Background
Organophosphorus (OP) nerve agents, such as sarin (O-isopropyl methylphosphonofluoridate), irreversibly inhibit the enzyme acetylcholinesterase (AChE). This inactivation of AChE causes a toxic accumulation of the neurotransmitter acetylcholine (ACh) and results in over-stimulation of muscarinic and nicotinic ACh receptors [
1‐
3]. Due to the continuous stimulation of muscles, glands, and central nervous system, victims exposed to these poisonous agents develop myosis, tightening of the chest, difficulty breathing, and a general loss of bodily functions. As symptoms progress, the victims suffer from convulsive spasms and seizures that can quickly progress to status epilepticus (SE), which has been strongly associated with brain damage in survivors, and death [
1,
3‐
6].
Current medical countermeasures against toxic levels of nerve agents can increase survival if administered within a short period of time following exposure, but they may not fully prevent neuropathology or functional impairment [
3,
6‐
9]. Although these countermeasures are readily available to soldiers in a combat setting, they are not accessible to the general public in case of a terrorist attack. The anticipated response time to treat civilian casualties exposed to nerve agent is estimated to be at least 30 min [
6], and many will have already initiated seizures by the time medical personnel arrive. Rapidly terminating nerve agent-induced seizures is critical because their duration and intensity have been directly linked to brain damage following exposure [
4,
10‐
12]. Previous studies have shown signs of neuropathology present within 20 min of seizure onset and have demonstrated the increased difficulty in terminating seizures lasting beyond 40 min [
6,
10]. Therefore, it is important to understand the mechanism of nerve agent-induced brain injury and identify treatments that are effective when administered after the initiation of seizures and the secondary responses that lead to brain injury.
We previously characterized the transcriptional response of rat piriform cortex following sarin exposure [
51]. We found that critical gene expression profile differences correlated with seizure induction and identified secondary responses that potentially lead to brain injury and cell death. In addition to the piriform cortex, other brain regions have been identified as sensitive to varying degrees to nerve agent exposure. These include the amygdala, hippocampus, septum, and thalamus [
3,
6,
10,
13]. To understand in greater detail the molecular responses of these brain regions to nerve agent, we utilized oligonucleotide microarrays to define the temporal transcriptional responses of these brain regions following sarin-induced seizure in a rat model. We then compared the transcriptional profiles of these four brain regions to the transcriptional response previously characterized in the piriform cortex to identify the common and unique molecular mechanisms significantly affected by sarin-induced seizure in these five sensitive brain regions.
Methods
Sarin exposure
Male Sprague-Dawley rats (350-500 g) were obtained from Charles River Laboratories (Wilmington, MA). They were housed in a temperature-controlled room with a 12-h light/12-h dark cycle and given food and water ad libitum. The research for this study was conducted at the United States Army Medical Research Institute of Chemical Defense (USAMRICD; Aberdeen Proving Ground, MD), which is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International. All of the animal procedures were approved by the Institute Animal Care and Use Committee at USAMRICD and conducted in accordance with the principles stated in the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996) and the Animal Welfare Act of 1966 (P.L. 89-544), as amended.
PhysioTel® F40-EET transmitters (Data Sciences International, St. Paul, MN) were surgically implanted into the animals to record bi-hemispheric cortical electroencephalogram (EEG) activity, body temperature, and gross motor activity throughout the study. After a two-week recovery period, the animals were challenged with 1 × LD50 sarin (108 μg/kg, sc) that was obtained and diluted in sterile saline at USAMRICD. One minute after seizure onset, animals were treated with atropine sulfate (2 mg/kg; Sigma-Aldrich, St. Louis, MO) and 2-pyridine aldoxime methylchloride (2-PAM; 25 mg/kg; Sigma-Aldrich), both administered in a single injection (im). Thirty minutes later, animals used for the 1-h to 24-h time points were given the anticonvulsant diazepam (10 mg/kg, sc; TW Medical Veterinary Supply, Austin, TX). Control animals received an equivalent volume of vehicle (saline), atropine sulfate, 2-PAM, and diazepam. Naïve animals received no injections.
