Pyridostigmine bromide exposure creates chronic, underlying neuroimmune disruption in the gastrointestinal tract and brain that alters responses to palmitoylethanolamide in a mouse model of Gulf War Illness.

Abstract

Gulf War Illness (GWI) is a chronic multisymptom illness that includes gastrointestinal disorders. Although the exact etiology of GWI is unknown, exposure to the drug pyridostigmine bromide (PB) is considered a major factor. Exposure to PB drives enteric neuroinflammation, promotes immunosuppression, and alters physiological functions of the colon in the short term but whether exposure to PB is sufficient to promote long term dysfunction is not known. Here, we tested whether exposure to PB is sufficient to drive long term changes that reflect GWI, and whether the endogenous anti-inflammatory mediator palmitoylethanolamide (PEA) is sufficient to reduce the detrimental effects of PB in the gut and brain of mice. Exposure to PB alone was not sufficient to cause major changes in neuromuscular transmission but did drive major changes by altering the effects of PEA. Calcium imaging data show that the mechanisms responsible include a shift in receptor signaling mediated by TRPV1, endocannabinoids, and peroxisome proliferator-activated receptors alpha (PPARα). Additional mechanisms include the development of glial reactivity and changes in enteric neurochemical coding and survival. PB and PEA caused major shifts in pro-inflammatory cytokines/chemokines in the brain and colon that persisted up to 5 months following exposure. Many of the effects of PB and PEA exhibit significant sex differences. Together, these results highlight novel mechanisms whereby PB promotes long-lasting changes in nervous system and immune function by inducing occult neuroplasticity that is revealed by subsequent exposure to unrelated drugs in a sex dependent manner.

1. Introduction

Many Veterans of the Persian Gulf War suffer from a chronic multisymptom illness commonly referred to as Gulf War Illness (GWI). GWI is characterized by the presence of certain chronic unexplained illnesses that are related to Gulf War service including chronic fatigue syndrome, fibromyalgia, functional gastrointestinal disorders, and undiagnosed illnesses such as weight loss, cardiovascular disease, neurological and psychological problems, and sleep disturbances. Functional gastrointestinal disorders (FGIDs) are among the most frequent and debilitating illnesses experienced by Gulf War veterans and are three times more common in Gulf War veterans than non-GW veterans (1). FGIDs are considered disorders of the brain-gut axis and commonly manifest with symptoms that include diarrhea, constipation, and abdominal pain (2). The pathogenesis of FGID in GWI is not understood, and there are no therapies available (3–5).

Although the exact etiology of GWI is unknown, exposure to the drug pyridostigmine bromide (PB) is considered a major factor (6, 7). PB is a reversible acetylcholinesterase (AChE) inhibitor that was administered as prophylactic protection against Iraqui nerve gas attack in a daily dose of aproximately 90 mg for seven days (8). Acetylcholine (ACh) is an important neurotransmitter in the autonomic nervous system (ANS) and the primary excitatory neurotransmitter in the enteric nervous system (ENS), which controls digestive functions. Acute exposure to PB increases the availability of ACh and the resulting contact time with muscarinic and nicotinic receptors (9, 10). In isolated intestinal preparations this leads to acute disruptions in the cholinergic pathway and produces massive, spontaneous contractions that eventually lead to a catastrophic failure of gut motility (11, 12). Further, PB has the potential to increase anti-inflammatory vagal cholinergic pathways that suppress proinflammatory cytokine relase from gut immune cells (13–16).

In prior work, we found that exposing mice to PB in a paradigm that reflects the actual exposure experienced by soldiers during the Gulf War disrupts gastrointestinal functions, and gastrointestinal and brain immune responses at 7 days and up to 30 days following exposure (7). PB drives these changes, in part, through effects on the ENS that involve the development of reactive gliosis. Here, we tested the hypothesis that reducing reactive gliosis with the drug palmitoylethanolamide (PEA) improves colonic dysfunction induced by PB. PEA is an endogenous N-acetylthanolamide that is produced by anti-inflammatory responses (17) and functions to reduce reactive gliosis in the brain and gut by activating glial peroxisome proliferator-activated receptors alpha (PPARα) receptors (18, 19). We tested our hypothesis using a mouse model where PB exposure matches the dose, time frame, and route of exposure experienced by soldiers in the Gulf War. Two months after PB exposure, mice were treated with PEA and subsequently assessed for in vivo and ex vivo gut functions, the structure and function of the ENS, and immune responses in the gut and brain. Surprisingly, our results show little lasting effects of PB treatment at 5 months, suggesting that PB, per se, does not directly cause long-term changes in GI function. However, prior exposure to PB dramatically altered the effects of PEA and combined exposure caused major changes in gut motility, altered the excitability, neurochemical coding, and survival of neurons and glia in the ENS, and dysregulated cytokines and chemokines in the colon and brain. The mechanisms responsible include a shift in receptor signaling mediated by transient receptor potential cation channels of the subfamily vanilloid type 1 (TRPV1), endocannabinoids, and PPARα. Together, these changes dramatically alter the effects of PEA; making PEA not only a poor treatment to counter the effects of PB, but also dangerous in this context. Thus, long-term changes in signaling mechanisms induced by exposure to PB could contribute to the symptoms of GWI by altering the effects of other unrelated drugs.

2. Materials and Methods

2.1. Animals

All experimental protocols were approved by the Michigan State University Institutional Animal Care and Use Committee (IACUC). C57BL/6 male and female mice were purchased from Charles River Laboratories (Hollister, CA) at five weeks of age. Mice were maintained in a temperature-controlled environment (Innocage system with ALPHA-dri bedding; Innovive, San Diego, CA) on a 12-h light:dark cycle with access to acidified water and a minimal phytoestrogen diet (Diet Number 2919; Envigo, Indianapolis, IN) ad libitum.

2.2. PB and PEA treatment

Pyridostigmine bromide (PB) was purchased from Sigma (St. Louis, MO). Mice (n = at least eight per group) were arbitrarily assigned to each group and administered PB dissolved in drinking water for 7 days. PB dosage (90 μg/mL) was chosen based on calculated conversions from known human dosing in the Gulf War (7, 8). Conversions from human to mouse based on body surface area yield a mouse dose of 90 μg/mL. Controls were given water only.

Palmitoylethanolamide (PEA) was purchased from Cayman Chemical (Ann Arbor, MI). At the beginning of the second month of the experiment, PEA was added to drinking water at a concentration of 0.07 mg/mL for one month. We chose this dose based on published data showing that PEA at a dose of 10 mg.kg (i.p.) is an effective dose to reduce reactive gliosis in the gut (18). PEA controls were given water for the whole month.

Animals were euthanized at 5 months. Samples of blood, colon, and brain were collected at the time of sacrifice. Macroscopic damage was assessed and scored as the colon was removed using a well characterized scale to quantify colonic length, fecal blood, and diarrhea (20).

