Comparison of neuropathology in rats following status epilepticus induced by diisopropylfluorophosphate and soman

Abstract

The increasing number of cases involving the use of nerve agents as deadly weapons has spurred investigation into the molecular mechanisms underlying nerve agent-induced pathology. The highly toxic nature of nerve agents restrict their use in academic research laboratories. Less toxic organophosphorus (OP) based agents including diisopropylfluorophosphate (DFP) are used as surrogates in academic research laboratories to mimic nerve agent poisoning. However, neuropathology resulting from DFP-induced status epilepticus (SE) has not been compared directly to neuropathology observed following nerve agent poisoning in the same study. Here, the hypothesis that neuropathology measured four days after SE is the same for rats exposed to DFP and soman was tested. Adult Sprague-Dawley rats were injected with soman or DFP to induce SE. Cortical electroencephalography (EEG) was recorded prior to and during soman-induced SE. EEG power analysis of rats administered soman revealed prolonged electrographic SE similar to that of rats that endure uninterrupted SE following injection of DFP. Rats that experienced soman-induced SE displayed less hippocampal neuroinflammation and gliosis compared to rats administered DFP. Seizure-induced weight change, blood-brain barrier (BBB) leakiness and neurodegeneration in most seizure sensitive limbic brain regions were similar for rats that endured SE following soman or DFP. The amalgamated pathology score calculated by combining pathological measures (weight loss, hippocampal neuroinflammation, gliosis, BBB integrity and neurodegeneration) was similar in rats administered the OP agents. These findings support use of the rat DFP model of SE as a suitable surrogate for investigating some, but not all delayed consequences produced by nerve agents.

Graphical Abstract

1. INTRODUCTION

Nerve agent poisoning is a medical emergency requiring administration of medical countermeasures for survival. Recently, there has been a burgeoning number of threats by terrorist organizations to induce mass casualties by the release of nerve agents in densely populated areas. Examples of the devastation caused by the use of such an agent amid a civilian population are evident in the 2013 Ghouta and 2017 Khan Shaykhun chemical attacks in Syria. Nerve agents, chemically classified as organophosphorus (OP) based agents, act by potently inhibiting acetylcholinesterase (AChE) via phosphorylation of a serine in the catalytic pocket, leading to the accumulation of acetylcholine at nicotinic and muscarinic synapses and an acute cholinergic crisis. A consequence of exposure to high levels of a nerve agent is a seizure that occurs within minutes and often progresses to status epilepticus (SE, an unremitting seizure or a series of seizures without regain of full consciousness lasting longer than 5 min). The current treatment to combat nerve agent poisoning is administration of atropine to reverse the effects of muscarinic activation, an oxime to reactivate AChE and a benzodiazepine to control seizure activity, all of which must be given soon after exposure for optimal effectiveness^{1, 2}. Even with antidote medical countermeasures, survivors of prolonged SE induced by nerve agents may develop neuropathology and permanent brain alteration as discussed in a recent review³. Understanding the mechanisms leading to brain injury after nerve agent exposure poses new challenges that require novel experimentation.

Exposure of rodents (mainly rats and guinea pigs) to nerve agents has been used to simulate the consequences of nerve agent poisoning in humans. However, due to the toxic nature of nerve agents their use in research in the USA is restricted to a limited number of organizations. On the other hand, closely related organophosphorus based agents like diisopropylfluorophosphate (DFP) are often administered to rodents in academic research laboratories to incite a similar cholinergic crisis as the nerve agents. DFP is a close structural and functional mimic of the highly reactive and poisonous nerve agent sarin (GB)⁴ and therefore, a rodent DFP model is expected to replicate many of the lethal or sublethal effects following exposure to nerve agents. However, the seizure-induced neuropathology that develops in rodent survivors days following administration of high levels of DFP has not been compared directly to neuropathology that develops following nerve agent poisoning in rodents in the same study. A collaboration was established with Dr. John McDonough and colleagues at the United States Army Medical Research Institute of Chemical Defense who have developed and optimized a rat model of soman poisoning to compare the effects of diisopropylfluorophosphate and soman-induced SE in rats. The goals were to: 1) identify model-dependent differences in the development of SE including cortical seizure activity, mortality and morbidity, and 2) characterize the neuropathological profile of the soman and DFP rat models 4 days after SE. The hypothesis that neuropathology measured 4 days after SE is the same for DFP and soman injected rats was tested. This study involves a direct comparison of two OP agents under similar experimental conditions in the same species with regard to seizure activity, mortality, weight loss, neuroinflammation, gliosis, blood-brain barrier (BBB) integrity and neurodegeneration. Furthermore, the results of the study provide important information on the consequences of nerve agent poisoning and the similarities in rodent OP-induced SE models.

2. MATERIALS & METHODS

2.1. Ethics Statement

All procedures concerning animals were approved by the Animal Care and Use Committees of Emory University and the United States Army Medical Research Institute of Chemical Defense and were conducted in accordance with the principles stated in the Guide for the Care and Use of Laboratory Animals, the Public Health Service Policy on the Humane Care and Use of Laboratory Animals, and the Animal Welfare Act of 1966 (P.L. 89–544), as amended. We selected adult male Sprague-Dawley rats because they are outbred avoiding the substrain differences often found with inbred mouse strains. Also, the early consequences of a cholinergic crisis following OP-poisoning have been characterized most thoroughly in male rats. All soman-treated rats included in the study received two injections of a vehicle (2% dimethylacetamide and 98% corn oil, sc) based on measured body weight (1 ml/kg) at 2 h and 26 h after SE onset. Similarly, all DFP rats (non-seizure controls and SE rats) received two injections of a vehicle (4% dimethylacetamide and 96% olive oil, sc) at 4 h and 24 h after SE onset. The original intention was to compare vehicle-injected rats to cohorts of rats administered a new investigational drug, however the experimental plan changed. There was no evidence that either vehicle altered seizure activity or the development of neuropathology. Therefore, we deemed the vehicle administered rats suitable for the study. The rats administered the investigational drug were not included in this study.

2.2. Electrode implantation in soman-treated rats

Rats were administered meloxicam (1 mg/kg, sc) 15–30 minutes prior to the start of surgery, and again approximately 24 hours after surgery, as an analgesic. EEG recording electrodes were implanted in rats anesthetized with isoflurane (3–5% for induction, 1–3% for maintenance) and placed in a stereotaxic instrument. Bupivacaine (2.5%) was injected intradermally in the scalp on either side of the intended incision site. To implant the electrodes, a longitudinal midline incision was made in the scalp, which was then retracted laterally to expose the skull. Burr holes were drilled through the skull for placement of screw electrodes that touched the surface of the cortex. An electrode was placed over each hemisphere of the parietal cortex, and a ground electrode was placed over the cerebellum. The headset was then secured in place with dental cement. Finally, the wound was closed with sutures, and rats were returned to their home cages for 6–7 days prior to testing.

