Oligodendrocyte involvement in Gulf War Illness

Abstract Low level sarin nerve gas and other anti‐cholinesterase agents have been implicated in Gulf War illness (GWI), a chronic multi‐symptom disorder characterized by cognitive, pain and fatigue symptoms that continues to afflict roughly 32% of veterans from the 1990–1991 Gulf War. How disrupting cholinergic synaptic transmission could produce chronic illness is unclear, but recent research indicates that acetylcholine also mediates communication between axons and oligodendrocytes. Here we investigated the hypothesis that oligodendrocyte development is disrupted by Gulf War agents, by experiments using the sarin‐surrogate acetylcholinesterase inhibitor, diisopropyl fluorophosphate (DFP). The effects of corticosterone, which is used in some GWI animal models, were also investigated. The data show that DFP decreased both the number of mature and dividing oligodendrocytes in the rat prefrontal cortex (PFC), but differences were found between PFC and corpus callosum. The differences seen between the PFC and corpus callosum likely reflect the higher percentage of proliferating oligodendroglia in the adult PFC. In cell culture, DFP also decreased oligodendrocyte survival through a non‐cholinergic mechanism. Corticosterone promoted maturation of oligodendrocytes, and when used in combination with DFP it had protective effects by increasing the pool of mature oligodendrocytes and decreasing proliferation. Cell culture studies indicate direct effects of both DFP and corticosterone on OPCs, and by comparison with in vivo results, we conclude that in addition to direct effects, systemic effects and interruption of neuron–glia interactions contribute to the detrimental effects of GW agents on oligodendrocytes. Our results demonstrate that oligodendrocytes are an important component of the pathophysiology of GWI.

