletransport proteins dynein light chain 2 and AP-2 complexsubunit mu, possibly indicative of increased glutamate vesi-cle transport.

2.3.4. Effect of sarin on calcium responseFollowing an OP-induced excessive glutamate response,N-methyl-D-aspartate ionotropic glutamate receptors andvoltage-gated calcium channels represent the major iontransporters responsible for the toxic elevation in intracellu-lar calcium levels (Choi, 1988). In accordance with aglutamate-dependent increase in intracellular calcium levels,MS detected upregulation of calcium binding proteins calre-tinin and calmodulin, as well as guanine nucleotide-bindingprotein G(o) subunit alpha, which are important for regula-tion of calcium ion transport. Whereas previous studies,using acute low dose sarin (0.5�LD50), demonstrated differ-ential expression of CAMKIIα transcripts in cortex, midbrain,brainstem, and cerebellum at longer time points (Damodaran2009); Damodaran et al., 2006a), higher doses of sarin(2�LD50) did not alter CAMKIIα levels (1.0870.27), as con-firmed by immunoblotting (0.9070.16, p¼0.61) (Fig. 3b). Inline with our data, previous studies that used AChE inhibitors(carbaryl, malathione, methyl- and ethyl-parathione) supportthe observed stable CAMKII protein levels and rather suggesta potential alteration in its activity (Pala et al., 1991).

2.3.5. Sarin-induced energy lossAs a result of increased intracellular calcium levels, OPexposure causes ATP depletion and cellular energy lossin vitro and in vivo (Chan et al., 2006; Gupta et al., 2001;Damodaran et al., 2006b). Indicative of OP-induced energydepletion in our study was a sarin-dependent upregulation ofessential energy metabolic proteins including ADP/ATP

translocase, mitochondrial aconitase, and cytochromes.Further indicative of an ongoing energy loss was an upregu-lation of ATPaseα1 (1.6570.08). Immunoblotting detection ofATPaseα1 in cytosolic fractions demonstrated a similar trendin the sarin (1.5170.19, p¼0.11) (Fig. 3c). In comparison, low-dose (0.5�LD50) sarin tested by the group of Damodaran et al.had a different impact on energy depletion that was limitedto changes in mRNA expression of cytochrome c and ATPa-seα1 subunits following 15 min to two hours exposure in rats(Damodaran et al., 2006b). In addition, the essential carbohy-drate metabolism protein pyruvate dehydrogenase E1demonstrated a compensatory upregulation.

Specific essential metabolic proteins were found down-regulated in our study, e.g., creatine kinase (Gupta et al.,2000), glyceraldehyde-3-phosphate dehydrogenase, and glyco-gen phosphorylase. Other downregulated metabolic proteinsincluded energy transport regulators, enzymes for steroidsynthesis and regulators of muscle contraction. Indicative ofa sarin-induced shift in energy balance towards energy losswas an increasing trend in MS-detected mitochondrialL-lactate dehydrogenase (1.1370.03) that was shown pre-viously during mevinphos-dependent energy depletion andnecrotic cell death of pheochromocytoma PC-12 cells (Chanet al., 2007).

2.3.6. Sarin effects on ROS responseTogether with an immediate impact on cell death signaling,acute and chronic OP exposure of humans and animals hasbeen previously linked to oxidative stress and ROS genera-tion. For instance, exposure to AChE inhibitors, e.g., chlorpyr-ifos, malathione, dimethoate, carbaryl, permethrin, anddichlorvos, has been associated with ROS-dependentincreases in vitamin and glutathione antioxidants, redoxstate imbalance, upregulation of lipid peroxidation, andalterations in ROS scavengers superoxide dismutase, cata-lase, glutathione S-transferase, glutathione peroxidase, andglutathione reductase (Akhgari et al., 2003; el-Sharkawy et al.,1994; Sharma et al., 2005). In this study, signs of sarin-dependent elimination of ROS were limited to MS-detecteddownregulation of glutathione S-transferase B (0.7570.14),and the lack of significant changes may be associated withrapid cellular events occurring prior to death.

2.3.7. Sarin-induced DNA damage and chromatin remodelingOP-induced DNA damage has been recorded in acute andchronic exposure settings. In a recent chronic OP exposurestudy, Ojha et al. reported induction of DNA damage in ratliver, kidney, spleen and brain (highest level of damage) 24 hpost treatment with chlorpyrifos, methyl parathion andmalathion (Ojha et al., 2013). In another study, acute (50 mgand 100 mg/kg BW daily for 1, 2, and 3 days) and chronic(1.12 mg and 2.24 mg/kg BW for 90 days) exposure to chlor-pyrifos resulted in a dose-dependent increase in DNA damagein rat liver and brain (Mehta et al., 2008). In accordance withthese prior studies, the MS profile displayed evidence of DNAdamage, including upregulation of DNA damage controlproteins, e.g., rab-7A, DNA damage binding protein-1, anddouble-strand break repair protein rad 21. MS also detectedan upregulation of nucleoside diphosphate kinase A, thatplays a role during energy reserve-related induction of

Fig. 3 – Confirmation of selective protein expression by Western blotting Western-blot analysis of brainstem lysates using(a) EAAT-1, (b) CAMKIIα, (c) ATPaseα1, (d) MBP, and (e) galectin-1 immunodetection. Graphs represent densitometry ofWestern-detected bands in upper panel normalized against the β-actin housekeeping control. Protein extracts of threeanimals (n¼3) were analyzed per group (error bars represent standard errors). Statistical analysis: Student's t-test, po0.05was used as a criterion for statistical significance.

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3 105

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3106

apoptosis, which supports evidence of pro-apoptotic signal-ing in response to high level sarin (Schlattner et al., 2009).

Besides DNA damage, OPs have been demonstrated tocause nuclear hypertrophy and chromatin dispersion in ratcerebrocortical and striatal neurons following 1 h subcuta-neous injection of soman (0.5, 0.8 and 1.0�LD50), resulting ininhibition of chromatin activity at 24 h (Martin et al., 1986).While the chromatin silencer histone deacetylase 5 demon-strated no changes, evidence of acute sarin-dependent chro-matin remodeling was given by MS-detected upregulation ofcore histone and linker proteins core histone macro-H2A,histone H1t, and histone H1.5, suggesting DNA damage-associated metabolic replacement of chromatin core andlinker histones, rather than organized regulation of chroma-tin activity.

2.3.8. Effect of sarin on heat shock protein responsePreviously, upregulation of HSP expression has been sug-gested to support correct assembly of proteins under stressconditions following acute OP exposure (Chan et al., 2007;Kim et al., 1999). For instance, HSP60 and HSP70 induce nitricoxide synthase I (NOSI) and protein kinase G-dependent anti-apoptotic signaling following acute bilateral low-dose mevin-phos microinjection (10 nmol in 50 nl) into the rostral ven-trolateral medulla of rats (Chan et al., 2007). NOS-dependentdelayed apoptosis was also observed following acute diiso-propylfluorophosphate (DFP, 1.8�LD50) exposure in rats (Kimet al., 1999).