2.3. Soman-induced SE and EEG recording

Adult male Sprague-Dawley rats were exposed to soman as previously described⁵. On the day of experiment the rats were weighed (between 302 and 395 g), placed in individual recording chambers and connected to the EEG recording system via their head piece plug. EEG signals were amplified with 1902 amplifiers, digitized with a Micro1401 data acquisition interface, recorded and analyzed using Spike2 software (all from Cambridge Electronic Design Limited, Cambridge, United Kingdom). Data channels were sampled at 512 Hz and digitally filtered with a high-pass 0.3 Hz filter, a low-pass 100 Hz, and a 60 Hz notch filter. Following 60 min of baseline EEG recording, rats were pretreated with 125 mg/kg asoxime chloride (HI-6), delivered intraperitoneally. Thirty minutes after pretreatment, 180 μg/kg of the nerve agent soman (GD) in saline was delivered subcutaneously. This dose and route elicited EEG seizure activity in 100% of rats studied as previously described⁵. Control rats were treated similarly except they were given sterile saline instead of soman. One minute after the soman challenge, rats received 2 mg/kg atropine methyl nitrate intramuscularly. Along with asoxime chloride, atropine methyl nitrate protected rats from systemic toxicity of soman without altering the development of seizures, allowing survival until development of neurological symptoms. The onset of EEG seizure activity was defined by the appearance of repetitive spikes and sharp waves with an amplitude >2x that of the baseline and a duration of at least 10 s. At 40 min after seizure onset, rats were treated with 0.45 mg/kg atropine sulfate admixed with 25 mg/kg 2-pyridine aldoxime methyl chloride (2-PAM, i.m.). They also received 1.8 mg/kg midazolam (MDZ) (i.m.). These treatments did not lead to seizure termination in any of the rats in this study. These drugs were given because they have historically improved survival but had no appreciable effect on neuropathology when administered at this time point. EEG activity was recorded for 6 h after the onset of SE. After this initial recording period, rats were returned to their home cage. On each of 4 subsequent mornings at approximately 0830 hours, rats were weighed and administered 5 ml of lactated Ringer’s solution subcutaneously if their weight was >10% below their pre-exposure weight. The rats were also reconnected to the EEG monitoring system for 30–90 minutes to record cortical activity. Spikes and areas of high spiking (ictal activity) were identified visually using the Spike2 sonogram feature. Microsoft Excel was used to obtain a running mean of EEG power in the 0–60 Hz band, defined as the square of the EEG amplitude signal, integrated over 300 sec periods. A script written in Python was created to remove large movement artifacts (single transients not of cerebral origin). The mean power was obtained for each rat and plotted as a function of time. All soman injected rats were euthanized 4 days after SE. One hemisphere was fixed for histology. The cortex and hippocampus was dissected from the other hemisphere and rapidly frozen. All brain tissues were shipped to Emory University for subsequent analysis.

2.4. DFP-induced SE

Adult male Sprague–Dawley rats (280–350 g body weight) were purchased from Charles River Labs (Wilmington, MA, USA) and housed in standard plastic cages (2 rats/cage) in a temperature-controlled room (22 ± 2° C) on a 12-hour reverse light–dark cycle. Food and water were provided ad libitum. On the day of organophosphorus exposure, the rats were weighed, placed individually into a plastic cage and moved into a ventilation hood. Awake rats were injected with pyridostigmine bromide (0.1 mg/kg in 0.9% saline, sc) followed twenty minutes later with atropine methyl nitrate (20 mg/kg in 0.9% saline, sc) and subsequently ten minutes later with diisopropylfluorophosphate (DFP) (D0879, Sigma) diluted in sterile distilled water (9.5 mg/kg, ip) as previously described^{6, 7}. This dose of DFP with the given route of administration results in prolonged SE lasting >5h without pharmacological intervention^6–8, as determined by cortical EEG recordings and modified Racine scoring. Control rats were treated similarly except they were given sterile water instead of DFP. SE was not interrupted pharmacologically in rats (termed “uninterrupted”), whose seizures eventually waned over 9–10 hr after SE onset. Each rat received a volume of injected compound based on measured body weight (1 ml/kg).

2.5. Behavioral scoring of seizure activity in DFP rat model

Following DFP exposure, rats presenting non-intermittent whole body clonic seizures that persist were declared to be in SE. The seizure activity was scored and recorded every 5 min for at least 1h using a modified Racine scale⁹ that was developed for scoring rat seizure behavior in our DFP model as previously reported^{6, 7, 10}. All rats were monitored for at least 6 h and persistent non-intermittent seizure activity was detected shortly after DFP exposure until the seizures waned on their own. The rats were then placed individually into clean plastic cages with fresh bedding, soft food and water and allowed to recover overnight. To maintain hydration, a single injection of lactated Ringer’s solution (2 ml, ip) was administered when the rats were placed into cages. All DFP-injected rats were euthanized 4 days after SE.

2.6. FluoroJade B histochemistry

Four days following DFP-induced SE rats were deeply anesthetized under isoflurane, transcardially perfused with saline to remove blood and decapitated. Rats that were exposed to soman and corresponding controls were deeply anesthetized with >75 mg/kg sodium pentobarbital and transcardially perfused with saline. The brains of all rats were removed rapidly and longitudinally bisected. One hemisphere of each brain was dissected and the cortex and hippocampus, rapidly frozen on dry ice and kept for RNA isolation and albumin western blots (below). The other hemisphere was fixed overnight (~24 hrs) in 4% paraformaldehyde for DFP or 10% formalin for soman, transferred to sucrose until they sank, and 40 μm coronal sections were prepared for immunohistochemistry and FluoroJade B (FJB) histochemistry as previously described^{7, 10}. For FJB staining the sections were mounted onto slides prior to staining. Briefly, hippocampal sections were immersed in 100% ethyl alcohol for 3 min followed by a 1 min change in 70% alcohol and a 1 min change in distilled water. The sections were transferred to a solution of 0.06% potassium permanganate for 15 min, rinsed for 1 min in distilled water and then transferred to a FJB staining solution (0.0004% FJB) for 30 min in the dark. Sections were rinsed with Tris-buffered saline (TBS), mounted onto clean slides and rapidly air dried. The slides were immersed in xylenes and coverslipped with D.P.X. (Aldrich Chem. Co., Milwaukee, WI) mounting media. FJB labeling was visualized using an AxioObserver A1 epifluorescence microscope equipped with an AxioCam MRc5 camera and pictures were obtained with the AxioVision AC 4.7 (Zeiss, Oberkochen, Germany) software.

2.7. Quantification of FluoroJade B labeled cells

Following FJB labeling, images of three hippocampal areas (i.e., hilus, CA1, CA3) were taken with a 5x objective lens using the AxioVision AC 4.1 software (Zeiss) from each of 4–6 sections in each rat, from the dorsal hippocampus between bregma −2.56 and −4.16 mm^{11, 12}. An observer blinded to the treatment of the animals and the experimental conditions counted the number of bright FJB positive cells in each area. The cell numbers were recorded and expressed as the mean number of positive FluoroJade B (+FJB) cells/section for each area in each rat. FJB staining was also quantified in the basolateral amygdala (BA), piriform cortex (PC), thalamus and the somatosensory cortex (SS). These regions were chosen for FJB quantification given their implication in temporal lobe epilepsy. All of the regions chosen for FJB quantification are highly interconnected with other limbic nuclei. Quantification of the number of +FJB cells in the amygdala, piriform cortex, thalamus and cortex was performed using ImageJ software (National Institute of Health, Bethesda MD, USA). In ImageJ the area of interest was selected and the +FJB cells were quantified using the “analyze particles” feature. The intensity threshold and the minimum and maximum cell size were initially determined in an empirical manner by an investigator blinded to the treatment conditions. Automatic quantification by ImageJ was carried out on 4–6 sections obtained from each animal. The data are presented as the mean ± standard error of the mean. The number of +FJB cells per section was also plotted for each individual rat.

2.8. Immunohistochemistry

Brains were prepared and sectioned as described above for FJB labeling. Immunohistochemistry was performed on free-floating sections as described previously with slight modifications⁷. Briefly, for fluorescence immunohistochemistry, the sections were blocked for 4 h in PBS containing 1% BSA, 10% goat serum and 0.3% Triton X-100 at 25°C. The sections were subsequently incubated in the primary antibody [rabbit anti-COX-2 (1:1000), Abcam; rabbit anti-cFos (1:1000), Abcam; rabbit anti-Iba1 (1:2000), Wako Chemicals, Richmond, VA] diluted in antibody dilution solution (ADS), which contains 0.1% gelatin and 0.3% Triton X-100 in PBS at 4°C overnight (18 h). After washing with ADS, the sections were incubated for 4 h at 25°C with Alexa Fluor fluorescent conjugated antibodies (all purchased from Molecular Probes, Eugene, OR) diluted to 1:500 in ADS. The blue-fluorescent Hoechst 33342 dye (Molecular Probes) was diluted 1:2000 in the ADS containing the fluorescent secondary antibody. After staining the sections were washed again with PBS, mounted onto slides, rapidly air dried, and cover slipped with FluoroGel mounting media. The fluorescence reactions were visualized using an AxioObserver A1 fluorescence microscope and pictures were obtained using the AxioVision AC 4.7 software. Subsequently, confocal images were taken with a BX51WI microscope equipped with a disk spinning unit (DSU) (Olympus, Center Valley, PA) and SlideBook4.2 software (Intelligent Imaging Innovations, Denver, CO). In control experiments, the sections were treated in a similar manner, except the primary antibodies were omitted. All negative control sections showed no staining (data not shown). All sections used for IHC were obtained from the dorsal hippocampus between bregma −2.56 and −4.16 mm¹¹.