the Iraqi invasion of Kuwait (Steele, Lockridge, Gerkovich, Cook, & Sastre, 2015;White et al., 2016). The cause of GWI is unknown, but exposure to low-level sarin nerve gas, pesticides and pyridostigmine bromide (PB), all acetylcholinesterase (AChE) inhibitors, have been linked to the etiology of GWI White et al., 2016). PB, a reversible AChE inhibitor, was ingested by GW veterans as a prophylactic against potential exposure to sarin nerve gas, an irreversible and lethal AChE inhibitor (Sullivan et al., 2003;White et al., 2016). Pesticides, predominantly organophosphates like chlorpyrifos and dichlorvos, were applied, often in excess, to protect against insects in the battle lines along the Mesopotamian marshes . Naturally, the primary focus of GWI research has been on the consequences of disrupting cholinergic synaptic neurotransmission. However, it is unknown how transient disruption of cholinergic synaptic neurotransmission can lead to chronic neurological deficits that have persisted for over two decades in GWI patients. Recent research has shown that neurotransmitters are also involved in communication between axons and myelinating glia. Here, we test the hypothesis that disrupting cholinergic neuro-glial communication via AChE inhibitors can impair oligodendrocyte development and function, and thereby contribute to the chronic pathophysiology of GWI.
In addition to synaptic release, neurotransmitters are released non-synaptically along axons through exocytosis at axonal varicosities (Wake et al., 2015) and through ion channels (Fields, 2011;Vizi & Lendvai, 2008). Recent research indicates that oligodendrocyte development and myelination are impaired when this axo-glial signaling is disrupted. For example, blocking vesicular release of the neurotransmitter glutamate from axons by Botulinum toxin treatment, inhibits local protein synthesis of the major protein in myelin, myelin basic protein (MBP), and impairs induction of myelination on electrically active axons (Wake, Lee, & Fields, 2011). Activity-dependent myelination has been primarily studied in the context of glutamatergic signaling (Fields, 2015;Kukley, Capetillo-Zarate, & Dietrich, 2007;Wake et al., 2011;Wake et al., 2015), but the neurotransmitter acetylcholine (ACh) has recently been suggested to influence oligodendrocyte progenitor cell (OPC) proliferation, differentiation, and myelination (De Angelis, Bernardo, Magnaghi, Minghetti, & Tata, 2012;Fields, Dutta, Belgrad, & Robnett, 2017). In the CNS, cholinergic neurons in the basal forebrain extend long-range axons to broadly innervate the entire cerebral cortex (Luchicchi, Bloem, Viaña, Mansvelder, & Role, 2014;Wu, Williams, & Nathans, 2014). Associated with arousal, focus, and emotional salience, ACh signaling has been found to modulate plasticity of synapses across brain regions including the hippocampus, hypothalamus, and visual cortex (Luchicchi et al., 2014;Picciotto, Higley, & Mineur, 2012). Receptors in cholinergic signaling include the ionotropic nicotinic acetylcholine receptors, and the G-protein coupled muscarinic receptors. The acetylcholinesterase enzyme (AChE) degrades ACh at cholinergic synapses to terminate synaptic transmission and thereby prevents neuronal hyperactivity and excitotoxicity. In contrast to synaptic transmission, far less is known about the effects of acetylcholine and AChE inhibitors on oligodendrocyte development and function.
The severity of GWI correlates with exposure to pesticides and PB in a dose-dependent manner (Steele, Sastre, Gerkovich, & Cook,2012;Wolfe, Proctor, Erickson, & Hu, 2002). Butyrylcholinesterase (BChE) is a nonspecific cholinesterase that hydrolyzes choline-based esters including toxicants such as organophosphate pesticides. A genetic variant of BChE, that encodes a less active form of BChE and hence is less adept at neutralizing GW toxicants, is one of the reported genetic risk factors for GWI (Steele et al., 2015). GW veterans with the atypical BChE gene are more susceptible to developing GWI upon exposure to GW pesticides. Together these findings highlight the important role of AChE inhibitors in the pathophysiology of GWI (Golomb, 2008).
Recent studies provide support for this previously unexplored hypothesis of involvement of myelinating glia in the pathophysiology of GWI. Brain imaging studies have reported white matter abnormalities in GW veterans (Chao, Zhang, & Buckley, 2015;Heaton et al., 2007) and such disruption has been associated with the key diagnostic symptoms of GWI: musculoskeletal pain (Rayhan et al., 2013;Van Riper et al., 2017), impaired attention (Janulewicz et al., 2017), disturbances of mood (Van Riper et al., 2017), and chronic fatigue (Rayhan et al., 2013). However, white matter is a complex tissue comprised of axons, astrocytes, oligodendrocytes, vascular cells, and microglia. Therefore, alterations in white matter detected by MRI could result from many types of cellular perturbations, including changes in axon number, diameter, tortuosity, vascular changes, alterations in astrocyte number or morphology, as well as direct effects on myelin. Furthermore, loss of myelin could be secondary to loss of axons, rather than a direct effect on oligodendrocytes.