2.3. Ca²⁺ imaging

Live samples of colonic myenteric plexus were prepared for Ca²⁺ imaging as previously described (21). Briefly, colonic segments were collected in ice-cold Dulbecco’s modified Eagle medium (DMEM) and transferred to Sylgard-coated open diamond shaped bath recording chambers. Tissue segments were opened along the mesenteric border and the mucosa, submucosa, and longitudinal muscle were removed by microdissection to expose the myenteric plexus. The resulting circular muscle myenteric plexus (CMMP) preparations were incubated for 15 minutes at room temperature in enzyme mixture consisting of 150 U/mL Collagenase type II and 1 U/mL Dispase (Life Technologies) dissolved in DMEM. CMMP were washed two times with DMEM and then incubated with DMEM (controls) or with 250 μM PB for 3 hours at 37°C (5% CO2, 95% air). CMMP preparations were rinsed with DMEM and then loaded with 4 μM Fluo-4 AM, 0.02% Pluronic F-127 and 200 μM water-soluble probenecid (Life Technologies) in DMEM for 45 minutes at 37°C (5% CO2, 95% air). CMMP were rinsed with DMEM and incubated with 200 μM probenecid in DMEM for 15 minutes to de-esterify before imaging. Drugs were bath applied in experiments. Specific drug application times and concentrations were 15 min for the cannabinoid CB₁ receptor antagonist SR141716A (0.5 μM) and the PPARα antagonist GW6471 (10 μM), 1 min for the vanilloid TRPV1 receptor antagonist capsazepine (1 μM) (Tocris; Minneapolis, MN), 3 hours for the Cx43 mimetic peptide 43Gap26 (20 μM) (Anaspec; Fremont, CA), and 5 min for tetrodotoxin (TTX, 1 μM) (Millipore Sigma; Burlington, MA). Images were acquired every 500 milliseconds through the 40x water-immersion objective (LUMPlanFI, 0.8 n.a.) of an upright Olympus BX51WI fixed stage microscope (Olympus, Tokyo Japan) using NIS-Elements software (version 4.5) and an Andor Zyla sCMOS camera (Andor, South Windsor, CT). Whole mounts were superfused with Krebs buffer (37°C) at 2–3 mL min⁻¹. PEA (10 μM) was dissolved in Krebs solution and bath applied for 30 seconds.

2.4. In vivo colonic motility

2.4.1. Pellet production

Fecal pellet output was measured as previously described (7, 22, 23). Briefly, mice were individually housed without water or food for 1 hour starting at 8:00 AM (Zeitgeber +2). Fecal pellets were collected, wet weight was measured immediately, and the dry weight was measured after an overnight dehydration at 60°C.

2.4.2. Colon bead assay

Distal colonic transit time was measured using glass beads (2 mm in diameter) as described previously (7, 24). Briefly, mice were lightly anesthetized with isofluorane, and a glass bead was inserted through the anus and gently pushed 2 cm orally by a customized syringe. The syringe was carefully withdrawn, and the bead expulsion latency was measured.

2.5. Ex-vivo colonic function

2.5.1. Contractility studies

Isometric muscle tension recordings were performed in circular muscular segments of the distal colon. Circular muscle rings were mounted in a tissue bath with oxygenated Krebs solution at 37°C (24). Each muscle ring was attached to an isometric force transducer and data were charted with LabChart 8 software (ADInstruments, Colorado Springs, CO) as described previously (25). Tissue segments were equilibrated for 20 minutes under 0.5 g initial tension. Electrical field stimulation (EFS, 20V, 1–30Hz) was applied through two platinum concentric electrodes and a GRASS stimulator (S88, GRASS telefactor, West Warwick, RI) to evoke neurogenic contractions and relaxations. EFS was applied to obtain neurogenic contractions. After EFS, maximal muscle contractions were obtained by adding carbachol (10 μM, a cholinergic agonist, Sigma). After being rinsed and stabilized, prostaglandin F2α (PGF_2α, 1 μM) was added to the bath and tissues were subjected to EFS to obtain neurogenic relaxation recordings (21).

2.5.2. Paracellular permeability of the colonic wall

Intestinal barrier function was assessed using Ussing chambers as described previously (26). Briefly, segments of distal colon were mounted in Ussing chambers (aperture 0.3 cm²; EasyMount Ussing Chamber system, Physiologic Instruments, San Diego, CA, USA), equilibrated for 20 minutes, and a cell-impermeant fluorescein-5-(and-6)-sulfonate (478.32 Da; Life technologies corporation, Carlsbad, CA) was added to the mucosal chamber (0.05 mg ml⁻¹). Samples from the serosal chamber were taken before the dye was added every 20 minutes for 2 hours (100 μL duplicates and buffer was replenished). Fluorescence intensity was measured on an Infinite M1000 PRO microplate reader (excitation/emission wavelengths 495/520 nm, Tecan Group Ltd, Mannedorf, Switzerland) using i-control™ microplate reader software (Tecan, version 1.6.19.2). Intestinal permeability was assessed from the slope of fluorescence values of the last four time points.

2.6. Whole-mount immunohistochemistry

Whole-mount preparations of mouse colonic myenteric plexus were prepared from segments of intestine preserved in Zamboni’s fixative as described previously (27). Whole mounts were rinsed three times for 10 minutes each with 0.1% TritonX-100 in phosphate-buffered saline (PBS-Triton) followed by a 45 minutes incubation in blocking solution (4% normal donkey serum, 0.4% Triton X-100 and 1% bovine serum albumin). Preparations were incubated with primary antibodies overnight at room temperature and secondary antibodies (Table 1) for 2 hours at room temperature before mounting. Fluorescent labeling was evaluated using the 20x or 40x objective (0.75 numerical aperture; Plan Fluor, Nikon, Melville, NY) of an upright epifluorescence microscope (Nikon Eclipse Ni) with a Retiga 2000R camera (QImaging, Surrey, BC, Canada) controlled by QCapture Pro 7.0 (QImaging) software (Brown et al. 2016, McClain 2015). Representative images were acquired through the 60x oil immersion objective (Plan-Apochromat, 1.42 numerical aperture) of an inverted Fluoview FV1000 confocal microscope (Olympus, Center Valley, PA, USA). Alexa Fluor 405, 488, 568, and 594 secondary antibodies were excited with 405 nm, 488 nm or 543 nm wavelengths and detected using SDM560 dichroic mirror and BA505–525 bandpass or BA560IF longpass filter sets (26).

Table 1.

Details of primary and secondary antibodies used in this study

Antibody	Source	Dilution	Catalog No.
Primary Antibodies
Goat anti-Calretinin	Swant, Switzerland	1:1000	CG1
Chicken anti-GFAP	Abcam, Cambridge, MA	1:1000	AB4674
Biotinylated anti-mouse HuC/D	Invitrogen, Carlsbad, CA*	1:200	A21272
Sheep anti-nNOS	Millipore, Billerica, MA	1:500	AB1529
Rabbit anti-M3 muscarinic receptor	Alomone	1:200	AMR-006
Secondary Antibodies
Dylight 405-conjugated streptavidin	Jackson Immuno, West Grove, PA	1:200	016-470-084
Alexa Fluor 488-Donkey anti-chicken	Jackson Immuno, West Grove, PA	1:200	703-545-155
Alexa Fluor 488-Donkey anti-sheep	Jackson Immuno, West Grove, PA	1:200	713-545-003
Alexa Fluor 568-Donkey anti-goat	Invitrogen, Carlsbad, CA*	1:200	A-11057
Alexa Fluor 594-Donkey anti-rabbit	Jackson Immuno, West Grove, PA	1:200	711-585-152

GFAP, glial fibrillary acidic protein; nNOS, neuronal nitric oxide synthase

^*Now Thermo Fisher Scientific, Waltham, MA.

