Cellular and humoral immune abnormalities in Gulf War veterans.

We examined 100 symptomatic Gulf War veterans (patients) and 100 controls for immunologic assays. The veterans and controls were compared for the percentage of T cells (CD3); B cells (CD19); helper:suppressor (CD4:CD8) ratio; natural killer (NK) cell activity; mitogenic response to phytohemagglutin (PHA) and pokeweed mitogen (PWM); level of immune complexes; myelin basic protein (MBP) and striated and smooth muscle autoantibodies; and antibodies against Epstein-Barr virus, cytomegalovirus, herpes simplex virus type 1 (HSV-1), HSV-2, human herpes Type 6 (HHV-6), and Varicella zoster virus (VZV). The percentage of T cells in patients versus controls was not significantly different, whereas a significantly higher proportion of patients had elevated T cells compared with controls. The percentage of B cells was significantly elevated in the patients versus the controls. The NK cell (NK) activity was significantly decreased in the patients (24.8 +/- 16.5 lytic units) versus the controls (37.3 +/- 26.4 lytic units). The percentage of patients with lower than normal response to PHA and PWM was significantly different from controls. Immune complexes were significantly increased in the patients (53.1 +/- 18.6, mean +/- SD) versus controls (34.6 +/- 14.3). Autoantibody titers directed against MBP and striated or smooth muscle were significantly greater in patients versus controls. Finally, the patients had significantly greater titers of antibodies to the viruses compared with the controls (p < 0.001). These immune alterations were detected 2-8 years after participation in the Gulf War. The immune alterations are consistent with exposure to different environmental factors. We conclude that Gulf War syndrome is a multifaceted illness with immune function alterations that may be induced by various factors and are probably associated with chronic fatigue syndrome.


Materials and Methods
Subjects. Immunosciences Lab., Inc. was approached by R. Haines of the U.S. Army requesting immune testing of 10th Army Unit Gulf War Veterans who had GWS with common symptoms (joint pain, fatigue, headache, memory or concentration difficulties, sleep disturbances, and rash). Haines, with the help of medical doctors, initially selected 50 symptomatic patients (group A), followed at a later date by an additional 50 symptomatic patients (group B). Medical records and health histories were not forwarded for us to review. Controls consisted of two groups: group C consisted of 50 asymptomatic vaccinated veterans of the 10th Army Unit who were not deployed to the Persian Gulf, and group D consisted of 50 healthy asymptomatic subjects sent to the laboratory by treating physicians for annual checkups. Patients ranged from 34 to 56 years of age and consisted of 74 males and 26 females. The controls were age and sex matched. Venous blood was received within 24 hr from different clinics from several states and immediately processed for immunologic testing. Groups A and C underwent immunologic testing in 1993. The testing of groups B and D occurred over an extended period from 1995 through 1999. The tests were done at no charge to the U.S. Army and as a result were limited to immune parameters presented below. The test requests were properly documented and kept in a confidential file. All subjects gave their informed consent and allowed inclusion of their data in this report without disclosure of their identity in the publication.
Lymphocyte subset enumeration. Direct immunofluoresence staining of cell surface antigen was accomplished using the Becton-Dickinson FACScan Immunocytometry system (San Jose, CA). The peripheral mononuclear cells are treated with monoclonal antibodies conjugated to fluorescein isothiocyanate (FITC) or phycoerythrin (PE). The blood samples were first treated with red blood cell lysing solution. Cells were then washed, stained with monoclonal antibody, and then analyzed with the FACScan flow cytometer. We used the following pairs of FITC-or PE-conjugated monoclonal antibodies: CD45/ CD14 for quality control check, CD3/CD4 for T-helper cells, CD3/CD8 for suppressor cells, CD3/CD19 for T and B cells, and CD3/CD16 plus CD56 for natural killer (NK) cells (NKHT3 + and NKHT3 -). Using these sets of monoclonal antibodies, we determined the percentage of positively stained cells for each marker pair and the percentage of double stained cells.
