BRAIN, BEHAVIOR, AND IMMUNITY ARTICLE NO.

12, 297–307 (1998)

BI980537

Cytokine Dysregulation Associated with Exam Stress in Healthy Medical Students Gailen D. Marshall, Jr.,*,1 Sandeep K. Agarwal,* Camille Lloyd,† Lorenzo Cohen,‡ Evelyn M. Henninger,* and Gloria J. Morris* *Division of Allergy & Clinical Immunology, Department of Internal Medicine, and †Department of Psychiatry, University of Texas Houston Medical School, Houston, Texas, 77030; and ‡Department of Behavioral Science, The University of Texas M. D. Anderson Cancer Center, Houston, Texas The mechanisms of stress-related immune alterations have not been fully elucidated. Cell-mediated immune responses as well as antibody and certain cytokines are reported as being suppressed during times of high stress. However, the role of suppression vs dysregulation has not been established in human stress models. The effect of exam stress on regulatory cytokines in 16 healthy medical students was assessed by measuring type-1 (IFN-γ) and type-2 (IL-10) cytokines from 72-h PHA/PMA-stimulated PBMC 4 weeks before and 48 h after exams. Results demonstrated decreased IFN-γ accompanied by increased IL-10 during exam stress that resulted in a decreased IFN-γ : IL-10 ratio. There was a significant correlation between the cytokine response to PHA/PMA and number and subjective adjustment to daily hassles. Additionally, students who reported greater levels of loneliness also reported greater numbers of and poorer subjective adjustment to hassles. The differences were consistent in both males and females but did not correlate with AM cortisol levels. Additionally, when individuals were grouped into high vs low preexam hassle levels, the type-1/type-2 shift in the IFN-γ : IL-10 ratio occurred in the low hassles group only. These data suggest that psychologically stressful situations shift type-1/type2 cytokine balance toward type-2 and result in an immune dysregulation rather than overall immunosuppression. This may partially explain the increased incidence of type-2-mediated conditions such as increased viral infections, latent viral expression, allergic/asthmatic reactions, and autoimmunity reported during periods of high stress.  1998 Academic Press Key Words: type-1/type-2 cytokines; cytokine dysregulation; exam stress.

INTRODUCTION

The effects of psychological stress on immune function have been intensely investigated (Kiecolt-Glaser, Glaser, Strain, Stout, Tarr, Holliday, & Speicher, 1986; Kiecolt-Glaser, Garner, Speicher, Penn, Holliday, & Glaser, 1984; Kiecolt-Glaser, Dura, Speicher, Trask, & Glaser, 1991; Esterling, Kiecolt-Glaser, Bodnar, & Glaser, 1994; Anderson, Farrar, Golden-Kreutz, Kutz, MacCallum, Courtney, & Glaser, 1998; Kiecolt-Glaser, Fisher, Speicher, Penn, & Glaser, 1987). Many chronic stressors have been associated with altered immune responses in humans. The postulated mechanism most often advanced to explain these results involves immunosuppression as a fundamental effect of stress (Glaser, Rice J., Sheridan, Fertel, Stout, Speicher, Pinsky, Kotur, Post, Beck, & Kiecolt-Glaser, 1987). Otherwise healthy populations experiencing life events such as bereavement (Schleifer, Keller, Camerino, Thornton, & Stein, 1983), marital discord (Kiecolt-Glaser et al., 1987), caregiving for a relative with chronic disease (Kiecolt-Glaser et al., 1991; Esterling et al., 1994; Pariante, Carpiniello, Orru, Sitzia, Piras, Farci, Del Giacco, Piludu, & Miller, 1997), and living 1 To whom reprint requests should be addressed at Division of Allergy & Clinical Immunology, University of Texas Houston Medical School, 6431 Fannin 4.044 MSB, Houston, TX 77030. 297 0889-1591/98 $25.00

Copyright  1998 by Academic Press All rights of reproduction in any form reserved.