Quantification of Iba1 immunohistochemistry was carried out using Pixcavator IA 5.1 (Intelligent Perception, Huntington, WV). The Iba1 positive images (200× total magnification) were taken from the CA3 region (stratum radiatum) of 4–6 dorsal hippocampal sections obtained from the brains of soman injected rats that experienced SE (n = 7–8) and rats that experienced DFP-induced SE (n = 9) 4 d earlier. The Iba1 stained images were uploaded to Pixcavator and the roundness of the microglia was measured using the equation:

$$\mathit{\text{roundness}}=4\pi \left(\frac{\mathit{\text{area}}}{\mathit{\text{perimete}}{\mathit{\text{r}}}^2}\right)\times 100$$

Roundness is defined as how close the shape of an object is to a perfect circle. In Pixcavator, a perfect circle has a roundness value of 100 and a line has a roundness value of 0. There are no units for the measurement of roundness. All images used to measure roundness contained 20 or more positively stained cells. For each section the average roundness was calculated and the data are presented as the average microglia roundness for each rat. The roundness of astrocytes is not reported due to technical problems experienced with the GFAP stained sections that compromised the accurate determination of the roundness of astrocytes.

2.9. RNA isolation and quantitative real-time polymerase chain reaction (qRT-PCR)

Total RNA was isolated using Trizol with the PureLink RNA Mini Kit (Invitrogen, Carlsbad, CA) from the frozen hippocampi. RNA concentration and purity were measured by an Epoch microplate spectrophotometer (BioTek, Winooski, VT) using the A260 value and the A260/A280 ratio, respectively. The first-strand complementary DNA (cDNA) synthesis was performed with 1 μg of total RNA, using a qScript cDNA superMix kit (Quanta Biosciences, Beverly, MA) in a reaction volume of 20 μl at 42 °C for 30 min. The reaction was terminated by heating at 85 °C for 5 min. The qPCR was performed as previously described^{7, 10}. Melting curve analysis was used to verify specificity of the primers by single-species PCR product. The 10 inflammatory mediators investigated are listed with the primer sequences in Table 1. Three housekeeping genes [β-actin, glyceraldehyde-3-phosphate (GAPDH), and hypoxanthine phosphoribosyltransferase 1 (HPRT1)] (Table 2) were used as an internal control. Samples without cDNA template served as the negative controls. Interleukin 1-beta (IL-1β), interleukin 6 (IL-6), and tumor necrosis factor alpha (TNFα) are produced and secreted by cells involved in both innate and acquired immunity to stimulate inflammation. C-C motif chemokine ligands 2, 3 and 4 (CCL2, CCL3 and CCL4) are chemokines that recruit peripheral leukocytes to sites of inflammation or injury. FBJ murine osteosarcoma viral oncogene homolog (cFos) is an immediate early gene that is activity dependent. Fos is often used as an indicator of neuronal discharge and thus seizure intensity¹³. Cyclooxygenase 2 (COX-2) is an intracellular rate-limiting enzyme in the production of prostanoids that contribute to the inflammatory response. Analysis of quantitative real time PCR data for each gene of interest was performed by subtracting the geometric mean of the three internal housekeeping genes from the measured cycle threshold value obtained from the log phase of the amplification curve to yield ΔCT for relative quantification. The fold change of each gene of interest was estimated for each animal 4 days after soman and DFP-induced SE relative to the amount of RNA found in the control animals using the 2^−ΔΔCT method¹⁴. All conditions for qRT-PCR were the same. Although the average fold change is shown in figures the ΔΔCT values were used for statistical comparison.

Table 1.

Real Time PCR primer sequences. The approved human gene nomenclature symbol is in parentheses if different from gene name.

Genes:	Forward Primer (sequence 5’−3’):	Reverse Primer (sequence 5’ −3’):
HPRT1	GGTCCATTCCTATGACTGTAGATTTT	CAATCAAGACGTTCTTTCCAGTT
β-ACTIN (ACTB)	CCAACCGTGAAAAGATGACC	ACCAGAGGCATACAGGGACA
GAPDH	GGTGAAGGTCGGTGTGAAC	CCTTGACTGTGCCGTTGAA
CCL2	CAGAAACCAGCCAACTCTCA	GTGGGGCATTAACTGCATCT
CCL3	TCCACGAAAATTCATTGCTG	AGATCTGCCGGTTTCTCTTG
CCL4	CATCGGAACTTTGTGATGGA	CACGATTTGCCTGCCTTTT
IL-1β (IL1B)	CAGGAAGGCAGTGTCACTCA	TCCCACGAGTCACAGAGGA
IL-6 (IL6)	AACTCCATCTGCCCTTCAGGAACA	AAGGCAGTGGCTGTCAACAACATC
TNFα (TNF)	CGTAGCCCACGTCGTAGC	GGTTGTCTTTGAGATCCATGC
COX-2 (PTGS2)	ACCAACGCTGCCACAACT	GGTTGGAACAGCAAGGATTT
cFos (FOS)	GGAATTAACCTGGTGCTGGA	TGAACATGGACGCTGAAGAG
CXCL10	GTGCTGCTGAGTCTGAGTGG	TTGCAGGAATGATTTCAAGTTTT
IL-15 (IL15)	CGATCTGGAAAATTGAAAGTC	CTGTACTCGTGCAAAATAACCTGT
GP91 Phox (CYBB)	TGTGACAATGCCACCAGTCT	TCTTGCATCTGGGTCTCCA
GFAP	CATCTCCACCGTCTTTACCAC	AACCGCATCACCATTCCTG
Iba1 (AIF1)	TCGATATCTCCATTGCCATTCAG	GATGGGATCAAACAAGCACTTC

Table 2.

CT values and geometric means of mRNA for three housekeeping genes obtained from the hippocampus in groups of rats for 4-day inflammatory mediator measurement. The number in parentheses represents the number of rats in each group.

Genes:	Non-seizure controls Soman (6)	Soman SE (8)	Non-seizure controls DFP (12)	DFP SE (11)
HPRT1	23.4 ± 0.3	23.9 ± 0.2	23.2 ± 0.3	22.6 ± 0.1
β-ACTIN	21.2 ± 0.3	20.9 ± 0.3	22.9 ± 0.3	22.5 ± 0.4
GAPDH	20.7 ± 0.3	20.6 ± 0.3	20.4 ± 0.3	20.1 ± 0.2
Geomean	21.8 ± 0.3	21.7 ± 0.2	22.1 ± 0.3	21.7 ± 0.2

Data are mean ± SEM

2.10. Western blot

Frozen cortices dissected from the half-brain of control rats and rats that experienced SE were homogenized on ice in 1.5 ml RIPA buffer (25 mM Tris HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS) containing a mixture of protease and phosphatase inhibitors (Roche Applied Science, Penzberg, Germany). The homogenates were placed on ice for 2 h and then centrifuged (14,000 ×g, 15 min, 4 °C). The protein concentration in the supernate was measured by Bradford assay (Thermo Fisher Scientific, Waltham, MA). The supernates (20 μg protein each) were resolved by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and electroblotted onto PVDF membranes (Millipore). Membranes were blocked with 5% (w/v) non-fat milk at 25 °C for 4 h and then incubated overnight at 4 °C with the primary antibody: rabbit anti-albumin (1:2000, Abcam, Cambridge, UK) and mouse anti-β-Actin (1:1000, Abcam). This procedure was followed by incubation with Near-InfraRed dye-conjugated secondary antibodies (IR-680 and IR-800; 1:2000, Li-Cor, Lincoln, NE) at room temperature for 2 h. The blots were scanned using a ChemiDoc MP imaging system (BioRad, Hercules, CA). The band intensities were quantified by ImageLab 6.0.1 (BioRad).

