Young male Sprague-Dawley rats, aged 3–4 months, were obtained from Harlan Sprague-Dawley Inc. (Indianapolis, IN, United States) for this study. A combined institutional animal care and use committee of the Central Texas Veterans Health Care System and the Texas A&M Health Science Center College of Medicine approved all the experiments performed in the current study. Upon arrival, animals were housed with ad libitum access to commercial rat chow and water. Following 12–15 days of acclimatization in the vivarium, animals were randomly assigned to the naïve control, vehicle (VEH) control, or GWI groups, which here onward will be referred to as “naïve controls,” “VEH-rats,” and “GWI-rats,” respectively. GWI-rats received daily exposure to GWIR-Cs, namely, PB, DEET, and PM, and mild stress (5 min of restraint stress) for 4 weeks. In contrast, VEH-rats received sterile water (the vehicle solution used to dissolve PB), and 70% alcohol (the vehicle solution used to dissolve PB).

The dosages of PB, DEET, and PM, and duration of stress were based on prior studies of rat models of GWI (Abdel-Rahman et al., 2002, 2004; Parihar et al., 2013; Hattiangady et al., 2014; Megahed et al., 2015). PB (Sigma, St. Louis, MO, United States) was administered via oral gavage, at a dose of 1.3 mg/kg in 500 μl of sterile water. DEET and PM (Chem. Service Inc., West Chester, PA, United States) were dissolved in 70% alcohol, at dosages of 40 and 0.13 mg/kg, respectively, and applied dermally on the back of the animal’s neck. To facilitate direct application on the skin, the fur on the back of the neck and the upper thoracic region were shaved off, the exposed skin was wiped with 70% alcohol, and 200 μl each of the DEET and PM solutions were sequentially spread over the shaved skin using a short plastic pipette. Each animal was subjected to 5 min of restraint stress in a rat restrainer (Stoelting Research Instruments, Wood Dale, IL, United States), as detailed in our previous study (Parihar et al., 2013). Because DEET and PM are toxic drugs, these chemicals were handled and disposed as chemical hazards. Research personnel preparing these chemical solutions avoided skin contact, ingestion and inhalation by preparing them in a fume hood and using personal protective equipment (PPE) such as lab coat, safety goggles, and chemical resistant gloves. The personnel also used PPE while treating animals inside a fume hood.

Each animal (n = 8/group) was deeply anesthetized with an overdose of isoflurane and decapitated, 6 months after the exposure to the GWIR-Cs and restraint stress. Following this, the brain was rapidly removed from the skull, snap frozen in dry ice, and stored at -80°C until further analysis. Following quick thawing, the entire hippocampus from both hemispheres of each brain was micro-dissected and processed for molecular biology studies, which comprised analyses of the expression of genes encoding proteins that are relevant for regulating oxidative stress, antioxidant activity, inflammation, and mitochondrial respiration, using qRT-PCR arrays.

Total RNA was first extracted from hippocampal samples belonging to GWI-rats and VEH-rats using RNeasy kit (Qiagen, Valencia, CA, United States; Cat#: 74104). We employed the manufacturer’s instructions, as previously detailed (Shetty et al., 2013). The concentration and quality of the RNA were determined by a Nanodrop spectrophotometer, using A260 and A260:A280 values, respectively (ThermoFisher Scientific, Wilmington, DE, United States). Total RNA (1 μg) was subsequently transcribed to cDNA using the RT2 First Strand Kit (Qiagen, Valencia, CA, United States; Cat# 330404) as previously detailed (Shetty et al., 2013). Briefly, 10 μl of genomic DNA elimination mixture was prepared using total RNA (2 μg for oxidative stress and 1 μg for inflammatory markers), 2 μl of genomic DNA elimination buffer, and nuclease free water. This mixture was incubated for 5 min at 42°C and chilled on ice. Following this, 10 μl of genomic DNA elimination mixture was gently mixed with 10 μl of RT cocktail, which contained 4 μl of BC3 (5X RT buffer 3), 1 μl of P2 (primer and external control mix), 2 μl of RE3 (RT enzyme mix 3), and nuclease free water. The cDNA synthesis reaction mixture was then incubated at 42°C for 15 min, before the reaction was stopped immediately by heating to 95°C for 5 min. Following the above reaction, 20 μl template cDNA was diluted with 91 μl of dH₂O and stored at -20°C until further analysis.