298

MARSHALL ET AL.

with a cancer diagnosis (Anderson et al., 1998) have been previously reported to have a suppression of natural killer cell cytotoxicity and lymphocyte proliferation to mitogens. Academic stress models have been used extensively to further investigate the effects of psychological stress on immune function in healthy young subjects (Kiecolt-Glaser et al., 1984; 1986). The stress associated with taking medical school examinations is an accepted model of stress and produces consistent findings. Exam stress in healthy medical students is associated with a suppression of natural killer cell cytotoxicity (Kiecolt-Glaser et al., 1984; 1986), suppression of mitogen- and antigen-stimulated lymphocyte proliferation (Glaser, Pearson, Bonneau, Esterling, Atkinson, & Kiecolt-Glaser, 1993; Dobbin, Harth, McCain, Martin, & Cousin, 1991; Glaser et al., 1993), decreased mitogen-stimulated interferon-γ (IFN-γ) production (Glaser et al., 1987; Dobbin et al., 1991), and increased latent viral reactivation (Glaser, Pearson, Jones, Hillhouse, Kennedy, Mao, & Kiecolt-Glaser. 1991; Glaser, Pearl, Kiecolt-Glaser, & Malarkey, 1994; Glaser et al., 1987). However, the mechanisms of these stress-related immune alterations have not been fully elucidated. Recent studies have demonstrated that host immune responses are under intricate control mechanisms that preserve homeostasis when functional and are responsible for a variety of disease states when dysfunctional (Mosmann & Sad, 1996). The human immune response is composed of a cell-mediated and a humoral component. The equilibrium established between these components allows the normal host to successfully combat various forms of intracellular and extracellular pathogens. The sometimes intricate balance between cell-mediated (CMI) and humoral responses to a particular antigenic challenge is regulated by cytokine production from subsets of CD4⫹ helper T-cells (TH) divided into subpopulations (TH1 and TH2) based upon their ability to produce specific patterns of cytokines (Mosmann, Cherwinksi, Bond, Giedlin, & Coffman, 1986; Romagnani, 1994; Mosmann & Sad, 1996; Del Prete, De Carli, Mastromauro, Biagotti, Macchia, Falagiani, Ricci, & Romagnani, 1991). TH1 cells support CMI, through the production of IFN-γ and tumor necrosis factor β (TNF-β), while TH2 cells support humoral immunity through the production of interleukin-4 (IL-4), IL-10, and IL-13. In situ, many other cells besides TH1 and TH2 can produce these cytokines which likely influence systemic immunoregulation. Thus the designation type-1 and type-2 cytokines is often preferred to reflect the cellular diversity of their origin (Romagnani, 1994; Croft, Carter, Swain, & Dutton, 1994; Mosmann & Sad, 1996). It has been reported that psychological stress alters CMI responses as well as certain antibody responses (Glaser et al., 1993; 1994; Kiecolt-Glaser et al., 1984; 1986; 1987; 1991; Esterling et al., 1994; Anderson et al., 1998). However, the role of suppression versus dysregulation of the type-1/type-2 cytokine balance has not been established in human stress models. Thus, this study was undertaken to determine whether a recognized human stress model (medical student exam stress) causes suppression or dysregulation of type-1/type-2 cytokines produced by peripheral blood mononuclear cells (PBMC) from healthy first-year medical students and what factors might be associated with the changes. METHODS

Subjects. Sixteen healthy, nonsmoking volunteers from the first-year medical school class at the University of Texas–Houston Medical School were invited to participate in the current investigation. The cohort (mean 23.7 ⫾ 3.7 years old, range 21–29 years old) included 6 males and 10 females.