2.11. Calculation of cumulative pathology

For pathology that includes multiple measures such as weight loss, hippocampal neuroinflammation, gliosis and neurodegeneration a single geometric mean was obtained from the multiple measures for each rat. For example, for weight loss the geometric mean was determined for the percent weight on days 1–4. For hippocampal neuroinflammation a single geometric mean was calculated for all 10 inflammatory mediators (fold change of control). The cumulative gliosis measure was determined by calculating a geometric mean of microglia roundness in the hilus, CA1 and CA3 regions. Subsequently, the total gliosis geometric mean was calculated using the induction of the glial markers GFAP and IBA1 (fold change of control) along with the calculated geometric mean for microglial roundness. For BBB leakiness the single albumin fold change of control value was used. A single geometric mean was calculated for FJB positive counts using all 7 limbic areas (hilus, CA1, CA3, somatosensory cortex, thalamus, amygdala and piriform cortex). The overall amalgamated pathological score is the geomean of the five pathological characteristics (% weight, hippocampal neuroinflammation, hippocampal gliosis, BBB leakiness and neurodegeneration).

2.12. Data analysis

Data obtained from albumin western blot and qRT-PCR are presented as box plots with individual points representing each rat for glial markers. Individual points for each animal are also shown for inflammatory mediators. Data obtained from microglia roundness and FJB quantification are presented as means ± standard error with individual points shown. Statistical analysis was performed with GraphPad Prism version 7 (GraphPad software, San Diego, CA). Student’s t test, Mann-Whitney test, or one-way ANOVA (with Bonferroni or Sidak’s posthoc tests) were performed as appropriate based on the questions asked and the comparisons made to examine differences of chemical effects. Student’s t test was used for comparisons made with parametric data and the Mann-Whitney test for comparisons made with nonparametric data for only two group data sets. The Bonferroni and Sidak posthoc was used to compare multiple selected pairs of treatment groups. The differences were considered to be statistically significant if p < .05. The Shapiro-Wilk test in Origin 9.4.2 (OriginLab, Northampton, MA) was used to test normality of the data. Experimental power analysis was performed with GPower3.1.9.2 (Universitat Kiel, Germany). Estimation statistics was performed for qRT-PCR mRNA fold change of inflammatory mediators from hippocampi using an internet application (http://www.estimationstats.com/#/) built for common use by Dr. Adam Claridge-Chang and Mr. Joses Ho. The application performs data analysis with a bootstrap estimation (DABEST) package in python¹⁵ with 5000 resampling, bias-corrected and accelerated bootstrap analysis to determine the confidence interval of differences between treatment groups. The mean differences between groups, the confidence interval from the bootstrapping and the p-values obtained using the Mann-Whitney test to compare treatments are presented in Table 3. The confidence intervals provide a plausible range for the mean of the group and predict the outcome of numerous replications of the experiment.

Table 3.

Estimation statistics for qRT-PCR mRNA fold change from the hippocampus with 5000 bootstrap iterations

Analyte/treatment	Mean Fold Change ± SEM	Mean difference (compared to DFP)	95% confidence interval	p-value
IL-1β
Soman (8)	0.9 ± 0.2	3.0	2.0 – 5.3	0.0003
DFP (11)	3.9 ± 0.7	-	-	-
TNFα
Soman (8)	1.6 ± 0.4	1.4	0.4 – 2.6	0.023
DFP (11)	3.0 ± 0.5	-	-	-
IL-6
Soman	2.6 ± 0.7	−0.1	−1.8 – 1.2	0.65
DFP	2.5 ± 0.3	-	-	-
COX-2
Soman	1.2 ± 0.3	1.7	0.7 – 2.7	0.007
DFP	2.9 ± 0.4	-	-	-
cFos
Soman	3.0 ± 0.5	−1.2	−2.2 – 0.06	0.11
DFP	1.8 ± 0.3	-	-	-
CCL2
Soman	26 ± 9	29	3.3 – 56.8	0.05
DFP	55 ± 11	-	-	-
CCL3
Soman	5 ± 1	9	5.5 – 14.6	0.001
DFP	14 ± 2	-	-	-
CCL4
Soman	3 ± 0.7	7	4.6 – 9.5	0.0004
DFP	10 ± 1	-	-	-
CXCL10
Soman	4.7 ± 2	7.5	1 – 12.3	0.009
DFP	12.3 ± 2	-	-	-
Gp91phox
Soman	8 ± 1.7	3	−1.23 – 7.2	0.15
DFP	11 ± 1.5
IL-15
Soman	1.6 ± 0.2	0.1	−0.5 – 0.8	0.90
DFP	1.7 ± 0.3	-	-	-

SEM = standard error of mean. P-values were obtained by a Mann-Whitney test comparing Soman to DFP.

3. RESULTS

3.1. Soman and DFP-induced status epilepticus in rats

Experiments were performed using two organophosphorus seizure models (DFP and soman) in rats to compare seizure-induced neuropathology. DFP was administered to 53 adult male Sprague-Dawley rats, of which 27 entered SE resulting in a 51% success rate (Figure 1B). As observed in previous studies, a large subset of rats injected with DFP succumbed within minutes to the effects of acute DFP exposure (26 rats), mainly by respiratory cessation. In this study, these rats are labeled “no SE” as they died prior to the onset of SE. In total, 42 of 53 rats either died within minutes of DFP administration, during the ensuing SE, or within 4 days after SE had waned (79% total mortality, i.e., the sum of mortality prior, during and post SE) (Figure 1C). However, more than half of the total mortality (49.1%) is attributed to rats that died from the acute toxic effects of DFP prior to the onset of SE.

Figure 1. Experimental paradigm of chemical administration in rat models of DFP and soman-induced SE.

A, for DFP all rats were administered pyridostigmine bromide and atropine methyl nitrate followed by a single injection of DFP to induce SE. The rats experienced uninterrupted SE and eventually the seizures waned over time. B, shown is the percent of rats entering SE following a single intraperitoneal injection of DFP. C, percent mortality before, during and after SE. Mortality prior to the onset of SE is attributed to acute respiratory arrest caused by DFP. D, for soman all rats were administered the asoxime chloride (HI-6) followed by a single injection of soman to induce SE. One minute after soman administration all rats were administered atropine methyl nitrate as a supporting agent to help the rats survive. Forty minutes after SE onset the rats were administered a combination of drugs [atropine sulfate, pralidoxime (2-PAM) and midazolam] to interrupt SE and aid in survival. E, shown is the percent of rats entering SE following a single subcutaneous injection of soman. F, percent mortality before and during SE induced by soman. Mortality prior to the onset of SE was absent in the soman model. The number of rats that experienced SE used for the 4-day experiments can be deduced by subtracting the total number of rats that died in panels C and F (both shades of gray) from the number of rats that experienced SE in panels B and E (white only), respectively. G, a single injection of DFP and soman causing SE resulted in significant weight loss over the next 4 days compared to non-seizure control rats. The weight loss observed in soman injected rats (blue symbols) was not different than the weight loss observed for DFP injected rats (red symbols) (p = 0.23 repeated measures two-way ANOVA). Similarly, the weight change for non-seizure control rats for soman and DFP were very similar (p = 0.59 repeated measures two-way ANOVA). Shown are the mean ± standard error of the mean. The dashed line indicates the original weight of the rats before any manipulation on day 0.

A separate cohort of rats was used for the rat soman model of SE. Soman was administered to 19 adult male Sprague-Dawley rats, of which 100% of the rats entered SE (Figure 1E). Unlike DFP, SE onset occurs within ~5 minutes of soman administration making it unlikely for rats to die prior to SE. On the other hand, the average latency to SE onset for DFP injected rats was ~45 minutes consistent with previous studies^6–8. Similar to experiments with DFP a subset of rats injected with soman died during SE. In fact, nearly all of the soman-induced mortality is attributed to rats that succumbed within 6 hr after SE onset. All rats that died following soman exposure died on experiment day 0. The total mortality for rats that entered SE was very similar for the two OP models [57.9% (11 of 19 rats) for soman vs. 59.3% (16 of 27 rats) for DFP] (Figure 1C, F).