Preparation of peripheral blood leukocytes. Leukocytes were prepared from heparinized peripheral venous blood by Ficol-Hypaque density gradient (Sigma Chemical Company, St. Louis, MO). Cells were washed three times with Hanks balanced salt solution and resuspended to a concentration of 10 7 cells/mL in a complete medium (CM) that consisted of RPMI-1640 supplemented with 10% fetal calf serum and 1% antibiotics (100 U penicillin and 100 µg/mL streptomycin). We examined purity of the cells by flow cytometry using CD45/CD14 monoclonal antibodies; purity was > 95%. Cells were used for different functional assays within 1 hr of isolation.
NK cell cytotoxicity assay. We used a standard 4-hr 51 Cr-release assay (Whiteside et al. 1990) to determine NK cell cytotoxicity. Briefly, we added 1 × 10 4 51 Cr-labeled K562 tumor target cells in 0.1 mL CM to different wells of a microtiter plate. Effector cells were then pipetted into triplicate wells for each dilution to give effector:target ratios of 6:1, 12:1, 24:1, and 48:1. After a 4-hr incubation at 37°C, the plates were centrifuged at 1,400 rpm for 4 min, and 0.1 mL of supernatant was collected from each well and placed in a gamma counter. The percentages of isotope released were calculated using the following formula: The results of NK cell assay for each effector:target ratio was expressed in terms of lytic units (LU), calculated as described by Whiteside et al. (1990).
Lymphocyte mitogenic assay. Lymphocytes were isolated and tested for mitogenic activity as described by Fletcher et al. (1992) and Maino et al. (1995). Briefly, 5 × 10 5 lymphocytes per 0.1 mL CM were cultured in flatbottom microtiter plate wells. Cells from patients and controls, as well as cells with known mitogenic stimulation, were cultured with or without an optimal concentration of either phytohemagglutin (PHA; 25 µg/mL; Grand Island Biological Company, Grand Island, NY) or pokeweed mitogen (PWM; 5 µg/mL; Sigma Chemical Company). After 48 hr of incubation, the cells were harvested and stained with CD69 monoclonal antibody conjugated to fluorescent dye and analyzed by flow cytometry. Wells with no mitogens added (negative control) provided information about the media and cells used in the assay so we could determine possible nonspecific modulatory activity. Values for patients and controls were compared with the daily negative and positive control for each assay. Results were calculated using the following formula: [2] We estimated three stimulation levels: low, < 75% of the number of control cells; normal, 75-125% of the number of control cells; and elevated, > 125% of the number of control cells.
Myelin basic protein antibody. The antibody to human myelin basic protein (MBP) was analyzed as previously described (Vojdani et al. 1992). Briefly, MBP (Sigma Chemical Co.) was checked for purity by polyacrylamide gel electrophoresis (Diebler et al. 1972). Antisera to MBP were induced in rabbits by repeated injection of the protein in complete Freund's adjuvant. Antibody activity in the rabbit sera and the patient and control samples was detected by adding different dilutions (1:100-1:10,000) of sera to wells of a microtiter plate previously coated with MBP. MBP (250 µg/mL) was dissolved in carbonate buffer (pH 9.6), and 200 µL of this solution was added to each well. After incubation, washing, and blocking with 1.5% bovine serum albumin plus gelatin, 200 µL of either diluted rabbit or human serum was added to the wells. Incubation was repeated for 1 hr at 25°C, and the sera were shaken out of the wells, which were then washed five times with wash solution. Next, 200 µL peroxidaseconjugated goat anti-rabbit or goat antihuman IgM (optimal dilution) was added to the appropriate well. After incubation and repeated washing, 200 µL substrate was added to each well. Plates were incubated for 1 hr at room temperature and read in a microtiter reader at 480 nm wavelength. We plotted a titration curve using rabbit antisera, and compared patient and control sera with this standard curve. Results are expressed by the ELISA (enzyme-linked immunosorbent assay) values.
Striated and smooth muscle antibody assay. We purchased skeletal muscle myoblast cell line CRL-1769 and smooth muscle cell line ATCC CRL-1444 from American Type Culture Collection (Rockville, MD). Cells were grown in Dulbecco's modified Eagle's medium (90%) and fetal bovine serum (10%). After 7 days in culture, the cells were harvested, sonicated, and used for coating ELISA plates. Each ELISA microplate well was coated with 100 µL cell lysate containing 10 µg protein in 0.1 M carbonate buffer (pH 9.5). Plates were incubated overnight at 4°C and then washed three times with 200 µL Tris-buffered saline containing 0.05% Tween 20 (pH 7.4). With the exception of goat anti-human IgG F(ab´) 2 , all other steps were similar to the MBP ELISA method described above. To detect nonspecific binding, we used all reagents except human serum in several control wells and coated some wells with different tissue antigens.
