Effects of exercise on immune functions of undernourished mice

Life Sciences, Vol. Printed in the USA

51, pp. 565-574

Pergamon Press

EFFECTS OF EXERCISE ON IMMUNE FUNCTIONS OF UNDERNOURISHED MICE S. M. Filteau, Robert A. Menzies, Thomas J. Kaido, Maureen P. O'Grady, John B. Gelderd* and Nicholas R.S. Hall Departments of Psychiatry and Behavioral Medicine, and Medical Microbiology and Immunology, University of South Florida College of Medicine, USF Psychiatry Center, 3515 East Fletcher Avenue, Tampa, Florida 33613 and *Department of Anatomy, College of Medicine, Medical Sciences Building, Texas A&M University, College Station, Texas 77843 (Received in final form June ii, 1992)

Summary. Regular moderate exercise may modulate the response to a stressor and thus improve immune functions in conditions commonly associated with immunodepression and elevated levels of stress hormones. For example, anorexia nervosa patients, many of whom engage in regular aerobic exercise, generally have normal immune function and viral disease resistance in spite of their severe undernutrition. To test the hypothesis that exercise can prevent undernutritioninduced immunodepression, mice were fed a nutritionally complete, semi-purified diet, either ad libitum or in restricted quantities to induce 25% loss of initial weight over 3 weeks. Half the animals from each dietary group were run on a treadmill for 30 min/day, 5 days/week. Exercise had no effect on several measures of nutritional status. Spleen weight and blastogenic response to lipopolysaccharide were significantly increased by exercise in undernourished mice. In vivo antibody response to sheep red blood cells, and in vitro splenic responses to concanavalin A and phytohemagglutinin were not significantly affected by exercise. Serum corticosterone level was increased by food restriction and significantly decreased by exercise in the undernourished mice. Within a treatment group there were no significant correlations between serum corticosterone level and any immune system measure. Hypothalamic concentration of uric acid was increased in food restriction groups and concentration of norepinephrine was increased in exercise groups. The results suggest that regular exercise may help prevent undernutrition-induced immunodepression, possibly through modulation of the stress response. Anecdotal reports suggesting that regular exercise can improve health and resistance to infectious disease have recently stimulated research on the effects of exercise on immune function. In healthy animals, regular moderate exercise has been shown to increase several immune functions, including antibody responses (I) and proliferative responses to mitogens (2). However, immunosuppressive effects of exercise training have also been reported (3, 4). The variable results presumably stem from a variety of experimental design factors including age of the animals (4), whether they are individually or group-housed (5), the particular immune response studied (2, 6), the intensity of the imposed exercise (7) and its timing with respect to the immunological testing (3,6). Several investigators have suggested that exercise-induced immunological changes may be related to changes in hormones associated with the stress response (8), such as increased levels of glucocorticoids (4,6,9) or g-endorphin (10). Thus, a very vigorous training program (7) or an exhaustive exercise 0024-3205/92 $5.00 + .00 Copyright © 1992 Pergamon Press Ltd All rights reserved.