Myelination proceeds in different brain regions at different times, but the process continues through adolescence into early adulthood.
In the prefrontal cortex (PFC), myelination continues during the early 20s (Miller et al., 2012), a demographic accounting for roughly 50% of deployed GW soldiers (Veterans Affairs, 2011). Additionally, there is a reserve pool of proliferative NG2+ cells in the adult brain, which have the potential to generate oligodendrocyte lineage cells throughout life (Nishiyama, Suzuki, & Zhu, 2014;Kang, Fukaya, Yang, Rothstein, & Bergles, 2010). Epidemiological data indicate that GW veterans who report impaired cognition as their prominent symptom were significantly younger than their GW veteran counterparts with nosymptoms or who experience primarily sensory symptoms (Gopinath et al., 2012). This pattern is consistent with possible involvement of disrupted PFC myelination in GWI and presents a compelling hypothesis for the neurological and cognitive impairments of GWI.
Here we test the hypothesis that acute exposure to AChE inhibitors affects oligodendrocyte proliferation, differentiation, survival, and myelination. These studies were carried out in an established Center for Disease Control (CDC) rat model of GWI (Koo et al., 2018) in combination with studies in cell culture. Key diagnostic features of GWI include musculoskeletal pain, impaired cognitive functioning, disturbances of mood, and debilitating fatigue; symptoms that have persisted over time (Binns et al., 2008;Maule et al., 2018;White et al., 2016). GWI animal models replicate many of these symptoms (Zakirova et al., 2016), including impaired working memory (Phillips & Deshpande, 2018) and social memory (Zakirova et al., 2016). The established animal model of GWI includes treatment with corticosterone (Cort) for 7 days before exposure to diisopropyl fluorophosphate (DFP), an irreversible AChE inhibitor used as a proxy for sarin nerve gas (Koo et al., 2018;O'Callaghan, Kelly, Locker, Miller, & Lasley, 2015;Zakirova et al., 2016). This necessitates studying the effects of Cort exposure independently, and together with DFP, on oligodendrocyte development and function. Interaction between these two agents is possible in influencing oligodendroglial biology. Stress and corticosterone have been shown to influence oligodendrocyte and myelin biology outside of the context of GWI. Corticosterone treatment has been shown to inhibit OPC proliferation (Alonso, 2000), promote OPC differentiation (Mann et al., 2008), and shorten the node of Ranvier length (Miyata et al., 2016). Inhibition of ACh signaling promotes remyelination in experimental autoimmune encephalomyelitis studies, an animal model of Multiple Sclerosis (MS), and in human MS clinical trials (Abiraman et al., 2015;Green et al., 2017;Li, He, Fan, & Sun, 2015;Liu et al., 2016;Mei et al., 2014;Welliver et al., 2018). Although this treatment is therapeutic for a demyelinating disease, disrupting ACh signaling may be detrimental to oligodendroglia in other contexts. In studies reported here, effects of DFP and Cort were investigated in adult rats in the GWI model, in the PFC, which is still undergoing myelination, and in subcortical white matter (corpus callosum), which in comparison to PFC is undergoing less active myelination.
Both DFP and Cort may act directly on oligodendroglia and indirectly by disrupting neuron-glia interactions. In addition to disrupting cholinergic signaling, these agents could have non-cholinergic actions or produce systemic effects, such as vascular and immune responses, that could have detrimental effects on myelinating glia. These alternatives were investigated using a combination of in vivo and in vitro studies. The results indicate that exposure to DFP, with and without Cort, disrupts oligodendrocyte development in the GWI animal model.
In vitro experiments using purified oligodendrocyte lineage cell monocultures, in the absence of detectible ACh, indicate that DFP and corticosterone have direct effects on oligodendroglial cell proliferation and survival, but these effects differ in important respects from those seen in the animal model of GWI. This finding distinguishes the consequences of the systemic and non-cholinergic effects of GW agents from their role in disrupting cholinergic signaling between axons and oligodendroglia as AChE inhibitors. The results of this study support the conclusion that oligodendrocyte biology is an important contributor to the pathophysiology of GWI and that GW agents impair cholinergic signaling between axons and myelinating glia but also have direct noncholinergic effects on these cells. Cort treatment in the GWI animal model has additional and, in some respects, counteracting effects to DFP on oligodendrocytes. The findings suggest possible therapeutic approaches to alleviate the chronic neurological symptoms in GW veterans from exposure to anticholinesterase agents during the GW.