2.7. Multiplex cytokine assay

Brain and colonic tissues collected at the time of sacrifice were homogenized in a Tris-Buffered Saline Tween (TBST) buffer to extract proteins. Cytokine/chemokine levels were evaluated using a bead-based multiplex immunoassay (Eve Technologies, Calgary, AB. Canada).

2.8. Solutions

Ca²⁺ imaging experiments were performed in modified Krebs buffer consisting of (in mmol/L): 121 NaCl, 5.9 KCl, 2.5 CaCl2, 1.2 MgCl2, 1.2 NaH2PO4, 10 HEPES, 21.2 NaHCO3, 1 pyruvic acid, 8 glucose (pH adjusted to 7.4 with NaOH). Muscle contractility and Ussing chamber studies were performed in normal Krebs buffer consisting of (in mmol/L): 117 NaCl, 4.7 KCl, 2.5 CaCl2, 1.2 MgCl2, 1.2 NaH2PO4, 25 NaHCO3 and 11 glucose. Live tissue was maintained in DMEM/F-12 nutrient mixture supplemented with HEPES and Glutamine (Life Technologies).

2.9. Chemical and Reagents

Chemicals and reagents were purchased from Sigma-Aldrich (St. Louis, MO), Tocris (Minneapolis, MN), Anaspec (Fremont, CA), Millipore Sigma (Burlington, MA), and Cayman Chemicals (Ann Arbor, MI). PB was dissolved in water 90 μg/mL and made available to mice over the 7 days ad libitum in drinking water. PEA was dissolved in ethanol (3.5%) and diluted in water to a final concentration of 0.07 mg/mL. PEA water was available ad libitum over 1 month. For in vitro experiments, PB was dissolved in Krebs at 250 μM and applied directly to CMMPs in recording chambers for 3 hours. PEA was dissolved in ethanol and diluted in Krebs to a final concentration of 10 μM. PEA was applied via a gravity-fed perfusion system in Ca²⁺ imaging experiments for 30 seconds. Inhibitors and antagonist drugs were dissolved in DMEM and bath applied. Carbachol (10 μM) was directly added to organ baths for isometric muscle tensions recordings.

2.10. Data analysis

Cell counts and ganglionic expression data were analyzed offline using Fiji software (National Institutes of Health, Bethesda, MD). Cell counts were performed using the cell counter plug-in of ImageJ software. Enteric neuron and glial cell numbers are presented as ganglionic packing density, which was calculated by tracing the ganglionic area and counting the number of HuC/D-immunoreactive neurons within the defined ganglionic area. The relative ganglionic expression of GFAP and M3 was measured by recording the mean gray values of GFAP and M3 fluorescence within a defined ganglionic area. Cell counts and ganglionic expression data were performed on a minimum of 10 ganglia per animal and averaged to obtain a value for that animal. The number of animals in each experiment is represented by n values, and Ca²⁺ imaging data are expressed as percent buffer control.

2.11. Statistical analysis

Data were analyzed with GraphPad Prism 7 (GraphPad Software, San Diego, CA) and are shown as mean ± standard error of the mean (SEM). Remaining data were analyzed by oneway and two-way, where appropriate, analysis of variance (ANOVA) with a Tukey post-test. P < .05 was considered statistically significant.

3. Results

3.1. Effects of pyridostigmine bromide (PB) and palmitoylethanolamide (PEA) on body weight, colonic length, and intestinal permeability

Exposure to PB is linked to the development of GWI. In prior work, we found that exposing mice to PB disrupts colonic function, and colon and brain immune responses (7). These changes were observed immediately following PB exposure and many persisted for up to one month. PB drives these changes, in part, through effects on the ENS that include the induction of reactive gliosis. Based on these data, we hypothesized that inhibiting reactive gliosis with the drug palmitoylethanolamide (PEA) would reduce the effects of PB. We tested this hypothesis by administering PB to male and female mice over 7 days to model human exposure during the Gulf War and subsequently treated a subgroup with PEA for one month beginning at the second month following exposure to PB (Figure 1A). Experiments were performed on animals at 5 months following the initial exposure of PB to study the long-term effects of PB and the potential benefits of PEA.

Figure 1. Experimental paradigm and effects of PB and PEA on body weight, colonic length, and intestinal permeability.

(A) Model of experimental timeline. Animals were exposed to PB for 7 days and to PEA for 1 month. (B) Model showing the mechanism of action for PB in the ACh pathway. (C-D) Body weight measurements for male (C) and female (D) mice treated with water (control, solid blue), PB (solid red), PEA (open purple), or both PB and PEA (open orange). Male mice exposed to PB gained less weight during the PEA treatment but recover to normal weights immediately after (C; p = 0.0421 compared to control, p = 0.0005 compared to PB alone). No differences in body weight were observed in females. (E) Summary data showing measurements of colonic length as an indicator of inflammation and/or fibrosis in males and females. (F) Summary data showing measurements of colonic permeability in tissue samples from males and females assessed by dye flux in Ussing chambers. No changes in colonic permeability were observed in male or female mice. Statistically significant comparisons are denoted by a vs Control, b vs PB, c vs Control + PEA, c vs PB + PEA. Two-way ANOVA, n = 7–8 animals per group.

Mice did not avoid water containing PB or PEA and all animals drank a comparable amount of water regardless of drug content (data not shown). PB and PEA did not affect body weight in male mice when singularly administered. Male mice exposed to PB gained less weight during the PEA treatment but recovered to normal weights immediately after (Figure 1C). PB and PEA did not affect body weight in female mice (Figure 1D). Mice exposed to PB, PEA or both did not exhibit overt macroscopic changes indicative of colitis such as bloody diarrhea, mucosal ulceration or changes in colonic length (Figure 1E). We also did not observe significant changes in colonic permeability (Figure 1F), suggesting that PB does not cause effects in the brain and other organs by altering colonic permeability. These results are in line with human GWI where FGIDs are observed in the absence of organic intestinal disease.