Behavioral observations were documented for each animal following exposure and placed in one of three categories (mild, moderate, or severe). The total was then calculated and graphed using the total number of toxic signs listed in the moderate (e.g., loss of posture, excessive salivation and/or lacrimation, and body tremors) and severe (e.g., complete loss of posture, clonic-tonic convulsions, and gasping) categories. These behavioral observations corresponded with the five stages of behavioral seizure intensity, which were rated using a modified Racine scale score [
14]: stage 0 = baseline behaviors, including resting, grooming, chewing, and sleeping; stage 1 = inactivity, unusual posture, piloerection, frozen posture, clumsy motion, and excessive grooming or chewing; stage 2 = oral tonus, head bobs, and body tremors; stage 3 = forelimb myoclonus, prostrate body extension, and salivation or lacrimation; stage 4 = loss of posture, whole body tremors, rigidity, body jerks, and forelimb myoclonus followed by rearing; and stage 5 = complete loss of posture, falling or generalized tonic-clonic convulsions, and gasping. Statistical significance between sarin-exposed seizing animals and their controls was calculated using Student's
t-test.
Animals were euthanized by decapitation at 0.25, 1, 3, 6, and 24 h after seizure onset. The amygdala, hippocampus, septum, and thalamus were immediately collected from each animal at the appropriate time point. Three animals were used for each experimental group (naïve, saline control, and sarin-exposed seizure) at each time point, with the exception of 1-h saline control, 3-h sarin-exposed seizure, and 24-h sarin-exposed seizure (n = 4). Each tissue was immediately snap-frozen in liquid nitrogen and stored at -80°C until use.
Sample preparation for microarray hybridization
Brain tissues were homogenized in RNeasy lysis buffer (QIAGEN, Valencia, CA) at three intervals of 30 sec each using the Mini-Beadbeater-96 (Biospec Products, Bartlesville, OK) and 6.35 mm stainless steel beads. Each homogenate was subsequently centrifuged for 10 min at 16,110 × g at room temperature, and the supernatant was transferred to a new microcentrifuge tube. Total RNA was then extracted and DNase I-treated using the RNeasy Mini Kit and RNase-Free DNase Set (QIAGEN) according to the manufacturer's protocol. The quantity and quality of the RNA was determined with a NanoDrop ND-1000 UV-vis spectrophotometer (Thermo Scientific, Wilmington, DE) and an Agilent Bioanalyzer (Agilent Technologies, Santa Clara, CA) throughout sample processing. Total RNA was processed for hybridization to GeneChip
® Rat Genome 230 2.0 oligonucleotide arrays (Affymetrix, Inc., Santa Clara, CA) using the BioArray Single-Round RNA Amplification and Biotin Labeling System (Enzo Life Sciences, Inc., Farmingdale, NY) as previously described [
15]. In brief, 1 μg (amygdala, hippocampus, and thalamus) or 500 ng (septum) of total RNA was used to generate first strand cDNA by using a T7-linked oligo(dT) primer. After second strand synthesis,
in vitro transcription was performed with biotinylated UTP and CTP for cRNA amplification. Biotinylated target cRNA generated from each sample was processed according to the manufacturer's protocol using an Affymetrix GeneChip Instrument System
http://affymetrix.com/support/technical/manual/expression_manual.affx as previously described [
15].
All microarray experiments were performed to comply with Minimal Information About a Microarray Experiment (MIAME) protocols and details can be found at the Gene Expression Omnibus (GEO) accessible through GEO Series accession number GSE28435. The data discussed in this publication have been deposited in the National Center for Biotechnology Information's Gene Expression Omnibus (GEO;
http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO
Series accession number GSE28435.