Immune complexes. We measured IgG, IgM, and IgA immune complexes by coating microtiter plates with anti-C1q. Addition of serum, second antibody, and substrate was similar to procedures for measuring MBP antibody by ELISA.
Determination of expected ranges. We obtained expected ranges as follows: • Earlier publications. • Suppliers of the CD markers (Becton-Dickinson) carried out in-house verification using blood from 100 healthy controls from which the means, SDs, and 95% confidence intervals (CIs) were determined (Babcock et al. 1987, Shearer et al. 2003). • The manufacturers of the immune complex and viral antibody test kits provided an expected range with their products (Abbas et al. 1994, Christenson et al. 1992, Matheson et al. 1990, Shehab and Brunell 1983). • We obtained expected ranges for NK cell activity from Whiteside et al. (1990) and confirmed them in our laboratory by performing the assay on 500 specimens from healthy subjects. • We determined expected ranges for PHA and PWM mitogenesis and for autoantibodies to MBP and striated and smooth muscle inhouse using 100 healthy controls for which means, SDs, and 95% CIs were calculated. Thus, the expected ranges are a combination of suppliers' recommendations and in-house quality control. Statistics. We used two-sided critical t-tests for comparison of independent means and Z-tests for comparison of independent proportions as described by Bourke et al. (1985).

Comparison of patients (groups A and B) and controls (groups C and D).
We analyzed the data to determine if differences existed between patient groups A and B and their controls for each tested immune parameter (data not shown). No significant differences were found; therefore, we combined patient groups A and B and control groups C and D for further statistical analysis. In addition, we observed no differences between males and females in any of the groups.
CD3 T cells. The percentages of CD3 T cells present in the peripheral blood of controls versus patients are presented in Table 1. The mean percentage of CD3 cells in patients (72.8 ± 10.3%, mean ± SD) was slightly higher compared with the controls (71.6 ± 7.3%), but the difference was not significant. However, the percentage of individuals that fell outside of the expected range of 53-79% CD3 cells was different between the two groups; 2% of controls and 9% of patients had < 53% CD3 cells, and 5% of controls and 30% of patients had > 79%. The critical Z-and p-values for variance from normal distribution were significant (p < 0.05).
CD19 B cells. The percentages of CD19 B cells present in the peripheral blood of controls versus the patients are presented in Table 1. The mean percentage of B cells in the patients (16.1 ± 6.0%, mean ± SD) was greater than that for the controls (11.5 ± 3.1%). The difference between the two means was highly significant (t = 11.402, p < 0.001). The distribution of B cells outside of the expected range (5-15%) was also different when the patients were compared with the controls. Forty-nine percent of patients and 5% of controls had > 15% B cells, whereas 0% of controls and 4% of patients had < 5% B cells. The critical Z-values for variance from normal distribution were significant (p < 0.01).
CD4:CD8 ratio. The CD4:CD8 (helper:suppressor) ratios for controls and patients are listed in Table 1. The ratio was significantly (p < 0.001) elevated in patients (2.23 ± 0.87, mean ± SD) compared with controls (1.74 ± 0.34). The Z-values for the percentage of patients with values outside the expected distribution of the CD4:CD8 ratio were significantly different compared with those of the controls (p < 0.001).
NK cell activity. The data obtained for NK lytic activity are presented in Table 2. The lytic activity was significantly less (p < 0.001) in patients (24.8 ± 16.5) than in controls (37.3 ± 26.4). The percentages of individuals with > 50 LU were not different between the two groups (p = not significant), whereas the p-value for the percentage of patients with < 20 LU was significant (p < 0.01).