| Mixed glial cell culture preparation
Primary rat mixed glial cell cultures were generated from P1-2 day old wild-type Sprague-Dawley rat pups. Briefly, pups were decapitated, and their cerebral cortices were isolated, minced, separated into a single cell suspension, and plated in T75 flasks. Mixed glial cultures were grown in Dulbecco's Modified Enriched Media (DMEM, ThermoFisher Scientific, Waltham, MA, Cat. No. 11995-065) that contained high glucose, L-glutamine, phenol red, and sodium pyruvate with 10% Fetal Bovine Serum (FBS, ThermoFisher Scientific, Cat. No. 16000-044) for 3 weeks at 37 C and 10% CO 2 .

| Oligodendrocyte progenitor cell purification
At 3-4 weeks post-dissection, flasks were shaken (180 rpm) for an hour at 37 C to remove microglia and dead cells, followed by a compete media change and an overnight shake under the same conditions. Media was collected from shaken flasks and plated onto two 6 cm tissue culture dishes per flask, for 15 min, to separate OPCs from heavier endothelial and astrocyte cells. Supernatant from 6 cm dishes was collected and centrifuged for 10 min at 1200 rpm. Cells were then plated onto 25 mm glass coverslips coated with 0.1 mg/mL poly-L-lysine (PLL) (Sigma-Aldrich, P9155) and 0.1 mg/mL poly-L-ornithine (PLO) (Sigma-Aldrich, P3655). Coverslips were used for calcium imaging 1-3 days post-plating. 80-90% of the cells on coverslips used for experiments were oligodendrocytes as confirmed by immunocytochemistry with the pan-oligodendrocyte marker, Olig2. Purified OPCs were grown in DMEM+10% FBS (described above) without additional growth factors. After 15 mins, 1 mL of HEPES buffer was added and the cells were then incubated a second time for 15 min in the dark at room temperature. Coverslip was washed one time for 10 min with HEPES buffer before use. MetaFluor Software (Molecular Devices) was used to image and measure fluorescence emission at 340 and 380 nm excitation wavelengths. Acetylcholine was diluted in HEPES buffer to 1 and 50 μM concentrations. Intracellular Fura-2 levels were calibrated using 10 μM A23187 with and without EGTA or Ca 2+ in buffer. Calcium concentrations were calculated from fluorescence levels using the equation derived by Grynkiewicz, Poenie, and Tsien (1985). While sampling, N was defined as a coverslip while n was defined as a cell.