566

Exercise and Immunity in Undernutrition

Vol. 51, No. 8, 1992

bout (3) may be stressors resulting in immunosuppression. On the other hand, a more modest level of exercise may modulate the response to stress (11) such that a given stressor would then be less immunosuppressive. This potential for exercise to modulate the response to stress and to mitigate stress-induced immunodepression would be particularly important in conditions in which immunodepression or stress already occurs. Recent reports concerning several such conditions have shown that regular moderate exercise inhibited tumor growth in hepatoma-implanted rats (12), increased natural killer cell activity in elderly women (13), and attenuated emotional distress and natural killer cell decrements in patients at risk for AIDS (14). Another condition which is generally associated with impaired immune function and increased incidence and severity of infectious disease is severe undernutrition (15). In this regard, it is of interest that severely undernourished anorexia nervosa patients, who commonly engage in regular exercise, are reported to have well-maintained immune functions and to be relatively free of viral infections (16-19). Therefore, the present study was conducted to determine whether a moderate exercise protocol could improve immune functions in severely undernourished mice. In addition, measurements of serum corticosterone levels and hypothalamic monoamines, neuroendocrine factors known to be involved in the response to many types of stressors (8), were included to determine whether any immune system changes induced by exercise could be associated with an altered stress response. Methods Animals and Diets. Young (4 weeks of age) female, outbred ICR mice (Harlan, Indianapolis, IN), were used since this work was conducted as part of a series of studies on anorexia nervosa, a disease which affects mainly young women. Animals were individually housed in hanging wire cages in a room maintained at 28°C with a 12/12 light-dark cycle (lights on 5:00 a.m.). For one week after arrival mice were acclimated to the facility and to the nutritionally complete powdered purified diet (AIN-76A with 50% starch, 50% glucose as the carbohydrate source - U.S. Biochemicals, Cleveland, OH). The experiment was run in three blocks, beginning on consecutive days. On each day, mice were divided into four treatment groups in a 2x2 design: ad libitum or restricted feeding, both with or without exercise. Feed restriction was designed to induce loss of 25% of initial body weight in an approximately linear fashion over the course of the 3 week experimental period. All mice were weighed daily and their food allotment given at 5:00 p.m., at the beginning of the dark period, in order to minimize changes in circadian rhythms due to imposition of meal-feeding (20). Exercise Protocol. Five days per week, during the light period, half the animals from each dietary group were exercised on a motorized treadmill (Columbus Instruments, Columbus, OH) in a room maintained at 24oC to reduce the possibility of heat stress. During the first week the exercise duration and intensity were increased from 0.2 meters/second for 5 minutes to 0.3 meters/second for 30 minutes and this higher level was maintained for the subsequent two weeks of the experiment. The treadmill was equipped with a shock grid to induce the animals to run but, since the animals learned quickly, it was necessary to use this, at a low setting, for only the first two days. All but one animal, who was excluded, adapted readily to the treadmill and could easily sustain the imposed amount of exercise. Unexercised mice were exposed at the same time to the 24°C lab in order to control for effects of the novel environment and handling. Mice were not exercised on either the day of sheep red blood cell (SRBC) immunization or the day of sacrifice. Once a week, in the 24oc lab, before and after an exercise session, rectal temperatures of all mice were measured using a probe for very small animals (Harvard Apparatus, South Natick, MA). At the end of the experimental period, between 9:00 and 10:00 a.m., mice were rapidly killed by decapitation in a separate room. Trunk blood was collected and hypothalamuses rapidly dissected on ice, weighed, transferred to Eppendorf tubes, frozen in liquid nitrogen and subsequently kept at 7 0 o c until monoamine analysis. Spleens, thymuses, hearts, and adrenal glands were dissected and weighed and the remainder of the carcasses frozen at -20oc for subsequent lipid analysis. Brain sections were frozen at -70°C until monamine analysis and the remainder of the carcases at -20oc for subsequent lipid analysis.