The Rat Oxidative Stress PCR Array used in this study (SABiosciences, Qiagen, Valencia, CA, United States; Cat#: PARN065Z) detected the expression of 84 key genes related to oxidative stress response, ROS metabolism, oxygen transport, and antioxidant activity in the hippocampus. A previously described (Shetty et al., 2013) qRT-PCR protocol was used. In brief, template cDNA (102 μl) from each hippocampal sample was mixed with 1,350 μl of 2X RT² qPCR master mix (SABiosciences, Qiagen, Valencia, CA, United States) and 1,248 μl of dH₂O to obtain a cocktail of 2,700 μl. This solution was then transferred to wells in a rat oxidative stress RT² Profiler^TM PCR array plate (25 μl/well), which were comprised of pre-dispensed gene specific primer sets. In addition to primers for the 84 genes related to oxidative stress, each array also included primers for detecting five housekeeping genes, namely, beta-actin, beta-2-microglubulin, hypoxanthine phosphoribosyl transferase 1, lactate dehydrogenase A, and ribosomal protein lateral stalk subunit P1, and three RNA and three PCR quality controls. The reactions were performed according to the manufacturer’s protocol using a CFX96 Real-Time system (Bio-Rad, Hercules, CA, United States). The PCR amplification was done using the following two-step cycling program: 10 min denaturizing at 95°C, 40 cycles of 95°C for 15 s, and 60°C for 1 min. At the end of the amplification assay, a melt curve analysis was conducted to evaluate the specificity of the reaction. The conditions for the melt curve were 95°C for 1 min, followed by incremental increases of 0.5°C per 5 min, from 65 to 95°C.

The Rat Mitochondrial Energy Metabolism PCR Array (SABiosciences, Qiagen, Valencia, CA, United States; Cat#PARN008Z) used in this study profiled the expression of 84 key genes in the hippocampus, which encoded proteins that are relevant for mitochondrial respiration, such as genes encoding components of the electron transport chain and oxidative phosphorylation complexes. The reactions were performed as described above. For data analyses, the Ct (threshold cycle) values of all wells were exported to an Excel spreadsheet and analyzed using web-based SABiosciences PCR array data analysis software. The 2^ΔCt values for each gene were compared between VEH-rats and GWI-rats (n = 4–5/group).

The template cDNA (1 μl) from each hippocampal sample was mixed with 12.5 μl of 2X RT² qPCR master mix, 1 μl of RT² primer (10 μM stock, SABiosciences, Qiagen, Valencia, CA, United States), and 10.5 μl of dH₂O to obtain a total volume of 25 μl. Primers relevant to selected genes (n = 35) encoding proteins involved in neurogenesis, learning and memory, and inflammation were used. Genes related to inflammation comprised the following: B-cell lymphoma 6 (Bcl6), complement component 3a receptor 1, chemokine (C-C-motif) ligand 5 (Ccl5, also known as Rantes), Ccl17, Ccl22, colony stimulating factor 2 (Csf2), interferon gamma (Ifng), insulin-like growth factor-binding protein 1 (Igf1), interleukin 1 alpha (IL1a), IL1b, IL2, IL4, IL6, IL10, IL12b, nuclear factor kappa B subunit 1 (Nfkb1), and tumor necrosis factor alpha (Tnfa). Genes that are relevant to neurodegeneration and/or cognitive and memory function included the following: mitogen-activated protein kinase 1 (Mapk1), Mapk3, protein kinase C alpha (Prkaca), glyoxylate reductase 1 homolog (N-pac), NMDA receptor regulated 2 (Narg2), nerve growth factor (Ngf), fibroblast growth factor-2 (Fgfb), vascular endothelial growth factor A (Vegfa), synaptotagmin-5 (Syt5), and forkhead box O3. Each array also contained two of three housekeeping genes glyceraldehyde-3-phosphate dehydrogenase, hypoxanthine phosphoribosyltransferase 1, and ribosomal protein lateral stalk subunit P1. The qRT-PCR reactions and data analyses were carried out as described above; 2^ΔCt values for each gene were compared between VEH-rats and GWI-rats(n = 6–8/group).