| Oligodendrocyte differentiation
OPCs were differentiated in N1 media with 0.2% FBS. OPCs were plated from flasks, as described above, in DMEM + 10% FBS for 24 hr and then switched to N1 media + 0.2% FBS for the remainder of the experiment.

| Astrocyte cultures
At 3-4 weeks post dissection, flasks were shaken overnight at 37 C and 1200 rpm to remove microglia and dead cells. Media was removed, and flasks were washed twice with sterile Earle's Balanced Salt Solution (EBSS). Trypsin, warmed to 37 C, was added to the flasks and trypsinization was stopped 10 min later with 1:1 addition of DMEM + 10% FBS. Cells were collected and centrifuged at 1200 rpm and then plated on PLL/PLO coated 25 mm glass coverslips.

| In vitro cell culture treatments
OPCs isolated from postnatal day 2 (P2) cerebral cortex were treated 24 hr after plating. Cells were fixed with 4% paraformaldehyde and immunocytochemistry was performed 72 hr following treatment.

| Pharmacological agents
Acetylcholine (Sigma), DFP (Sigma), Corticosterone (Sigma) and Calcium Ionophore A23187 (Sigma) were used in the study. Ethanol (0.6%) was used as the corticosterone vehicle for in vivo and in vitro experiments. Doses used for each experiment are described in the results section.

| Tissue sectioning
Rats were sacrificed by decapitation and brains were rapidly removed.
One hemisphere was frozen for protein analysis and the other was post-fixed in 4% paraformaldehyde overnight and cryopreserved in 30% sucrose for up to 4 days. Following adequate cryopreservation, brains were embedded in optimal cutting temperature (OCT) embedding media (Fisher Healthcare). Embedded tissue was cyrosectioned into 14 μm thick sagittal sections.

| Immunohistochemistry
Tissue sections were brought to room temperature and then rinsed one time with PBS to dissolve the embedding medium OCT. The sections were then incubated in Citrate buffer, pH 6, at 94 C for 10 min to promote epitope retrieval. The tissues sections were then rinsed three times with PBS containing 0.1% TritonX-100 and blocked for 1 hr in blocking buffer (5% normal goat serum and 0.1% Triton X-100 in PBS). Sections were incubated overnight at 4 C with the primary antibodies. Primary antibodies used include: APC/CC1 (Millipore, Mouse) at 1:500; Olig2 (Millipore, Rabbit) at 1:500; Ki67 (Abcam, Rabbit) at 1:500; Olig2, (Millipore, Mouse) at 1:100. The following day, the tissue sections were rinsed three times with PBS containing 0.3% TritonX-100 and then incubated for 2 hr at room temperature with secondary antibodies (listed in the Immunocytochemistry methods section) at 1:200 dilution. Tissue sections were then rinsed three times with PBS and mounted on coverslips using mounting medium (Vectashield Antifade Mounting Medium with DAPI). Ki67 staining was performed on animals fixed 24 hr after treatment. CC1 staining was performed on animals fixed 21 days after treatment. A fluorescent light microscope with AxioCam MRm was used to acquire 10 images per region per animal at 40 X magnification.

| Acetylcholine release assay
Cells were plated onto coverslips as described above and incubated with 150 μL buffer or DFP at 37 C. Conditioned buffer was collected, and flash frozen after 4 hr. Buffer was analyzed with a Choline/Acetylcholine fluorometric assay kit (Abcam). To measure total levels of acetylcholine and choline ([Ach + Ch]), AChE enzyme was used in the buffer. To measure choline levels excluding acetylcholine, ([ACh]), DFP was added to the buffer with no AChE.