In parallel with a non-significant upregulation in mito-chondrial HSP60 (mitochondrial chaperonin 60, CPN60), pro-teomic profiling detected upregulation of apoptotic signalingproteins, including rab-7A and proteins of the 14-3-3 andGTP-binding protein family, suggesting lack of HSP60-dependent anti-apoptotic signaling. Similar to Chan et al.(2007), MS detected downregulation of inducible HSP90, e.g.,HSP90α, HSP90β and the Hsp90 co-chaperone Cdc37, thatserves as stabilizers for the HSP90 proteins. Acute sarin ex-posure also caused downregulation of constitutivelyexpressed HSP70 (HSPA9). These results suggest that theprotein degradation-protective HSP response is degraded bythe high-dose effect of sarin. In accordance with a sarin-dependent degradation of HSPs, MS detected upregulation ofpolyubiquitin C, probable E3 ubiquitin-protein ligase and78kDa glucose-regulated protein, which further supportsactivation of pathways for targeted protein degradation.

2.3.9. Induction of cell death vs. regenerative pathways bysarinBesides a direct toxic effect, OP-associated activation of CAMKII,protein kinase C, and c-Jun N-terminal kinase signaling canmediate destabilization of neurofilaments and cytoskeletal pro-teins, e.g., tubulin, microtubule-associated protein-2 and -tau,kinesin motor proteins, GFAP, MBP, and vimentin, leading todisruption of axonal transport and axonal degeneration(Choudhary et al., 2001; Gearhart et al., 2007; Terry et al., 2007).MS demonstrated a sarin-dependent downregulation of MBP(0.5570.12) and myelin proteolipid protein (0.7170.31), both ofwhich are essential components of axon myelin sheets. Immu-noblotting of the cytosolic fraction confirmed downregulationof MBP (0.6370.04, po0.05) (Fig. 3d). Further evidence of

sarin-induced cytotoxicity was given by MS-detected downregu-lation of tubulins (tubulin-α1B, β2c, -β3, and -β4) andmicrotubule-associated protein tau. In addition, the microfilament assemblyprotein microtubule-associated serine/threonine protein kinase,and the cytoskeletal proteins septin-5, profilin-2, and myelinprotein 2, were upregulated following sarin exposure, indicativeof compensatory structural protein replacement. While MSdetected an upregulation in the neuronal plasticity proteinneuronal growth regulator 1 (NEGR, neurotractin) (Schaferet al., 2005), a lack of ongoing axonal regeneration was indicatedby the MS-observed downregulation (0.4470.04) in the β-galactoside binding protein galectin-1 (Endo, 2005; Horie et al.,2005) which was confirmed by immunoblotting of the cytosolicfraction (0.7570.06, po0.05) (Fig. 3e).

In a prior study by Damodaran et al. under low dose sarin(0.5�LD50), spatial and temporal differences in GFAP andvimentin transcript levels were observed in the brainstemversus cortex, cerebellum, midbrain, and spinal cord, follow-ing 1 h to 7 days post treatment which was suggested toresult from increased astrocyte dysfunction and loss in thisarea (Damodaran et al., 2002). Together with a non-significantdownregulation of GFAP (0.4970.25) MS detected a down-regulation in the astrocyte death marker and calcium bindingprotein S100-B (0.6370.19) that is important for GFAP stabi-lization supporting sarin brain degenerative effects (Ziegleret al., 1998).

2.4. Reactome FI network analysis

To gain insight into the detailed proteomic interaction net-work following sarin exposure, the MS data was furtheranalyzed by FI network predictions using the Reactome FIapplication of Cytoscape 3.0.2 (Fig. 4) (Milacic, 2012; Wu et al.,2010) Only 115 entries of the MS-identified 198 UniProt/Swis-sProt accession numbers were matched with GeneWiki nameentries using the BioMart portal. Of the 115 GeneWiki namesentered into the Reactome FI application of Cytoscape 3.0.2under the 2012 algorithm for “gene sets” using “Fetch FIannotations” and “use linker genes” for network prediction,only 86 original nodes plus 18 predicted nodes (104 totalnodes) were found for the total query. Of the 104 total nodes,24 original nodes plus 12 predicted nodes belonged to the MSupregulation cluster (above 1.2-fold expression) and 14 originalnodes plus 6 predicted nodes belonged to MS downregulationcluster (below 0.83-fold expression). Predicted general linkernodes included the epidermal growth factor receptor EGFR, theKlotho β-glucuronidase, BMP antagonist DAND5, HSP90AA,the signaling regulators GRB2, MAPK, SRC, and PAFAH1B. Inthe MS upregulation cluster, predicted linker nodes includedthe synapse-associated protein DLG1, mitochondrial ADP/ATPtranslocase SLC25A4, signaling kinase PRKACA, HSP8, theproteasome protein PSMC2, cell cycle protein CDC42, and aseries of nuclear factors e.g., TFAP2A, HNF4A, GTF2H2, ERCC2,RAD21, SINA3. The MS downregulation cluster includedsynapse-associated protein DLG4, amyloid precursor protein(APP), chaperonin CCT5, microtubule motor protein dyneinDYNC1LI2, and signaling proteins HRAS, and NFκB.

According to the Cytoscape 3.0.2 Reactome FI biologicalnetwork cluster prediction for “GO Biological Process”, fivenetwork clusters (Fig. 4) were identified from the 104 total

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3 107

nodes by the spectral partitioning-based network clusteringalgorithm (Newman, 2006). The biggest predicted modulewith 35 green nodes was related to cellular metabolism andmaintenance including enzymes for energy production, andproteins related to mRNA/protein synthesis and degradation.The second predicted module with 32 blue nodes wasassociated with synaptic transmission and cell stressresponse, including EGF/NGF/IFNγ signaling, protein trans-port, actin cytoskeletal organization, and ROS response. Thethird predicted module with 22 purple nodes encompassedcytoskeletal rearrangement, transport and axon guidanceincluding microtubule organization and synaptic transmis-sion. The fourth predicted module with 8 brown nodes wasrelated to cellular energy rescue response including gluta-mate/malate biosynthesis, carbohydrate and lactate metabo-lism. The last module with 7 blue-green nodes includedregulation of calcium response, cell cycle, gene expressionand chromatin remodeling.

3. Discussion

3.1. General observations

Overall, the MS-identified broad-spectrum proteomic profilefor acute high dose sarin exposure is in agreement with priorOP genomic response profiling studies. For instance, simila-rities were observed with regard to acute OP-induced gluta-mate-mediated excitotoxicity (Lallement et al., 1991;O’Donnell et al., 2011; Chan et al., 2006; Solberg and Belkin,1997), calcium overload (Choi, 1988), ATP depletion andcellular energy loss (Gupta et al., 2001), increase in ROSdefense (Akhgari et al., 2003; el-Sharkawy et al., 1994;Sharma et al., 2005; Verma and Srivastava, 2001), DNAdamage (Mehta et al., 2008; Ojha et al., 2013), chromatinremodeling (Martin et al., 1986), HSP response (Chan et al.,2007; Kim et al., 1999), and targeted protein degradation(ubiquitination). Correspondingly, MS proteomic profilingalso matched previously observed signs of OP-induced celldeath including destabilization of neurofilaments and cytos-keletal proteins, leading to disruption of axonal transport andaxonal degeneration (Choudhary et al., 2001; Gearhart, et al.,2007; Terry et al., 2007).

