Human Leukocyte Antigen (HLA) Alleles Prevent Metabolically-Induced Inflammation and Cerebrocortical Thinning in Gulf War Illness

Page 16 of 27 Human Leukocyte Antigen (HLA) Alleles Prevent Metabolically-Induced Inflammation and Cerebrocortical Thinning in Gulf War Illness Peka Christova1,2, Lisa M. James1,2,3, Adam F. Carpenter1,4, Scott M. Lewis1,4, Brian E. Engdahl1,5, Apostolos P. Georgopoulos1,2,3,4* 1Brain Sciences Center, Department of Veterans Affairs Health Care System, Minneapolis, MN, 55417, USA 2Department of Neuroscience, University of Minnesota Medical School, Minneapolis, MN 55455, USA 3Department of Psychiatry, University of Minnesota Medical School, Minneapolis, MN 55455, US 4Department of Neurology, University of Minnesota Medical School, Minneapolis, MN 55455, USA 5Department of Psychology, University of Minnesota Medical School, Minneapolis, MN 55455, USA


Introduction
Nearly thirty years after the 1990-1991 Persian Gulf War, Gulf War era veterans have continued to suffer from chronic, diffuse symptoms commonly referred to as Gulf War Illness (GWI) 1,2 . Although respiratory, gastrointestinal, and dermatological systems are involved in the GWI symptom constellation, most of the symptoms such as fatigue, pain, and neurological, cognitive and mood disruptions implicate primarily the central nervous system and the brain, in particular. Consequently, considerable research has investigated the nature and causes of brain abnormalities associated with GWI.

GWI and brain atrophy
Numerous functional and structural brain abnormalities associated with GWI have been identified 3,4 . Among the most consistent findings regarding brain abnormalities has been evidence of brain atrophy in veterans with GWI [5][6][7][8][9][10][11][12] . We recently reported substantial (~10%) atrophy in veterans with GWI relative to healthy veterans 5 , in areas proximal to the most permeable regions of the blood brain barrier (BBB) 13 . Disruption of the BBB has been shown to be associated with atrophy in an animal model of GWI in which rats were exposed to a combination of stress restraint and chemicals (nerve agent prophylactics and insecticides) that Gulf War veterans may have encountered in theater 14 . Brain morphology is likely to be compromised by a breakdown in BBB integrity resulting from chronic inflammation [15][16][17] , which is common in GWI.

GWI: Lack of immunogenetic protection
Mounting evidence from human [18][19][20][21] , in vitro [22][23][24][25][26] , and in silico 27 studies implicates persistent antigens stemming from lack of immunogenetic protection in GWI and GWassociated inflammation. Specifically, six Class II human leukocyte antigens (HLA), which play an essential role in host protection via antibody production, have been shown to distinguish healthy GW veterans from those with GWI, and the number of these protective HLA alleles corresponds with GWI symptom severity in a dose dependent manner 18 . Furthermore, these protective HLA alleles have been shown to influence brain function in Gulf War veterans 19 and protect against brain atrophy 21 characteristic of veterans with GWI 5 . Given the instrumental role of Class II HLA in antibody production against foreign antigens via CD4+ T cells, these studies point to lack of immunogenetic protection against foreign antigens leading to antigen persistence and, consequently, inflammation in veterans with GWI 18,21 . We have speculated that persistent foreign antigens from vaccines administered to Gulf War veterans lacking immunogenetic protection against them may underlie GWI. To that end, in vitro studies have demonstrated harmful effects of serum from GWI veterans on neural cultures that is ameliorated with serum from healthy Gulf War veterans 22 , pooled antibodies 23 , and antibodies against Gulf War vaccine antigens [24][25][26] .

GWI and inflammation
Several studies have documented evidence of inflammation in veterans with GWI 28-31 . C-reactive protein (CRP), a non-specific biomarker of inflammation, has been shown to be elevated in GWI [28][29][30][31] , and has been implicated in GWI symptoms involving the brain, in particular 28 . Although historically viewed as a marker of inflammation that arises in response to inflammatory cytokines such as interleukin 6, evidence also suggests that CRP may also play a causal role in inflammation 32 . Thus, CRP may both signal and potentiate inflammation, thereby increasing BBB permeability. Indeed, systemic inflammation has been shown to predict BBB dysfunction 33 . Furthermore, inflammation has been intimately connected to metabolic disorders 34 which are common among Gulf War veterans 35 .