| Immunoblotting
To extract proteins from tissue and cell culture, samples were lysed in RIPA buffer (Sigma Aldrich) with protease inhibitor cocktails (Complete Mini EDTA-free Protease Inhibitor Cocktail, Sigma Aldrich).
Lysate was mixed with LDS sample buffer (Thermo Fisher Scientific) and electrophoresed in a 4-12% Bis-Tris Gel (Invitrogen) for 2 hr at 150 V in MOPS-SDS running buffer (Thermo Fisher Scientific). The samples were transferred to PVDF membrane (Immobilon-P, Millipore) overnight at 4 C in Tris-Glycine transfer buffer (Thermo Fisher Scientific).
Membranes were blocked in blocking buffer, containing TBS (10 mM Tris-HCl, pH 7.5, 0.9% NaCl), 0.1% (vol/vol) Triton X-100 and 5% (wt/vol) bovine serum albumin (MP Biomedicals) or 5% nonfat dry milk (American Bio), for 1 hr at room temperature (RT). The appropriate primary antibody was diluted in blocking buffer and incubated overnight with the PVDF membrane at 4 C. Primary antibodies used were: GAPDH (Cell Signaling Tech, Rabbit) used at 1:4000; NSE (Abcam, Rabbit) used at 1:2000; GAPDH (Encor, Mouse) used at 1:4000 dilution; MBP (Millipore, Rabbit) used at 1:1000; GFAP (Invitrogen, Rabbit) used at 1:500, GAP43 (Millipore Sigma Aldrich, Rabbit) used at 1:1000, NG2 (Abcam, Mouse) used at 1:1000. The PVDF membrane was washed four times, 15 min each, in washing buffer, TBS (10 mM Tris-HCl, pH 7.5, 0.9% NaCl) and 0.1% (vol/vol) Triton X-100. The corresponding secondary antibody, diluted in blocking buffer, was then incubated with the PVDF membrane for 2 hours at RT. Secondary antibodies that were used include ECL Anti-Mouse IgG Horseradish Peroxidase-linked F(ab')2 fragment or ECL Anti-Rabbit IgG Horseradish Peroxidase-linked F(ab')2 fragment. Chemiluminescent substrate was applied for 10 min (SuperSignal West Pico Plus, Thermo Scientific). Membranes were quantified with densitometry using Image J software and normalized to NSE loading control. In reported bar graphs of data, each treatment condition was normalized to saline control levels.    We conclude that AChE expressed on OPCs is functional and that its inhibition in the presence of ACh leads to increased AChdependent Ca 2+ signaling, which could influence OPC development and function.

| Direct effects of GW agents on oligodendrocyte biology
Any effect on oligodendrocyte development and function in the GWI animal model or in GW veterans could be due to direct effects on   Figure S1a) and Cort ( Figure S1b) to determine the most appropriate treatment concentration to avoid toxicity (defined by Olig2+ cell counts). Based on the dose response curves, we adopted an OPC treatment paradigm using 1 μM Cort, 1 μM DFP, and a combined Cort+DFP condition, as used in the GWI animal model. OPCs in growth medium were treated 24 hr after plating and were examined via immunocytochemistry 72 hr following treatment.
The data showed that Cort treatment significantly decreased OPC proliferation (Ki67+ Olig2+ cells) (Figure 3a, 10.6 vs. 2.6% respectively, To interpret whether the effects of DFP were due to cholinergic or non-cholingeric signaling we measured the levels of ACh in the culture serum. We previously showed that neither OPCs nor astrocytes secrete ACh. We measured no detectable ACh in the media containing 10 or 0.2% FBS used for OPC proliferation and differentiation, respectively, using a fluorometric assay with a sensitivity of ≥100 pmol. Since there was no detectable ACh in either culture medium, the observed responses were due to direct and non-cholinergic effects of Cort and DFP. Taken together, the data demonstrates that Cort inhibits proliferation and drives maturation of OPCs by direct action on these cells, while DFP increases maturation marginally, but also stimulates apoptosis. In combination with Cort, the toxicity of DFP in vitro acts to counter the promaturation effects of Cort. To test if ACh alone could directly affect MBP expression, we treated oligodendrocytes with ACh in vitro. We found that elevated ACh was not sufficient to increase MBP expression in cell-culture ( Figure S2a Bar graphs are total cell counts (N = 5, n = 50). X-axis is drawn at the saline control value. DFP had no effect on the frequency of proliferating oligodendrocytes in the corpus callosum (χ 2 [1, n = 3,729] = 0.240, p = 0.624). Cort treatment was associated with significantly fewer proliferating oligodendrocytes (χ 2 [1, n = 4,624] = 60.080, p < 0.001). Cort+DFP treatment was associated with significantly fewer proliferating cells than saline condition (χ 2 [1, n = 2425] = 6.910, p = 0.009). (f)