MS was also able to expand previous findings with regardsto the chromatin remodeling response, including DNAdamage-associated metabolic replacement of chromatin coreand linker histones. In comparison to previous OP geneexpression studies, our study also delineated distinctchanges in MS-detected protein expression in associationwith the used 2�LD50 dose of sarin and the time of analysisat 15–25 min post exposure. For instance, MS profilingdetected an upregulation in essential proteins of mitochon-drial energy metabolism following high dose sarin exposure,including ATPaseα1, ADP/ATP translocase, aconitase, andcytochromes. In contrast, low doses of sarin (0.5�LD50)seemed to have a different impact on energy depletion thatwas limited to changes in mRNA expression of cytochrome cand ATPaseα1 subunits following 15 min to 2 h of exposure inrats (Damodaran et al., 2006b). Besides temporal and dose-related effects, potential differences in protein versus mRNA

expression studies can be explained by the complexity ofprotein metabolism, including functional modifications andprotein trafficking/storage, rather than immediate proteindegradation. While MS is able to quantitate hundreds ofproteins in a single analysis, there are many proteinsexpressed in too low an abundance that is below detectionlimits and certain types of proteins and peptides that aremore easily ionized and thus more easily detected thanothers. Membrane proteins, in particular, are difficult todetect due to their hydrophobicity. The non-appearance of aprotein in the MS is by no means indicative that the protein isnot in the cell or changed. Correspondingly, this studyincluded an analysis of cellular protein networks, such asReactome FI, which has the ability to predict proteins thatwould fall under the MS detection limit, in order to give amore comprehensive picture of the actual protein responsesthat occur following insult.

3.2. Reactome FI network analysis

With regards to the FI network predictions, the five generatedmodules of (1) cellular metabolism and maintenance,(2) synaptic transmission and cell stress response, (3) cytos-keletal rearrangement, transport and axon guidance, (4)cellular energy rescue response, and (5) calcium response,gene expression and chromatin remodeling match the directGO term analysis of MS-detected proteins very well. Only 115entries of the MS-identified 198 UniProt/SwissProt accessionnumbers were matched with GeneWiki name entries usingthe BioMart portal, which was the highest number of matchesin comparison to other search engines such as UniProt(http://www.uniprot.org/docs/pathway; UniProt Consortium)that only matched 97 entries. The almost 50% mismatchwhen converting UniProt/SwissProt accession numbers intoGeneWiki names can be explained by the fact that not everyprotein/protein fragment has been associated with a pub-lished Wiki article and associated GeneWiki name (Goodet al., 2011). In addition, GeneWiki names only cover humangenes while the UniProt/SwissProt accession numbers coverprotein information from several species including guineapig. The Reactome FI Network of Cytoscape covers close to50% of known human proteins that are associated withcancer and other human diseases, which might explain alack of network integration for some sarin toxicity identifiedproteins (Wu et al., 2010). When entered into the Reactome FIapplication of Cytoscape 3.0.2 under the 2012 algorithm for“gene sets” using “Fetch FI annotations” and “use linkergenes” for network prediction, the 115 GeneWiki namesresulted in 86 original nodes that were linked by enrichmentwith 18 additionally predicted linker nodes resulting in aninteraction network of 104 total nodes.

The 18 additionally predicted linker nodes give furtherinsight into potential regulatory pathways following sarinexposure. For instance, the inserted KL linker node (Klotho β-glucuronidase) demonstrated direct links to the predictedlinker nodes growth factor receptor bound protein (GRB2),HRAS, and MAPK signaling kinases. KL also demonstrateddirect connections to the sodium/potassium ATPase trans-port proteins ATP1B1 and ATP1A3, and nodes of the ubiquitincluster. While presently known functions of KL include

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3108

regulation of tissue sensitivity to insulin and oligodendrocytematuration and myelination (Chen et al., 2013), the ReactomeFI prediction reveals novel functions of KL as a modulator ofATP-dependent ion channel transporters, MAPK signalingand ubiquitination. In addition, the EGF receptor was pre-dicted as a major linker node together with its predicteddownstream effectors GRB2, DAND5, HRAS, MAPK, SRC, andoriginal node PP2R1A protein kinase. The EGF receptor nodealso demonstrated direct connections to the original node celldivision cycle CDC37, the predicted node transcriptionalregulator TFAP2A, and nodes of the HSP and ubiquitinclusters. The inclusion of the EGF receptor as a predictedlinker node is supported by its function in control of cellsurvival responses. For instance, in a recent study of acuteintoxication with DFP (9 mg/kg BW, i.p.), pretreatment withthe neuregulin-epidermal growth factor isoform (NRG-EGF)significantly reduced neuronal injury at 24 h post-DFP injec-tion (Li et al., 2012). Further downstream, Reactome FI pre-diction included the linker node of the BMP antagonistDAND5, which demonstrated direct connections with nodesfor metabolic proteins TP1, GLUD1, LDHB, the predicted nodePRKACA, nodes of the HSP cluster and the nuclear factorHNF4A. It has been shown that chlorpyrifos modulates BMP-dependent neurite outgrowth, and such a finding supportsthe model prediction (Howard et al., 2005).

In the MS upregulation cluster, predicted linker nodesincluded the synapse-associated protein DLG1 (SAP97) thatdemonstrated direct connections to the calmodulin CALM1node for calcium-dependent signaling, and also connectionswith the glutamate transporter node SLC1A3 (EAAT-1) andthe predicted node CDC42. In addition, mitochondrial ADP/ATP translocase SLC25A4 was predicted as a linker node thatdemonstrated connections with the mitochondrial adeninenucleotide carrier SLC25A6 and adenylate kinase AK1.Another downstream-predicted node in the MS upregulationcluster was PRKACA that demonstrated direct interconnec-tions with the guanine nucleotide binding protein GNAO1and the clathrin adaptor complex protein AP2M1 with nodesof the YWHA (14-3-3) binding proteins. PRKACA cluster alsodemonstrated common interconnections with the predictednode CDC42 and nodes of the ubiquitin cluster. Further, inaccordance with the role of protein kinases in cell deathsignaling following acute OP poisoning (Bloch-Shildermanet al., 2005) the Reactome FI program predicted the PRKACAkinase linker node, HSP8 and the proteasome PSMC2 linkernode as well as a series of several nuclear regulatory linkernodes including TFAP2A, HNF4A, GTF2H2, ERCC2, RAD21,and SINA3.

In the MS down regulation cluster, DLG4 (SAP90/PSD95)was predicted as a linker node that demonstrated commoninterconnections with nodes of the GDP dissociation inhibitor(GDI2), and the proteolipid protein PLP2. In addition, DLG4demonstrated direct connections with the predicted signalinglinker node HRAS that was further connected to the cell cycleprotein CDC37. The prediction of the synapse-associatedprotein linkers DLG1 in the up regulation cluster and DLG4in the down regulation cluster is in line with their role inlocalization and function of glutamate receptors and potas-sium channels (Fujita and Kurachi, 2000). A regulatory role forsynapse-associated proteins in response to sarin is further

supported by previous findings that demonstrated an upregulation of PSD-95 following 2 h of low dose sarin (Damo-daran et al., 2006a). In addition, APP was predicted as a linkernode in the MS down regulation cluster that demonstrateddirect interconnections with HSP90AB1, and the predictedlinker node NFκB that was further interconnected with theeffector proteins S100B. APP also demonstrated a commoninterconnection to DLG4. The inclusion of the APP linker nodeis supported by literature that proposes the use of AChEinhibitors (such as tacrine and its derivatives) besides itscurrent use in Alzheimer therapy (Fernandez-Bachiller et al.,2012) for inhibition of APP-dependent neurodegeneration inOP toxicity interference therapy (Lorke et al., 2014).