Six months after the exposure to GWIR-Cs and stress, GWI-rats and naïve controls (n = 8/group) were subjected to terminal anesthesia with isoflurane and intracardiac perfusion with 4% paraformaldehyde solution in phosphate buffer. Following this, the brain was carefully removed from the skull of each animal and post-fixed in 4% paraformaldehyde for approximately 16 h at 4°C. Next, brain tissues were treated with 30% sucrose solution in phosphate buffer until it sank to the bottom, and 30-μm-thick cryostat sections were cut coronally through the entire septo-temporal axis of the hippocampus. Sections were collected serially in 24-well plates filled with phosphate buffer. Every 20th section through the hippocampus was then processed from randomly chosen GWI-rats and naïve controls (n = 5/group) for Nrf2 and NeuN dual immunofluorescence. In brief, sections were first processed for Nrf2 immunofluorescence, which comprised sequential incubation of sections in 10% donkey serum for 30 min and rabbit anti-Nrf2 solution (1:500; Abcam, Cambridge, MA, United States) overnight. Following a wash in phosphate buffer, the sections were incubated in donkey anti-rabbit AF594 solution (1:200, ThermoFisher Scientific, Waltham, MA, United States) for 1 h and thoroughly washed with phosphate buffer thereafter. Next, sections were incubated in mouse anti-NeuN solution (1:1000; MilliporeSigma, Darmstadt, Germany) overnight, washed in phosphate buffer, and incubated in donkey anti-mouse AF488 solution (1:200; ThermoFisher Scientific) for 1 h. The sections were then thoroughly washed in phosphate buffer, and mounted on slides using SlowFade antifade mounting medium (ThermoFisher Scientific). Thereafter, 2-μm-thick Z-sectioning analysis was performed using confocal microscopy to visualize dense NeuN and speckled Nrf2 expression. Since Nrf2 expression was prominently seen in the CA3 subfield of the hippocampus in both GWI-rats and naïve controls, the percentage of CA3 pyramidal neurons displaying neuronal translocation of Nrf2 was quantified using multiple sections in each animal. In addition, since the number of Nrf2 speckles within nuclei varied in different groups, we also quantified the percentage of neurons expressing robust Nrf2 (having more than two speckles) among those expressing neuronal Nrf2. The values were compared between naïve control rats and GWI-rats (n = 5/group).

Six months after the exposure to GWIR-Cs and stress, additional GWI-rats and naïve controls (n = 8/group) were deeply anesthetized with an overdose of isoflurane. The heart was exposed and blood was quickly collected from the right atrium. Following this, the brain was rapidly removed from the skull, snap frozen in dry ice, and stored at -80°C. The serum was extracted from the collected blood using standard procedures and stored at -80°C. The entire hippocampi from both hemispheres were micro-dissected following quick thawing, and the samples were processed to measure Nrf2 levels through an ELISA. Serum samples were processed for measuring the relative concentration of various pro-inflammatory cytokines and chemokines, and MDA.

Each hippocampal sample was weighed and lysed in 750 μl of ice-cold tissue extraction buffer (Life Technologies, Carlsbad, CA, United States) with protease inhibitors (Sigma–Aldrich Corp., St. Louis, MO, United States), using a sonic dismembrator for 10 s. The lysed solution was centrifuged twice (1,500 ×g at 4°C for 15 min) to avoid contamination with cellular debris. Aliquots of the supernatant solution were stored at -80°C until further use. The protein concentration of different samples was measured using a Pierce BCA reagent kit (ThermoFisher Scientific). We measured Nrf2 concentration in hippocampal lysates using the ELISA method, according to manufacturer instructions (Signosis, Santa Clara, CA, United States; Cat#: TE-0027). Briefly, hippocampal lysates were diluted to 10 μg/100 μl and added to each well of a 96-well plate that was pre-immobilized with the Nrf2 consensus sequencing oligonucleotide, which resulted in binding of activated Nrf2 in the whole cell lysate to the oligonucleotide. Samples were incubated with 50 μl of a specific antibody against an Nrf2 subunit for 1 h at room temperature (RT), washed, incubated with 50 μl of horseradish peroxidase-conjugated secondary antibody for 45 min at RT, washed again, and incubated with 50 μl of substrate solution for 10–30 min or until positive wells turned blue. Thereafter, 25 μl of stop solution was added, which transformed the color of the solution from blue to yellow. The yellow colored reaction product was measured using a Micro plate reader at 450 nm. The amount of Nrf2 in the lysate was directly proportional to the intensity of color. The values were compared between naïve control rats and GWI-rats (n = 6/group).