| Effects of GW agents on oligodendrocytes in corpus callosum
The PFC is implicated in GWI and it is still undergoing myelinating in young adults. The major white matter tract in rodents is the corpus callosum. We therefore analyzed the effects of GW agents on oligodendrocyte lineage cells residing in grey matter PFC versus white matter corpus callosum. The effects of DFP and Cort on proliferation and maturation of OPCs differed in the corpus callosum compared with the PFC. This is not unexpected, given the differences in the cellular composition and environment in these two regions. In the corpus callosum, 19.9% of Olig2+ cells were proliferating, immature OPCs, but in the PFC more cells were in this state (25.6%), as this grey matter region is still undergoing active myelination. In the corpus callosum, Cort ( Figure 5c, e, f, 19.9 vs. 12.8%, respectively, χ 2 [1, n = 4624] = 60.080, p < 0.001), but not DFP (Figure 5b, e, f, 19.9 vs. 19.4%, respectively,

| GWI animals have increased myelin basic protein levels in subcortical white matter
Changes in the integrity of white matter tracts has been found to underlie the key cognitive and sensory impairments in GW veterans (Rayhan et al., 2013;Van Riper et al., 2017). Therefore, we sought to understand next how the changes in oligodendrocyte development observed in the corpus callosum of the GWI animal model translates to myelin formation.
By using immunoblot to measure myelin basic protein (MBP) levels in subcortical white matter, we found that GWI animals co-treated with Cort and DFP had significantly increased levels of MBP in subcortical white matter as early as 12 hr posttreatment (Figure 6a, e, 1.0 ± 0.071 vs. 1.600 ± 0.078 relative MBP expression, respectively, ANOVA Histological analysis showed that at later time points, Cort and Cort + DFP increased the number of mature oligodendrocytes in the corpus callosum (Figure 5k), which is consistent with increased subcortical MBP levels measured by immunoblot. The observed changes in protein expression in vivo were detected as early as 12 hr following treatment, which is too early to be caused by changes in OPC proliferation and differentiation, as the cell cycle of OPCs is~25-30 hr (Durand, Gao, & Raff, 1997). Thus, the early increase in MBP protein must reflect an increase in its synthesis. Immunoblot analysis of a marker of OPCs supports this conclusion. We measured OPC membrane protein marker NG2 by immunoblot analysis after 12, 24, and 72 hr Alternatively, neuroinflammation and injury can alter axonal sprouting (Chen & Zheng, 2014), which would create new axonal branches to be myelinated, thus stimulating synthesis of MBP. To test this hypothesis, we measured the levels of growth associated protein 43 (GAP43), a marker for axonal sprouting (Benowitz & Routtenberg, 1997) and found that GAP43 levels in subcortical white matter did not change after any treatment condition at 12 hr ( Figure S4a Additionally, to test if changes in astrocyte biology were responsible for increased subcortical MBP levels, we measured the levels of the astrocytic protein, GFAP, by immunoblot. GFAP levels remained unaffected by treatment conditions at 12 hr ( Figure S5a,  posttreatment. This is consistent with studies showing GFAP expression is not altered up to 3 days posttreatment with GW agents (O'Callaghan et al., 2015). However, at 21 days post treatment, GW agents significantly affected GFAP expression ( Figure S5d with decrease in GFAP expression with chronic exposure to Cort (O'Callaghan et al., 1991;Nichols, Osterburg, Masters, Millar, & Finch, 1990) and DFP (Gupta & Abou-Donia, 1995) reported in other studies.
The results do not support axonal sprouting, early OPC proliferation or gliosis as the source of increased MBP. In conclusion, immunoblot of subcortical white matter indicates persistently elevated MBP protein levels in subcortical white matter, potentially via elevated MBP protein production at early timepoints (i.e., 24 hr) and by an increase in the number of mature oligodendrocytes in the corpus callosum at later timepoints (21 days). Further, the finding that exogenous ACh on an in vitro monoculture of OPCs is insufficient to change MBP protein levels ( Figure S2) suggests the necessity of interactions with axons as axons provide physical support and an appropriate substrate for myelination to promote MBP production in oligodendrocytes (Wake et al., 2011).
Together these results show effects on OPC proliferation, survival, maturation, and increased myelin basic protein expression in the subcortical white matter resulting from treatment with Cort and DFP in the GWI animal model. These effects differ in different brain regions and in cell culture because of the differential contribution of cholinergic and non-cholinergic effects of GW agents and differences in cellular environments of grey and white matter regions where these oligodendroglial populations reside (Table 2).