Interestingly, the Reactome FI program predicted a linkerbetween DLG4 and the GDP dissociation inhibitor GDI2 in theMS downregulation cluster. When looking at DLG-, EGFR-,and KL-signaling, MAPK was predicted as a common linkernode that plays a major role in OP-dependent neuronal death(Ki et al., 2013). Directly targeting DLG to inhibit excessiveglutamate receptor assembly and ion channel signaling(Godreau et al., 2004), or upregulation of GDI2 or related GDIsfor interference with glutamate-related vesicle transport(Huang et al., 2004), may help in reduction of OP-inducedglutamate excitotoxicity, prevention of cellular calcium influxand MAPK-dependent cell death signaling.

In summary, the MS-identified proteomic profile for acutehigh dose sarin exposure revealed a broad spectrum of proteinsinvolved in glutamate-mediated excitotoxicity, calcium over-load, energy depletion responses and compensatory carbohy-drate metabolism, increases in ROS defense, DNA damage andchromatin remodeling, HSP response, targeted protein degrada-tion (ubiquitination) and cell death response by direct GObiological process categorization. With regards to the sarin-dependent effect on respiration, our study supports the poten-tial interference of sarin with CO2/H

þ sensitive chemoreceptorneurons of the brainstem retrotrapezoid nucleus (RTN) thatsend excitatory glutamergic projections to the respiratorycenters. Besides being sensitive to activation by CO2/H

þ, theseneurons have recently been demonstrated to be sensitive tomodulation by astrocytes gap junction hemi-channel releasedATP and Ca2þ (Sobrinho et al., 2014). While no changes in gapjunction connexin structural composition was observed, ourstudy supports a potential sarin-dependent Ca2þ overload atgap junctions through upregulation of gap junction associatedCa2þ binding protein calretinin and calmodulin. In addition, theupregulation of the astrocyte glutamate transporter EAAT1 anda downregulation of the Kþ/Cl- co-transporter (KCC2) mightindicate a sarin-dependent interference with direct synaptictransmission (Dallas et al., 2007; Uvarov et al., 2007). Also, theobserved upregulation in cellular energy metabolic proteinssuch as ADP/ATP translocase, mitochondrial aconitase, cyto-chromes, and ATPaseα1 indicate a potential change in ATPlevels that might have an impact on ATP-dependent inductionof neurotransmission from the RTN chemoreceptor neurons.While quantitative MS/MS is still limited in detecting everymodulated cellular protein, the Reactome FI network enrich-ment analysis is able to construct networks by linking knowl-edge database-derived protein nodes with actually MS-detectedprotein nodes for prediction of biological pathway interconnec-tions that play a role during the brainstem proteomic responses

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3 109

following high dose sarin exposure. For instance, the identifiedlinker proteins KL, and synapse-associated proteins DLG1 andDLG4 might be used as future therapeutic targets to modulatelong-term pathological changes in NMDA and AMPA glutamatereceptor expression that might interfere with respiration-related synaptic transmission (Dubal et al., 2015; Jourdi andKabbaj, 2013).

In conclusion, the gained information will help in betterunderstanding participation of this region in the dysregula-tion of respiration for the development of novel counter-measures against fatal sarin poisoning.

4. Experimental procedures

4.1. Animal study

Protocols of animal experiments were reviewed and approvedby both the Battelle Institutional Animal Care and UseCommittee (IACUC) and the USAMRMC Animal Use andReview Office. The study used a total of eight female guineapigs (n¼8) that were divided into two groups: (A) naïvecontrol group (n¼4) that did not receive any treatment orvehicle, and (B) animals exposed to sarin in 0.9% saline bysubcutaneous injection (n¼4). The sarin challenge was2�LD50¼80 μg/kg. Guinea pigs were continuously monitored,and clinical signs were recorded at these times after chal-lenge: between 0 to 5 min, and at 15 and 30 min followingexposure. For measurement of blood AChE activity, one mlsamples of blood were taken via catheter before sarin injec-tion, 45715 s after the challenge dose (interim) and immedi-ately after individual time of death from trunk blood followedby storage in ethylene-diamine-tetra-acetic acid (EDTA) at�20 1C. Upon death, which occurred in three exposed ani-mals at approximately 15 min and 1 exposed animal atapproximately 25 min, brains were immediately dissectedand halved sagittally, by evenly dividing brain halves forprocessing by (a) proteomics analysis or (b) total brain AChEdetermination. For proteomics analysis the brainstem wasdissected followed by flash freezing in liquid nitrogen andstorage at �80 1C. For determination of total brain AChEactivity brain halves were placed in 0.1 M sodium phosphatebuffer (pH 7.4) on ice, followed by weight determination andmincing with scissors in ice cold buffer, visible blood wasdrained away, a 5 mL aliquot of cold homogenization bufferwas added per gram of brain tissue, and the tissue washomogenized and stored at �80 1C.

4.2. AChE activity assay

Blood and brain AChE activity were determined in 96 wellplates using a modified Ellman assay on a Hitachi 912 clinicalanalyzer (Ellman 1961). Iso-OMPA (tetra-isopropyl- pyropho-sphoramide) was used at 4 mM to inhibit butyrylcholinester-ase (BChE). The Hitachi 912 automatically combined 300 μL of0.25 mM dithionitrobenzoic acid (DTNB) in a 50 mM phos-phate buffer (pH 7.2), 50 μL of a 7.1 mM acetylthiocholine(ATC), and 6μL sample in each assay under optimized linearhydrolysis conditions. These conditions gave a final reactionmixture of 42.1 mM phosphate, 0.21 mM dithionitrobenzoic

acid (DTNB), 1.0 mM ATC and sample. Measured kinetic AChEactivity is expressed as μmol min�1 mL�1, or internationalunits per milliliter (U/ml) for blood samples and units permilligram (U/mg) for brain tissue samples. The change inabsorbance was calculated by subtracting the backgroundmeasurement (600 nm) from the primary nitrobenzoic acidabsorption wavelength (480 nm). This bichromatic approachis used to correct for sample colorimetric interferences. Thefinal AChE activity is calculated by multiplying the Hitachivalue with dilution factor. At least four samples per groupwere analyzed. Statistical analysis was performed using one-way ANOVA (blood) or student's t-test (brain), and po0.05was used as a criterion for statistical significance.

4.3. Protein isolation for mass spectrometry andimmunoblotting

For generation of MS tissue lysates, 12.5 mg of brainstemwere minced and placed in 200 ml of lysis buffer (8 M urea,50 mM triethyl ammonium bicarbonate (TEAB), and 75 mMNaCl), and sonicated at high power for 5 s to lyse the tissue.For generation of lysates for immunoblotting, 25 mg ofbrainstem were weighed and placed in a pre-chilled Douncehomogenizer with 2 ml of isolation buffer (150 mM NaCl,50 mM Tris, 1% NP-40). Homogenates were transferred to atube and centrifuged at 1000� g for 10 min at 4 1C and thesupernatants were stored at �80 1C. To prepare solubleprotein cytosolic fractions for immunoblotting, 25 mg ofbrainstem was processed according to the Subcellular Frac-tionation kit from Pierce (Thermo Scientific, Rockford, IL).