Systemic inflammation and neurodegeneration
Considerable evidence suggests a robust association between CRP and chronic inflammatory and neurodegenerative conditions including cardiovascular disease, type 2 diabetes mellitus, age-related macular degeneration, hemorrhagic stroke, Alzheimer's disease, and Parkinson's disease 36 . In addition, elevated CRP levels have been associated with brain atrophy even in healthy people [37][38][39][40][41] , highlighting the contribution of inflammatory markers on brain morphology independent of conditions that are typically associated with neurodegeneration. Although the mechanisms remain to be fully elucidated, there is evidence that CRP stimulates production of proapoptotic cytokines including interleukin-1β (IL-1β), tumor necrosis factor-α (TNFα), and reactive oxygen species 32 . Elevated peripheral inflammation has been associated with declines in cognitive function, an effect presumed to be mediated by the effect of inflammation on brain morphology 39 . BMI is also inherently related to inflammation because adipose tissues release inflammatory cytokines, which promote further metabolic disruptions in a vicious cycle 34 .

The current study
Separate lines of research have demonstrated that GWI is associated with elevated CRP, metabolic conditions, and with brain atrophy, and that protection from GWI is conferred by presence of specific Class II HLA alleles. The aim of the present study was to evaluate the association between CRP, body mass index (BMI), and brain morphometry in veterans with GWI with regard to HLA. We have proposed that exposure to foreign antigens coupled with a genetic vulnerability that precludes their elimination results in persistent antigens and, ultimately, inflammation, cell damage, and atrophy 18 ,21, 42 . Recent advances in morphometric analyses permit distillation of cortical volume into its component parts, surface area and cortical thickness, allowing for finer grain examination of the influence of inflammation on these two variables independently.

Participants
Sixty-three veterans with GWI (59 men and 4 women; age 55.2 ± 1.1 y, mean± SEM) gave informed consent to imaging acquisitions in accordance with the Declaration of Helsinki. GWI status was determined using a self-report symptom checklist that permits classification as GWI case or control according to the Center for Disease Control 1 and the Kansas criteria 2 , the case definitions recommended by the Institute of Medicine 43 . The Center for Disease Control definition requires one or more symptoms in at least two domains that include fatigue, pain, or mood and cognition. The more restrictive Kansas criteria requires that veterans report moderate to severe symptoms in at least 3 of 6 domains: fatigue, pain, neurological/cognitive/mood, skin, gastrointestinal, and respiratory. All GWI veterans in the present study met both case definitions. Consistent with the Kansas criteria case definition, none of the participants had co-occurring medical or psychiatric conditions that could account for GWI symptoms or impair reporting 2 , including veterans with schizophrenia, bipolar disorder, drug abuse or alcoholism, neurological conditions and inflammatory rheumatological conditions. Individuals with traumatic brain injury were also excluded from the study. The study was approved by the University of Minnesota and Minneapolis VA Health Care System Institutional Review Boards.

Body Mass Index (BMI)
BMI was computed using the height and weight of the participant (BMI = kg/m 2 ).

CRP
Non-fasting peripheral venous blood samples were collected for evaluation of high sensitivity C reactive protein and analyzed using standard procedures by the Minneapolis VAHCS Clinical Laboratory.

HLA Genotyping
DNA isolation was carried out from 3 ml of whole blood drawn in EDTA tubes, using a commercially available kit (ArchivePure cat. 2300730) from 5Prime (distributed by Fisher Scientific or VWR) with an expected yield of 50-150µg of DNA. The purified DNA samples were sent to Histogenetics (http://www.histogenetics.com/) for high-resolution HLA Sequence-based Typing (SBT; details are given in https:// bioinformatics.bethematchclinical.org/HLA-Resources/ HLA-Typing/High-Resolution-Typing-Procedures/ and https://bioinformatics.bethematchclinical.org/WorkArea/ DownloadAsset.aspx?id=6482). Their sequencing DNA templates are produced by locus-and group-specific amplifications that include exon 2 and 3 for class I (A, B, C) and exon 2 for class II (DRB1, DRB3/4/5, DQB1, and DPB1) and reported as Antigen Recognition Site (ARS) alleles as per ASHI recommendation 44 .