Microarray data analysis
Raw signal intensities from each GeneChip
® were imported into Partek Genomics Suite v6.4 (Partek, Inc., St. Louis, MO) along with those from the piriform cortex samples [
51]. The signal intensities were normalized using the robust multiarray averaging (RMA) algorithm [
16]. Normalized data for all five brain regions (amygdala, hippocampus, piriform cortex, septum, and thalamus) were analyzed by principal component analysis (PCA) [
17] to identify patterns in the dataset and highlight similarities and differences among the samples. The major sources of variability identified within the dataset were used as grouping variables for analysis of variance (ANOVA). The calculated
p-value and geometric fold change for each probeset identifier were imported into Ingenuity Pathways Analysis (IPA; Ingenuity
® Systems,
http://www.ingenuity.com) to identify the canonical pathways, biological functions, and networks of genes significantly affected by sarin-induced seizure. Biological functions are categories that genes are classified into based on their cellular or physiological role in a healthy or diseased organism. Genes may be classified into more than one biological function. A canonical pathway is a well-established signaling or metabolic pathway that is manually curated on the basis of published literature. Canonical pathways are fixed prior to data input and do not change upon data input. Networks are distinct from canonical pathways in that they are built
de novo from input data based on known molecular interactions identified in the published scientific literature. To identify canonical pathways that were most significant to the dataset, molecules that met the designated
p-value cutoff (≤ 0.05) and were associated with a canonical pathway in Ingenuity's Knowledge Base were considered for the analysis. The significance of the association between the dataset and the canonical pathway was measured in two ways: 1) A ratio of the number of molecules from the data set that mapped to the pathway divided by the total number of molecules that mapped to the canonical pathway was displayed. 2) Fisher's exact test was used to calculate a
p-value determining the probability that the association between the genes in the dataset and the canonical pathway was explained by chance alone. To determine networks of genes significantly affected by sarin exposure, molecules were overlaid onto a global molecular network developed from information contained in Ingenuity's Knowledge Base. Networks of molecules were then algorithmically generated based on their connectivity. The Functional Analysis of a network identified the biological functions and/or diseases that were most significant to the molecules in the network. The network molecules associated with biological functions and/or diseases in Ingenuity's Knowledge Base were considered for the analysis. Right-tailed Fisher's exact test was used to calculate a
p-value determining the probability that each biological function and/or disease assigned to that network is due to chance alone.
Multiplexed RT-PCR
The GenomeLab Gene Expression Profiler (GeXP; Beckman Coulter, Inc., Brea, CA) genetic analysis system was used to measure the expression levels of 21 differentially expressed cytokines or chemokines (see Additional File
1) by multiplexed RT-PCR to validate the microarray data. Primers were designed using the eXpress Designer module of the GenomeLab eXpress Profiler software, with each primer consisting of 20 nucleotides of gene-specific sequence as well as a universal primer sequence. RT-PCR product sizes ranged from 151 to 351 nt with a 7-nt minimum separation size between each fragment (see Additional File
1). The custom multiplexed panel also contained glyceraldehyde 3-phosphate dehydrogenase (GAPDH) for normalization and an internal control gene (kanamycin resistance, Kan
r).
RNA samples used in the microarray experiment and the GenomeLab GeXP Start Kit (Beckman Coulter, Inc.) were used for the RT-PCR reactions according to the manufacturer's protocol. The custom multiplex was first optimized by reverse primer dilution to attenuate the gene signals that were close to or above the linear detection limit of the GeXP system detector (130,000 RFU in raw data or 120,000 RFU in analyzed data) and to balance the signal of each peak within the multiplex reaction. The final concentrations of the reverse primers within the multiplex are shown in Additional File
2. Fifty nanograms of total RNA was reverse transcribed with the optimized reverse primer multiplex. Subsequently, 9.3 μl of cDNA from each RT reaction was transferred to the PCR reaction mix containing 20 nM of the forward primer set multiplex. All experiments included "no template" (i.e. without RNA) and "no enzyme" (i.e. without reverse transcriptase) negative controls to confirm the absence of peaks at the expected target sizes.
The fluorescently-labeled PCR products were diluted 1:20 in 10 mM Tris-HCl (pH 8), and 1 μl of each dilution was added to 38.5 μl sample loading solution along with 0.5 μl DNA size standard-400 (GenomeLab GeXP Start Kit). The GeXP system was then used to separate the amplified PCR products based on size by capillary gel electrophoresis and to measure their fluorescent dye signal strength in arbitrary units (A.U.) of optical fluorescence, which is the fluorescent signal minus background. The multiplexed RT-PCR data were initially analyzed using the Fragment Analysis module of the GenomeLab GeXP system software, followed by the eXpress Analysis module of the eXpress Profiler software. First, the length or size of the products was determined using the Fragment Analysis module. The fragment data, peak height, and peak area information was then imported into the analysis module of the eXpress Profiler software where the fragments were compared to the expected PCR product sizes to identify each transcript.
The expression of each gene within a sample was normalized to GAPDH expression to minimize inter-capillary variation, and the normalized intensity of each replicate (n ≥ 3) was used to calculate an average intensity of each sample group (i.e. control or sarin-induced seizure at each time point). The fold expression difference between control and sarin-induced seizure samples was then evaluated for all genes at each time point and compared to the fold expression changes obtained by microarray analysis.