All rats that survived SE were monitored and weighed daily the next four days. Weight change following SE often correlates with functional recovery in rats after DFP or soman^{6, 8, 16, 18–21}. On days 1 and 2 rats that experienced SE regardless of whether they were injected with soman or DFP lost a similar percent of body weight (~11% on day 1 and ~17% on day 2; Figure 1G). Similarly, but to a lesser degree rats that were injected with DFP but did not experience SE displayed a very similar weight change. On day 3 DFP-injected rats that experienced SE showed an increase in weight that continued on day 4 (Figure 1G). By contrast, rats experiencing SE following administration of soman continued to lose weight on days 3 and 4 (Figure 1G). On day 4 the percent of pre-SE body weight was 87 ± 4% for DFP-SE rats, n = 11 and 79 ± 3% for soman-SE rats, n = 8. However, there was no significant difference in the overall weight change over the four days for rats that experienced soman -induced SE compared to DFP injected rats (p = 0.23, two-way ANOVA repeated measures) (Figure 1G), suggesting a similar impact of SE on functional recovery in both models. It should be noted that none of rats in the DFP-SE and soman-SE groups reached the weight loss endpoint of ≥ 25% and ≥ 30% of body weight from baseline, respectively.

3.2. Cortical seizure activity following soman exposure

The cortical EEG temporal profile for rats that experience uninterrupted SE following DFP exposure has been reported¹⁰. Adult male rats were instrumented with bilateral cortical electrodes prior to EEG recording and soman exposure. The aim of this experiment was to compare the cortical EEG profile for rats that experience soman-induced SE to the cortical EEG data previously obtained with DFP. On the day of soman exposure, rats were connected to the EEG instrument and their baseline brain activity was recorded for 2 h prior to soman. All rats tested displayed normal cortical activity prior to any drug administration as determined by the low amplitude, mixed frequency and shape of the waveforms (Figure 2A–C). Cortical EEG was recorded continuously for 6 h after soman administration in all rats. Exposure to soman led to SE with the appearance of seizures consisting of large amplitude and high frequency spikes (>2x the baseline prior to drug treatment) persisting for more than 10 sec. All rats received a single injection of soman and the latency to SE onset was very similar (Figure 2A-C). Forty minutes after SE onset the rats were administered a combination of atropine sulfate (0.45 mg/kg), 2-PAM (25 mg/kg) and midazolam (1.8 mg/kg) intramuscularly. This combination of countermeasures reduced EEG amplitude in all rats, but did not terminate electrographic SE as there was no apparent return to basal level during the 6 h recording. As a result, SE was prolonged in the soman injected rats. EEG power in the 0–60 Hz band integrated in 300 sec epochs over the entire 8h EEG recording was obtained for every soman injected rat. A plot of the mean EEG power as a function of time revealed persistent high frequency seizure activity in all rats consistent with prolonged SE lasting for several hours (Figure 2D). The cortical EEG temporal profile for rats that experienced soman-induced SE in this study was very similar to the profile for rats that experienced DFP-induced SE that was not terminated pharmacologically¹⁰. The remaining results below compare the delayed effects such as neurodegeneration, neuroinflammation, gliosis, BBB breakdown and weight loss that develop because of the prolonged SE experience in soman and DFP injected rats.

Figure 2. Cortical EEG recording following soman-induced SE.

Cortical electroencephalography (EEG) activity was recorded for 8 hr prior and during SE induced by exposure to soman. Representative EEG traces (in blue) from the cortical recording of three adult male rats (A1, B, C) showing increased spike activity just after exposure to soman (black triangle) that develops into SE (denoted by the bar at the top). The black dot represents the time of injection of a combination of drugs [atropine sulfate (0.45 mg/kg, im), 2-PAM (25 mg/kg, im) and midazolam (1.8 mg/kg, im)] to attenuate SE and promote survival. Below the raw EEG trace is a sonogram of the spike activity obtained in Spike2. The colors of the sonogram indicate the spectral power density in decibels (dB) at the indicated frequency. A2, shows magnified 5-minute sections of the recording taken from the baseline prior to soman (a), at the peak intensity of SE (b) and at the end of the recording (c) in A1. A3, shows a 1 min interval of the recording during SE. The area within the black outlined box in panel b of A2 was magnified and shown below. D, the EEG power in the 0–60 Hz bandwidth averaged over 300 sec epochs during the 8 h period for soman injected rats (n = 8 rats). The dashed line at the bottom indicates baseline power before soman administration. Data are the mean ± standard error of the mean.

3.3. Seizure-induced upregulation of cyclooxygenase-2

In rats COX-2 is induced within minutes following the onset of OP-induced SE and remains elevated for days after SE is terminated^{6, 7, 10, 16, 22}. To compare the extent of COX-2 induction in the two models, rats that experienced soman or DFP-induced SE were euthanized 4 d after SE onset and one brain hemisphere was processed for immunohistochemistry. Fluorescence immunohistochemistry performed on coronal hippocampal sections (40 μm) obtained from non-seizure control rats revealed a low basal level of COX-2 in the hippocampus, piriform cortex and the amygdala (Figure 3A). Rats that experienced soman-induced SE or uninterrupted SE induced by DFP displayed robust upregulation of COX-2 in neurons 4 d following SE onset (Figure 3B). COX-2 was upregulated in both models in the same seizure sensitive limbic brain regions such as the CA3 cell layer of the hippocampus, the piriform cortex and the basolateral amygdala (Figure 3B). The number of cells displaying seizure-induced COX-2 upregulation in the CA3 region of the hippocampus and the basolateral amygdala appeared lower in rats that experienced soman-induced SE compared to rats that received DFP (Figure 3B).

Figure 3. Induction of inflammatory mediators 4 days after DFP and soman-induced SE in rats.

Images of COX-2 staining (green) in 40 μm sections obtained from a control (non-seizure) rat (A) and rats that experienced SE following soman and DFP administration (B) all euthanized on day 4. Images were obtained from the cornu ammonis 3 (CA3) region of the hippocampus, piriform cortex (PC) and basolateral amygdala (BLA) at 50x (left panels) and 200x (enlarged right panels) total magnification and are representative of five sections each from three rats per treatment. Scale bar = 125 μm and 30 μm (left to right). The white boxes in the basolateral amygdala shows COX-2 staining magnified in the panels on the right. The arrows indicate cells where COX-2 is upregulated in neurons of rats that experience SE compared to non-seizure controls. C, qRT-PCR was used to quantify the change in abundance of mRNAs of inflammatory mediators from non-seizure control rats and rats that experienced SE induced by DFP (red) or soman (blue) 4 d earlier. The mRNA fold change of cyclooxygenase 2 (COX-2), tumor necrosis factor alpha (TNFα), chemokine (C-C motif) ligands 2, 3 and 4 (CCL2, CCL3 and CCL4), C-X-C motif chemokine 10 (CXCL10), Interleukin-6 (IL-6) and the NADPH oxidase 2 (gp91phox) remained elevated above control levels in the hippocampus of rats that experienced SE. Although the fold changes are shown statistical analysis was carried out using the ΔΔCT values. Hippocampal inflammation 4 d after DFP-induced SE was significantly higher compared to rats that experienced soman-induced SE (p =.001, repeated measures ANOVA with Sidak’s multiple comparisons test). Images of cFos staining (red) are shown in a hippocampal section (40 μm) obtained from a control rat (D1) and rats that experienced DFP (D2) and soman-induced SE (D3) all euthanized on day 4. The images are representative of three sections from at least three rats in each group taken at 100x total magnification. Scale bar = 350 μm. CA1, cornu ammonis 1; DG, dentate gyrus; SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum. E, the average mRNA fold change of cFos in the hippocampus of rats that experienced SE was elevated above non-seizure controls (black dashed line). However, the induction of cFos mRNA in the hippocampus was not different for rats that experience soman-induced SE (blue) compared to DFP injected rats (red), p = 0.1, t-test using ΔΔCT values. The symbols represent each individual rat within the group.