Mitogen stimulation. The results of PHA and PWM stimulation of peripheral lymphocytes in the controls and patients are summarized in Table 2. The mean percentage of stimulation by either PHA or PWM was not different between controls and patients. However, the distribution of stimulation values did differ between the two groups. For PHA, more GWS patients (32%) had a stimulation index < 75% of expected (p < 0.01) but 4% of patients had a stimulation index > 125% (p < 0.05). For PWM stimulation, an increased percentage of patients had an index < 75% (p < 0.01), whereas the difference between the percentages of patients (4%) and controls (1%) with values > 125% were not significant.
Autoantibodies. The observations for IgM anti-MBP in the controls and patients are shown in Table 3. The mean ELISA units for IgM antibodies were significantly (p < 0.001) greater in patients (45.9 ± 35.8) than in controls (28.4 ± 13). The Z-value for the percentage of patients with > 50 ELISA units was significant (p < 0.001). The mean value for the patients with IgM titers > 50 ELISA units was 75.7 ± 15.9.
Antibodies to muscle (striated and smooth). The results for IgG antibodies to both smooth and striated muscle are shown in Table 3. The mean IgG titers observed in the patients (42.8 ± 72.3) were significantly (p < 0.001) greater than those for the controls (15.9 ± 11.4).
Immune complexes. The observations on immune complexes found in the peripheral blood of controls and patients are summarized  in Figure 1. Immune complexes were significantly (p < 0.001) elevated in the patients (52.1 ± 18.5 mEq/mL) compared with controls (34.5 ± 14.3 mEq/mL) ( Figure 1A). In addition, the percentage of patients with immune complexes > 50 mEq/mL was significantly higher than that for controls (p < 0.01) ( Figure 1B).

Antibodies to viruses.
The results obtained on antibodies to viruses are presented in Table 4. Mean titer levels for each virus were significantly (p < 0.001) higher in the patients than in controls.
The differences in the mean titer levels between the patients and controls resulted from a disproportionate number (percentage) of patients who had elevated titers to each virus. For example, the mean IgM antibodies to EBV viral capsid antigen (VCA) was 724 ± 401 ELISA units for the 56 individuals that fell above the maximum expected titer of 300 ELISA units. Similar observations were evident for each of the remaining viruses. Z-values for the percentage of patients with elevated titers were significantly different from those for controls ( Figure 2).

Discussion
The patients and the controls that underwent immunologic testing were from two different time periods, 1993 versus 1995-1999. In addition, the two sets of controls were different: one consisted of asymptomatic U.S. Army soldiers who did not serve in the Gulf War, and the other consisted of healthy individuals undergoing annual checkups. Therefore, we deemed it necessary to determine if there were any differences between the groups. t-Tests comparing the means of each immune parameter revealed that no differences existed between the two sets of patients as well as the two sets of controls (data not shown), permitting the pooling into patients (n = 100) versus controls (n = 100). This analysis revealed two interesting observations: a) GWS patients had immunologic alterations approximately 2-8 years after participation in the Gulf War; and b) the asymptomatic soldiers' immunologic findings were no different from those of healthy individuals undergoing an annual checkup.
When we analyzed the data on basic laboratory testing (CBC, chemistry, T 3 , T 4 , TSH, RF, and other measures, the differences between patients and controls were not significant. Therefore, we conclude that such tests are of little value in diagnosing GWS.