CYTOKINE DYSREGULATION WITH EXAM STRESS

299

Procedures. The first year of medical school at this institution consists of 5-week blocks ending in a 3-day exam period. Subjects participated during the second week of the fall semester (preexam) and 2 days after completion of the first block of exams approximately 4 weeks later (postexam). Subjects were asked to avoid consuming alcohol at least 48 h prior to each visit. At both visits, subjects underwent a brief physical exam, had venous blood drawn, and completed a series of self-report questionnaires. To control for diurnal variation, both visits were between 6 and 8 AM. The study was approved by the Committee for the Protection of Human Subjects at the University of Texas–Houston Health Science Center. Informed consent was obtained from each subject prior to enrollment in the study. Isolation of PBMC. Venous blood was collected in heparin-coated tubes (Becton– Dickinson, San Jose, CA) and diluted with an equal volume of RPMI 1640 medium (Life Technologies, Gaithersburg, MD). PBMC were obtained using Ficoll–Hypaque (Pharmacia Biotech, Piscataway, NJ) density gradient centrifugation. PBMC were harvested, washed twice with Ca⫹2- and Mg⫹2-free HBSS (Life Technologies), and resuspended in RPMI 1640 containing 10% human AB serum (Biocell, Rancho Dominguez, CA), 90 U/ml penicillin, 90 µg/ml streptomycin, and 2 mM L-glutamine (all from Sigma, St. Louis, MO). Quantitation of PBMC was determined by an automated hematology analyzer (Coulter) and viability was determined by trypan blue dye (Sigma) exclusion. Mitogen cultures. Cultures of 1 ⫻ 106 PBMC in 1.0 ml of medium were stimulated with 10 µg/ml phytohemagglutinin (PHA; Sigma) and 1 ng/ml phorbol myristate acetate (PMA; Sigma) for 72 h at 37°C and 5% CO2. The combination of PHA and PMA maximizes levels of IFN-γ and IL-10 in culture supernatants (Marshall, Wen, Abrams, & Umetsu, 1993). Supernatants were harvested and frozen at ⫺70°C for future cytokine determination. Clinical laboratory assessment. At each visit, venous blood was collected in an EDTA-coated tube for determination of complete white blood cell count with differential, hematocrit, hemoglobin, and platelets. In addition, venous blood was collected in a serum tube for blood chemistry analyses including electrolytes, renal function tests, liver function tests, and nutritional assessment. These assays were performed by the clinical laboratories at Hermann Hospital (Houston, TX). Cytokine analyses. Cytokine levels in culture supernatants were determined by sandwich-antibody ELISA using paired monoclonal antibodies specific for IFN-γ (Genzyme Cambridge MA) and IL-10 (Pharmingen, San Diego CA). ELISAs were read on an Emax ELISA plate reader (Molecular Devices, Sunnyvale, CA) and concentrations calculated based upon standard curves using the SOFTmax for Windows software package (Molecular Devices). Sensitivities of the ELISAs are 3.0 and 9.3 pg/ml. Cortisol analyses. Plasma cortisol levels were determined using a commercial ELISA kit (Neogen, Lexington, KY). Briefly, cortisol was extracted from plasma into an organic phase with ethyl ether. After allowing the ethyl ether to evaporate, the cortisol was resuspended in the kit buffer and analyzed by quantitative competition ELISA. ELISAs were read on an Emax ELISA plate reader (Molecular Devices) and concentrations calculated based upon standard curves using the SOFTmax for Windows software package (Molecular Devices). Self-report measures. To assess the extent and source of possible stress, subjects answered a battery of self-report questionnaires. Subjects completed a modified version of the Schedule of Recent Experiences (Holmes & Rahe, 1967). This inventory

300

MARSHALL ET AL.

measured both the frequency of life change events and their perceived impact over the past 6 months. They also completed the Daily Hassles Scale (Kanner, Coyne, Schaefer, & Lazarus, 1981) to measure the frequency and subjective adjustment to negative or irritating daily events over the past month. The Perceived Stress Scale (PSS) (Cohen, Kamarck, & Mermelstein, 1983) measures perceptions of ongoing stress. Subjects were asked to rate their subjective stress over the past month. The PSS has good reliability (.84 to .86) and is correlated with depression status and other stress outcome measures. Symptom reporting was measured using the Brief Symptom Inventory (BSI) (Derogatis, 1977). This widely used symptom inventory yields an overall distress score as well as subscale scores reflecting somatic distress, anxiety, depression, hostility, and five other symptom categories. Subjects were asked to indicate which of the 53 symptoms they experienced as bothersome during the preceding month and to rate the intensity of symptom distress (Derogatis, 1977). The revised UCLA Loneliness Scale was included to provide a brief subjective measure of the adequacy of interpersonal contacts (Russel, Peplau, & Cutrona, 1980). Patients were also asked to complete diet summaries for the previous 3 days and a questionnaire assessing degree of regular physical activity. Harvard step test. To assess physical conditioning level, the submaximal Harvard Step Test was utilized as previously described (Sharkey, 1990). Briefly, each subject was asked to rest for a period of 5 min. A resting pulse rate was determined. Then the subject stepped up and down on a bench (15 3/4 inches high) at a rate of 22 1/2 steps per min for 5 min. The subjects rested for 5 min and a postexercise pulse rate was determined. Scoring and conditioning levels were determined from tables provided by Sharkey (1990). Statistics. All data are displayed as mean ⫾ standard error of the mean (SEM). Y error bars on charts and graphs represent SEM. Comparisons between pre- and postexam utilized the paired t test. Exploratory analyses used two-tailed Pearson correlation to examine the association between self-report measures and cytokine levels. Due to the lack of a normal distribution, a logarithmic transformation was conducted on the cytokine measures for these analyses. Finally, additional exploratory analyses examined type-1/type-2 cytokine production when individuals were categorized into high vs low preexam hassles groups by median split and statistical analyses utilized paired t test. RESULTS