VOI. 51, NO. 8, 1992

Exercise and Immunity in Undernutrition

567

SDlenocvte Preparation. Spleens were minced in RPMI 1640 medium containing 10% fetal bovine serum, 100 units/ml penicillin, 100 ~tg/ml streptomycin and 2 mM L-glutamine (complete medium). The cell suspension was washed twice. Lymphocyte Blastogenesis. Spleen cells in complete medium were seeded at 2.5x105 cells/well in 96-well flat-bottomed microtitre plates. Medium alone or medium with an optimal concentration of concanavalin A (Con A, 5 ~g/ml) or phytohemagglutinin (PHA, 5 I.tg/ml) or Salmonella enteriditis lipopolysaccharide (LPS, 1 p.g/ml; all mitogens from Sigma, St. Louis, MO) was added, and the plates were incubated for 72 hours at 37°C and 5% CO2. Tritiated thymidine (0.5 p.Ci/well) was added for the final 6 hours incubation. Samples were tested in triplicate. Antibody Response to SRBC. Five days before the end of the feeding period mice were injected with 2x108 SRBC in 0.1 ml phosphate-buffered saline. Hemagglutinin titres were measured by standard methods and expressed as the reciprocal of the base 2 log of the maximum serum dilution exhibiting spreading of SRBC. Measures of Nutritional Status and Serum Corticosterone Levels. Serum albumin was measured by the bromcresol green method using a commercial kit (Sigma, St. Louis, MO). Carcasses were made into homogeneous samples for lipid analysis (21) by lyophilization for 4 days, grinding in a conventional coffee grinder and then finishing drying in a l l0°C oven overnight. Serum corticosterone levels were determined using a commercial kit (ICN Biomedicals, Costa Mesa, CA). Measurement of Hypothalamic Monoamines and Metabolites. Hypothalamus samples were homogenized by sonication (Heat Systems, Inc.) with the micro tip at half maximum power for 10 seconds in 0.4 ml of 0.1 M HC104 containing 0.1 mM EDTA and N-methyldopamine as an internal standard. The samples were injected directly into the HPLC system after centrifugation. The HPLC system consisted of a C- 18 column (Keystone spherisorb, ODS 1) and mobil phase of 0.1 M NaH2PO 4 (pH 3.15-3.2) containing 0.1 mM Na2EDTA, 0.2-0.4 mM sodium octylsulfonate and 26% CH3CN (v/v) degassed with helium. The colunm is maintained at 30-32 ° C and pumped at 1.0 ml/min with the electrochemical detectors (Bio Analytical systems) set at potentials of 0.8 and 0.95 Statistics. Data was subjected to 2x2 analysis of variance. The day of assay was included as a significant covariate in the analyses of blastogenic responses only. Where significant F values were obtained, t-test comparisons were made between the two non-exercised groups to determine the effect of food restriction and between each exercised group and its respective non-exercised control to determine the effect of exercise. Where necessary, data was normalized by log transformation before analysis of variance. For hemagglutinin titres which could not be normalized by transformation, the Kruskal-Wallis non-parametric test was used (22). Statistical significance was defined as p < 0.05 throughout. Results All animals survived the imposed treatments, but on autopsy it was noted that 5 of the 15 restricted, non-exercised mice exhibited grossly evident discoloration throughout the gut lumen. Histological examination was inconclusive since it was unplanned and tissues were not fixed until after the carcasses had been frozen at -20°C. However, the discoloration was apparently due to severe bleeding of the gut wall. Data from the five affected mice did not appear to differ from that of other animals in the restricted, non-exercised group, therefore it was included in the analysis. Such intestinal discoloration was not seen in any animals from other experimental groups. Table I shows body and organ weights and several measures of nutritional status for the experimental animals. Exercise did not affect body weight changes of either well-nourished or undernourished mice. Food intake was decreased by exercise in the ad libitum-fed animals but the amount of food needed to induce 25% weight loss in the undernourished mice was unaffected by exercise. Organ weights are expressed per gram body weight in order to facilitate comparisons between groups of mice with very different final weights. Heart weight was significantly increased by exercise in the well-nourished mice, as has been shown by others to occur with aerobic training

568

Exercise

and Immunity

in Undernutrition

Vol.

51, No. 8, 1992

TABLE I Body and Organ Weights, Food Intakes, Serum Corticosterone Levels and Measures of Nutritional Status I Ad libiturn No exercise

Ad libitum Exercise

Restricted No exercise

Restricted Exercise

E.M.S.2 2.583 4.637 42.959

Initial weight (g) Final weight (g) Food intake (g/21 days)

22.2 27.3 71.7

22.0 25.8 66.5 b

21.7 15.9 a 32.7 a

22.7 16.9 31.9

Spleen weight (mg/g) 3 Thymus weight (mg/g) 3 Heart weight (mg/g) 3

2.89 2.49 3.96

2.66 2.19 4.62 b

1.74 a 0.77 a 6.28 a

2.58 b 0.86 5.79

0.461 0.222 0.592

Serum corticosterone (ng/ml) Serum albumin (g/dl) Carcass lipid (%)4

314 4.1 25.2

329 4.3 27.0

664 a 3.9 5.8 a

451 b 3.7 7.7

61239 0.274 17.81

1 2 3 4

N = 12-15 per group for most data, 8-12 for serum albumin, and 10 for carcass lipid. Error mean square mg/g body weight % of carcass wet weight

a

Significanteffectofdiet; restricted, no exercise different ~om adlibitum, no exe~ise, t-test, p<0.05. Significant effect of exe~ises; diffe~nt ~om non-exercised o f s a m e dietary group, t-test, p<0.05.

b

in rodents (12). The increased heart weight per gram body weight in the undernourished rodents simply reflects the lower body weight of these animals since absolute heart weights did not differ among groups. Similarly, adrenal weights were unaffected by experimental treatments (data not shown). Spleen and thymus weights as a proportion of body weight were decreased by undernutrition, as has been frequently shown to occur in severe malnutrition (23). Exercise significantly increased spleen weight, but not thymus weight in the food-restricted animals. Carcass lipid levels were significantly decreased by undernutrition but serum albumin levels were unaffected, indicating depletion of energy stores but maintenance of visceral protein. These results suggest that energy was the main limiting factor in the diet of the food-restricted mice. Exercise had no effect on either carcass lipid or serum albumin. Serum corticosterone levels are presented in Table I. All values are high for rodents at this time of day (24), suggesting the animals were stressed. Individual housing, which was necessary for accurate control and measurement of food intake, may have contributed to these high values (25). In addition, mice were handled twice daily for weighing and feeding and for exercise or control exposure to the 24°C lab. Therefore, they had learned to associate entrance of lab personnel into the room with food or exercise which might explain the high corticosterone levels at the time of sacrifice. This interpretation, rather than technical problems, is considered reasonable since