3.2. Exposure to PB alone is not sufficient to induce long-term changes in colonic motility

In prior work, we showed that PB causes acute disruptions in colonic motility but whether these changes persist over longer periods of time is not known. To answer this question, we conducted a longitudinal study over 5 months to assess the long-term effects of PB on GI functions (Figure 2). We assessed colonic motility in vivo in male and female mice by measuring fecal pellet output, fecal pellet water content, and colonic bead expulsion time during the 7-day PB treatment and for every two weeks thereafter for 5 months. Female animals exposed to PB exhibited increased fecal pellet production [162.5%, F (30, 432) = 1.693; p = 0.0008; Figure 2B] and delayed distal colonic transit at 4.5 months [777.7%, F (30, 401) = 2.617; p = 0.0002; Figure 2F]. PB alone did not affect fecal pellet production and bead expulsion time in male mice (Figure 2A, 2E) or fecal pellet water content in either males or females (Figure 2C–D). We conducted isometric muscle tension recordings in segments of colon to measure specific changes in neuromuscular transmission that might underlie the effects of PB on gut motility (Figure 3). Data from these recordings show that PB exposure altered neurogenic relaxation in males [33.8%, F (21, 318) = 0.7208; p = 0.0348; Figure 3B] but did not alter neurogenic contractility in males (Figure 3A) or neurogenic contractility or relaxations in females (Figure 3D–E). PB exposure had no significant effect on contractions driven by directly stimulating intestinal smooth muscle with carbachol (10 μM) in male or female mice, indicating that changes in neuromuscular transmission mainly reflect alterations to enteric neurons (Figure 3C, 3F). These data show that PB exposure over a time course relevant to GW veterans has the potential to produce long-lasting changes in gut motility in mice, but that most of the effects of PB are relatively subtle.

Figure 2. Effects of PB and PEA on colonic motility in vivo.

(A-F) Data showing the effects of PB and PEA on colonic motility in male (left; A, C, E) and female (right; B, D, F) mice. Measurements of fecal pellets produced per hour (A, B), fecal fluid content (C,D), and colonic bead expulsion time (E, F). Statistically significant comparisons are denoted by a vs Control, b vs PB, c vs Control + PEA, c vs PB + PEA. Two-way ANOVA, n = 7–8 animals per group.

Figure 3. Effects of PB and PEA on colonic motor functions ex vivo.

Data from isometric muscle tension recordings showing the effects of PB and PEA on ex vivo colonic motor functions at 5 months in male (A-C) and female (D-F) mice. Frequency (Hz)–response curves for neurogenic contractions are shown in panels A and D, and relaxations in panels B and E. Responses to the cholinergic agonist carbachol are shown in panels C and F. Males and females exhibited a decreased in neurogenic relaxations in PB (B) and PB + PEA treated mice, respectively (E; p = 0.0326 compared with control). No differences were observed in neurogenic or carbachol driven contractions in male (A, C) and female (D, F) mice. Statistically significant comparisons are denoted by a vs Control, b vs PB, c vs Control + PEA, c vs PB + PEA. Two-way ANOVA, n = 7–8 animals per group.

3.3. Exposure to PB alters the effects of PEA on colonic motility

Reactive gliosis in the ENS and CNS contributes to the pathophysiology of mouse GWI models that involve PB (7, 28). Gliosis disrupts nervous system homeostasis and reducing gliosis could improve gastrointestinal and immune disruptions driven by PB. Certain drugs such as the PPARα agonist PEA reduce reactive gliosis in the intestine (18, 19) but whether they provide therapeutic benefit against the long-term effects of PB is not known. We tested this concept by exposing mice to the drug PEA at two months following exposure to PB. PEA reduced pellet production in both male and female mice that were exposed to PB [81.7%, F (30, 430) = 3.271; p = 0.0004 in male, 66.1%, F (30, 432) = 1.693; p =0.0178 in female, Figure 2A–B] but did not significantly alter fecal pellet production in control animals. PEA also decreased fecal fluid content in male and female mice exposed to PB [29.3%, F (30, 384) = 0.9982; p = 0.0009 in male, 23.7%, F (30, 350) = 1.021; p = 0.0073 in female, Figure 2C–D] but did not affect fecal fluid content in control animals. PEA significantly delayed colonic transit in male mice exposed to PB [577.7%, F (30, 398) = 1.217; p = 0.0008; Figure 2E], while PEA alone was sufficient to delay colonic transit in females [86.9%, F (30, 401) = 2.617; p = 0.0006; Figure 2F]. These data show that prior exposure to PB can disrupt gut motility by altering the impact of anti-inflammatory drugs such as PEA.

Our results above suggest that PB exposure followed by exposure to PEA alters key gut functions and show that these alterations begin immediately after beginning the PEA treatment and persist for at least two months. We conducted isometric muscle tension recordings in segments of colon to understand changes in neuromuscular transmission that might underlie the effects of PB and PEA on gut motility (Figure 3). PB impaired neurogenic relaxations in male mice (Figure 3B), and PEA impaired neurogenic relaxations in female animals exposed to PB [37.7%, F (21, 295) = 0.6277; p = 0.0326; Figure 3E] but PEA did not cause significant changes in contractility in either sex (Figure 3A, 3D). Exposure to PEA had no significant effect on contractions driven by carbachol in male or female mice, indicating that changes in neuromuscular transmission mainly reflect alterations to enteric neurons (Figure 3C, 3F). Together, our ex vivo (Figure 3) and in vivo (Figure 2) motility data show that exposure to PB has lasting effects that alter the impact of PEA on intestinal motor functions. Further, differential effects in males and females suggest that the mechanisms underlying the effects of PB and PEA on gut motility are sex-dependent.

3.4. Cellular mechanisms underlying the effects of PB and PEA in the enteric nervous system

Colonic motility is regulated by intrinsic neural circuitry in the myenteric plexus composed of enteric neurons and glia. PB and PEA alter the activity of enteric neurons and glia (7, 29) but the mechanisms underlying these effects and potential interactions are not understood. We conducted calcium (Ca²⁺) imaging recordings in the myenteric plexus to understand the cellular mechanisms affected by PB and PEA in enteric neurons and glia. In these experiments, we focused on potential receptor systems for PEA including PPARα, cannabinoid 1 receptors (CB1), and TRPV1 (30). Whole mount preparations of myenteric plexus were incubated in the presence or absence (control) of PB (250 μM) for 3 hrs and subsequently challenged with PEA (10 μM). PEA drove large Ca²⁺ responses with relatively slow kinetics in myenteric neurons and glia in control samples (Figure 4A, B, E, F). Under these conditions, glial and neuronal responses to PEA were almost entirely blocked by tetrodotoxin [TTX; 87.2% reduction in glia, F (13, 364) = 3.826; p<0.001; 83.5% reduction in neurons, F (13, 125) = 2.904; p<0.001], indicating a requirement for neuronal signaling. Neuronal responses were completely blocked by the TRPV1 antagonist capsazepine and almost entirely blocked by the glial connexin-43 (Cx43) channel mimetic peptide 43Gap26 [85.6% reduction, F (13, 125) = 2.904; p = 0.011]. Glial responses were partially blocked by the CB1 antagonist SR141716A [46.9% reduction, F (13, 364) = 3.826; p = 0.0001] and not significantly reduced by other inhibitors. Thus, PEA drives intercellular communication between myenteric neurons and glia in control samples through mechanisms that require the activation of neuronal TRPV1 channels, glial signaling involving Cx43, and endocannabinoid signaling.