Subject characteristics. There were no differences observed in the diet of the subjects from the baseline to the postexam period as determined by diet diaries. Subjects did not report consumption of any alcohol within 48 h prior to either visit. Furthermore, there were no abnormalities and/or interval changes in the hemoglobin, hematocrit, total protein, albumin, uric acid, or lactate dehydrogenase indicating no gross changes in nutritional status (data not shown). The Harvard Step Test did not reveal any change in physical conditioning status during the postexam period relative to preexam (data not shown). Finally, subjects did not report any major changes in their physical activity between the two data points. One subject was excluded from all of the cytokine analyses because of technical problems with the ELISA assay and a second subject was excluded from the exploratory analyses because the Hassles score was determined to be a statistical outlier. Finally, two subjects did not complete the Brief Symptom Inventory and were excluded from the analyses utilizing the BSI as a self-report measure.

CYTOKINE DYSREGULATION WITH EXAM STRESS

301

FIG. 1. The effect of exam stress on the type-1/type-2 cytokine balance. Cytokines assayed in supernatants from 72-h PHA/PMA-stimulated cultures of PBMC from healthy medical students. (A) IFN-γ (pre- vs postexam, p ⫽ n.s. by paired t test). (B) IL-10 (pre- vs postexam, p ⫽ .003). (C) IFN-γ: IL-10 ratio (pre- vs postexam, p ⫽ .007.) n ⫽ 15 subjects.

Effect of exam stress on PHA-induced cytokines from PBMC. PBMC from subjects were stimulated with PHA and PMA for 72 h. Mitogen-stimulated PBMC from subjects during the postexam period produced slightly less IFN-γ (type-1 cytokine) than PBMC from the preexam period (pre, 44183 ⫾ 5573 pg/ml; post, 35993 ⫾ 4537 pg/ml; p ⫽ n.s.; Fig. 1A). However, production of the type-2 cytokine, IL-10,

302

MARSHALL ET AL.

TABLE 1 Self-Report Measures Self-report measures

Preexam a

Schedule of recent experiences Subjective adjustment to schedule of recent experiencesa Daily hassles Scale b Subjective adjustment to daily Hassles scale b Perceived Stress Scalea R-UCLA Loneliness Scalea Brief Symptom Inventory: General Symptoms Index b

7.3 48.1 21.5 29.7 25.5 6.4 56.9

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

0.9 8.1 3.1 4.4 1.7 0.4 2.8

Postexam 7.7 61.7 20.1 30.0 28.4 6.8 58.4

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

0.9 7.6* 3.4 4.8 2.1 0.5 3.0

n ⫽ 16 subjects. n ⫽ 15 subjects. * p ⬍ .05.

a b

was markedly increased during the postexam period (pre, 799 ⫾ 108 pg/ml; post, 1493 ⫾ 230 pg/ml; p ⫽ .003; Fig. 1B). Similar patterns of IFN-γ and IL-10 production were observed in both males and females (data not shown). The relative ratio of IFN-γ :IL-10 (Fig. 1C) decreased during the postexam period (30 ⫾ 5) compared to the preexam period (73 ⫾ 15; p ⫽ .007). There were no differences observed in plasma cortisol levels during the postexam period relative to the preexam period (data not shown). Self-report measures. Scores from the self-report measures are listed in Table 1. The only significant increase in distress during the postexam period was the Subjective Adjustment to the Schedule of Recent Experiences (pre, 48.1 ⫾ 8.1; post, 61.7 ⫾ 7.6; p ⬍ .05). Several scales demonstrate trends of increased distress during exams; however, they did not reach statistical significance. Correlations between stress and cytokine production. Even though adjustment to life events at 6 months was the only self-reported stress measure to significantly change from preexam to postexam, several psychosocial measures were correlated with cytokine levels (Table 2). The number of hassles and subjective adjustment to hassles were positively correlated with IL-10 levels (number of hassles pre, r ⫽ ⫺.85, TABLE 2 Correlation of Type-1/Type-2 Cytokine Production with the Daily Hassles Scale IFN-γ

Number of hassles (Preexam) Subjective adjustment to hassles (Preexam)

IL-10

IFN-γ: IL-10 ratio

Preexam

Postexam

Preexam

Postexam

Preexam

Postexam

⫺0.06

⫺0.20

0.85***

0.66**

⫺0.63*

⫺0.52*

⫺0.13

0.11

0.75**

0.63*

⫺0.62*

⫺0.56*

Note. Mitogen-stimulated IL-10 production by PBMC before and after exams was positively correlated and IFN-γ: IL-10 ratio before and after exams was negatively correlated with the number of hassles and the subjective adjustment to the Daily Hassles Scale. Values listed above are correlation coefficients (r). n ⫽ 14 subjects. * p ⬍ .05. ** p ⬍ .01. *** p ⬍ .001.