3.4. Seizure-induced hippocampal neuroinflammation

Rats that experienced SE after administration of soman and DFP were euthanized 4 d after the onset of SE and one hippocampus from each rat was processed for RNA extraction and purification. qRT-PCR was carried out to measure the expression of a selected panel of 10 inflammatory mediators (Table 1), a majority of which were previously shown to be upregulated in rats 4 days following uninterrupted DFP-induced SE⁷. Rats that experienced uninterrupted SE following injection of DFP displayed a more robust increase in the average mRNA level of inflammatory mediators investigated in the hippocampus compared to soman injected rats (11 ± 5 average fold induction for DFP treated rats, n = 11 rats vs. 5 ± 2 average fold induction for soman treated rats, n = 8 rats; p = .001, repeated measure ANOVA with Sidak’s multiple comparisons test using ΔΔCT values) (Figure 3C). However, no difference was detected in expression of the housekeeping genes (β-actin, GAPDH and HPRT1; Table 2). Individually, mRNA levels in the hippocampus of DFP-SE rats were 55-fold above non-seizure controls for CCL2, 14-fold for CCL3, 10-fold for CCL4, 12-fold for CXCL10, 3-fold for TNFα and 11-fold for NOX2 (gp91phox) (Figure 3C). By comparison, for rats that experienced SE following soman administration mRNA levels were 26-fold above non-seizure controls in the hippocampus for CCL2, 5-fold for CCL3, 3-fold for CCL4, 5-fold for CXCL10, 1.6-fold for TNFα and 8-fold for NOX2 (gp91phox) (Figure 3C). For individual comparisons a significant difference in mRNA upregulation comparing soman treated rats to DFP administered rats was observed for COX-2 (2.9 ± 0.4 fold induction for DFP treated rats, n = 11 rats vs. 1.2 ± 0.3 fold induction for soman treated rats, n = 8 rats; p = 0.007, Mann-Whitney test, estimation statistics) (Figure 3C, Table 3). Similarly, the SE-induced upregulation of IL-1β was lower in rats that experienced soman-induced SE (3.9 ± 0.7 fold induction for DFP treated rats, n = 11 rats vs. 0.9 ± 0.2 fold induction for soman treated rats, n = 8 rats; p = 0.0003, Mann-Whitney test, estimation statistics) (Figure 3C, Table 3).

FBJ murine osteosarcoma viral oncogene homolog (cFos) is an immediate early gene that is activity dependent. Fos is often used as an indicator of neuronal discharge and Fos staining can indicate seizure activity or intensity¹³. cFos staining can indicate whether hippocampal activity remained elevated 4 d following SE. Fluorescence immunohistochemistry performed on coronal hippocampal sections (40 μm) obtained from rats that experienced DFP and soman-induced SE 4 d earlier revealed high expression of the cFos protein in the CA1 cell layer of the hippocampus(Figure 3D). On the other hand, cFos levels are low in non-seizure controls rats (Figure 3D). The level of cFos mRNA was also induced in the hippocampus of a majority of rats that experienced soman and DFP-induced SE. Although the mean induction of cFos appeared higher for rats injected with soman, due to high variation in the data no significant difference was detected comparing DFP-SE rats to soman-SE rats (1.8 ± 0.3 fold change for DFP-SE rats, n = 11 vs. 3 ± 0.5 fold change for soman-SE rats, n = 8; p = 0.1, student’s t-test using ΔΔCT values) (Figure 3E). Taken together, these data suggest that exposure to soman or DFP and the subsequent SE induces an inflammatory cascade involving COX-2 in rats. However, neuroinflammation measured on day 4 was lower in soman injected rats.

3.5. Seizure-induced gliosis

Experiments were performed to investigate gliosis in rats that experienced soman and DFP-induced SE. Intense immunostaining of Iba1 (a microglia marker) was observed in seizure sensitive brain regions in coronal sections of rats that experienced SE but not in non-seizure controls (Figure 4A, B). qRT-PCR was used to quantify changes in mRNA of astrocytic GFAP and microglial Iba1 in rat hippocampi 4 d after SE. Changes in the mRNA levels of these markers can be indicative of the level of astrogliosis and microgliosis. qRT-PCR revealed that mRNA of GFAP and Iba1 were upregulated above non-seizure control levels in the hippocampus of rats that experienced SE 4 d earlier (Figure 4C). The average induction of GFAP mRNA was significantly higher in rats that experienced SE following soman compared to rats that were injected with DFP (17.3 ± 3 fold above control for soman, n = 8 rats vs. 5.8 ± 1 fold above control for DFP, n = 10 rats; p = .01, one-way ANOVA with Sidak’s multiple comparisons test using ΔΔCT values) (Figure 4C). The average level of Iba1 mRNA in the hippocampus appeared higher for rats administered soman compared to rats that experienced SE following DFP administration [17.9 ± 9 fold above control (n = 8 rats) for soman treated rats vs. 6 ± 0.6 fold above control (n = 10 rats) for DFP treated rats] (p = 0.7, one-way ANOVA with Sidak’s multiple comparisons test using ΔΔCT values) however, statistical significance was not attained due to the high variability in the soman group (Figure 4C).

Figure 4. Gliosis induced by DFP and soman 4 days after SE.

Images taken at 50x total magnification of ionized calcium-binding adapter molecule 1 (Iba1) staining (green/red) in 40 μm sections from a non-seizure control rat (A) and rats that experienced DFP and soman-induced SE (B) all euthanized on day 4. The images shown are a representative of five sections per rat for each treatment in the cornu ammonis 1 (CA1), piriform cortex (PC) and basolateral amygdala (BLA) revealing induction of Iba1 in rats that experience SE. White dashs outline the principal neuron layer. SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum. Scale bar = 125 μm. C, induction of GFAP and IBA1 mRNA in the hippocampus of rats that experienced DFP (red; n = 10 rats) and soman-induced SE (blue; n = 8 rats) all euthanized on day 4. The symbols represent each individual rat within the group. The black horizontal bar within the symbols (middle) represents the mean. Although fold change is shown statistical analysis was performed on the ΔΔCT values by one-way ANOVA with Sidak’s multiple comparisons test (DFP-SE vs. Soman-SE). D, confocal images taken at 200x total magnification of Iba1 staining (green/red) in the CA3 region (SR, stratum radiatum) of hippocampal sections (40 μm) obtained from a non-seizure control rat and rats that experienced DFP and soman-induced SE all euthanized on day 4. Scale bar = 50 μm. The arrows indicate typical resting microglia in controls and activated microglia following SE. E, quantification of microgliosis defined by the roundness of the cells measured in Pixcavator IA. The average roundness of microglia from DFP injected rats (n = 9) and soman injected (n = 7–8) rats were compared by two-way ANOVA with Tukey’s multiple comparisons test. The average microglia roundness in the CA1 of one rat in the soman-SE group was identified as an outlier and removed from the analysis reducing the number to 7.

High magnification images of fluorescent Iba1 immunohistochemistry performed on coronal hippocampal sections taken from half brains of rats 4 d after SE revealed higher microglia cell body swelling in the CA3 region (stratum radiatum) of the hippocampus of rats that experienced SE compared to non-seizure control rats (Figure 4D). Iba1 is a calcium-binding adapter protein that exists in the cytoplasm of microglia; Iba1 labeling in brain section can be used to reveal microglia morphology. Quantification of microgliosis was performed by determining the morphological roundness of microglia as a feature of glial activation as the cell body of microglial cells swell during gliosis. Using Pixcavator IA the average roundness of microglia in the CA3 region of non-seizure controls rats was 10.2 ± 0.5 (n = 6 rats) whereas, the average roundness of microglia in the CA3 region of rats that experienced SE was 15.9 ± 0.7 (n = 8 rats) for soman and 17.9 ± 0.5 (n = 9 rats) for DFP (p = 0.03, two-way ANOVA with Tukey’s multiple comparisons test). The average roundness of microglia was also significantly lower in the CA1 and hilus for rats injected with soman compared to rats that experienced uninterrupted DFP-induced SE (Figure 4E, Supplemental Figure 1). Taken together, these data suggest that microgliosis as defined by the amoeboid appearance of the microglia is higher in rats that experience uninterrupted SE after DFP administration compared to rats injected with soman.