4.4. Mass spectrometry analysis

MS protein samples were further processed by addition of5 mM tris (2-carboxylethyl) phosphine hydrochloride (TCEP)for 60 min and 15 mM iodoacetamide (IAA) for 30 min in thedark for reduction and alkylation of cysteines. 3 ml of eachsample were taken to perform protein concentration usingthe Bradford assay from Bio-Rad (Hercules, CA). Approxi-mately 125 to 150 mg of total protein was diluted to achievea 2 M urea concentration and digested overnight at 37 1C with4 mg of trypsin, followed by quenching with 1% formic acid.Samples were desalted using a C18 peptide trap fromMichrom (Auburn, CA). The desalted samples were vacu-fuged, and placed in 50 mM TEAB for iTRAQ 4-plex labelingaccording to the manufacturers’ instructions (Applied Biosys-tems, Life Technologies, Grand Island, NY). To allow forcross-sample comparison 10 ml from each sample werepooled together to generate an internal standard. Sampleswere labeled in the following mass/charge isobaric tagscheme: experimental (115), control (116), and internal stan-dard (117). The sample supplementation with equal amountsof an internal standard derived from portions of each sampleallows for cross sample comparison by facilitation of max-imum quantitation across all proteins. Based upon theBradford assay results, equal amounts of the 115 and 116labeled samples were pooled and diluted to 100 ml with watercontaining 0.1% formic acid, and placed in the auto-samplerfor injection. Each sample was analyzed in three LC–MS/MStechnical replicates. 1.5 ml of the peptide mixture was

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3110

separated on a Waters BEH C18 capillary column prior toonline analysis using a 240-min increasing gradient of acet-onitrile with 0.1% formic acid. Following elution from thecolumn, ions were generated using 1.4 kV on a coatedcapillary tip in a New Objective source and entered into theLTQ-Orbitrap Velos (Thermo Fisher, San Jose, CA). In the MS,a full scan was taken in the LTQ, followed by data-dependentMS/MS analysis of the top 8 peaks. MS/MS analysis includedcollision-induced dissociation (CID) in the LTQ for structuralinformation and higher-energy collisional dissociation (HCD)in the Orbitrap at 30,000 resolution for quantitation. Toenable low abundance detection, peptide analysis was exe-cuted under dynamic exclusion. The generated.RAW files

Fig. 4 – Reactome FI network analysis of sarin exposure. Cellcompartment diagram of Cytoscape 3.0.2 Reactome FIbiological network prediction. Round nodes represent MSoriginally identified proteins, and squares representpredicted linker nodes that are not part of the original input.Red lining represent upregulated nodes and black liningrepresent downregulated protein nodes. Nodes without redor black lining represent proteins that lacked changes inexpression. The built in Reactome FI spectral partitioning-based network clustering algorithm (Newman, 2006)assigned “-4“ for activating/catalyzing, “-|“ for inhibition,“-“ for FIs extracted from complexes or inputs, and “—“ forpredicted FIs. According to “GO Biological Process” networkcluster classification green nodes represent cellularmaintenance and energy metabolism; blue nodes representsynaptic transmission and cell stress recognition/signaling,purple nodes represent cytoskeletal rearrangement,transport, and axon guidance; brown nodes representcellular energy rescue response; blue-green nodes representcalcium response, regulation of cell cycle, gene expressionand chromatin remodeling. The most likely cellularcompartment localization of nodes predicted by “GO CellComponent” network function analysis is indicated.

were searched together for each sample using a mammalianprotein database in Proteome Discoverer 1.3 (Thermo Fisher,San Jose, CA) combined with a decoy database consisting ofthe reverse sequences of the proteins in the database for falsepositive analysis using the SEQUEST algorithm. A masstolerance of 0.4 Da was used for the parent scan and 0.8 Dawas used for the CID spectra, while 25 ppm was used for theHCD spectra. A sole guinea pig database was not used due tothe only a small percentage of the guinea pig proteome beingcharacterized in SwissProt. Variable modifications forN-terminal and lysine side chain iTRAQ labeling, methionineoxidation and cysteine carboxymethylation were included inthe database search. Proteins were considered identified iftwo or more separate peptides were found with po.05according to the peptide validator node in Proteome Disco-verer 1.3., and iTRAQ ratios were calculated by separatenormalization against the control and the internal standardbased upon the protein median using the Quantitation nodein Proteome Discoverer. All errors are given as the standarderror of the mean (SEM). Isotopic correction factors of 0.59%(�1) and 4.64% (þ1) were used for 115, 1.38% (�1) and 2.78%(þ1) were used with 116, and 2.13% (�1) and 2.69% (þ1) wereused with 117 iTRAQ reagents. Following protein identifica-tion and quantitation protein GO terms were identifiedthrough UniProt Knowledge Base. In addition to the SwissProt accession numbers, guinea pig accession numbers weredetermined through NCBI BLAST analysis with a necessary90% reference protein homology as matching criteria.

4.5. Western blot analysis

Protein extracts (25 μg total protein) were separated by SDS-polyacrylamide gel electrophoresis on 10% SDS gels (BioRad,Hercules, CA) followed by Western blotting of selected pro-teins showing abundance change from the LC/MS-MS. Proteinblots on nitrocellulose or PVDF membranes were incubatedseparately with the following primary antibodies to identifythe following proteins: rabbit polyclonal CaMKIIα (bs5046R,Bioss, Woburn, MA) at 1:500; rat monoclonal myelin basicprotein (MBP, NB600-717, Novus Biological, Littleton, CO) at1:1000); rabbit polyclonal galectin-1 (PA5-30706, ThermoScientific, Rockford, IL) at 1:1000; rabbit polyclonal GLUT-1/GLAST-1/EAAT-1 (LS-A9447, LifeSpan BioSciences, Seattle,WA) at 1:1000; mouse monoclonal sodium potassiumATPaseα1 (sc-21712) at 1:500; goat polyclonal EAAT-2 (N-19)(sc7760) at 1:250; goat polyclonal EAAT-3 (C-20) (sc-7761) at1:250; and goat polyclonal VGLUT-2 (N-12)/SLC17A6 (sc-26026) at 1:250 from Santa Cruz Biotechnology Santa Cruz,CA. β-actin (CST 4967L, Danvers, MA) was used as a loadingcontrol at 1:1000.

The corresponding secondary antibodies coupled to alka-line phosphatase (Promega, Madison, WI) were used at 1:5000dilutions. Western Blue Stabilized Substrate for AlkalinePhosphatase (Promega, Madison, WI) was used for colori-metric detection of antibody-specific bands. Stained mem-branes were scanned with an Epson Perfection V500 PhotoScanner followed by densitometry using NIH ImageJ 1.47v.Densitometry included measurement of same-size area meanintensities from antibody bands of interest divided by thecorresponding bands for β-actin for normalization. To

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3 111

normalize changes in expression, the average protein densityof three control samples was set to one-fold. Three samplesper group were analyzed with errors listed as SEM. Statisticalanalysis was performed using student's t-test, and po0.05was used as a criterion for statistical significance (Fig. 3).

4.6. Reactome FI network analysis

UniProt/SwissProt accession numbers were converted intoGeneWiki names via the BioMart portal webpage (http://cen-tral.biomart.org/converter/#!/ID_converter/gene_ensembl_con-fig_1). GeneWiki names were entered into the Cytoscape 3.0.2.program through the Reactome FI3.0.1beta application (Wuet al., 2010) under the 2012 algorithm for “gene sets” using“Fetch FI annotations” and “use linker genes” for networkprediction. Following network prediction, spectral partition-based network clustering was used, and network modules wereclassified by screening for “GO Biological Process.” The networkfunctions were further analyzed by the “GO Cell Component”followed by manual dragging of nodes to the most likelylocalization within a self-designed cell compartment diagram.For comparison of protein pathway node distribution withinnetwork clusters, either 1) all of the MS-identified and con-verted GeneWiki names were entered into the Reactome FInetwork analysis, or 2) GeneWiki names in the MS upregula-tion cluster (using cutoff at or above 1.2-fold expression), or 3)GeneWiki names in the MS downregulation cluster (usingvalues equal or below 0.83-fold expression), were queried byconsidering the number of detected proteins and the totalexperimental expression range from 0.30 to 3.16-fold (Fig. 4).