CYTOKINE DYSREGULATION WITH EXAM STRESS

303

p ⬍ .001; number of hassles post, r ⫽ ⫺.66, p ⬍ .01; subjects assessment to hassles pre, r ⫽ ⫺.75, p ⬍ .01; subjective assessment to hassles post, r ⫽ ⫺.63, p ⬍ .05) and negatively correlated with IFN-γ : IL-10 ratio (number of hassles pre, r ⫽ ⫺.63, p ⬍ .05; number of hassles post, r ⫽ ⫺.52, p ⬍ .05; subjects assessment to hassles pre, r ⫽ ⫺.62, p ⬍ .05; subjective assessment to hassles post, r ⫽ ⫺.56, p ⬍ .05). Subjects who reported more hassles and a greater subjective adjustment to hassles preexam had higher IL-10 levels and a lower IFN-γ :IL-10 ratio at both preexam and postexam (Table 2). Finally there were no correlations between plasma cortisol levels and type-1/type-2 cytokine production (data not shown). To investigate the relationship between the Daily Hassles Scale and stress-related type-1/type-2 cytokine alterations, the number of hassles and the subjective adjustment to the Daily Hassles Scale at the preexam visit was divided at the median (Fig. 2). Both groupings resulted in the same classification of subjects into preexam low and high hassles groups. There was again a slight decrease in IFN-γ levels with exam stress in both the low hassles (pre, 47919 ⫾ 7746 pg/ml; post, 34034 ⫾ 4635 pg/ ml; p ⫽ n.s.) and the high hassles group (pre, 42156 ⫾ 9419 pg/ml; post, 39818 ⫾ 8614 pg/ml; p ⫽ n.s.). IL-10 production was also increased in both the low hassles group (pre, 541 ⫾ 107 pg/ml; post, 1175 ⫾ 389 pg/ml; p ⫽ n.s.) and the high hassles group (pre, 1039 ⫾ 161 pg/ml; post, 1813 ⫾ 280 pg/ml; p ⫽ .03). Of note, the high hassles group produced significantly more IL-10 than the low hassles group during the preexam period (high, 1039 ⫾ 161 pg/ml; low, 541 ⫾ 107 pg/ml; p ⫽ .02). When the two groups were analyzed for changes in IFN-γ:IL-10 ratio with exam stress (Fig. 2C), only the low hassles group showed this effect (pre, 109 ⫾ 24; post, 39 ⫾ 7; p ⫽ .03) while the high hassles group did not demonstrate a significant difference (pre, 43 ⫾ 10; post, 23.5 ⫾ 6; p ⫽ n.s.), even though the high hassles group had a significantly lower IFN-γ : IL-10 ratio during the preexam period compared to the low hassles group (high, 43 ⫾ 10; low, 109 ⫾ 24; p ⫽ .02). DISCUSSION

The association between psychological processes, neuroendocrine function, immune responses, and health continues to generate intensive interest and research effort across the spectrum of psychoneuroimmunology. The behavior most commonly studied involves stress, an alteration of the host homeostatic state (Kusnecov & Rabin, 1994). The most common immunological effect of chronic stress suggests immunosuppression and the most common neuroendocrine effect invokes the hypophyseal– pituitary–adrenal axis and/or sympathetic nervous system activation. In the current investigation, we utilized a reliable model of psychological stress, the academic exam stress model (Glaser et al., 1987), to demonstrate decreased IFNγ production accompanied by increased IL-10 production by PBMC isolated from healthy medical students during medical school exams. This resulted in a decrease in the IFN-γ :IL-10 ratio during the postexam period compared to the preexam period suggesting that psychological stress associated with medical student exams produces a shift in the type-1/type-2 cytokine balance toward a type-2 response. The number of daily hassles and subjective adjustment to hassles before the exam was positively correlated with IL-10 levels and negatively correlated with IFN-γ :IL-10 ratio before and after exam. Similar to the effects of exam stress, number of hassles and adjustment to hassles were associated with type-1/type-2 cytokine shift. The other selfreport measures did not show statistically significant correlations with cytokine lev-

304

MARSHALL ET AL.