MRI image processing
Data were processed by a 704-core High Performance Computing system (CentOS 6.5 Linux, Rocks 6.1.1), Matlab R2016 (64 bit), and the Human Connectome Project (HCP; humanconnectome.org) pipeline, including the FreeSurfer (FS; http://surfer.nmr.mgh.harvard.edu) HCP version. Full details were presented previously 5 . Briefly, we used a modified version of FS, implemented in the structural HCP pipeline, which utilizes both T1-w and T2-w images to eliminate uncertainty and improve surface reconstruction 45 .
FS, a widely used tool, provides reliable brain measurements validated by manual 46,47 and histological analysis 48 . FS's surface-based stream reconstructs the cortical surface by using models of the boundary between the white and gray matter and between the gray matter and cerebrospinal fluid (this is the pial surface) 49 -51 . This algorithm uses intensity and continuity information to generate a continuous cortical ribbon, and creates a mesh of hundreds of thousands of triangles that recover the geometry and the topology of the pial surface and gray/ white boundary. Cortical thickness is calculated as the closest distance from the gray/white boundary and pial surface at each vertex on the tessellated surface in both hemispheres 52 . The FS procedure allows submillimeter resolution that is higher than the voxel resolution of the original data. The cortical ribbon is then parcellated into regions of interests (ROIs) based on gyral and sulcal structures of the Desikan-Killiany atlas 53 . Each resulting image was visually inspected for inaccuracy. Mean thickness, cortical volume, and surface area were extracted for each brain ROI from each individual participant's brain image.

Brain areas
The volumes, surface area and thickness of 68 cerebrocortical Regions of Interest (ROIs; 34 left and 34 right) were analyzed. Their names are given in the Appendix.

MRI Acquisition
As mentioned in the Methods above, there were two kinds of MRI acquisitions, one using an 8-channel system and the other a 15-channel system. Participants were distributed similarly In the HLA-p0 group, 13 and 28 participants were imaged with the 8-and 15-channel equipment, respectively, whereas in the HLA-p1 group those counts were 5 and 17, respectively; there was no significant association between the HLA groups and the MRI method of acquisition (P = 0.564, Fisher's exact test for the 2x2 table).

Comparisons of non-neural measures between the 2 HLA groups
Gender. The HLA-p0 group comprised 37 men and 4 women, whereas the HLA-p1 group comprised 22 men; the two groups did not differ significantly with respect to the gender composition (P = 0.288, Fisher's exact test for the 2x2 table).
CRP ( Figure 2). CRP was appreciably lower in the HLA-p1 group but the difference did not reach statistical significance: The mean ± SEM ln(CRP) for the HLA-p0 group was 0.81 ± 0.15, and that for the HLA-p1 group was 0.44 ± 0.20. These means did not differ significantly (t-test, t [61] = 1.46, P = 0.149). (Figure 3). In the HLA-p0 group, ln(CRP) was positively and significantly correlated with BMI (r = 0.409, P = 0.006, N = 41). In contrast, there x 22 HLA-p1 brains = 1496 values. In addition, for each participant, the following data were available: gender, age, HLA group (presence or absence of a GWI protective allele), BMI, and CRP. CRP values were skewed to the right and were (natural) log-transformed to normalize their distribution 28 . Standard statistical methods were used to analyze the data, including t-test, Fisher's exact test for 2x2 tables, and partial correlation analyses. All statistical analyses were done using the IBM-SPSS statistical package (version 23).   Association of CRP with cortical surface. No statistically significant partial correlations were found between ln(CRP) and cortical surface in either HLA group (HLA-p0, controlling for age, gender, and eTIV: r p = −0.002, P = 0.903, DF = 2783; HLA-p1, controlling for age and eTIV: r p = −0.018, P = 0.405, DF = 1492). (Figure 4). A highly statistically significant partial correlation was found between ln(CRP) and cortical thickness in the HLA-p0 group (r p = −0.063, P = 0.00092, DF = 2783, controlling for age, gender, and eTIV). In contrast, no significant partial correlation was found for the HLA-p1 group (r p = 0.045, P = 0.081, DF = 1492, controlling for age and eTIV; notice the positive correlation). A more detailed examination of the ln(CRP) vs. thickness relation in the HLA-p0 group revealed that a better fit was a power fit, i.e. a linear relation in a log-log scale. This makes sense, since it implies that even very high levels of CRP will never result in zero thickness. The results for the log-log relation were: r p = −0.07, P = 0.000208, DF = 2783, controlling for age, gender, and eTIV). The association of cortical thickness (lnCT) and CRP (lnCRP) remained significant after accounting for BMI (r p = −0.069, P = 0.000284, DF = 2782, controlling for BMI, age, gender, and eTIV). BMI was not significantly associated with cortical thickness in either group (HLA-p0 group: r = −0.023, P = 0.215, N = 2788; HLA-p1 group: r = −0.037, P = 0.157, N = 1496).