Figure 4. Effects of PEA on the activity of enteric neurons and glia.

Summary data (A, E) and representative traces from calcium (Ca²⁺) imaging recordings of enteric glial (A-D) and neuronal (E-H) activity stimulated by PEA (10 μM). Effects in control samples are shown on the left and those in samples pretreated with PB (250 μM; 3h) are shown on the right. SR141716A = CB1 receptor antagonist (0.5 μM); GW6471 = PPARα antagonist (10 μM); capsazepine = TRPV1 antagonist (1 μM); 43Gap26 = connexin-43 hemichannel mimetic peptide (20 μM); tetrodotoxin (TTX) = voltage gated sodium channel antagonist (1 μM). Each trace represents an individual glial cell (B-D) or neuron (F-H). Statistically significant comparisons are denoted by a vs Control, b vs PB. One-way ANOVA, n= 5–63 glia, n= 3–24 neurons.

Glial responses to PEA were significantly reduced in samples incubated with PB [61.0% decrease, F (13, 364) = 3.826; p<0.001; Figure 4A, C]. Neuron responses also tended to decrease but changes in neuronal peak responses did not reach significance (Figure 4E, G). As in control samples, neuron and glial responses to PEA were blocked by TTX in samples exposed to PB, indicating that the requirement for neuronal signaling was maintained under these conditions. However, the receptor systems involved were significantly different. Neuronal responses to PEA were nearly completely blocked by the CB1 antagonist SR141716A in samples exposed to PB [96.6% reduction, F (13, 125) = 2.904; p<0.001; Figure 4E, H]. The reduced glial responses to PEA following PB were not altered by SR141716A alone. Instead, the reduction in glial responses to PEA following PB was corrected by blocking PPARα receptors with GW6471 [106.5% of control, F (6, 165) = 3.161; p<0.0001]. Prior exposure to PB increased the sensitivity of glial responses to PEA to blockade by 43Gap26 but did not alter the blockade of neuronal responses to PEA by 43Gap26. Interestingly, neuronal response to PEA were no longer blocked by capsazepine in samples treated with PB (Figure 4E). These data indicate that exposure to PB reduces neuron-glia communication evoked by PEA and shifts the mechanisms involved from a TRPV1 pathway to a pathway that requires neuronal endocannabinoid signaling and glial PPARα.

3.5. Effects of PB and PEA on reactive gliosis and neuron survival in the myenteric plexus

PEA acts through PPARα receptors to block reactive gliosis in the brain and the gut (18, 19) and reduces the glial expression of pro-inflammatory proteins such as inducible nitric oxide synthase (iNOS), the Ca²⁺ binding protein S100β, glial fibrillary acidic protein (GFAP), prostaglandin E2 (PGE2), cyco-oxygenase-2 (COX2) and tumor necrosis factor α (TNFα) (18). We assessed whether PEA affected glial reactivity at 5 months following exposure to PB by measuring GFAP expression and analyzed myenteric neuron survival with HuC/D labeling. Immunohistochemical characterization of M3 muscarinic acetylcholine receptor expression and markers of major excitatory and inhibitory neuron subtypes was conducted to assess enteric neurochemical coding. PEA decreased GFAP expression in the colonic myenteric plexus of male (Figure 5A, D) and female (Figure 5B, D) mice regardless of treatment with PB. Interestingly, both female groups treated with PB exhibited a significant increase of GFAP expression when compared to their counterparts [Control and Control + PEA) (F (3, 49) = 3.545; PB: 27.9%, p = 0.0269; PB + PEA: 69.5%, p = 0.0292; Figure 5B, D]. Surprisingly, female mice treated with PB + PEA exhibited a 15.4% increase in myenteric neuron numbers [F (3, 74) = 0.2356; p = 0.0060; Figure 5B, E]. No changes in neuron numbers were observed in males (data not shown). Similarly, female mice treated with PEA exhibited a 69.1% increase in the expression of muscarinic M3 receptors in neurons and this effect was exacerbated to 123.2% by prior exposure to PB [F (3, 27) = 1.845; p = 0.0001; Figure 5C, F]. Similar effects on M3 expression were not observed in male animals (data not shown). Both male and female mice treated with PEA exhibited an increase in the proportion of excitatory neurons labeled with calretinin [F (3, 26) = 2.531; p = 0.0012; 65.4% male and F (3, 21) = 2.160; p = 0.0027; 112.0% female; Figure 6A–B, E] and a corresponding decrease in inhibitory neurons expressing neuronal nitric oxide synthase (nNOS) [F (3, 17) = 1.471; p = 0.0014; 25.1% male and F (3, 20) = 2.017; p = 0.0147; 28.1% female; Figure 6C–D, F]. PB alone had minor effects on neurochemical coding that did not reach significance. These results support the conclusion that PEA decreases reactive gliosis but also induces significant changes in neurochemical coding that involve an increase in the expression of M3 acetylcholine receptors, an increase in the proportion of excitatory neurons, and a decrease in inhibitory neurons. These changes likely contribute to the observed disruption in colonic motor functions.

Figure 5. Exposure to PB drives reactive gliosis and enteric neurodegeneration.

(A,B) Representative confocal images of myenteric ganglia showing immunolabeling for glial fibrillary acidic protein (GFAP, green, glia) and HuC/D (blue, neurons) in samples from male (A) and female (B) control animals, and animals treated with PB, PEA, or both at 5 months post exposure to PB. (C) Representative confocal images of myenteric ganglia showing immunolabeling for GFAP (green), HuC/D (blue), and muscarinic M3 receptors (magenta) in female mice at 5 months. (D-F) Summary data showing the quantification of GFAP immunoreactivity in male and female (D) mice. GFAP expression is reduce in male (D; p = 0.0115) and female (D; p = 0.0269) mice treated with PEA. All female mice treated with PB exhibit higher expression of GFAP compared to their control counterparts (D; p = 0.0269, p = 0.0292). (E) Quantification of myenteric neuron density in female mice. Female treated with PB + PEA exhibit an increase in neuron packing density (E; p = 0.006). (F) Quantification of M3 receptor immunofluorescence labeling in myenteric ganglia. M3 expression is higher in female mice exposed to PEA (F; p = 0.0001). Statistically significant comparisons are denoted by a vs Control, b vs PB, c vs Control + PEA, c vs PB + PEA. Scale bars in A-C: 30 μm. One-way ANOVA, n = 7–8 animals per group.

Figure 6. PEA affects the expression of excitatory and inhibitory enteric neurons in the myenteric plexus.