Discussion
In this study, we performed gene expression profiling to assess the temporal transcriptional changes associated with sarin-induced seizure. Comparison of the transcriptional profiles of five nerve agent-sensitive rat brain regions (amygdala, hippocampus, piriform cortex, septum, and thalamus) indicated the presence of a robust inflammatory response in sarin-exposed seizing animals. In agreement with previously reported findings [
13,
18‐
26], we observed a rapid activation of the innate immune response that persisted throughout the 24-h time period examined. Therefore, we predict that pro-inflammatory cytokines and their signaling pathways could potentially mediate some of the molecular and structural changes observed after nerve agent-induced seizure activity. Because current countermeasures do not fully prevent neuropathology, particularly in scenarios where treatment is delayed (e.g., civilian terrorist attack), identifying therapeutic targets that mediate the cascade of secondary events leading to brain damage and functional impairment is critical.
Our findings, which are in agreement with previous studies of soman-induced brain damage [
1,
11,
18,
20,
27], indicate that sarin toxicity correlates with the development and duration of seizures. Significant changes in gene expression were seen in all five brain regions following seizure occurrence with the greatest effects in the piriform cortex, which is in agreement with the findings of Lemercier at al. [
28] and McDonough et al. [
10]. However, it should be noted that a higher vulnerability to nerve agent-induced brain damage has been observed in other cerebral regions of other animal models. These highly susceptible brain regions include the thalamus in mice [
11], the amygdala in guinea pigs [
27], and the frontoparietal cortex and cerebellum in monkeys [
29,
30]. This variability may be due to interspecies differences in functional anatomy and/or differences in study design. However, the five brain regions examined in our study are nerve agent sensitive regions of the rat brain. Therefore, we examined each of these brain regions for common nerve agent-induced pathways to explore molecular mechanisms involved in nerve agent-induced brain injury.
Gene ontology analysis revealed numerous canonical pathways that were significantly altered by sarin-induced seizure in the five regions of the rat brain examined in this study. The pathways that were significantly affected in all five brain regions are known to be associated with an inflammatory response and include many pro-inflammatory cytokines. These pathways include ATM signaling, CD40 signaling, IL-10 signaling, IL-6 signaling, MIF regulation of innate immunity, role of double-stranded PKR in interferon induction and antiviral response, toll-like receptor signaling, and TREM1 signaling. Furthermore, pro-inflammatory cytokines were among the de novo networks identified as most significantly affected in all five brain regions from sarin-exposed seizing animals. Two of the top six networks were associated with an inflammatory response. One network of genes was built around TNF-α as a central node and the other around IL-6 as a central node.
Pro-inflammatory cytokines are known to mediate cellular communication and play a significant role in the pathological processes involved in various brain diseases, such as SE [
31‐
35]. Although picomolar or low nanomolar ranges of cytokines enhance neuronal survival, higher concentrations have deleterious effects on neuronal viability [
36]. The findings presented in this study are in agreement with the observed inflammatory response previously reported following nerve agent intoxication [
13,
18‐
26], as well as in other models of seizure induction [
33,
37]. Although these studies all show an up-regulation of inflammatory cytokines, it should be noted that the expression patterns of individual cytokines vary from study to study. These variations likely stem from differences in the nerve agent tested (sarin vs soman), animal models used (rats vs mice), pharmacological treatments given, and, maybe most importantly, seizure intensity [
24]. For example, Williams et al. [
19] and Dhote et al. [
24] employed quantitative real time-PCR to characterize the temporal response of inflammatory cytokines following soman-induced seizures in rats and mice, respectively. Williams and colleagues showed an initial up-regulation of TNF-α mRNA in the hippocampus, piriform cortex, and thalamus at their earliest time point of 2 h after exposure. This was subsequently followed by an increase in IL-1β and IL-6 mRNA at 6 h after exposure. The findings of Dhote et al. [
24] also indicated a significant induction of inflammatory cytokines in the whole cortex and hippocampus but not in the cerebellum. Cytokine activation was seen as early as 1 h after exposure in the cortex with a peak response between 6 and 24 h. However, cytokine up-regulation was delayed to 6 h after exposure in the hippocampus with peak expression levels observed between 24 and 48 h. Our data also show an acute up-regulation of inflammatory cytokines, but the temporal expression patterns are slightly different. We observed a significant increase in IL-1β, IL-6, and TNF-α expression levels as early as 0.25 h following seizure onset in all five brain regions examined. Of the three cytokines listed, IL-1β expression peaked the earliest following seizure onset and dropped to near control levels in the hippocampus, septum, and thalamus by 24 h, whereas it appeared to be leveling out in the amygdala and increasing in the piriform cortex at 24 h. It is known that IL-1β induces its own synthesis [
38], and this positive feedback loop could enhance and extend the IL-1β response [
33]. Therefore, we hypothesize that this positive feedback loop could be a contributing factor in the greater neuropathology seen in the piriform cortex of the rat model. The peak in IL-1β expression is followed by IL-6 and TNF-α, which both peaked at 3 h in all five brain regions. Moynagh et al. [
39] showed that IL-1β stimulates the expression of NF-κB, which in turn activates a variety of genes involved in the inflammatory response in CNS diseases [
33]. Therefore, our data appear to support the role of NF-κB in the neuropathology resulting from nerve agent-induced seizures as it shows a sharp induction by 3 h after seizure onset in all brain regions examined (data not shown) along with IL-6 and TNF-α expression. TNF-α expression decreased to near control levels by 24 h in all five brain regions. However, IL-6 expression decreased at 6 h in all five brain regions and then increased at 24 h after seizure onset in all regions except the thalamus.