Vol. 51, No. 8, 1992

Exercise and Immunity in Undernutrition

569

TABLE II Hypothalamic Monoamines and Their Metabolites 1

Metabolite or ratio

Ad libitum No exercise

Ad libitum Exercise

Restricted No exercise

Restricted Exercise

E.M.S.

Norepinephrine MHPG 2 MHPG/norepinephrine

0.775 0.135 0.174

1.060 b 0.146 0.138

0.840 0.134 0.164

0.900 0.126 0.139

0.01513 0.0030 0.0049

Serotonin 5-HIAA2 5-HIAA/serotonin

0.806 0.603 0.837

0.868 0.656 0.802

0.978 0.746 0.805

0.830 0.802 0.978

0.1850 0.1553 0.0748

Dopamine Dopac HVA2 Dopac/dopamine HVA/dopamine

0.497 0.219 0.192 0.468 0.402

0.557 0.225 0.191 0.418 0.367

0.465 0.204 0.206 0.451 0.465

0.461 0.218 0.209 0.498 0.480

0.0167 0.0036 0.0019 0.0100 0.0160

Uric acid

0.195

0.214

0.304

0.00993

.335 a

Values are I.tg/g tissue wet weight. Sample sizes are 9-12 per group. MHPG = Methoxyhydroxyphenylglycol; 5-HIAA = 5-hydroxy indole acetic acid; HVA = homovanillic acid. From ANOVA using log transformed data. Values in the table are antilogs of log means. Significant effect of diet; restricted, no exercise different from ad libitum, no exercise, t-test, p<0.05. Significant effect of exercise; different from non-exercised of same dietary group, t-test, p<0.05.

measurement in the same assay of serum corticosterone levels of a group of previously untouched Sprague-Dawley rats, used in a different experiment but killed similarly to the mice in this experiment, produced normal, unstressed serum hormone levels (20 + 14 ng/ml, mean + S.D., n = 8). In spite of the high corticosterone levels in well-nourished mice, the levels in undernourished mice were significantly higher. Exercise decreased serum corticosterone level in food-restricted mice but did not affect this measure in the ad libitum-fed animals. Concentrations of hypothalamic monoamines, their metabolites and ratios of metabolites are shown in Table II. The values are similar to those found by others for concentrations of monoamines in rodent hypothalamus (26,27). Few significant group differences were seen other than an increase in uric acid levels in food-restricted mice and an increase in norepinephrine in exercised mice. There was a significant effect of diet on the ratio of homovanillic acid to dopamine in the analysis of variance but this effect was not significant when individual group means were compared. Mouse rectal temperatures, taken at weekly intervals before and after the daily session of exercise or transport to the 24oc lab, are shown in Figure 1. In the food-restricted mice, temperatures taken before exercise decreased with time, and the exercised group had lower temperatures after weeks one

570

Exercise

and Immunity

in Undernutrition

Vol.

51, No. 8, 1992

and two than did the non-exercised group. The stress of being in the exercise lab induced small temperature increases in non-exercised mice. Large temperature increases were seen after exercise, particularly in food-restricted mice whose rectal temperatures were then in the same range as those of well-nourished mice.

41 40

d

b,d

b,d

b,d

39

°C 38

[]

before exercise

[]

after exercise

37

36 ALAL,ExR R,Ex -

Day 7

AIAL,ExR R,Ex -

Day 14

AIAL.ExR R,Ex Day 21

FIG. 1 Weekly Mouse Rectal Temperatures Before and After Exercise Sessions 1

a,c

b,d

N = 14-15 per group. Error mean squares from the ANOVA's were, before and after exercise, respectively: for day 7, 0.293 and 0.059; day 14, 0.475 and 0.131; day 21, 1.462 and 1.492. Significant effect of diet; restricted no exercise different from ad libitum, no exercise, before (a) and after (c) the exercise session, t-test, p<0.05. Significant effect of exercise; different from non-exercised of same dietary group, before (b) and after (d) the exercise session, t-test, p<0.05.