One of the fundamental issues facing the field of immunotoxicology is the degree of perturbation to a given parameter after exposure to a xenobiotic that translates into a significant health risk. A number of methods have evolved to address the question of how a chemical or drug affects the ability of the immune system to resist challenge by pathogens or abnormal host cells (e.g., tumor). These include both functional (i.e., activity of one or more specific cell types) as well as host resistance assays. They were designed to evaluate the overall changes in the functional integrity of the immune system associated with certain chemicals, pharmaceuticals, and biotherapeutic molecules (Cornacoff et al. 1995;Loveless et al. 1997;Luster et al. 1988Luster et al. , 1993Smialowicz et al. 1995). Therefore, analysis of circulating lymphocyte subpopulations reveals that alterations do exist in increased T-cell numbers, B-cell numbers, and CD4:CD8 ratio in these patients (Table 1). An increased CD4:CD8 ratio indicates an increase in inducer helper CD4 relative to CD8 suppressor/cytotoxic cells. These observations are compatible with previous reports on GWS (Zhang et al. 1999) as well as those regarding lymphocyte subset alterations after exposure to xenobiotics (Baj et al. 1994;Zahalsky 1991, 1992;Thrasher and Broughton 2001;Vojdani et al. 1992Vojdani et al. , 1993. Further assessment of the immune system of GWS patients was done by measurement of T-and B-cell function and NK cell activity. Although the mean values for mitogenic response to PHA and PWM were not different in patients versus controls, the percentage of individuals with abnormally low mitogenesis was significantly different between the two groups ( Table 2). Such an analysis is appropriate because immune dysregulation can result in either increased or decreased mitogenic response in some individuals. Mitogen stimulation is used clinically to assess cellular immunity in patients with immunodeficiency, cancer, and autoimmunity, as well as to assess the immunotoxic potential of drugs and chemicals in humans and experimental animals (Smialowicz 1995). Similarly, assessment of NK lytic activity also gives valuable information. The lytic activity of NK cells (Table 2) was significantly lower in the GWS patients (24.8 ± 16.5) versus the controls (37.3 ± 26.4). This was also reflected by an elevated number of GWS patients (47%) with low lytic activity of NK cells. Decreased numbers of NK cells have been reported in GWS individuals with chronic fatigue (Zhang et al. 1999). Furthermore, decreased mitogenesis and decreased NK activity have been reported after exposure to xenobiotics and in chronic fatigue (Heuser and Vojdani 1997;Levine et al. 1998;Racciatti et al. 2001;Vojdani et al. 1992Vojdani et al. , 1993Whiteside and Friberg 1998). Thus, the observed changes in the functional status of T-, B-, and NK cells in GWS patients are consistent with immune alterations in humans after exposure to xenobiotics as described above.
The most critical but open question in immunosurveillance is whether information on differences between individuals with abnormal NK, T-, and B-cell function can predict future development of cancer or autoimmune diseases. Indications of the significant role for NK cells in preventing the development of cancer in both mice and humans have been reported (Imai et al. 2000;Wilson et al. 2001). A prospective cohort study among a Japanese general population showed that medium and high cytotoxic activities were associated with reduced cancer risk, whereas low activity was associated with an increased risk (Imai et al. 2000). Wilson et al. (2001) Article | Immune abnormalities in Gulf War veterans Environmental Health Perspectives • VOLUME 112 | NUMBER 8 | June 2004   demonstrated suppressed NK cell activity with altered host resistance in mice. NK cell activity was incrementally decreased by intravenous administration of antibody to NK surface receptors. After a ≥ 80% decrease in spontaneous NK activity, a challenge with ≥ 1 × 10 3 B16F10 melanoma cells resulted in increased tumor burden in the lungs (Wilson et al. 2001). Furthermore, when challenged with 1 × 10 5 melanoma cells, tumor burden was not increased until spontaneous NK activity had been decreased by ≤ 50-60%. Altered host resistance is a function of both the magnitude of the decrease in NK activity and the magnitude of the challenge to the host (Wilson et al. 2001).
Increased levels of autoantibodies to various organs in humans, including MBPs, have been reported after exposure to toxic chemicals Zahalsky 1991, 1992;Thrasher et al. 1987Thrasher et al. , 1990Thrasher et al. , 1993Vojdani et al. 1992Vojdani et al. , 1993. Thus, the increased incidence of antibodies to MBP and striated and smooth muscle (Table 3) in the GWS patients is suggestive of autoimmunity, possibly resulting in tissue injury from toxic chemical exposure (Cooper et al. 2002;Griem et al. 1998). The presence of IgM antibodies to MBP appears to indicate that an active process involving release of these self-antigens is occurring up to 8 years after injury. Central nervous system injury has been reported in research animals exposed to pyridostigmine bromide, DEET (N,N-diethyl-m-toluamide), and permethrin (Abou-Donia et al. 2001) and in some GWS patients, in particular, ALS (Haley 2003;Horner et al. 2003). The observations presented here suggest that additional studies are needed on neural damage and/or axonal demyelinization in symptomatic Gulf War veterans. Neural antigen and MBP antibodies have been reported in patients with neurologic disorders (Terryberry et al. 1998;Willison and Yuki 2002), including ALS (Rogers et al. 1996), autism (Vojdani et al. 2002), and multiple sclerosis (Vojdani et al. 2003). Therefore, our results are consistent with the excess of ALS among Gulf War veterans reported by Haley (2003).