We used a rapid and sensitive rat cytokine array to measure16 cytokines/chemokines in the serum in a high-throughput manner, as per manufacturer instructions (Signosis; Cat#: EA4006). Briefly, serum samples (1:5 diluted) were dispensed to wells (100 μl per well) of 96-well plates that were coated with specific cytokine capture primary antibodies and incubated for 2 h at RT with gentle shaking. Following three washes with wash buffer, 100 μl of diluted biotin-labeled detection antibody mixture was added to each well and incubated for 1 h at RT with gentle shaking. Then, samples were washed and 100 μl of diluted streptavidin-HRP conjugate was added to each well and incubated for 45 min at RT. After three washes, 100 μl substrate was added to each well and incubated at RT. One hour later, 50 μl of stop solution was added to each well, which transformed the color of the solution from blue to yellow. The yellow reaction product was measured using a Micro plate reader at 450 nm. The concentration of the various cytokines and chemokines in the serum was directly proportional to the intensity of color in this procedure. The values for each cytokine or chemokine were compared between naïve control rats and GWI-rats (n = 6/group).

The level of MDA, which is a by-product of lipid peroxidation (a mechanism of cellular injury), in the serum was detected using an ELISA kit (LSBio, Seattle, WA, United States; Cat #: LS-F28018). Briefly, each well of the 96-well plate was pre-coated with an antigen-specific capture antibody. Diluted serum and known standards were added to each well followed by a fixed quantity of HRP-conjugated target antigen. The free antigen in the sample competed with the HRP conjugated antigen for binding to the capture antibody. Unbound antigen was washed away. A 3,3′,5,5′-tetramethylbenzidine substrate was then added which reacted with the HRP enzyme resulting in color development. Sulfuric acid stop solution was added to terminate the color development reaction. Then, the absorbance of the well was measured at a wavelength of 450 nm using a plate reader. A standard graph was plotted and the MDA concentration in unknown samples was computed from the best-fit curve. The values were expressed as ng/ml and compared between naïve control rats and GWI-rats (n = 6/group).

Prism software was used for all statistical analyses. Unpaired, two-tailed, Student’s t-test was employed for all comparisons between VEH and GWI or Naïve and GWI groups. When standard deviations were statistically significant between the two groups, a non-parametric test (Mann–Whitney test) was employed. Numerical data were presented as mean ± SEM and a p-value less than 0.05 was considered as statistically significant.

The expression of a multitude of genes (n = 27) that typically respond to increased oxidative stress (referred to as oxidative stress-response genes) was significantly increased in the hippocampus, at 6 months after exposure to GWIR-Cs and mild stress (Figures 1A,B1–B27), which implied long-term changes in the expression of these genes. The genes were as follows: alsin rho guanine nucleotide exchange factor 2 (Als2; [B1]), apolipoprotein E (ApoE; [B2]), catalase (Cat; [B3]), cathepsin B (Ctsb; [B4]), 24-dehydrocholesterol reductase (Dhcr24; [B5]), dual oxidase 2 (Duox2; [B6]), excision repair cross-complementation group 6 (Ercc6; [B7]), glutamate-cysteine ligase catalytic subunit (Gclc; [B8]), glutathione peroxidase 3 (Gpx3; [B9]), Gpx4 (B10), Gpx6 (B11), Gpx7 (B12), heme oxygenase 1 (Hmox1; [B13]), isocitrate dehydrogenase (Idh1; [B14]), myeloperoxidase (Mpo; [B15]), NAD(P)H quinone dehydrogenase 1 (Nqo1; [B16]), nudix hydrolase 1 (Nudt1; [B17]), parkinsonism associated deglycase (Park7; [B18]), peroxiredoxin1 (Prdx1; [B19]), Prdx2 (B20), Prdx6 (B21), prion protein (Prnp; [B22]), proteasome subunit beta 5 (Psmb5; [B23]), sequestosome 1 (Sqstm1; [B24]), thioredoxin reductase 1 (Txnrd1; [B25]), Txnrd2 (B26), and Selenoprotein P Plasma 1 (Sepp1; [B27]).