Supplemental data

The supplementary material includes an excel spreadsheetof the raw MS data and information needed for the Cytos-cape Reactome FI application which was used to generate s2 and 4.

Funding information

This work was supported by DoD Contract no. SP0-700-00-D-3180 from the Defense Threat Reduction Agency, Fort Belvoir,VA 22060-6201.

Acknowledgments

The authors wish to recognize the statistical analysis assis-tance of Nancy A. Niemuth and the scientific support staff atBattelle Biomedical Research Center.

Appendix A. Supplementary materials

Supplementary data associated with this article can be foundin the online version at http://dx.doi.org/10.1016/j.brainres.2015.03.041.

r e f e r e n c e s

Akhgari, M., Abdollahi, M., Kebryaeezadeh, A., Hosseini, R.,

Sabzevari, O., 2003. Biochemical evidence for free radical-induced lipid peroxidation as a mechanism for subchronic

toxicity of malathion in blood and liver of rats. Hum. Exp.

Toxicol. 22 (4), 205–211.Bajgar, J., 2005. Complex view on poisoning with nerve agents and

organophosphates. Acta Med. (Hradec Kralove) 48 (1), 3–21.Bhardwaj, S., Musalgaonkar, N., Waghmare, C., Bhattacharya, B.K.,

2012. Single dose exposure of sarin and physostigmine

differentially regulates expression of choline acetyltransferase

and vesicular acetylcholine transporter in rat brain. Chem. Biol.

Interact. 198 (1–3), 57–64, http://dx.doi.org/10.1016/j.cbi.2012.05.002.

Bloch-Shilderman, E., Kadar, T., Levy, A., Sahar, R., Rabinovitz, I.,

Gilat, E., 2005. Subcellular alterations of protein kinase C

isozymes in the rat brain after organophosphate poisoning.

J. Pharmacol. Exp. Ther. 313 (3), 1082–1089, http://dx.doi.org/

10.1124/jpet.105.083469.Carey, J.L., Dunn, C., Gaspari, R.J., 2013. Central respiratory failure

during acute organophosphate poisoning. Respir. Physiol.

Neurobiol. 189 (2), 403–410, http://dx.doi.org/10.1016/j.

resp.2013.07.022.Chan, J.Y., Chan, S.H., Dai, K.Y., Cheng, H.L., Chou, J.L., Chang, A.Y.,

2006. Cholinergic-receptor-independent dysfunction of

mitochondrial respiratory chain enzymes, reduced

mitochondrial transmembrane potential and ATP depletion

underlie necrotic cell death induced by the organophosphate

poison mevinphos. Neuropharmacology 51 (7–8), 1109–1119,http://dx.doi.org/10.1016/j.neuropharm.2006.06.024.

Chan, J.Y.H., Cheng, H.-L., Chou, J.L.J., Li, F.C.H., Dai, K.-Y., Chan, S.

H.H., Chang, A.Y.W., 2007. Heat shock protein 60 or 70

activates nitric-oxide synthase (NOS) I- and inhibits NOS II-

associated signaling and depresses the mitochondrialapoptotic cascade during brain stem death. J. Biol. Chem. 282,

4585–4600, http://dx.doi.org/10.1074/jbc.M603394200.Chen, C.D., Sloane, J.A., Li, H., Aytan, N., Giannaris, E.L., Zeldich, E.,

Hinman, J.D., Dedeoglu, A., Rosene, D.L., Bansal, R., Luebke, J.I.,

Kuro-o, M., Abraham, C.R., 2013. The antiaging protein Klothoenhances oligodendrocyte maturation and myelination of the

CNS. J. Neurosci. 33 (5), 1927–1939, http://dx.doi.org/10.1523/

JNEUROSCI.2080-12.2013.Choi, D.W., 1988. Calcium-mediated neurotoxicity: relationship to

specific channel types and role in ischemic damage. TrendsNeurosci. 11 (10), 465–469.

Choudhary, S., Joshi, K., Gill, K.D., 2001. Possible role of enhanced

microtubule phosphorylation in dichlorvos induced delayed

neurotoxicity in rat. Brain Res. 897 (1-2), 60–70.Dallas, M., Boycott, H.E., Atkinson, L., Miller, A., Boyle, J.P., Pearson,

H.A., Peers, C., 2007. Hypoxia suppresses glutamate transport in

astrocytes. J. Neurosci. 27 (15), 3946–3955, http://dx.doi.org/

10.1523/JNEUROSCI.5030-06.2007.Damodaran, T.V., 2009. Molecular and Transcriptional Responses

to Sarin Exposure. Handbook of Toxicology and ChemicalWarfare Agents.

Damodaran, T.V., Patel, A.G., Greenfield, S.T., Dressman, H.K., Lin,

S.M., Abou-Donia, M.B., 2006a. Gene expression profiles of the

rat brain both immediately and 3 months following acute

sarin exposure. Bioche. Pharmacol. 71, 497–520, http://dx.doi.org/10.1016/j.bcp.2005.10.051.

Damodaran, T.V., Greenfield, S.T., Patel, A.G., Dressman, H.K., Lin,

S.K., Abou-Donia, M.B., 2006b. Toxicogenomic studies of the

rat brain at an early time point following acute sarin exposure.

Neurochem. Res. 31 (3), 367–381, http://dx.doi.org/10.1007/

s11064-005-9023-5.

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3112

Damodaran, T.V., Bilska, M.A., Rahman, A.A., Abou-Doni, M.B.,2002. Sarin causes early differential alteration and persistentoverexpression in mRNAs coding for glial fibrillary acidicprotein (GFAP) and vimentin genes in the central nervoussystem of rats. Neurochem. Res. 27, 407–415.

Dubal, D.B., Zhu, L., Sanchez, P.E., Worden, K., Broestl, L., Johnson,E., Ho, K., Yu, G.Q., Kim, D., Betourne, A., Kuro, O.M., Masliah,E., Abraham, C.R., Mucke, L., 2015. Life extension factor klothoprevents mortality and enhances cognition in hAPPtransgenic mice. J. Neurosci. 35 (6), 2358–2371, http://dx.doi.org/10.1523/JNEUROSCI.5791-12.2015.

Eason, M.P., 2013. Sarin exposure: a simulation case scenario.South Med. J. 106 (1), 55–62, http://dx.doi.org/10.1097/SMJ.0b013e31827cd12d.

el-Sharkawy, A.M., Abdel-Rahman, S.Z., Hassan, A.A., Gabr, M.H.,el-Zoghby, S.M., el-Sewedy, S.M., 1994. Biochemical effects ofsome insecticides on the metabolic enzymes regulatingglutathione metabolism. Bull. Environ. Contam. Toxicol. 52 (4),505–510.

Endo, T., 2005. Glycans and glycan-binding proteins in brain:galectin-1-induced expression of neurotrophic factors inastrocytes. Curr. Drug Targets 6 (4), 427–436.

