Background
Sarin (O-isopropyl methylphosphonofluoridate) is a toxic organophosphorus (OP) nerve agent that was first discovered on October 10, 1938, by German scientists who were originally tasked with synthesizing more potent pesticides [
1]. The production and stockpiling of sarin and other chemical warfare agents (CWAs) was banned by the Chemical Weapons Convention of 1993. However, OP nerve agents still remain a threat in armed conflicts and terrorist attacks, such as the terrorist sarin gas attack on the Tokyo subway in 1995 by members of the Japanese Uhm-Shinrikiu cult; the attack resulted in injuries to more than 5,500 civilians and 12 deaths [
2,
3]. CWAs are likely to be a weapon of choice for many other terrorist organizations because they are relatively accessible or simple to produce, easy to transport, and can be delivered in mass quantities [
4,
5].
Like other OP nerve agents, sarin irreversibly inhibits acetylcholinesterase (AChE), causing an accumulation of acetylcholine (ACh) at cholinergic synapses. This ACh buildup results in a cholinergic crisis due to overstimulation of muscarinic and nicotinic receptors in the central and peripheral nervous system, including the neuromuscular junction [
4,
6,
7]. A victim exposed to these CWAs initially experiences symptoms such as myosis, tightening of the chest, difficulty breathing, and a general loss of bodily functions. As symptoms progress, the victim suffers from convulsive spasms and seizures, which can lead to death if left untreated [
4,
6‐
10].
Current medical countermeasures to nerve agent intoxication include an anti-muscarinic (e.g., atropine) that blocks excess ACh at muscarinic receptors to alleviate parasympathetic overstimulation, an oxime (e.g., 2-pyridine aldoxime methylchloride, 2-PAM) to reactivate inhibited AChE molecules, and an anticonvulsant such as diazepam [
6‐
8,
11]. These therapeutics increase survival if administered within a short period of time following exposure, but they may not fully prevent neurological damage [
2,
6,
10,
12‐
14]. Previous studies have shown that the development of long-lasting seizure activity following nerve agent exposure is highly correlated with the occurrence of brain damage [
6,
15]. Survivors of nerve agent poisoning can experience long-term neurological and behavioral outcomes months or years following exposure [
2]. Previous findings of Scremin et al. [
16] revealed that sarin-exposed rats showed behavioral abnormalities up to 16 weeks post-exposure. To date, most of our understanding on this issue comes from studies performed on survivors of the Tokyo subway attack, and most of these findings encompass only the psychiatric sequelae due to the high prevalence of post-traumatic stress disorder [
3]. More recently, Loh and colleagues [
5] reported on the long-term cognitive sequelae of a soldier exposed to sarin gas by means of an improvised explosive device (IED) while he was deployed to Iraq in 2004. Testing performed ten months following exposure revealed that the victim suffered from reduced information processing speed, poor focused attention, and difficulty in motor coordination. Despite these studies, the long-term neurologic sequelae of nerve agent exposure are still unclear. Therefore, the molecular effects and biological pathways involved in nerve agent-induced neurodegeneration need to be examined to determine drug treatments that would be effective when administered after the onset of seizures and secondary responses that lead to brain injury.
Gene expression profiling is an effective approach for elucidating chemical toxicant mechanisms of action [
17,
18]. Oligonucleotide microarrays are commonly used to simultaneously measure the mRNA levels of thousands of genes in a cell and are capable of detecting even subtle changes in gene expression. Gene expression profiling has been successfully used to investigate the mechanisms of toxicity and resulting effects of various CWAs [
19‐
24].
We performed gene expression profiling of the piriform cortex, one of the regions in the central nervous system to show massive early-onset tissue pathology from nerve agent-induced seizures [
10,
13,
25], following sarin exposure in a rat model to identify the molecular effects involved in nerve agent-induced neurodegeneration. In the present manuscript, we have focused on the transcriptional responses observed in sarin-exposed seizing animals. The gene expression alterations seen in sarin-exposed non-seizing animals will be the focus of a future manuscript. Consistent with previous studies, gene ontology analysis revealed a strong inflammatory response following nerve agent-induced seizure onset [
22‐
24,
26‐
31]. Therefore, we have identified pro-inflammatory cytokines as potential molecular targets for the development of effective neuroprotectants following nerve agent exposure.
Discussion
Despite extensive efforts to prevent their use, OP nerve agents remain a viable threat to soldiers in times of war as well as to the civilian population in the event of a terrorist attack. It is well-established that the acute toxicity of nerve agents is the result of AChE inhibition; however, the molecular mechanisms and biological pathways involved in the resulting neurodegeneration following seizure onset are poorly understood. To help determine these molecular events, animals were challenged with a 1 × LD
50 dose of sarin, and microarray analysis was used to identify gene expression changes in the rat piriform cortex over a 24-h time period following seizure onset. Using our exposure model, roughly 50% of the sarin-exposed animals developed seizure activity with a mean latency of 10.2 min, as indicated by behavioral assessment and electrocortical activity. In this study, we analyzed gene expression changes in the piriform cortex of seizing animals because it has been identified as one of the regions in the central nervous system to show massive, early-onset tissue pathology following nerve agent-induced seizures [
10,
13,
25]. In agreement with previous studies [
6,
15,
26,
28,
36], we found major gene expression profile differences correlated with seizure induction and identified a strong inflammatory response that could potentially lead to brain injury and cell death. The transcriptional responses of sarin-exposed non-seizing animals in this dataset will be the focus of a future manuscript.