The GWS patients in the present study were found to have significantly elevated circulating immune complexes (Figure 1). Increased immune complexes have not been previously reported to our knowledge in GWS patients. Circulating immune complexes are formed by excessive antigen antibody reaction. Immune complexes have been implicated in numerous immunopathologic conditions, including systemic lupus erythematosus, rheumatoid arthritis, glomerulonephritis, and infectious induced inflammation (Abbas et al. 1994). Deposition of immune complexes can occur from cell-or tissue-specific antibodyantigen reactions, resulting in organ injury and/or immune complex diseases (Bigazzi et al. 1986). Thus, it would appear from these observations on increased immune complexes in the patient population in the present study that inflammation and autoimmune reactions may exist in a subgroup of patients with GWS. Circulating immune complexes containing IgG, IgM, and IgA antibodies can generate a variety of substances associated with muscle damage and the acute phase response, activating the classic pathway of complements (Sorensen et al. 2003).
Finally, we tested the GWS patients to determine if an increase in antibodies to several HSV types was present (Table 4, Figure 2). The data clearly show that significantly increased Article | Vojdani and Thrasher 844 VOLUME 112 | NUMBER 8 | June 2004 • Environmental Health Perspectives   antibody titers occurred in the GWS patients compared with the controls for each virus tested (EBV, CMV, HSV-1, HSV-2, HHV-6, and VZV; Table 4). When the observations were limited to only affected individuals, the increased titers to each virus was even more evident ( Figure 2). Exactly when the viral infections occurred cannot be determined from these data. However, the increased IgM antibodies to EBV VCA suggest that reactivation of EBV is probably occurring and may involve the other viruses. To our knowledge, there has been no other report regarding increased viral antibodies in GWS patients. In addition, infections with Mycoplasma fermentans, Mycoplasma hominis, and Mycoplasma penetrans must also be considered (Vojdani and Franco 1999).
Although we did not perform polymerase chain reaction to detect DNA of these viruses, the presence of viral antibodies and mycoplasma DNA may be related to dysregulation of the immune system found in this study.

Conclusion
The observations in these GWS patients suggest that an alteration in the number and function of T and B cells and NK cell activity has occurred that may be associated with service in the Persian Gulf. Furthermore, the presence of antibodies to MBP and striated and smooth muscle, increased immune complexes, and increased antibody titers to several DNA viruses indicate that viral reactivation and/or an active inflammatory immune process may be ongoing in some GWS patients (Ferguson andCassaday 2001/2001;Patarca 2001;Rook and Zumia 1997). Based on these observations and earlier reports by others, we believe that GWS is a multifactorial disease caused by exposure to a variety of environmental conditions, for example, xenobiotics, vaccinations, and other stressor-related conditions of the Gulf War environment as summarized in Figure 3. We believe that the outlined multiple factors along with genetic susceptibility due to polymorphism of PON1, loss of neuropathy target esterase, glutathione S-transferase, cytochrome P450 enzymes, or other factors may affect some individuals, resulting in immune dysregulation (Haley et al. 1999;Loewenstein-Lichtenstein et al. 1995;Shields 1994;Whatt et al. 2000). These immune functional alterations reported herein may cause viral reactivation and induction of proinflammatory cytokines, resulting in symptoms similar to those of chronic fatigue and fibromyalgia, as well as other symptoms of GWS (Ferguson andCassaday, 2001/2002;Patarca 2001;Rook and Zumia 1997;Zhang et al. 1999). The variation in individual susceptibility to environmental stresses and toxicants is a new discipline (toxicogenomics), initiated at the National Institute of Environmental Health Sciences, that studies the relationship between genes and environmental stressors (Waters et al. 2003). This new knowledge of toxicogenomics may enable us to answer why, upon exposure to these environmental factors, some soldiers developed GWS and others did not. Finally, it appears that additional studies involving asymptomatic deployed Gulf War veterans versus symptomatic Gulf War veterans would be beneficial in further understanding the immunologic observations presented herein.