3.6. Blood-brain barrier leakiness after SE

The blood-brain barrier (BBB) is a physical barrier that restricts macromolecules from entering the central nervous system from the periphery and is important for maintaining homeostasis. The integrity of the BBB may be compromised because of prolonged SE. Leakiness of the BBB can be determined by perfusing animals with a buffered saline solution sufficient to remove all traces of blood from the brain, and subsequently measuring the extravasation of serum albumin in the brain parenchyma. Upon removal of blood from the brain, serum albumin is expected to be low or absent in the brain parenchyma of an animal with an intact BBB. Serum albumin levels were elevated above non-seizure controls in the cortex of rats 4 d after soman-induced SE measured by western blot (Figure 5A), which is consistent with breakdown of the BBB. The average normalized fold increase in serum albumin detected in the cortex was 2.9 ± 0.4 (n = 8) fold above non-seizure controls for soman injected rats (Figure 5B). Albumin levels were also high in the brain parenchyma of rats that experienced uninterrupted SE after injection of DFP (4 ± 0.9 fold above non-seizure controls). No significant difference was detected in the levels of albumin comparing DFP-SE to soman-SE rats (p = 0.69, one-way ANOVA with Sidak’s multiple comparisons test), suggesting a similar degree of BBB leakiness in the two rat OP models of SE (Figure 5B).

Figure 5. Blood-brain barrier leakiness.

A, the amount of serum albumin in the cortex 4 days after soman-induced SE was used to assess the integrity of the blood–brain barrier as all rats were perfused with sterile saline to completely remove blood from all tissues. The albumin levels (green bands, above) in the cortex of non-seizure controls (labeled with C) or rats that experienced SE induced by soman (labeled with S) were measured by western blot with GAPDH (red bands, below) used as a loading control. B, the band intensity of albumin was normalized to the housekeeping GAPDH. The symbols represent each individual rat within the group. p = 0.69, one-way ANOVA with Sidak’s multiple comparisons test (comparing DFP-SE vs soman-SE). The western blot data in panel B displaying albumin and GAPDH levels in DFP non-seizure controls (n = 8) and rats that experienced uninterrupted SE with DFP (n = 11) was previously shown in a recently published review¹⁶.

3.7. Seizure-induced neurodegeneration

Experiments were performed to determine the extent of neuronal injury using FJB staining performed on coronal hippocampal sections (40 μm) taken from the brains of rats euthanized 4 d after SE onset following DFP or soman exposure. Neurodegeneration was detected by positive FJB staining in the CA1, CA3 and hilus regions of the hippocampus of rats that experienced SE (Figure 6A). To determine whether hippocampal neurodegeneration was similar following soman and DFP, FJB positive cells were counted in the hilus, CA1 and CA3 by an observer blinded to the treatment of the rats. In CA1 the total number of FJB positive cells per section was similar in rats that experienced soman-induced SE compared to rats administered DFP (153 ± 28 FJB positive cells per hippocampal section for soman treated rats, n = 8 rats vs. 168 ± 42 positive cells per hippocampal section for DFP injected rats, n = 9 rats; p = .9, Mann-Whitney test) (Figure 6A, C). There was also no difference in neurodegeneration observed in the CA3 of soman treated rats compared to DFP injected rats (80 ± 33 FJB positive cells per hippocampal section for eight soman treated rats vs. 67 ± 19 positive cells per hippocampal section for nine DFP injected rats), (p = 1, Mann-Whitney test) (Figure 6A, C). Analysis of FJB staining in the hilus revealed significantly less neurodegeneration in soman-SE rats compared to DFP-SE (37 ± 5 FJB positive cells per section for soman treated rats, n = 8 vs. 60 ± 2 FJB positive cells per section for DFP-injected rats, n = 9) (p < .0001, Mann-Whitney test) (Figure 6A, C). Overall, the numbers of FJB positive principal neurons in the hippocampus were similar in rats that experienced SE following soman and DFP exposure.

Figure 6. Neurodegeneration in limbic brain regions.

A, representative images of positive FluoroJade B (FJB) staining in hippocampal sections (40 μm) in the cornu ammonis 1 (CA1), cornu ammonis 3 (CA3) and hilus regions 4 days after soman (top) and DFP-induced SE (bottom) in rats. SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum. Scale bar = 200 μm. B, representative images of positive FluoroJade B staining in coronal sections (40 μm) in the thalamus, basolateral amygdala (BLA) and piriform cortex regions 4 days after soman (top) and DFP-induced SE (bottom) in rats. The images were taken at 50x total magnification. The images are representative of 5 dorsal hippocampal sections per rat. The white dashed shapes outline the BLA. Scale bar = 200 μm C, the average number of injured neurons per section 4 d after SE in three dorsal hippocampal regions (hilus, CA1 and CA3) of rats that experienced soman-induced SE (blue bar, open squares, n = 8 rats) and rats that experienced DFP-induced SE (red bar, open circles, n = 9 rats) (p < 0.0001 in the hilus, p = 0.9 in CA1 and p = 1 in CA3, by Mann-Whitney test comparing soman to DFP). The bars show the mean ± standard error of the mean. The symbols represent each individual rat within the group. D, the average number of injured neurons per section 4 d after SE in the somatosensory cortex (SS), thalamus, piriform cortex (PC) and basolateral amygdala (BLA) of rats that experienced soman-induced SE (blue bar, open squares, n = 8 rats) and rats that experienced DFP-induced SE (red bar, open circles, n = 9 rats) (p = 1 in the somatosensory cortex, p = 0.9 in the thalamus, p = 0.04 in the piriform cortex and p = 0.3 in the basolateral amygdala, by student’s Mann-Whitney test comparing soman to DFP).

Positive FJB staining was observed and quantified in other seizure sensitive brain regions including the somatosensory cortex, thalamus, piriform cortex and basolateral amygdala. Robust neurodegeneration was detected in these brain regions in all rats that experienced SE (Figure 6B, D). However, in the piriform cortex FJB positive cells in rats exposed to soman was significantly lower compared to rats that experienced SE following DFP administration (62 ± 16 FJB positive cells per section for soman treated rats, n = 8 rats vs. 124 ± 20 positive cells per section for DFP-SE rats, n = 9 rats; p = 0.04, Mann-Whitney test) (Figure 6B, D). All other areas investigated (thalamus, somatosensory cortex and basolateral amygdala displayed a similar average number of FJB positive cells per section in all SE rats regardless of whether they were injected with soman or DFP (Figure 6D). Positive FJB cells were absent in non-seizure control rats that were administered sterile water instead of DFP or rats that were administered sterile saline instead of soman (not shown). Taken together, these data indicate that neurodegeneration detected by FluoroJade B is present in seizure sensitive brain regions and the distribution and number of degenerating cells differ in particular seizure sensitive brain nuclei for rats that experience SE following soman compared to DFP.

3.8. Comparison of combined pathology

To determine whether there is an overall difference in pathology of DFP versus soman injected rats the geometric mean of 5 pathological characteristics that occur as a consequence of prolonged SE was determined for each rat that experienced SE. The pathological geometric mean values ranged from 8 to 21 (Figure 7). No obvious difference was detected in the overall pathology for rats that experienced soman-induced SE compared to rats that experienced uninterrupted SE after DFP injection (15 ± 1 for soman-SE rats, n = 8 vs. 17 ± 1 for DFP-SE rats, n = 9; p = 0.09, student’s t-test) (Figure 7).

Figure. 7. Comparison of amalgamated pathology.

A geometric mean was obtained using 5 hallmark characteristics of OP-induced SE (weight loss, hippocampal neuroinflammation, gliosis, BBB integrity and neurodegeneration) for rats that experienced uninterrupted DFP-induced SE (n = 9 rats) and rats administered soman (n = 8 rats). The data are the mean and standard error of the mean. The symbol represents each individual rat within the group. p = 0.09 by student’s t-test.

4. DISCUSSION

Exposure to high levels of a nerve agent can result in death within minutes. Recently, there has been an increase in cases involving the use of nerve agents as deadly weapons. Nerve agents are extremely toxic due to their ability to rapidly and effectively inactivate the enzyme acetylcholinesterase. The high toxicity of nerve agents like soman limit their use in academic research laboratories investigating the underlying mechanisms of seizure-induced neuropathology that develop following nerve agent poisoning. Instead, other less toxic organophosphorus based agents like diisopropylflourophosphate (DFP) are often used as a surrogate in academic research laboratories that are not associated with the military to mimic nerve agent poisoning in rodents. DFP is a recommended surrogate for nerve agent poisoning for several reasons. DFP is structurally similar to the nerve agents sarin and soman. The mechanism of action of DFP (bind, inhibit and age cholinesterases) is similar to nerve agents and like the nerve agents exposure to high levels of DFP often leads to the development of seizures, status epilepticus, death if left untreated, and neuropathology in survivors¹⁶. DFP ages acetylcholinesterase more slowly compared to soman, however exposure to high levels of DFP eventually results in potent inhibition of acetylcholinesterase, a cholinergic crisis and SE in rats^{6, 7, 10, 16, 23–30}. Here, seizure-induced neuropathology was compared in adult male rats administered soman or DFP to produce SE. Five major neuropathological hallmarks (weight loss, hippocampal neuroinflammation, gliosis, BBB integrity and neurodegeneration) of OP-induced SE were compared. A single total neuropathological score for each rat was achieved by combining the five measures. Although there were differences detected in the five pathological measures, the combined pathology analysis revealed that rats that experience DFP-induced SE shared a similar pathology as soman injected rats. Collectively, the data support use of the rat DFP model of SE as a suitable surrogate for investigating some, but not all delayed neuropathology produced by nerve agents.