Six different genes encoding proteins that are involved in ROS metabolism were upregulated in the hippocampus at 6 months post-exposure to GWIR-Cs and mild stress (Figures 2A,B1–B6), suggesting the chronic presence of enhanced ROS in the hippocampus. This comprised the following genes: flavin containing monooxygenase 2 (Fmo2; [B1]), superoxide dismutase 2 (Sod2; [B2]), Sod3 (B3), copper chaperone for superoxide dismutase (Ccs; [B4]), stearoyl-CoA desaturase 1 (Scd1; [B5]), and uncoupling protein 2 (Ucp2, [B6]). There were also increased expression of several genes involved in oxygen transport in the hippocampus at 6 months post-exposure to GWIR-Cs and stress (Figures 2C,D1–D4). These genes were the following: cytoglobin (cygb; [D1]), dynamin 2 (Dnm2; [D2]), intraflagellar transport 172 (Ift172, [D3]), and solute carrier family 38 member 1 (Slc38a1; [D4]).

Many genes (n = 20) encoding proteins that are involved in antioxidant activity were upregulated in the hippocampus at 6 months post-exposure to GWIR chemicals and mild stress, further supporting the chronic presence of enhanced oxidative stress in the hippocampus (Figure 3A). The up-regulated expression of 12 of these genes are illustrated in Figure 1, which included Cat (B3), Ctsb (B4), Gpx3, Gpx4, Gpx6, and Gpx7 (B9-B12), Mpo (B15), Prdx1, Prdx2 and Prdx6 (B19-B21), and Txnrd1 and Txnrd2 (B25-B26). The enhanced expression of the antioxidant gene Sod3 is illustrated in Figure 2B3. Finally, there were seven additional genes with increased expression (shown in Figure 3), namely, adenomatous polyposis coli (Apc; [B1]), glutathione reductase (Gsr; [B2]), glutathione S-transferase kappa 1 (Gstk1, [B3]), glutathione S-transferase pi 1 (Gstp1; [B4]), Prdx5 (B5), prostaglandin-endoperoxide synthase 1 (Ptgs1; [B6]), and sulfiredoxin 1 (Srxn1; [B7]).

Many genes (n = 26) involved in mitochondrial respiration were up-regulated in the hippocampus, 6 months post-exposure to GWIR-Cs and mild stress (Figures 4A1–A3). These comprised genes encoding proteins that are important for the function of complexes I, II, IV, and V of the mitochondrial respiratory chain (Figures 4B1–B9,C,D1–D5,E1,E2). The nine genes with altered expression pertaining to complex I were NADH dehydrogenase [ubiquinone] 1 alpha subcomplex 8 (Ndufa8, [B1]), Ndufa9 [B2], Ndufa11 [B3], NADH dehydrogenase [ubiquinone] 1 beta subcomplex 3 (Ndufb3, [B4]), Ndufb5 [B5], NADH dehydrogenase [ubiquinone] Fe-S protein 6 (Ndufs6, [B6]), Ndufs7 [B7], Ndufs8 [B8], and NADH dehydrogenase [ubiquinone] flavoprotein 1 (Ndufv1, [B9]). The only gene that was modified in relation to complex II function was succinate dehydrogenase complex, subunit B, iron sulfur (Sdhb; [C]). The five genes associated with complex IV (cytochrome C oxidase – the terminal component of mitochondrial respiratory chain) were cytochrome C oxidase subunit Vb (Cox5b; [D1]), cytochrome C oxidase subunit V1a polypeptide 1 (Cox6a1; [D2]), cytochrome C oxidase subunit VIIa polypeptide 2 like (Cox7a2L; [D3]), surfeit 1 (Surf1; [D4]), and cytochrome C-1 (Cyc1; [D5]). The two genes relevant to complex V (mitochondrial ATP synthase, the enzyme that converts ADP to ATP), which had an altered expression, were ATP synthase, H+ transporting, mitochondrial F1 complex, alpha subunit 1 (Atp5a1; [E1]) and Atp5b (E2).