Data showing the proportion of calretinin (A-B) or neuronal nitric oxide synthase (nNOS; C-D) immunoreactive enteric neurons in male and female animals following exposure to PB and PEA. (A-B) Representative confocal images show myenteric ganglia labeled with antibodies against GFAP expression (GFAP, green), anti-Hu antibodies to identify enteric neurons (Hu, blue), and the excitatory neuron marker calretinin (calretinin, magenta). (CD) Representative confocal images show myenteric ganglia labeled with antibodies against the inhibitory neuron marker nNOS (nNOS, green) and anti-Hu antibodies to identify enteric neurons (Hu, blue). (E-F) Summary data showing quantification of the proportions of myenteric neurons expressing calretinin and nNOS in males and females. Exposure to PEA elevate the proportion of calretinin in neurons in males (p<0.01) and females (E; p<0.01). Exposure to PEA elevate the proportion of nNOS in neurons in males (p = 0.0014) and female (F; p = 0.0147). (a vs Control, b vs PB, c vs Control + PEA, c vs PB + PEA). *Asterisk indicate cells positive for calretinin, nNOS and Hu. Scale bar in A-D: 30μm. One-way ANOVA, n = 7–8 animals per group.

3.6. Exposure to PB and PEA alters immune profiles in the colon and brain

Bi-directional signaling between the immune and nervous systems is mediated, in part, by cholinergic transmission. To understand how alterations to cholinergic signaling during exposure to PB impacts peripheral and central immune responses, we assessed the expression of pro- and anti-inflammatory cytokines and chemokines in the colon and brain of mice after 5 months of PB exposure. PB treatment alone did not cause significant changes in the expression of colonic cytokines/chemokines in either sex (Table 2; Figure 7). In the brain, exposure to PB decreased the expression of IL-1α [12.9%, F (3, 25) = p = 0.0044] and IL-17 [50%, F (3, 21) = 0.5463; p = 0.0146] in male mice (Table 3; Figure 8A, C) but had no significant effect on brain cytokines in females.

Table 2.

Summary of cytokines/chemokines in the colon affected by PB and PEA treatment

Colon
Male
Chemokine/Cytokine	Expression	Treatment
G-CSF	Decreased (p<0.05)	−PB [90μg/mL] + PEA [0.07mg/mL] (vs. PB [90μg/mL])
MIG	Increased (p<0.05)	−PB [90μg/mL] + PEA [0.07mg/mL] (vs. PB [90μg/mL])
RANTES	Increased (p<0.05)	−Control + PEA [0.07mg/mL] (vs. PB [90μg/mL])
Female
Chemokine/Cytokine	Expression	Treatment
none

Figure 7. Exposure to PB and PEA alters colonic cytokines/chemokines.

Data from multiplex cytokine/chemokine arrays showing the effects of exposure to PB on colonic immune responses in male and female mice. (A) Heat maps show the fold change in cytokines and chemokines in males and females. Summary data show significant differentially regulated cytokines and chemokines in males. After 5 months, male mice treated with PB followed by PEA exhibit a decrease in granulocyte colony stimulation factor (G-CSF; B; p = 0.0161) and an increase in expression of MIG (also known as CXCL9; B; p = 0.025) compared to animals treated with PB alone. Male mice exposed to PEA alone show higher levels of RANTES (also known as CCL5; B; p = 0.0034) than animals treated with PB. No significant changes were observed in female mice. (a vs Control, b vs PB, c vs Control + PEA, c vs PB + PEA). One-way ANOVA, n = 7–8 animals per group.

Table 3.

Summary of cytokines/chemokines in the brain affected by PB and PEA treatment

Brain
Male			Female
Chemokine/Cytokine	Expression	Treatment	Chemokine/Cytokine	Expression	Treatment
IFN-γ	Increased (p=0.0507)	−Control + PEA [0.07mg/mL] (vs. control)	IL-9	Increased (p<0.05)	−Both PEA treated (vs. PB [90μg/mL])
IL-1α	Decreased (p<0.05, p<0.01)	−PB [90ug/mL] −Control + PEA [0.07mg/mL] −PB [9μg/mL] + PEA [0.07mg/mL] (vs. control)	IL-12(p70)	Decreased (p<0.05, p<0.01)	−Both PEA treated (vs. control and PB [90μg/mL])
IL-1β	Increased (p<0.05)	−Both PEA treated (vs. control and PB [90μg/mL])	MIP-1α	Increased (p<0.05, p<0.01)	−Both PEA treated (vs. PB [90μg/mL])
IL-2	Decreased (p<0.05)	−Control + PEA [0.07mg/mL] (vs. control)
IL-12(p70)	Decreased (p<0.01)	− PB [90μg/mL] + PEA [0.07mg/mL] (vs. control)
IL-17	Decreased (p<0.05)	PB [90μg/mL] (vs. control)
MIP-1α	Increased (p<0.05)	−Both PEA treated (vs. control and PB [90μg/mL])
VEGF	Increased (p<0.05)	−Control + PEA [0.07mg/mL] (vs. PB [90μg/mL])

Figure 8. Exposure to PB and PEA alters brain cytokines/chemokines.

Data from multiplex cytokine/chemokine arrays showing the effects of exposure to PB and PEA on brain immune responses in male (A, C) and female (A-B) mice. (A) Heat maps show the fold change in cytokines and chemokines in males and females. Summary data show significant differentially regulated cytokines and chemokines in females (B) and males (C). (a vs Control, b vs PB, c vs Control + PEA, c vs PB + PEA). One-way ANOVA, n = 7–8 animals per group.

PEA treatment produced much more profound alterations in peripheral and central immune responses. PEA treatment increased colonic proinflammatory cytokines/chemokines such as MIG (CXCL9) [41.61%, F (3, 24) = 1.747; p = 0.025] and RANTES [32.3%, F (3, 26) = 1.348; p = 0.034] and decreased colonic anti-inflammatory cytokine G-CSF [87.7%, F (3, 24) = 6.501; p = 0.0161] in male mice (Table 2; Figure 7A, B). These alterations were not observed in female mice (Figure 7A). In the brain, PEA exposure caused an increase in proinflammatory cytokines/chemokines including IL-1β [Control + PEA: 67.9%, PB + PEA: 58.1%, F (3, 27) = 5.994; p<0.0102], MIP-1α [Control + PEA: 76.6%, PB + PEA: 71.5%, F (3, 27) = 39.31; p = 0.0124], and VEGF [31.5%, F (3, 26) = 4.854; p = 0.0427; Figure 8A, C] and a decrease in IL-1α [F (3, 25) = 0.4735; Control + PEA: 95.7%, p = 0.0044; PB + PEA: 98.7%, p = 0.033], IL-2 [14.1%, F (3, 25) = 3.286; p = 0.0317], and IL-12(p70) [62.2%, F (3, 25) = 1.074; p = 0.0048] in male mice (Table 3, Figure 8A, C). In female animals, PEA caused changes in brain cytokines/chemokines that were reflected by increased expression of IL-9 [Control + PEA: 16.3%, PB + PEA: 17.1%, F (3, 28) = 3.481; p = 0.0211] and MIP-1α [F (3, 27) = 5.234; Control + PEA: 51.2%, p = 0.0068; PB + PEA: 64.1%, p = 0.0150] and decreased IL-12 [F (3, 27) = 0.7315; Control + PEA: 27.1%, p = 0.0387; PB + PEA: 25.8%, p = 0.007; Table 3, Figure 8A, B]. Together, these data show that PEA disrupts immune signaling in the brain and intestine and that PB alone has little significant lasting effects.