Prior to comparing the neuropathology, we determined whether soman and DFP injected rats shared a similar SE experience. Although diazepam was administered to a subset of rats to interrupt or terminate DFP-induced SE in a recent study¹⁰, the EEG profile of the soman injected rats closely resembled that of rats that experienced uninterrupted DFP-induced SE as the initial bout of SE extended for greater than 5 hours. In the rat DFP model of SE rats that experienced a longer duration of seizure activity displayed a greater neuropathology¹⁰ consistent with a separate but similar study³¹ demonstrating that a longer duration of the initial SE causes more intense sequelae. Therefore, we compared the seizure-induced neuropathology 4 days after SE induced by soman and DFP. Investigation of neuropathology 4 days post SE allowed sufficient time for the neuropathology to develop and the rats to recover from any lingering seizure activity from the initial bout of SE. Significant differences in neuropathology were observed in the two rodent models of SE. For example, all rats administered soman experienced SE whereas only 51% of rats injected with DFP experienced SE (Figure 1B, E). It should be noted that cortical EEG activity was used to identify SE for soman injected rats whereas in this study SE was declared in rats administered DFP by behavioral observation using a modified Racine scale. It is common that some rats administered DFP fail to experience SE^{6–8, 10, 16}. In the current study, the acute mortality following intraperitoneal DFP administration was high, but within the normal variation detected in our laboratory over the past 10 years.

Weight loss and regain was similar in soman and DFP injected rats that experienced SE 4 days earlier. Rats that experienced SE following administration of DFP displayed a trend of weight regain on days 3 and 4 post SE whereas rats exposed to soman continued losing weight daily, giving the appearance that adult rats functionally recover faster from DFP-induced SE compared to rats injected with soman. However, it should be noted that the variation in the weight change for the DFP injected rats was high and thus no statistical significance was attained comparing the weight change of the SE rats.

Another difference between rats that experienced soman-induced SE and rats that endured uninterrupted SE following injection of DFP is that the overall inflammatory burst detected in the hippocampus was slightly higher in DFP injected rats as measured by qRT-PCR four days after SE. This suggests that the return of the inflammatory mediators to baseline levels may occur faster following soman-induced SE compared to DFP or the inflammatory response is slightly impaired in soman injected rats. As a result, the DFP-SE model might be more suitable for investigating neuroinflammation following OP-induced SE.

Microgliosis and astrogliosis were prominent in the hippocampus four days after soman and DFP-induced SE consistent with previous studies^{24, 32} demonstrating OP-induced SE results in a robust glial response in the days after SE. The induction of Iba1 mRNA was similar among rats that experienced SE regardless of whether they were exposed to soman or DFP (Figure 3). However, the roundness of microglia following SE was significantly higher in all three hippocampal regions (hilus, CA1 and CA3) in rats that experienced DFP-induced SE compared to soman injected rats. The higher gliosis is consistent with a higher level of inflammatory mediators in rats that experienced DFP-induced SE. Microglia can remain activated for a long period following SE. Perhaps microglial activation did not reach a plateau and was still increasing on day 4 in rats administered soman. To resolve this issue it is necessary to obtain a more complete temporal profile of microglial activation beyond 4 days caused by soman and DFP-induced SE.

BBB breakdown and neurodegeneration measured on day 4 were nearly identical in soman and DFP injected rats that experienced SE. All rats that experienced SE displayed similar elevated levels of serum albumin in the cortex four days after SE compared to non - seizure controls, which is consistent with a leaky BBB. The similar level of serum albumin detected in the cortex of DFP and soman injected rats suggest that the lower level of neuroinflammation observed in soman injected rats is not caused by attenuation of breakdown of the BBB. However, it is important to mention here that it is unknown when the BBB becomes leaky or restored following OP-induced SE. Perhaps soman-induced SE breakdown and/or repair of the BBB occurs earlier than following DFP accounting for the differences in neuroinflammation and gliosis. In the future, it will be important to investigate the time-dependent breakdown and restoration of the BBB following OP-induced SE in rats.

We recently demonstrated that neuroinflammation is associated with the ensuing neurodegeneration and breakdown of the BBB in adult rats following DFP-induced SE^{6, 8, 16}. In the current study, of all brain regions investigated, a significant difference in neurodegeneration was observed in the hilar subregion of the hippocampus and the piriform cortex, which displayed less neurodegeneration following soman injection. Hippocampal hilar neurons die shortly after SE onset (within 5–6 hr)¹⁶. Perhaps neurons in the hilus and the piriform cortex die faster after soman-induced SE compared to DFP resulting in fewer FJB positive cells observed 4 days after SE in the soman model. Nevertheless, the overall degree of neurodegeneration observed was high in all SE rats and is likely a result of the prolonged SE experience. The high number of FJB positive cells in limbic brain regions in the current study is consistent with previous studies demonstrating robust neurodegeneration in seizure sensitive brain regions after soman and DFP-induced SE^{6, 7, 32–40}.

5. CONCLUSION

Exposure to high levels of OP agents causes SE, which if left untreated results in neuropathology in surviving rats. Comparison of the pathology that develops days after soman and DFP-induced SE revealed similarities and differences in rats. The models differed in the degree of hippocampal neuroinflammation, gliosis and neurodegeneration in the piriform cortex and the hilus. These differences are important and should be further investigated with a more precise temporal profile. However, the other pathological measures were very similar between soman and DFP injected rats. When combined the overall pathology scores did not differ for both OP-based agents. Therefore, the rat DFP model of SE is a suitable surrogate for investigating various delayed consequences produced by nerve agents. In the future, it will be important to determine whether other OP based agents like pesticides display a similar magnitude of seizure-induced pathology. This information could be useful for the development of agent independent effective medical countermeasures to improve survival and mitigate the ensuing brain injury following OP poisoning.

Highlights.

• Electrographic status epilepticus is similar for rats administered soman and DFP

• Neuroinflammation and gliosis is lower after soman-induced SE compared to DFP

• Weight change, BBB leakiness and cell death is similar for soman and DFP-induced SE

• The overall SE-induced pathology was similar in rats administered soman and DFP

ABBREVIATIONS

SE

status epilepticus

GFAP

glial fibrillary acidic protein

IBA1

ionized calcium-binding adapter molecule 1

CA1

Cornu Ammonis 1

CA3

Cornu Ammonis 3

CT

cycle threshold

Con

control

COX-2

cyclooxygenase 2

cFos

FBJ murine osteosarcoma viral oncogene homolog

DFP

diisopropyl fluorophosphate

OP

organophosphorus compound

IP

intraperitoneal

SC

subcutaneous

FJB

FluoroJade B

qRT-PCR

quantitative real time polymerase chain reaction

IL-1β

interleukin −1β

TNFα

tumor necrosis factor alpha

CXCL10

C-X-C motif chemokine ligand 10

CCL2

chemokine (C-C motif) ligand 2

CCL3

chemokine (C-C motif) ligand 3

CCL4

chemokine (C-C motif) ligand 4

IL-6

interleukin 6

GAPDH

Glyceraldehyde 3-phosphate dehydrogenase

HPRT1

hypoxanthine phosphoribosyltransferase 1

gp91phox (NOX2)

glycosylated NADPH oxidase

IHC

immunohistochemistry

DFP-SE

DFP-induced status epilepticus

soman-SE

soman-induced SE