FIG. 2. Type-1/type-2 cytokine balance in groups divided at the median of Daily Hassles Scale. Cytokines assayed in supernatants from 72-h PHA/PMA-stimulated cultures of PBMC from healthy medical students divided into high and low hassles groups at the median score. (A) IFN-γ ( p ⫽ n.s. by paired t test for both groups). (B) IL-10 (pre- vs postexam, low hassles, p ⫽ n.s.; high hassles, p ⫽ .03.; low vs high hassles group during the preexam period, p ⫽ .02). (C) IFN-γ: IL-10 ratio (pre- vs postexam, low hassles, p ⫽ .03; high hassles, p ⫽ n.s.; low vs high hassles group during the preexam period, p ⫽ .02). Open bars represent preexam period. Solid bars represent postexam period. n ⫽ 14 subjects.

CYTOKINE DYSREGULATION WITH EXAM STRESS

305

els, but several trends were noted which would possibly be significant with a larger sample size. Finally, we demonstrated that PBMC subjects with higher preexam hassles scores did not demonstrate a significant shift in the type-1/type-2 cytokine ratio toward a type-2 response compared to PBMC from subjects with lower preexam hassles scores, although at the preexam point they already had a significantly lower IFN-γ:IL-10 ratio indicative of a shift toward type-1 cytokines. Taken together, these data suggest that the psychological stress associated with medical student exams produces immune dysregulation rather than an overall suppression. Previous studies using the medical student exam stress model have demonstrated a suppression of cytokine (IFN-γ, IL-2) production in mitogen-stimulated cultures; however, the concurrent production of type-2 cytokines such as IL-10 was not investigated in these studies (Glaser et al., 1987; Dobbin et al, 1991). Our data confirmed the previous findings showing a decrease in IFN-γ during medical exams compared to a time of lower stress (Glaser et al., 1987; Dobbin et al., 1991), but extend the results by demonstrating increases in IL-10 production during the same stress period. IL-10, also called cytokine regulatory factor, can downregulate IFN-γ production and thus overall cellular immune responses (Mosmann & Sad, 1996). Imbalances in the IFN-γ :IL-10 ratio (i.e., decreased IFN-γ and/or increased IL-10) may be one of the mechanisms whereby exposure to exam stress produces a decrease in cellular immunity. Interestingly, in individuals who were initially shifted toward a type-2 cytokine response (high hassles group), further shift due to exam stress was much less than those with an initial type-1 predominant response (low hassles group). These data are consistent with previous reports from our laboratory showing a type-2 predominant cytokine response to be resistant to further change by in vitro exposures to corticosteroid levels commonly reported in plasma of stressed individuals (Agarwal & Marshall, 1998b). These data suggest that the level of stress at baseline can have major effects on immune changes observed during the stressor period itself. Finally, the immune changes were independent of changes in serum cortisol. The lack of correlation between serum cortisol and immune function has been reported by other investigators (Glaser et al., 1994). Timed periodic (i.e., 6–24 h) urinary or even salivary cortisol, instead of random plasma sampling, may be necessary to see differences in smaller sample sizes. The relevance of these findings centers on the different therapeutic and/or prophylactic approaches that may be useful for individuals experiencing chronic stress. Imbalances in type-1/type-2 cytokine production have been associated with several immune-based diseases. Decreased type-1 cytokine responses can result in decreased CMI and a resulting susceptibility to increased viral, fungal, and mycobacterial infections (Romani, Pucetti, Mencacci, Cenci, Spaccapelo, Tonnetti, Grohmann, & Bistoni, 1994; Barnes, Lu, Abrams, Wang, Yamamura, & Modlin, 1993). In contrast, humoral immunodeficiencies result in increased susceptibility to many bacteria and certain parasites (Urban, Maliszewski, Madden, Katona, & Finkelman, 1995; Finkelman, Madden, Cheever, Katona, Morris, Gately, Hubbard, Gause, & Urban, 1994). Furthermore, allergen-specific type-2 cytokine responses play an important role in hypersensitivity diseases, such as allergic rhinitis and asthma (Robinson, Hamid, Bentley, Ying, Kay, & Durham, 1993). Therefore the stress-related immune alterations reported in this study are dysregulation rather than pure immunosuppression. The small sample size makes it difficult to explore the specific cause(s) of the cytokine alterations and it is not possible to

306

MARSHALL ET AL.

examine interactions between variables. We are currently extending the sample size of medical students to more fully address these issues. Further studies are in progress to examine other human stress models and define the mechanisms of immunological changes associated with specific psychosocial variables. ACKNOWLEDGMENTS These studies were supported in part by Grant NGT-9-25 from the National Aeronautics and Space Administration.