Studies have shown that nerve agent-induced seizures initially result from AChE inhibition and overstimulation of cholinergic receptors, which is followed by an increase of excitatory amino acids (EAAs), such as glutamate [
1‐
3]. This glutamatergic hyperactivity causes an opening of N-methyl-D-aspartate (NMDA) calcium channels and a subsequent increase in intracellular calcium [
1,
40,
41], which in turn initiates signaling cascades that cause neuronal death. Based on our findings and those of other investigators, it appears that inflammatory signaling pathways are an important component of nerve agent-induced brain injury; however, the molecular mechanisms by which nerve agent-induced seizures produce acute neuroinflammation or how this phenomenon contributes to the ensuing neuropathology following exposure is still unclear.
In vitro studies have shown that pro-inflammatory cytokines play a role in glutamate toxicity as they inhibit glial cells from taking up excess extracellular glutamate [
42]. Cytokines are thought to further enhance this glutamatergic hyperactivity by increasing NMDA receptor activity [
43], which promotes excitotoxic neuronal cell death [
44,
45]. Another mechanism by which pro-inflammatory cytokines could contribute to nerve agent toxicity involves blood-brain barrier (BBB) function. The brain is normally isolated from the peripheral immune system via the BBB; however, a neurotoxic insult, such as a convulsant dose of sarin, can induce both a local and peripherally recruited inflammatory response. BBB damage has been seen in many neurodegenerative diseases and animal models of seizure [
46], and studies have shown that pro-inflammatory cytokines including TNF-α, IL-1β, IL-6, and interferon-λ are implicated in the regulation of BBB permeability [
47‐
49]. For example, IL-1β can affect the permeability of the BBB via disruption of the tight-junction organization or production of nitric oxide and matrix metalloproteinases in endothelial cells [
49]. The data from our sarin exposure model indicate that there could be a breakdown of the BBB as well. This was indicated by the up-regulation of inter-cellular adhesion molecule-1 (ICAM1) and E-selectin in all five brain regions. These molecules are thought to be linked to signal transduction cascades leading to junctional reorganization as they can interact with the actin cytoskeleton, which in turn is an indicator of infiltration of peripheral leukocytes into damaged brain regions through the BBB [
47]. We also observed a decrease in occludin expression levels, which is one of the main components of tight junctions, indicating a possible loss of tight junction integrity. The findings of our study support those of Abdel-Rahman et al. [
50] who demonstrated that a toxic dose of sarin can break down the BBB and speculated that this disruption plays an important role in sarin-induced cell death in the motor cortex, hippocampus, and cerebellum. Furthermore, Damodaran et al. [
22,
23] identified numerous BBB-related genes that were altered at 15 min and persisted until three months following 0.5 and 1.0 × LD
50 sarin exposure.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KDS collected and processed brain regions for microarray and GeXP analysis, analyzed all data, and drafted the manuscript. LAL participated in developing and coordinating the study. CLR conducted behavioral assessments of the animals. JLM dissected brain regions from all animals. JFD participated in the brain dissections, study design, data analysis, and helped draft the manuscript. All authors read and approved the final manuscript.