Association of CRP with cortical thickness (CT)
Comparison of cortical thickness between the two HLA groups. Cortical thickness, adjusted for age, gender, eTIV, and ln(CRP) was overall higher in the HLA-p1 than the HLA-p0 group but this difference did not reach statistical significance; the mean ± SEM of ln(CT) for the HLA-p0 group was 0.911 ± 0.003, N = 41, and for the HLA-p1 group 0.916 ± 0.005, N = 22; P = 0.346). was no significant correlation in the HLA-p1 group (P = 0.309).

Comparisons of association of CRP with neural measures between the 2 HLA groups
The association of CRP with neural measures (cortical volume, surface and thickness) was evaluated by computing the partial correlation coefficient between ln(CRP) and each measure above, controlling for estimated total intracranial volume (eTIV), age and gender, as needed (e.g. gender was entered as covariate only for the HLA-p0 group, since the HLA-p1 group comprised only men). We found the following.
Association of CRP with cortical volume. No statistically significant partial correlations (r p ) were found between ln(CRP) and cortical volume in either HLA group  inflammation and obesity has been referred to as a "chicken or the egg" question as activation of the immune system can promote weight gain and obesity, and obesity can perpetuate inflammation 34 . Furthermore, CRP and BMI have each been linked to reduced cortical thickness 66, 67 , as has expression of genes specific to astrocytes and microglia, the brain's resident immune cells 68 . Here, CRP, a marker of inflammation, was highly significantly associated with BMI and cortical thinning, but only in veterans lacking a protective HLA allele; BMI was not associated with cortical thinning in either group. This suggests that persistent antigens resulting from lack of immunogenetic protection against them may contribute to both inflammation and obesity, the latter of which may promote additional inflammation. In contrast, there were no significant associations between CRP, BMI, and cortical thickness in GWI veterans possessing 1 protective HLA allele. Taken together, these findings raise the possibility of at least two inflammatory drivers in GWI -one stemming from persistent antigens and a separate mechanism that includes metabolically-induced inflammation. It also raises the possibility that the varied symptoms associated with GWI may be partially attributable to differences in HLA composition. For example, cognitive symptoms may be more prominent in those lacking HLA protection due to reduced cortical thickness stemming from persistent antigens. Next, we discuss environmental exposures that may contribute to antigen persistence in GWI, highlighting the role of vaccines.

In-theater exposure and GWI
GWI is typically attributed to deployment-related exposures either alone or in combination with psychological stress or other exposures. Commonly investigated intheater exposures includes chemical nerve agents (e.g., sarin) and widespread use of pyridostigmine bromide (PB) to protect against possible nerve agent attacks, frequent use of pesticides, and airborne contaminants from oil well fires. While implicating in-theater exposures has provided convenient explanations for the excess symptomatology seen in GWI veterans, relative to nondeployed GW veterans or those deployed to other regions, the lack of documentation on actual in-theater exposures limits causal connections regarding exposure to any specific agent and GWI. Furthermore, despite decades of study it remains unclear how exposure to such toxicants could contribute to GWI symptoms 30 years later. Take, for instance, sarin. A number of military personnel were exposed to low-level sarin following the destruction of an Iraqi ammunition dump, in Khamisiyah, Iraq. Although sarin is extremely toxic, it is short-lived inside the body where it is quickly metabolized and excreted through urine. Thus, while some deleterious effects may be expected following sarin (and other toxicant exposure), it is unclear how sarin could contribute to progressive effects nearly 30 years later given the rapid metabolism and short half-life. A condition known as organophosphate-induced

Discussion
Here we investigated the effect of GWI-protective HLA alleles on the association between inflammation, BMI, and brain morphometry in veterans with GWI. The results demonstrated a highly significant association between CRP with BMI and cortical thickness, but not surface area or cortical volume, in GWI veterans lacking one of the protective Class II HLA alleles. There were no such associations in veterans possessing one of the protective alleles. These findings suggest that persistent antigens stemming from lack of HLA protection influence metabolic and inflammatory systems, contributing to cortical thinning. These results add to the literature highlighting the role of inflammation in GWI, narrow the focus of GWI risk factors to those that are known to contribute to inflammation, and point towards potential treatment avenues aimed at reducing inflammation. We discuss below GW-related exposures, highlighting the potential role of persistent antigens on GWI inflammation.