REFERENCES Agarwal, S. K., & Marshall, G. D. (1998a). In vivo alteration in type-1 and type-2 cytokine balance: A possible mechanism for elevated total IgE in HIV-infected patients. Hum. Immunol. 59, 99–105. Agarwal, S. K., & Marshall, G. D. (1998b). Type-1, not type-2 cytokines are sensitive to dexamethasone. FASEB J. 12, A764–A764(abstract). Anderson, B. L., Farrar, W. B., Golden-Kreutz, D., Kutz, L. A., MacCallum, R., Courtney, M. E., & Glaser, R. (1998). Stress and immune responses following surgical treatment for regional breast cancer. J. Natl. Cancer Inst. 90, 30–36. Barnes, P. F., Lu, S., Abrams, J. S., Wang, E., Yamamura, M., & Modlin, R. L. (1993). Cytokine production at the site of disease in human tuberculosis. Infect. Immun. 61, 3482–3489. Cohen, S., Kamarck, T., & Mermelstein, R. F. (1983). A global measure of perceived stress. J. Health Soc. Behav. 24, 385–396. Croft, M., Carter, L., Swain, S. L., & Dutton, R. W. (1994). Generation of polarized antigen-specific CD8 effector populations: Reciprocal action of IL-4 and IL-12 in promoting type-2 vs. type-1 cytokine profiles. J. Exp. Med. 180, 1715–1728. Del Prete, G., De Carli, M., Mastromauro, C., Biagotti R., Macchia D., Falagiani P., Ricci M., & Romagnani, S. (1991). Purified protein derivative of Mycobacterium tuberculosis and excretory–secretory antigen(s) of Toxicara canis expand in vitro human T cells with stable and opposite (type 1 T helper or type 2 T helper) profile of cytokine production. J. Clin. Invest. 88, 346–350. Derogatis, L. R. (1977). SCL-90-R: Administration scoring and procedures manual I. Clinical Psychometrics Research: Baltimore, MD. Dobbin, J. P., Harth, M., McCain, G., Martin, R. A., & Cousin, K. (1991). Cytokine production and lymphocyte transformation during stress. Brain Behav. Immun. 5, 339–348. Esterling, B. A., Kiecolt-Glaser, J. K., Bodnar, J. C., & Glaser, R. (1994). Chronic stress, social support, and persistent alterations in natural killer cell responses to cytokine in older adults. Health Psychol. 13, 291–298. Finkelman, F. D., Madden K. B., Cheever, A. W., Katona, I. M., Morris, S. C., Gately, M. K., Hubbard, B. R., Gause, W. C., & Urban, J. F. (1994). Effects of IL-12 on immune responses and host protection in mice infected with intestinal nematode parasites. J. Exp. Med. 179, 1563–1572. Frankenhauser, M., von Wright, M. R., Collins, A., von Wright, J., Sedvall, G., & Swahn, C. G. (1978). Sex differences in psychoneuroendocrine reactions to exam stress. Psychosom. Med. 40, 334–343. Glaser, R., Pearl, D. K., Kiecolt-Glaser, J. K., & Malarkey, W. B. (1994). Plasma cortisol levels and reactivation of latent-Epstein–Barr virus in response to examination stress. Psychoneuroendocrinology 19, 765–772. Glaser, R., Pearson, G. R., Bonneau, R. H., Esterling, B. A., Atkinson, C., & Kiecolt-Glaser, J. K. (1993). Stress and the memory T-cell response to the Epstein–Barr virus in healthy medical students. Health Psychol. 12, 435–442. Glaser, R., Pearson, G. R., Jones, J. F., Hillhouse, J., Kennedy, S., Mao, H., & Kiecolt-Glaser, J. K. (1991). Stress-related activation of Epstein–Barr virus. Brain Behav. Immun. 5, 219–232. Glaser, R., Rice J., Sheridan, J., Fertel, R., Stout, J., Speicher, C. E., Pinsky, D., Kotur, M., Post, A., Beck, M., & Kiecolt-Glaser, J. K. (1987). Stress-related immune suppression:Health implications. Brain Behav. Immun. 1, 7–20. Holmes, T. H., & Rahe, R. H. (1967). The social readjustment scale. Psychosom. Res. 11, 213–218.