Fernandez-Bachiller, M.I., Perez, C., Monjas, L., Rademann, J.,Rodriguez-Franco, M.I., 2012. New tacrine-4-oxo-4H-chromenehybrids as multifunctional agents for the treatment ofAlzheimer’s disease, with cholinergic, antioxidant, and beta-amyloid-reducing properties. J. Med. Chem. 55 (3), 1303–1317,http://dx.doi.org/10.1021/jm201460y.

Fujita, A., Kurachi, Y., 2000. SAP family proteins. Biochem.Biophys. Res. Commun. 269 (1), 1–6, http://dx.doi.org/10.1006/bbrc.1999.1893.

Gearhart, D.A., Sickles, D.W., Buccafusco, J.J., Prendergast, M.A.,Terry Jr., A.V., 2007. Chlorpyrifos, chlorpyrifos-oxon, anddiisopropylfluorophosphate inhibit kinesin-dependentmicrotubule motility. Toxicol. Appl. Pharmacol. 218 (1), 20–29,http://dx.doi.org/10.1016/j.taap.2006.10.008.

Gilley, J.A., Kernie, S.G., 2011. Excitatory amino acid transporter 2and excitatory amino acid transporter 1 negatively regulatecalcium-dependent proliferation of hippocampal neuralprogenitor cells and are persistently upregulated after injury.Eur. J. Neurosci. 34 (11), 1712–1723, http://dx.doi.org/10.1111/j.1460-9568.2011.07888.x.

Godreau, D., Neyroud, N., Vranckx, R., Hatem, S., 2004. MAGUKs:beyond ionic channel anchoring First published on LesMAGUK: au-dela de l’accrochage des canaux ioniques. Med.Sci. (Paris) 20 (1), 84–88, http://dx.doi.org/10.1051/medsci/200420184.

Good, B.M., Howe, D.G., Lin, S.M., Kibbe, W.A., Su, A.I., 2011.Mining the Gene Wiki for functional genomic knowledge. BMCGenomics 12 (603)http://dx.doi.org/10.1186/1471-2164-12-603.

Gupta, R.C., Milatovic, D., Dettbarn, W.D., 2001. Depletion ofenergy metabolites following acetylcholinesterase inhibitor-induced status epilepticus: protection by antioxidants.Neurotoxicology 22 (2), 271–282.

Gupta, R.C., Goad, J.T., Milatovic, D., Dettbarn, W.D., 2000. Cholinergicand noncholinergic brain biomarkers of insecticide exposure andeffects. Hum. Exp. Toxicol. 19 (5), 297–308.

Horie, H., Kadoya, T., Sango, K., Hasegawa, M., 2005. Oxidizedgalectin-1 is an essential factor for peripheral nerveregeneration. Curr. Drug Targets 6 (4), 385–394.

Howard, A.S., Bucelli, R., Jett, D.A., Bruun, D., Yang, D., Lein, P.J., 2005.Chlorpyrifos exerts opposing effects on axonal and dendriticgrowth in primary neuronal cultures. Toxicol. Appl. Pharmacol.207 (2), 112–124, http://dx.doi.org/10.1016/j.taap.2004.12.008.

Huang, C.C., You, J.L., Wu, M.Y., Hsu, K.S., 2004. Rap1-induced p38mitogen-activated protein kinase activation facilitates AMPAreceptor trafficking via the GDI.Rab5 complex. Potential role in(S)-3,5-dihydroxyphenylglycene-induced long term

depression. J. Biol. Chem. 279 (13), 12286–12292, http://dx.doi.

org/10.1074/jbc.M312868200.Hulet, S.W., Sommerville, D.R., Miller, D.B., Scotto, J.A., Muse, W.T.,

Burnett, D.C., 2014. Comparison of sarin and cyclosarin

toxicity by subcutaneous, intravenous and inhalation

exposure in Gottingen minipigs. Inhal. Toxicol. 26 (3), 175–184.Jourdi, H., Kabbaj, M., 2013. Acute BDNF treatment upregulates

GluR1-SAP97 and GluR2–GRIP1 interactions: implications for

sustained AMPA receptor expression. PLoS One 8 (2), e57124,

http://dx.doi.org/10.1371/journal.pone.0057124.Khan,W.A., Dechkovskaia, A.M., Herrick, E.A., Jones, K.H., Abou-Donia,

M.B., 2000. Acute sarin exposure causes differential regulation of

choline acetyltransferase, acetylcholinesterase, and acetylcholine

receptors in the central nervous system of the rat. Toxicol. Sci. 57

(1), 112–120.Ki, Y.W., Park, J.H., Lee, J.E., Shin, I.C., Koh, H.C., 2013. JNK and p38

MAPK regulate oxidative stress and the inflammatory

response in chlorpyrifos-induced apoptosis. Toxicol. Lett. 218

(3), 235–245, http://dx.doi.org/10.1016/j.toxlet.2013.02.003.Kim, Y.B., Hur, G.H., Shin, S., Sok, D.E., Kang, J.K., Lee, Y.S., 1999.

Organophosphate-induced brain injuries: delayed apoptosis

mediated by nitric oxide. Environ. Toxicol. Pharmacol. 7,

147–152, http://dx.doi.org/10.1016/S1382-6689(99)00006-X.Klein-Rodewald, T., Seeger, T., Dutschmann, M., Worek, F.,

Morschel, M., 2011. Central respiratory effects on motor nerve

activities after organophosphate exposure in a working heart

brainstem preparation of the rat. Toxicol. Lett. 206 (1), 94–99,

http://dx.doi.org/10.1016/j.toxlet.2011.07.005.Lallement, G., Carpentier, P., Collet, A., Pernot-Marino, I.,

Baubichon, D., Blanchet, G., 1991. Effects of soman-induced

seizures on different extracellular amino acid levels and on

glutamate uptake in rat hippocampus. Brain Res. 563 (1–2),

234–240.Li, Y., Lein, P.J., Liu, C., Bruun, D.A., Giulivi, C., Ford, G.D., Tewolde, T.,

Ross-Inta, C., Ford, B.D., 2012. Neuregulin-1 is neuroprotective in a

rat model of organophosphate-induced delayed neuronal injury.

Toxicol. Appl. Pharmacol. 262 (2), 194–204, http://dx.doi.org/10.1016/

j.taap.2012.05.001.Lorke, D.E., Nurulain, S.M., Hasan, M.Y., Kuca, K., Petroianu, G.A.,

2014. Prophylactic administration of non-organophosphate

cholinesterase inhibitors before acute exposure to

organophosphates: assessment using terbufos sulfone. J.

Appl. Toxicol. 34 (10), 1096–1103, http://dx.doi.org/10.1002/

jat.2939.Martin, L.J., Doebler, J.A., Wall, T.J., Shih, T.M., Anthony, A., 1986.

Brain neuronal chromatin responses in acute soman

intoxicated rats. Neurochem. Res. 11, 1203–1215, http://dx.doi.

org/10.1007/BF00965948.Mehta, A., Verma, R.S., Srivastava, N., 2008. Chlorpyrifos-induced

DNA damage in rat liver and brain. Environ. Mol. Mutagen. 49,

426–433, http://dx.doi.org/10.1002/em.Newman, M.E., 2006. Modularity and community structure in

networks. Proc. Natl. Acad. Sci. USA 103 (23), 8577–8582, http://dx.

doi.org/10.1073/pnas.0601602103.O’Donnell, J.C., McDonough, J.H., Shih, T.M., 2011. In vivo

microdialysis and electroencephalographic activity in freely

moving guinea pigs exposed to organophosphorus nerve

agents sarin and VX: analysis of acetylcholine and glutamate.