Many significant molecular changes were seen at our earliest time point of 0.25 h after seizure onset. Unlike the later time points, a number of these changes were seen in metabolic pathways, such as those involved in the metabolism of glutamate, inositol, and phospholipids. The number of significantly altered genes and canonical pathways increased over time following seizure induction and appeared to represent an inflammatory response, which was seen as early as 0.25 h. This is suggested by the appearance of biological functions such as cell-mediated immune response (0.25 h, 1 h, 3 h, 6 h), immune cell trafficking (0.25 h, 1 h, 6 h, 24 h), inflammatory response (0.25 h, 1 h, 6 h, 24 h), and immunological disease (1 h, 3 h, 24 h) among the 25 biological functions most significantly altered following sarin-induced seizure. Further support of this statement is provided by the number of inflammatory-related pathways that were significantly altered at each time point, such as IL-10 signaling (0.25-24 h), PPAR signaling (0.25-24 h), IL-6 signaling (1-24 h), and TREM1 signaling (1-24 h). Thus, even at our latest time point of 24 h, inflammatory functions and pathways were still significantly altered by sarin-induced seizure.
In addition to our analysis of molecular effects at each time point, we also performed a two-way interaction ANOVA using exposure (saline vs. sarin) and time to obtain an overall view of significant molecular effects resulting from sarin-induced seizure during the 24-h time course. In agreement with the analyses performed at each individual time point, we identified biological functions associated with an inflammatory response (e.g., inflammatory response, inflammatory disease, immune cell trafficking, and immunological disease) among the most significant responses. Additionally, the most significant canonical pathways identified were all related to an inflammatory response. These included IL-6 signaling, IL-10 signaling, TREM1 signaling, MIF regulation of innate immunity, type I diabetes mellitus signaling, p38 MAPK signaling, toll-like receptor signaling, and acute phase response signaling. In our assessment of the five de novo networks of genes most significantly modulated by sarin-induced seizure over the 24-h time course, we identified those associated with cell-to-cell signaling and interaction, inflammatory response, cellular movement, cell death, cellular development, cellular function and maintenance, cell morphology, cellular growth and proliferation, lipid metabolism, and nervous system development and function.
The findings presented in this study support the temporal model proposed by McDonough and Shih [
6] that links nerve agent-induced seizures to resulting neuropathology. In this three-phase model, seizure initiation is a cholinergic phenomenon that lasts from the time of exposure to approximately 5 min after seizure onset. It has been reported that a convulsant dose of soman immediately inhibits brain cholinesterase with maximum inhibition within 10 min and a large increase in ACh concentration at the time of seizure initiation. Furthermore, previous studies have shown an immediate induction of AChE mRNA expression levels in the rat brain following sarin exposure [
37,
38]. If seizures are not immediately stopped, a transition phase occurs 5-40 min post-exposure where other neurotransmitter systems are perturbed. During this phase, the level of excitatory amino acids (EAAs), such as glutamate, increases and potentiates seizure activity [
6]. Our findings at 0.25 h after seizure onset support this transition phase of the model. D-glutamine and D-glutamate metabolism and glutamate receptor signaling were two of the canonical pathways significantly altered immediately after seizure onset. Previous studies have shown an increase in choline (Ch), a precursor for ACh, 15-30 min after nerve agent exposure [
39], a time in which seizure activity has already been initiated, due to increased hydrolysis of phospholipids [
40], and is supported by the presence of phospholipid degradation among the significant pathways at this early time point. The most significantly altered pathway at 0.25 h was inositol metabolism. Inositol works closely with Ch as a primary component of cell membranes. It is necessary for normal nerve and brain function as it is required for proper action of several neurotransmitters, such as ACh and serotonin. Studies have shown that membrane phosphoinositide (PI) is hydrolyzed following the activation of neurotransmitter receptors, such as N-methyl-D-aspartate (NMDA), to yield inositol 1,4,5-triphosphate (IP
3), a second messenger that transmits signals from the receptor into the cell by releasing calcium from non-mitochondrial intracellular stores [
6,
8,
41]. This leads to the last phase of the model, which is predominantly a noncholinergic phenomenon starting approximately 40 min after seizure onset with the presence of prolonged epileptiform activity. It is proposed that this excess influx of calcium is the ultimate cause of neuropathology following nerve agent exposure as it can hyperactivate enzymes such as lipases, proteases, endonucleases, kinases, or phosphatases that can cause damage to cell membranes, cytoskeleton, or organelle structure and function [
41,
42].