CYTOKINE DYSREGULATION WITH EXAM STRESS

307

Kanner, A. D., Coyne, J. C., Schaefer, C., & Lazarus, R. S. (1981). Comparison of two modes of stress management: Daily hassles and uplifts versus major life events. J. Behav. Med. 4, 1–39. Kiecolt-Glaser, J. K., Dura, J. R., Speicher, C. E., Trask, O. J., & Glaser, R. (1991). Spousal caregivers of dementia victims: Longitudinal changes in immunity and health. Psychosom. Med. 53, 345–362. Kiecolt-Glaser, J. K., Fisher, L., Speicher, C. E., Penn, G. M., & Glaser, R. (1987). Marital quality, marital disruption, and immune function. Psychosom. Med. 49, 13–34. Kiecolt-Glaser, J. K., Garner, W., Speicher, C. E., Penn, G. M., Holliday, J. E., & Glaser, R. (1984). Psychosocial modifiers of immunocompetence in medical students. Psychosom. Med. 46, 7–14. Kiecolt-Glaser, J. K., Glaser, R., Strain, E. C., Stout, J. C., Tarr, K. L., Holliday, J. E., & Speicher, C. E. (1986). Modulation of cellular immunity in medical students. J. Behav. Med. 9, 5–21. Kusnecov, A. W., & Rabin, B. S. (1994). Stressor-induced alterations of immune function: Mechanisms and issues. Int. Arch. Allergy Immunol. 105, 107–121. Marshall, J. D., Wen, Y., Abrams, J. S., & Umetsu, D. T. (1993). In vitro synthesis of IL-4 by human CD4⫹ T cells requires repeated antigenic stimulation. Cell. Immunol. 152:18–34. Mosmann, T. R., Cherwinksi, H., Bond, M. W., Giedlin, M. A., & Coffman, R. L. (1986). Two types of murine helper T cell clone. Definition according to profiles of lymphokine activities and secreted proteins. J. Immunol. 136, 2348–2356. Mosmann, T. R., & Sad, S. (1996). The expanding universe of T-cell subsets: Th1, Th2 and more. Immunol. Today 17, 138–146. Pariante, C. M., Carpiniello, B., Orru, M. G., Sitzia, R., Piras, A., Farci, A. M. G., Del Giacco, G. S., Piludu, G., & Miller, A. H. (1997). Chronic caregiving stress alters peripheral blood immune parameters: The role of age and severity of stress. Psychother. Psychosom. 66, 199–207. Robinson, D., Hamid, Q., Bentley, A., Ying, S., Kay, B., & Durham, S. R. (1993). Activation of CD4⫹ T cells, increased Th2 type cytokine mRNA expression, and eosinophil recruitment in BAL after allergen inhalation challenge in patients with atopic asthma. J. Allergy Clin. Immunol. 92, 313– 324. Romagnani, S. (1994). Lymphokine production by human T cells in disease states. Ann. Rev. Immunol. 12, 227–257. Romani, L., Pucetti, P., Mencacci, A., Cenci, E., Spaccapelo, R., Tonnetti, L., Grohmann, U., & Bistoni, F. (1994). Neutralization of IL-10 up-regulates nitric oxide production and protects susceptible mice from challenge with Candida albicans. J. Immunol. 152, 3514–3521. Russel, D., Peplau, L. A., & Cutrona, C. E. (1980). The revised UCLA Loneliness Scale: concurrent and discriminant validity evidence. J. Personal Soc. Psychol. 39, 472–480. Schleifer, S., Keller, S., Camerino, M., Thornton, J. C., & Stein, M. (1983). Suppression of lymphocyte stimulation following bereavement. J. Am. Med. Assoc. 250, 374–377. Sharkey, B. J. (1990). Aerobic fitness. In B. J. Sharkey (Ed.), Physiology of Fitness, pp. 255–263. Human Kinetics: Champaign. Urban, J. F., Maliszewski, C. R., Madden, K. B., Katona, I. M., & Finkelman, F. D. (1995). IL-4 treatment can cure established gastrointestinal nematode infections in immunocompetent and immunodeficient mice. J. Immunol. 154, 4675–4684. Received July 31, 1998