In addition to the above genes, nine other genes relevant to mitochondrial/other organelle function displayed enhanced expression (Figure 5). These included the following: ATPase, H+ transporting, lysosomal accessory protein 1 (Atp6ap1; [A1]), ATPase, H+ transporting, lysosomal V0 subunit A2, involved in acidification of cell organelles (Atp6voa2; [A2]), phospholysine phosphohistidine inorganic pyrophosphate phosphatase, involved in respiratory electron transport (Lhpp; [A3]), pyrophosphatase [inorganic] 1, important for phosphate metabolism of cells (Ppa1; [A4]), solute carrier family 25 member 10, one of mitochondrial membrane carrier proteins (Slc25a10; [A5]), solute carrier family 25 member 20, another mitochondrial membrane carrier protein (Slc25a20; [A6]), uncoupling protein 1, one of mitochondrial anion carrier proteins (Ucp1; [A7]), ubiquinol-cytochrome C reductase binding protein, participates in the transfer of electrons when ubiquinone is bound (Uqcrb; [A8]), and ubiquinol-cytochrome C reductase core protein I, involved in metabolism (Uqcrc1; [A9]).

Many genes linked to inflammation, namely, Nfkb1, IL1a, Csf2, Bcl6, and IL6 were upregulated in the hippocampus at 6 months after exposure to GWIR-Cs and stress (Figures 6A1–A5). The increased expression of these pro-inflammatory genes was associated with the decreased expression of anti-inflammatory genes IL4 and IL10 (p < 0.01–0.0001, Figures 6B1,B2). Further, several genes associated with neurodegenerative disorders and/or cognitive and memory function also had an altered expression. These included the increased expression of Ngf, Fgfb, Mapk1, Mapk3, Prkaca, N-pac, and Narg2 (Figures 6C1–C7), and the decreased expression of Vegfa (Figure 6C8).

Using confocal microscopy and dual immunofluorescence, we examined if NeuN + neurons in the hippocampus expressed nuclear Nrf2 (activated Nrf2) (Figure 7). Since the frequency of neurons with nuclear Nrf2 was greatest in the CA3 hippocampal in GWI-rats, we chose this subfield for quantification. In naïve controls, neurons expressed nuclear Nrf2 infrequently. Further, neurons positive for Nrf2 displayed minimal Nrf2 expression (one or two speckles) in the nucleus (Figures 7A1,A2,C1). In comparison, more CA3 pyramidal neurons expressed nuclear Nrf2 in GWI-rats (Figure 7B1). The overall amount of nuclear Nrf2 in these neurons was also greater, which was typically evidenced by three or more speckles of positive staining (Figure 7B2). GWI-rats had significantly more neurons with nuclear Nrf2 (Figure 7C1, p < 0.001) and a greater fraction of neurons with robust nuclear Nrf2 (i.e., >three speckles of variable size, Figure 7C2, p < 0.05). To validate the activation of Nrf2 in the hippocampus, we performed an ELISA for activated Nrf2 using hippocampal tissue lysate. The overall amount of activated Nrf2 was greater in the hippocampus of GWI-rats, in comparison to naïve controls (p < 0.0001, Figure 7C3). Thus, we confirmed that chronic GWI was associated with considerable activation of Nrf2 in the hippocampus.

Many cytokines and chemokines were elevated in the serum, 6 months after exposure to GWIR-Cs and stress (Figure 8). These were the following: Tnfa, IL1b, IL1a, transforming growth factor beta (Tgfb), macrophage inflammatory protein 1 alpha (Mip1a, also called CCL3), IL6, interferon gamma-induced protein 10 (IP10, also called Cxcl10), stem cell factor, a cytokine that binds to c-KIT receptor (Scf), Vegf, Fgfb, IL5, and IL15 (Figures 8A1–A12). Cytokines/chemokines that were unchanged included macrophage chemoattractant protein 1 (Mcp1), Ccl5 (or Rantes), Ifng, and leptin. The concentration of MDA (a by-product of lipid peroxidation and increased oxidative stress) was also greatly enhanced in the serum of GWI-rats (Figure 7B). Thus, multiple cytokines and chemokines were upregulated, at a systemic level, in association with the increased oxidative stress in a rat model of chronic GWI.