4. Discussion

GWI is a chronic multisymptom illness of unknown etiology that affects Veterans of the Persian Gulf War (6, 8, 31). The most widely held hypothesis is that acute exposure to the anti-nerve gas drug PB produced long-lasting effects, but the underlying mechanisms remain unclear (8, 32). In prior work, we found that exposing mice to PB in a paradigm that models human exposure in the Gulf War is sufficient to cause acute disruptions in GI motility, ENS function, and peripheral and central immune responses that persist for up to one month (7). PB drives these changes, in part, by inducing reactive gliosis in the ENS. Here, we tested whether exposure to PB is sufficient to cause long-lasting changes in GI functions, peripheral nervous system functions, and peripheral and central immune profiles, and whether reducing reactive gliosis with the drug PEA would provide therapeutic benefit. Surprisingly, our data show that PB, alone, had little overt long-term effects and that those that were present were subtle. These effects were revealed by the subsequent exposure to PEA. Following exposure to PB, PEA drove profound changes in GI motility, ENS function, and immune responses through mechanisms that involve a switch in enteric signaling mediated by TRPV1, endocannabinoids, and PPARα, altered neurochemical coding, and changes in glial phenotype that were sex-dependent. Together, these observations show that exposure to PB induces long-term changes in receptor signaling pathways that alter the effects of subsequent drugs. These mechanisms could underlie the development of diverse chronic symptoms in GWI.

Multiple lines of evidence link exposure to PB with the subsequent development of GWI in Veterans (8, 32). PB likely contributed to the frequent development of FGIDs in GWI by disrupting cholinergic transmission and altering the gut-brain axis (1). Acetylcholine (ACh) is the main excitatory neurotransmitter in the autonomic nervous system (ANS) and exerts its actions through nicotinic (nAChRs) and muscarinic (mAChRs) receptors before being degraded by AChE and recycled (11, 33, 34). Tight control of excitatory transmission is important for the coordination of enteric reflexes in the intestine, parasympathetic neuroeffector functions, and ganglionic transmission throughout the ANS. In vitro data show that disrupting ACh degradation with PB induces short-term changes in motor function in the guinea-pig ileum that include the emergence of large, spontaneous contractions (11). Yet how PB might lead to long-term changes in motor function is not understood.

In prior work, we found that the PB mouse model of GWI used here displays acute increases colonic motor functions that persist up to 30 days following initial exposure (7). This is consistent with human data showing that diarrhea is common in veterans with GWI (1). Surprisingly, our data from the current study show that these changes do not persist for longer periods up to 5 months. These data suggest that PB, per se, is not sufficient to induce long-term changes in GI motility. However, major changes emerge when animals exposed to PB are subsequently given the drug PEA. PEA is considered an anti-inflammatory compound and is widely used as a dietary supplement. PEA had no effect on GI motility in PB naïve animals, suggesting that PB creates persistent, occult changes in motility circuits that create unexpected impacts of subsequent drugs. PEA drove significantly altered intestinal motility in male and female mice that were exposed to PB and decreased fecal pellet production, reduced fluid content in pellets, and delayed colonic transit. While these effects and not necessarily consistent with diarrhea observed in GWI veterans, they could relate to abdominal pain, cramping, and moderate or multiple gastrointestinal symptoms, which are also common in GWI veterans (1). These results also support prior observations showing that PEA decreases intestinal motility in the context of inflammation (35). Thus, some of the underlying changes in neurotransmission produced by exposure to PB may share similarities with those caused by acute inflammation. In this regard, it is not surprising that the animal model used here exhibited a decrease in motility since most mouse models of colitis exhibit slowed motility following resolution while humans with FGIDs may present with either constipation or diarrhea. Our prior data support the concept of inflammation associated with PB by showing that the acute effects of PB involve major changes to immune profiles in the intestine (7). These results suggest that neuroplasticity driven by PB likely increases susceptibility to developing neurological and immune disorders such as FGIDs following exposure to normally innocuous challenges. This is important because most mouse GWI models use a combination of insults such as PB, permethrin, and stress (36, 37) and there is little understanding of how PB specifically increased susceptibility to developing GWI.

Our data show that mechanisms underlying the persistent effects of PB include changes to key receptor signaling pathways involved in intercellular communication in the ENS. In naïve tissues, PEA excites enteric neurons and glia through mechanisms that involve CB1 receptors, TRPV1 channels, and intercellular signaling mediated by glial Cx43 hemichannels and neuronal action potentials. PEA can, in fact, act via multiple mechanisms (38) that include directly activating PPAR-α (39) and “entourage effects” whereby PEA can produce indirect receptor-mediate effects (40–42). PEA can also indirectly promote CB1 and CB2 receptors signaling by inhibiting enzymes responsible for endocannabinoid degradation (41, 43). Similar mechanisms contribute to the activation of TRPV1 channels since they are also modulated by endocannabinoids (44, 45). Our results support prior observations showing that the effects of PEA on motility involve CB1 receptor signaling (35, 46). Our Ca²⁺ imaging data show that CB1-TRPV1 pathways are required for PEA to activate neuron-glia signaling in the myenteric plexus under normal conditions. However, more prominent contributions of signaling through PPAR-α, CB1, and Cx43 hemichannels emerge in samples treated with PB. This shift in signaling mechanisms is consistent with the development of reactive gliosis in inflammation where PPAR-α plays an important role in regulating glial activation (18). Neural signaling evoked by PEA became entirely dependent upon CB1 in samples exposed to PB, suggesting that PB increases the prominence of endocannabinoid signaling in enteric circuits. This would be expected to decrease excitability in enteric neurocircuits since presynaptic CB1 receptors act to dampen neurotransmission. Although the mechanisms differ, PEA acts through neuron-glia communication in either condition.

The cellular mechanisms involved in the effects of PB on colon motility include the development of neuroinflammation. Enteric neuroinflammation involves the development of reactive gliosis and neurodegeneration (47) and our data show that PB promotes the development of neuroinflammation in the ENS of female mice. PB-induced changes that reflect neuroplasticity include altered neurochemical coding, neuron numbers, and reactive gliosis. Elevated GFAP expression is a characteristic of reactive gliosis (48–51) and is linked to the production of danger cues in the ENS (20). GFAP expression remained elevated 5 months after PB treatment in female animals. These results extend our previous observations that showed elevated GFAP expression at 30 days following exposure to PB. Interestingly, reactive gliosis was not evident immediately after the 7 days PB treatment, suggesting that neuroinflammatory processes driving reactive gliosis continue to evolve after the initial exposure to the drug (7, 52). Given that reactive gliosis is a key mechanism of neuroinflammation in the ENS (20, 47, 53, 54) and our data show reactive gliosis persisting long after the initial exposure to PB, reactive gliosis can be considered an ongoing mechanism that contributes to chronic symptoms following exposure to PB. Importantly, enteric glia drive neuroinflammation through interactions with visceral sensory neurons (53) and these processes could contribute to abdominal pain experienced in veterans with GWI (1, 54).