Arch. Toxicol. 85 (12), 1607, http://dx.doi.org/10.1007/s00204-

011-0724-z.Milacic, M., Haw, R., Rothfels, K., Wu, G., Croft, D., Hermjkaob,H.,

D’Eustachio, P., Stein, L., 2012. Annotating cancer variants and

anti-cancer therapeutics in reactome. Cancers (Basal) 8; 4(4),

1180-1211 10.3990/cancers4041180.Ojha, A., Yaduvanshi, S.K., Pant, S.C., Lomash, V., Srivastava, N., 2013.

2013. Evaluation of DNA damage and cytotoxicity induced by

three commonly used organophosphate pesticides individually

b r a i n r e s e a r c h 1 6 1 1 ( 2 0 1 5 ) 1 0 1 – 1 1 3 113

and in mixture, in rat tissues. Environ. Toxicol. 28 (10), 543–552,http://dx.doi.org/10.1002/tox.20748.

Pala, I., Vig, P.J., Desaiah, D., Srinivasan, A., 1991. In vitro effects oforganophosphorus compounds on calmodulin activity. J. Appl.Toxicol. 11 (6), 391–395.

Ralphe, J.C., Segar, J.L., Schutte, B.C., Scholz, T.D., 2004.Localization and function of the brain excitatory amino acidtransporter type 1 in cardiac mitochondria. J. Mol. Cell Cardiol.37 (1), 33–41, http://dx.doi.org/10.1016/j.yjmcc.2004.04.008.

Rickett, D.L., Glenn, J.F., Beers, E.T., 1986. Central respiratoryeffects versus neuromuscular actions of nerve agents.Neurotoxicology Spring 7 (1), 225–236.

Schafer, M., Brauer, A.U., Savaskan, N.E., Rathjen, F.G.,Brummendorf, T., 2005. Neurotractin/kilon promotes neuriteoutgrowth and is expressed on reactive astrocytes afterentorhinal cortex lesion. Mol. Cell Neurosci. 29 (4), 580–590,http://dx.doi.org/10.1016/j.mcn.2005.04.010.

Schlattner, U., Tokarska-Schlattner, M., Ramirez, S., Bruckner, A.,Kay, L., Polge, C., Epand, R.F., Lee, R.M., Lacombe, M.L., Epand,R.M., 2009. Mitochondrial kinases and their molecularinteraction with cardiolipin. Biochim. Biophys. Acta 1788 (10),2032–2047, http://dx.doi.org/10.1016/j.bbamem.2009.04.018.

Sharma, Y., Bashir, S., Irshad, M., Gupta, S.D., Dogra, T.D., 2005.Effects of acute dimethoate administration on antioxidantstatus of liver and brain of experimental rats. Toxicology 206(1), 49–57, http://dx.doi.org/10.1016/j.tox.2004.06.062.

Sheldon, A.L., Robinson, M.B., 2007. The role of glutamatetransporters in neurodegenerative diseases and potentialopportunities for intervention. Neurochem. Int. 51 (6-7),333–355, http://dx.doi.org/10.1016/j.neuint.2007.03.012.

Sobrinho, C.R., Wenker, I.C., Poss, E.M., Takakura, A.C., Moreira, T.S.,Mulkey, D.K., 2014. Purinergic signalling contributes tochemoreception in the retrotrapezoid nucleus but not thenucleus of the solitary tract or medullary raphe. J. Physiol. 592 (6),1309–1323, http://dx.doi.org/10.1113/jphysiol.2013.268490 Pt.

Solberg, Y., Belkin, M., 1997. The role of excitotoxicity inorganophosphorous nerve agents central poisoning. TrendsPharmacol. Sci. 18, 183–185.

Terry Jr., A.V., Gearhart, D.A., Beck Jr., W.D., Truan, J.N.,Middlemore, M.L., Williamson, L.N., Bartlett, M.G.,Prendergast, M.A., Sickles, D.W., Buccafusco, J.J., 2007. Chronic,intermittent exposure to chlorpyrifos in rats: protractedeffects on axonal transport, neurotrophin receptors,

cholinergic markers, and information processing.J. Pharmacol. Exp. Ther. 322 (3), 1117–1128, http://dx.doi.org/10.1124/jpet.107.125625.

Torres-Altoro, M.I., Mathur, B.N., Drerup, J.M., Thomas, R.,Lovinger, D.M., O’Callaghan, J.P., Bibb, J.A., 2011.Organophosphates dysregulate dopamine signaling,glutamatergic neurotransmission, and induce neuronal injurymarkers in striatum. J. Neurochem. 119 (2), 303–313, http://dx.doi.org/10.1111/j.1471-4159.2011.07428.x.

Tsuchida, S., Satoh, M., Kawashima, Y., Sogawa, K., Kado, S., Sawai,S., Nishimura, M., Ogita, M., Takeuchi, Y., Kobyashi, H., Aoki, A.,Kodera, Y., Matsushita, K., Izumi, Y., Nomura, F., 2013. Applicationof quantitative proteomic analysis using tandem mass tags fordiscovery and identification of novel biomarkers in periodontaldisease. Proteomics 13 (15), 2339–2350, http://dx.doi.org/10.1002/pmic.201200510.

Uvarov, P., Ludwig, A., Markkanen, M., Pruunsild, P., Kaila, K., Delpire,E., Timmusk, T., Rivera, C., Airaksinen, M.S., 2007. A novel N-terminal isoform of the neuron-specific K-Cl cotransporter KCC2.J. Biol. Chem. 282 (42), 30570–30576, http://dx.doi.org/10.1074/jbc.M705095200.

Wu, G., Feng, X., Stein, L., 2010. A human functional proteininteraction network and its application to cancer dataanalysis. Genome Biol. 11 (5), R53, http://dx.doi.org/10.1186/gb-2010-11-5-r53.

Yanagisawa, N., Morita, H., Nakajima, T., 2006. Sarin experiencesin Japan: acute toxicity and long-term effects. J. Neurol. Sci.249 (1), 76–85, http://dx.doi.org/10.1016/j.jns.2006.06.007.

Yi, J.H., Hazell, A.S., 2006. Excitotoxic mechanisms and the role ofastrocytic glutamate transporters in traumatic brain injury.Neurochem. Int. 48 (5), 394–403, http://dx.doi.org/10.1016/j.neuint.2005.12.001.

Yokoyama, K., Yamada, A., Mimura, N., 1996. Clinical profiles ofpatients with sarin poisoning after the Tokyo subway attack.Am. J. Med. 100 (5), 586.

Young, J.F., Gough, B.J., Suber, R.L., Gaylor, D.W., 2001. Correlationof blood cholinesterase levels with toxicity of sarin in rats.J. Toxicol. Environ. Health Part A 62 (3), 161–174, http://dx.doi.org/10.1080/009841001458280.

Ziegler, D.R., Innocente, C.E., Leal, R.B., Rodnight, R., Goncalves, C.A.,1998. The S100B protein inhibits phosphorylation of GFAP andvimentin in a cytoskeletal fraction from immature rathippocampus. Neurochem. Res. 23 (10), 1259–1263.