Differential Associations of CRP on Brain Morphometry
Several studies have documented morphometric alterations associated with GWI including reduced brain volume 4, 5 . Here we focused on the cerebral cortex, analyzing its volume, surface area, and thickness. The current study documents the association between CRP and reduced cortical thickness in GWI veterans lacking protective HLA alleles, but no association with surface area or cortical volume. This dissociation of effects is consistent with findings that changes in cortical gray matter volume are almost exclusively driven by changes in surface area 54,55 , and that surface area and volume are underlain by distinct genetic influences 56 and reflect different aspects of brain architecture and distinct cellular mechanisms 55 . In addition, cortical thickness and surface area are negatively correlated 57 . Cortical thickness is a more sensitive measures of gray matter alteration than volume 58 , and many neurological and psychiatric conditions are associated with variations in cortical thinning [59][60][61][62][63][64] , leading some to suggest cortical thickness is a candidate biomarker for differential diagnosis and disease prognosis 65 . The current study provides evidence of cortical thinning in veterans with GWI who lack immunogenetic protection; longitudinal studies are warranted to examine the association of cortical thickness with GWI disease prognosis.

Metabolic and Immune System Interactions and Cortical Thinning
Considerable research highlights a robust association between the immune and metabolic systems such that inflammation is unequivocally causally associated with metabolic system disruptions which promote further inflammation 34 . Indeed, the association between chronic neurotoxicity (OPICN) has been described in the literature, and is characterized by neuropathological, neurological, neuropsychological, and neurobehavioral alterations as a result of a large single exposure or repeated small sarin exposures 69 . Notably, the characteristics are similar to those seen in GWI and some have considered OPICN as a possible explanation for GWI 70 . OPICN may indeed be related to GWI, particularly the excess of GWI seen in deployed veterans relative to their non-deployed and other-deployed counterparts; however, to date, the mechanisms of OPICN progressive effects remain unclear 71 and the contribution of sarin and other nerve agents to GWI is largely inconclusive. Others have concluded that pesticides and/or PB are causally associated with GWI 4,72 . Similar to nerve agent exposure, use of pesticides and PB may play a contributory role in GWI symptomatology; however, neither OPICN, pesticides, nor use of PB can account for GWI cases observed in non-deployed veterans. It is notable that up to 36% of veterans who never deployed (depending on which GWI case definition was used) report GWI-consistent symptom profiles 4,72 , suggesting that the constellation of symptoms referred to as GWI cannot be solely attributable to in-theater exposure.

Vaccines and GWI
Both deployed and non-deployed GW veterans received multiple vaccinations in preparation for the war and those vaccines have become another widely investigated risk factor for GWI. Reports indicate that new recruits during the GW era received up to 17 vaccines during the first two weeks of basic training 72 . Several studies have concluded that vaccinations are associated with GWI, and have documented evidence of GWI in non-deployed GW veterans and in other-deployed veterans who received vaccines, albeit at lower rates than deployed GW veterans 2,73,74 . Furthermore, a dose-response relationship has been reported between the number of vaccines and severity of health symptoms 31 . In addition to standard vaccines (e.g., yellow fever, cholera, hepatitis B, etc.), the Gulf War was the first time that a large number of military personnel (150,000 deployed and non-deployed veterans) received the anthrax vaccine, the safety of which has been called into question 75 . As Nass 75 discusses, an unusually high rate of adverse events was reported for anthrax vaccine relative to other vaccines, with vaccine recipients developing at least 2 of the following 3 symptoms: fatigue, muscle or joint pains, and cognitive or emotional impairment, symptoms consistent with the CDC case definition of GWI 1 . Nass 75 concludes that statements that existing scientific evidence does not support an association between anthrax vaccine and GWI are simply not true. Indeed, given the incidence of GWI in non-deployed veterans who received preparatory vaccinations, it seems quite likely that anthrax and potentially one or more of the 17 antigens administered are implicated in GWI. To that end, recent in vitro studies have documented deleterious effects of GWI serum on cell cultures and reversal of those effects with the concomitant addition of antibodies against antigens contained in the vaccines administered to GW veterans [24][25][26] .