Since the development of the model proposed by Shih and McDonough, many studies have shown that there is also an increase in pro-inflammatory mRNA and protein expression following nerve agent exposure that lasts hours-to-days after exposure [
24,
27,
29‐
31]. Furthermore, there has been increasing evidence over the past several years implicating inflammatory reactions in the pathogenesis of several neurodegenerative disorders, such as Alzheimer's disease, Parkinson's disease, multiple sclerosis, and epilepsy [
43,
44]. Studies using various seizure models have shown an increase in cytokine mRNA and protein expression levels within 30 min following seizure induction in brain regions involved in seizure onset and spread [
45‐
47]. Therefore, it is likely that this late phase of the model involves neuroinflammatory processes that lead to neuropathology following nerve agent exposure.
Pro-inflammatory cytokines, such as IL-1β, IL-6, and TNF-α, are expressed at very low levels in healthy brain tissue but are rapidly induced following insult. In our study, we observed a significant increase in pro-inflammatory gene expression as early as 0.25 h following sarin-induced seizure onset, and this inflammatory response was still present at our latest observed time point of 24 h. In support of this finding, Damodaran et al. [
22,
23] and Chapman et al. [
29] also observed an increase in cytokine expression following sarin-induced seizure activity. Damodaran et al. [
22,
23] used microarray analysis to study gene expression profiles 0.25 h after sarin exposure. As with our study, they reported numerous changes in gene expression profiles immediately following sarin exposure with cytokines being among the significantly altered signal transduction pathways. Chapman and colleagues [
29] used midazolam to control seizure duration and monitored protein expression levels of IL-1β, IL-6, TNF-α, and prostaglandin E2 (PGE2) in the hippocampus and cortex at 2, 4, 6, 8, 24, 48, and 144 h and 30 days following 5 or 30 min of seizure activity. They observed a significant increase in cytokine expression starting at their earliest time point of 2 h and peaking at 2-24 hr following sarin, with the greatest increase in animals subject to 30 min of seizure activity.
Further support of our findings is provided by studies that show a neuroinflammatory response to soman. Svensson et al. [
26,
28] have previously shown increases in IL-1β mRNA and protein levels following soman exposure. In addition, Dhote et al. [
30] and Williams et al. [
27] used quantitative RT-PCR to analyze the neuroinflammatory gene response following a convulsant dose of soman (1.6 × LD
50). Dhote et al. [
30] showed an increase of IL-1B, TNF-α, IL-6, inter-cellular adhesion molecule-1 (ICAM-1), and suppressor of cytokine signaling (SOCS) 3 mRNA in the whole cortex at 0.50, 1, 2, 6, 24, 48, and 168 h following soman exposure, which confirmed the earlier findings of Williams et al. [
27] where they observed an initial up-regulation of TNF-α mRNA at 2 h post-exposure followed by an increase in IL-1β and IL-6 mRNAs 6 h later. Johnson and Kan [
31] have recently quantified the protein levels of these cytokines in vulnerable brain regions following soman-induced seizure onset. They reported a significant increase in IL-1β, IL-6, and TNF-α protein levels between 10-18 hrs after the mRNA peak expression levels. Dillman et al. [
24] used oligonucleotide arrays to analyze gene expression profiles of rat hippocampi at 1, 3, 6, 12, 24, 48, 72, 96, and 168 h following exposure to a convulsant dose of soman. In agreement with our findings, they observed an increasing alteration in gene expression profiles over the first 24 h following soman exposure. Within this time frame, they identified a strong inflammatory response with the presence of immunological and inflammatory disease among the most significant biological processes altered, and the most significant canonical signaling pathways including p38 MAPK, toll-like receptor, IL-6, and IL-10. During the later phase of their time course, they observed a shift in expression that resembled an injury response (24-96 h), which was subsequently followed by a recovery phase at their latest time point of 168 h. The similarities during the first 24 of our study and the study done by Dillman and colleagues [
24] lead us to believe that we would also observe a similar shift in gene expression that would involve molecular processes and pathways involved in an injury and recovery phase. However, further studies analyzing gene expression profiles over a longer time period are needed to confirm this same mechanism of action following sarin-induced seizure. Angoa-Pérez and colleagues [
11] recently studied the effects of soman on the expression of cyclooxygenase-2 (COX-2), which is the initial enzyme in the biosynthetic pathway of pro-inflammatory prostaglandins (PGEs) and a factor that has been implicated in seizure initiation and propagation. They found that the induction of COX-2 expression and subsequent production of PGEs correlated with seizure intensity in the rat brain from 4 h to 7 d, suggesting that these molecules could play a role in neuronal degeneration well after the cholinergic and glutamatergic response. Angoa-Pérez and colleagues hypothesize that seizures occurring in response to a PGE overload would likely not respond to the standard treatment of anticholinergics and benzodiazepines, indicating that other therapeutics, such as COX-2 inhibitors, should be added to prevent or minimize neuropathology that occurs in the later phase of the McDonough and Shih model.
Authors' contributions
KDS collected and processed brain regions for microarray and multiplexed PCR 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.