Based on the observation of reactive gliosis, we postulated that controlling this process with the anti-inflammatory drug PEA would provide some therapeutic benefit to gut dysfunction induced by PB. PEA is an endogenous N-acylethanolamide that is produced as part of the endogenous anti-inflammatory responses (17) and PEA has received considerable attention as an anti-inflammatory agent in animal models of inflammatory bowel disease (55). In the nervous system, PEA activates PPAR-α receptors and reduces reactive gliosis, decreases the production of NO by reactive glia, and reduces glial expression of pro-inflammatory proteins such as iNOS, S100B, GFAP, prostaglandin E₂ (PGE₂), cyco-oxygenase-2 (COX2), and TNFα (18, 19). In agreement, PEA did reduce GFAP expression in both male and female mice in our study. Interestingly, PEA also affected neurochemical coding and survival. PEA reduced the abundance of inhibitory neurons expressing nNOS, increased the abundance of excitatory neurons expressing calretinin, and increased the overall number of neurons identified by HuC/D labeling in female treated with both PB and PEA. This shift in neurochemical coding would be expected to increase excitatory neuromuscular transmission and this would be consistent with the presentation of diarrhea in humans. In mice, we observed an overall decrease in motility, and this is consistent with mouse models of intestinal inflammation where a decrease in nitrergic neurons causes slowed motility (20, 53). Changes in neurochemical coding may reflect ongoing neuroplasticity and/or neuroinflammatory processes but why female mice treated with PB and PEA appear to exhibit an increased abundance of enteric neurons is not clear. In the brain, PEA does enhance neurogenesis and improve neuron survival by inducing brain-derived neurotrophic factor (BDNF) expression in astrocytes (56–58). It is possible that similar processes underlie the effects observed here and this would be an interesting topic to address in future work.

Immune disruption was evident in samples of colon and brain in our study. In prior work, we found that PB causes acute immunosuppression in the colon and brain that potentially involves T lymphocytes, mast cells, and monocytes (7). Interestingly, these immunosuppressive effects appear to persist in the brain of male mice up to 5 months following exposure to PB while effects in the colon are more transient. Key proinflammatory cytokines/chemokines produced by T lymphocytes such as IL-17, and by macrophages and neutrophils such as IL-1α were reduced. It is possible that the broad immunosuppressive actions of PB involve enhanced activation of cholinergic anti-inflammatory reflexes mediated by the vagus nerve. In support, activation of the vagus by nicotinic receptor agonists or acetylcholinesterase inhibitors reduces inflammation by suppressing cytokine synthesis (10, 16, 59, 60). It is unlikely that PB causes these effects through direct influences in the brain given its limited ability to cross the blood-brain barrier (71) and the more likely explanation is that PB affects immune responses in the periphery where the vagus is considered the primary neural regulator of anti-inflammatory tone (62). Here, ACh suppresses innate immune mechanisms by activating alpha7 nicotinic receptors expressed by macrophages and other innate immune cells to inhibit the production of pro-inflammatory cytokines including TNF-α and IL-1β (10, 15, 63–65). Our current data suggest that these mechanisms continue to promote an immunosuppressive environment within the male brain up to 5 months following exposure to PB.

PEA had significant immunomodulatory effects in the brain and colon that did not necessarily depend on prior exposure to PB. These effects suggest that PEA has direct immunomodulatory effects on normal mechanisms in healthy animals. Male mice exhibited decreases in G-CSF in the colon and IL-1α, IL-2, and IL-12(p70) in the brain while female mice exhibited decreased IL-12 in the brain. These cytokines/chemokines are normally expressed by mast cells (66, 67) which participate in innate host defense reaction and are found in peripheral tissues innervated by small nerve fibers and within the endoneurial compartment, where they orchestrate inflammatory processes (68). Mast cells express PPAR (69) and PEA inhibits mast cell activation, degranulation, and the expression/release of nerve growth factor (NGF) (70–72) in a PPAR-α-dependent manner (18). Thus, it is possible that changes in brain cytokines reflect altered signaling by mast cells. PEA may also exert anti-inflammatory actions in the intestine through indirect activation of CB1 and CB2 receptors (35, 73). However, many cytokines/chemokines were also increased such as MIG (CXCL9) and RANTE (CCL5) in the male colon, IL-1β, MIP-1α, and VEGF in the male brain, and IL-9 and MIP-1α in the female brain. Many of these cytokines are also produced by mast cells (IL-1B, VEGF, IL-9), but are also expressed by other immune cells such as monocytes, macrophages, and lymphocytes (74–83). PEA stimulates the phagocytosis of pathogens by macrophages and microglia and it is possible that a shift in macrophage/microglia phenotype could explain the increase in these cytokines/chemokines (84, 85).

Many of the effects of PB and PEA on intestinal motility and immune responses were sex dependent. These observations extend our prior findings which showed that the effects of PB are sex-dependent and that PB has a more pronounced acute effect in males, but a more pronounced chronic effect in females (7). PB primarily affects ACh availability and prolongs its actions at nicotinic and muscarinic receptors. Interestingly, the female sex hormones progesterone and 17-estradiol act as functional blockers of several nicotinic receptor subtypes (86, 87). Hormonal blockade of nicotinic receptor could explain why females developed slower evolving responses to PB than males. Likewise, clinical data from a study where PEA was used to treat neuropathic pain associated with lumbosciatalgia showed that pain was only improved in women (88), suggesting that PEA acts through different mechanisms in males and females. Our data support prior observations showing that PEA affects signaling through cannabinoid receptors and it is well-known that cannabinoid receptors are modulated by hormones and sex (88). Understanding the basis of sex differences cause by PB is important to develop better treatments for GW Veterans, especially since women comprised nearly 7 percent of the 700,000 of the military personal in the GW and GWI is more common in female GW veterans than their male counterparts (89–91).

The results of this study identify novel mechanisms whereby exposure to PB in a paradigm that replicates what was experience by soldiers in the GW drives long-lasting changes in the neural control of gastrointestinal functions, gut immunity, and immune responses in the brain. Importantly, our data show that many of the longer lasting effects of PB are occult and are only made evident by the subsequent exposure to other drugs. Changes in intercellular signaling mechanisms between enteric neurons, glia, and immune cells are affected by exposure to PB and alter the actions of unrelated drugs that are otherwise beneficial. These mechanisms could play a central role in lasting neurological problems in individuals with GWI. Many of the effects of the drugs used here are sex dependent understanding the sex-specific mechanisms that underlie the pathogenesis of GWI may yield important insights that lead to novel therapies.

Highlights:

• Pyridostigmine bromide (PB) exposure drives occult changes in neuro-immune mechanisms.

• PB drives sex-dependent, long-term changes in gut function by altering the effects of subsequent drugs.

• Underlying mechanisms include enteric neuro- and glioplasticity, and enteric signaling mediated by TRPV1, endocannabinoids, and PPARα.

• Shifts in central and peripheral immune responses persist up to 5 months following exposure.