Vaccines, Persistent Antigens, and Inflammation
In discussing the role of vaccines in GWI, the Institute of Medicine 76 noted that adverse effects of vaccinations could be due to the toxicity of the antigen in the vaccine and/or from stimulation of the immune system (the intended goal of immunization) that may result in immune enhancement, hypersensitivity, or an immune-mediated pathological response. They continue to describe acute effects (e.g., rash) as well as sustained effects including tissue damage due to inflammation that tends to be seen in the lung, kidney, joints, and brain in animal studies.
Finally, the IOM report notes that genetic inheritance strongly influences the immune response to immunization and to actual infection, and explains the high variability in response to vaccinations. HLA genes, in particular, are highly related to immune response. Antibody production hinges on a match between epitopes derived from an antigen and an individual's HLA profile. In the absence of an HLA-antigen match, the antigen persists 21,42 , contributing to inflammation and ultimately tissue damage. As previously mentioned, there is mounting evidence documenting inflammation in veterans with GWI [28][29][30][31] . Furthermore, we identified 6 HLA alleles that were present in healthy GW veterans but were absent or significantly less frequent in those with GWI, indicating that GWI is associated with genetic lack of protection 18 . Thus, we suspect that GWI is a result of persistent antigens that could not be eliminated due to HLA-antigen incongruence, resulting in inflammation and ultimately neurodegeneration. Notably, HLA variability has been specifically implicated in regards to variability in antibody production to anthrax vaccine 77 . Similarly, in silico analyses highlight variability in the binding affinity of GWI vaccine antigens to protective HLA, thereby differentially affecting antibody production 27 .
We have studies underway aimed at identifying persistent antigens in veterans with GWI with the goal of ultimately eliminating them (thereby reducing inflammation and slowing progression of cortical thinning) via personalized immunotherapy. Findings from two recent in vitro studies in our lab demonstrated that serum from veterans with GWI results in detrimental changes to cell morphology in neural cultures; however, the damaging effects were neutralized with the addition of serum from healthy Gulf War veterans 22 or human immunoglobulin G 23 . Consistent with the Persistent Antigen hypothesis 21,42 , the neutralizing effects are presumably due to ability of antibodies present in serum from healthy Gulf War veterans and in pooled IgG to eliminate persistent antigens in veterans with GWI. Subsequent in vitro studies have documented evidence of adverse of GWI serum on cell cultures and reversal of those effects with the concomitant addition of antibodies against antigens contained in the vaccines administered to GW veterans [24][25][26] . Taken together, these findings highlight the fact that harmful persistent antigens circulating in the blood partly contribute to inflammation and ultimately cortical thinning in GWI. We anticipate that elimination of persistent antigens would reduce those effects.

Limitations
The findings from the present study must be considered in light of limitations. First, a small number of HLA alleles were investigated in the present study; thus, the effects of other HLA alleles on metabolically-induced inflammation and cerebrocortical thinning are unknown. The six Class II HLA alleles investigated here were identified from among 144 Class I and Class II alleles in a prior study of GWI 18 . These six alleles were shown to discriminate veterans from GWI from healthy GW controls, and exhibited a dosedependent relationship with GWI symptoms such that fewer protective alleles were associated with more GWI symptoms. Other alleles did not confer that protection and are, therefore, considered non-protective. Subsequent studies have demonstrated various protective effects of these alleles as it relates to GWI [19][20][21][22][23][24][25][26][27] . This study extends those findings to demonstrate additional protection of these six Class II HLA alleles on metabolically-induced inflammation and cerebrocortical thinning. Second, the veterans in the present study had elevated BMI which is known to be associated with inflammation via release of cytokines 34,78,79 ; however, there were no difference in BMI between veterans with and without protective alleles, and BMI did not account for the observed association between CRP and cerebrocortical thinning in veterans without protective HLA alleles. There is the possibility that this group of GW veterans with elevated BMI are not representative of GWI veterans in general although that is unlikely given that several other studies of Gulf War veterans have similarly found elevated BMI 30,31,80,81 . Finally, the study sample was relatively small. Future studies with larger samples are warranted to corroborate the findings and evaluate generalizability to the broader GWI population.

Conclusion
In the present study we demonstrated that inflammation is highly associated with BMI and cortical thinning in veterans with GWI who lack protective HLA alleles, but not in those who possess a GWI-protective HLA allele. These findings suggest that inflammation due to the persistence of foreign (potentially vaccine-related) antigens resulting from lack of HLA protection contributes to GWI. Future studies aimed at eliminating persistent antigens may be effective in reducing the varied symptoms associated with GWI, particularly those involving the brain. 1 Left banks of the superior temporal sulcus