| DISCUSSION
In this study we have investigated the plasticity of myelinating glia in the context of white matter abnormalities in GWI and demonstrated that biology and development of oligodendrocyte lineage cells are significantly affected by exposure to AChE inhibiting agents and the stress hormone, corticosterone. We conclude that DFP, a sarin nerve gas surrogate, decreases maturation of OPCs when acetylcholine signaling is present. DFP also decreases proliferation of oligodendrocytes in the PFC, a region with a higher percentage of proliferating oligodendrocytes than in the corpus callosum.
Our finding that DFP decreases maturation of oligodendrocytes is consistent with the current literature about the role of ACh signaling on oligodendrocyte development (Fields et al., 2017). In MS human clinical trials and MS animal models, inhibition of muscarinic receptors and consequent inhibition of ACh signaling is shown to promote remyelination (Abiraman et al., 2015;Green et al., 2017;Li et al., 2015;Liu et al., 2016;Mei et al., 2014;Welliver et al., 2018). Therefore, our finding that elevated ACh signaling decreases the number of mature oligodendrocytes is consistent with the literature. In our noncholinergic in vitro studies, DFP increased the frequency of mature oligodendrocytes and also significantly increased apoptosis in OPCs.
Another factor to consider, is that the CDC GWI animal model includes chronic Cort exposure prior to DFP exposure. In terms of oligodendrocyte and white matter biology, we find that Cort alone has robust protective effects by promoting maturation of oligodendrocytes and decreasing proliferation. These findings are consistent with previous literature exploring the effects of Cort on oligodendrocyte biology (Alonso, 2000;Miyata et al., 2016).
Given the somewhat antagonistic interactions of DFP and Cort, we find that the Cort+DFP condition reveal region specific effects on oligodendrocyte biology. With Cort+DFP treatment, we find that in the PFC, OPCs are pushed into a more proliferative and less mature state. In contrast, with the same treatment paradigm in the corpus callosum, OPCs are pushed to a less proliferative and more mature state.
This could be due to the relatively more mature population of oligodendrocytes in the corpus callosum as well as differences in ACh availability in the two regions. The importance of regional differences is highlighted by the reported abundance of cognitive rather than sensory symptoms in the younger GW veterans (Gopinath et al., 2012), implicating the involvement of PFC, a region where myelination continues into the third decade of life (Miller et al., 2012). The decrease in mature oligodendrocytes identified in the PFC suggests that the PFC is especially vulnerable to GW agents; a finding that is consistent with the neuropsychological impairments presented by GW veterans (Janulewicz et al., 2017;Sullivan et al., 2003;Sullivan et al., 2018). Our cell-culture studies more closely match the data from the corpus callosum where Cort+DFP promotes maturation of oligodendrocytes. Given that there is no detectable ACh in our OPC monocultures, we find that promotion of maturation by Cort overwhelms the anticholinesterase effect of DFP.
MBP protein was elevated in subcortical white matter. The increase in MBP levels with Cort+DFP co-treatment, could be due to promotion of the oligodendrocyte lineage toward a more mature state by Cort in the corpus callosum. However, an increase in MBP does not necessarily imply a healthier and more functional white matter (Kristensson et al., 1986). Increased MBP mRNA transcripts have been reported during periods of demyelination, providing evidence for a compensatory increase in MBP levels in response to pathology (Kristensson et al., 1986). A transient increase in MBP mRNA, occurring within 6 hr, has been previously reported as an oligodendroglial cellular response to injury (Bartholdi & Schwab, 1998). This argument is underscored by studies showing that DFP exposure of rats increases myelin decompaction while having no effect on the g-ratio or white matter volume (Naughton et al., 2018). Therefore, it is important to note the complicated relationship between MBP levels, oligodendrocyte maturity, and myelin integrity.
The results of the quantitative histological analysis of oligodendroglia in the PFC of the GWI rat model is consistent with RNA-seq data in mice showing that combined Cort and DFP treatment decreases the fraction of mRNA transcripts associated with mature oligodendrocytes in the PFC (Ashbrook et al., 2018), but our results are not consistent with this gene profiling study in other respects.
Cort treatment alone did not alter the number of mature oligodendrocytes in the PFC, but mRNA transcripts associated with oligodendrocytes are reportedly reduced in this condition (Ashbrook et al., 2018).
Also, the frequency of mature oligodendrocytes in the PFC decreased in response to DFP treatment, but mRNA transcripts associated with mature oligodendrocytes remained unchanged under this condition (Ashbrook et al., 2018). These discrepancies may be explained by the fact that gene and protein abundance are indirect indicators of cell numbers, and gene expression and protein levels are influenced by physiological conditions. Alternatively, methodological or species differences could also account for discrepancies between mRNA profiling and histological analysis.
Based on our findings we predict that GW veterans would have decreased white matter integrity varying by brain region. These predictions are supported by the brain imaging studies of GW veterans (Bierer et al., 2015;Rayhan et al., 2013;Van Riper et al., 2017). The published neuroimaging data reflect both increased and decreased myelin integrity depending on the myelin track analyzed and methodological differences in measurement. For example, previous GWI imaging data with MRI has shown that axial diffusivity in the right inferior fronto-occipital fasciculus, a white matter tract that links cortical regions involved in fatigue, pain, emotional and reward processing, and the right ventral attention network in cognition, is significantly increased in GW veterans and correlate with the severity of pain and fatigue (Rayhan et al., 2013). In veterans with post-traumatic stress disorder (PTSD), increased structural integrity has been reported in the cingulum bundle, a white matter tract connecting the right amygdala and anterior cingulate cortex (Bierer et al., 2015). Importantly, an equal number of studies have identified decreased myelin integrity depending on the brain region. It has been shown in veterans with GWI and chronic pain that there is a lower white matter integrity across multiple brain regions including the frontal gyrus, corpus callosum, and precentral gyrus (Van Riper et al., 2017). GW veterans with PTSD also display significantly reduced mean diffusivity in the right, but not left cingulum (Bierer et al., 2015). Our studies show that DFP treatment of OPC monoculture also promotes toxicity. Taken together, the data and available evidence suggest that the effects of AChE inhibition, corticosterone exposure, and their combined treatment, on oligodendrocyte biology and white matter vary depending on the brain region and cell environment, reflecting underlying differences in ACh availability and cellular composition between regions.

| CONCLUSION
Our study shows that impairment of oligodendrocyte biology is an important aspect of the pathophysiology of GWI. We have identified that DFP, an analog to sarin nerve gas, reduces the frequency of differentiated oligodendrocytes across multiple brain regions. Our data shows that Cort, used in the CDC animal model, antagonizes the effects of DFP, as Cort alone increases the frequency of differentiated oligodendrocytes. With co-treatment of Cort and DFP, we find a lower frequency of CC1+ oligodendrocytes in the PFC and higher frequency of CC1+ oligodendrocytes in the corpus callosum. The cell count data in the corpus callosum is supplemented and corroborated by elevated MBP levels in the sub cortex. These differences highlight the heterogenous responses of oligodendrocytes to agents implicated in GWI and used in the GWI animal model. Similar heterogeneity is reflected in brain imaging studies and in the wide range of symptoms experienced in GWI. Taken together, these findings suggest therapeutic avenues where restoring the endogenous cholinergic signaling required for normal oligodendrocyte cell biology and function may potentially alleviate the chronic symptoms of veterans with GWI. This study also suggests that civilian exposure to AChE inhibitors, such as commercial pesticides, may have chronic effects on white matter, especially during childhood and early adolescence, when the brain is at its most plastic.