Splenic lymphocyte blastogenic responses are illustrated in Figure 2. No group differences were seen except that the response to LPS of undernourished mice was significantly increased by exercise. There was some concern that intestinal damage may have led to increased uptake of gram negative bacteria which may have influenced the splenic response to LPS. However, in three of the five mice which exhibited intestinal bleeding, a shortage of spleen cells precluded measurement of the LPS response, and in the other two animals, the responses were within the range of the rest of the group. When blastogenic responses were calculated per spleen, rather than per 2.5x105 cells, in order to account for differences in spleen weight, the total spleen response to all mitogens was decreased by undernutrition (results not shown). However, the response to LPS remained the only one significantly increased by exercise in the undernourished mice. Within each treatment group, no significant correlations were seen between serum corticosterone level and any of the immune system measures. Although several significant correlations between particular immune functions and monamine concentrations were found, these showed no consistent trends and appeared to be random, resulting from the number of comparisons made (results not shown).

Vol. 51, No. 8, 1992

Exercise and Immunity in Undernutrition

571

110-

[] [] []

100 "

AL AL Ex a

9080

60 50 40 30 20 10

Medium

PHA

Con A

LPS

FIG. 2 Splenic L y m p h o c y t e Blastogenic R e s p o n s e s l , 2 1 2

*

A L = ad libitum, no exercise; A L E x = ad libitum, exercise; R = restricted, no exercise; R E x = restricted, exercise. Error means squares were, for medium control, 1.57x106; for PHA, 7.02x108; for Con A, 2.33x109; for LPS, 6.76x107. Sample size is 14 per group except for restricted, no exercise where N=12. Error bars are calculated as "/EMS/n. Different from restricted, no exercise, t-test, p<0.05.

TABLE III Hemagglutinin Titres Against Sheep Red Blood Cells I Mean Ad libitum, no exercise Ad libitum, exercise Restricted, no exercise Restricted, exercise 1

9.0 8.6 8.0 8.8

No. of Mice (13) (14) (14) (12)

Analysis by Kruskal-Wallis non-parametric test. Mean rank sums were 32.0, 26.8, 19.2 and 30.9 for groups in the order listed above.

572

Exercise

and I m m u n i t y

in U n d e r n u t r i t i o n

Vol.

51, No.

8, 1992

Discussion The results presented above support the hypothesis that aerobic exercise can protect against the impaired immune function and infectious disease resistance of severe undernutrition. The most notable effect of exercise was complete prevention of the intestinal bleeding seen in one third of the undernourished, non-exercised animals. In addition, exercise increased spleen weight and blastogenic response to LPS in the food-restricted mice. The effects of exercise on immune functions were modest which was possibly a function of the particular exercise and undernutrition protocols chosen. The choice of protocols was based on a preliminary experiment in which exercise during undernutrition tended to increase thymus weight and splenic responses to Con A and PHA. It should be mentioned that the animals' activity in their home cages was not quantitated and thus it is possible that 24 hour activity of the non-exercised mice was not less than that of the exercised mice. However, rodents do not normally sustain running for periods of 30 minutes. The increased heart weight in the exercised well-nourished mice was evidence of a physiological difference induced by the exercise protocol. The intensity of exercise used was designed to be great enough to have such measurable physiological effects but not so great as to be itself a severe stressor. The lack of increase of serum corticosterone levels in exercised compared to non-exercised mice suggests that the exercise protocol was not so stressful as to elevate corticosterone either 24 hours after exercise or in anticipation of exercise on the day of sacrifice. However, possibly a slightly more intense exercise regimen - either greater running speed or longer duration - would have had greater physiological and immunological effects while still avoiding the damaging effects of a severe stressor. A further explanation for the only modest effects of exercise is that the nutritional deprivation was not severe enough to cause large depressions in the immune functions measured. Similar undernutrition protocols used by us in other studies have induced depressions in Con A response and antibody response to SRBC in addition to the alterations seen in this study. Such depressions were not seen in the present work, possibly because of variations among outbred mice. The exercise protocol may have had no effect on immune functions which were already normal in unexercised mice, for example, in the well-nourished mice in the present work. It may be relevant that the positive exercise effects seen in the undernourished animals were generally seen in the measures (e.g. prevention of intestinal bleeding) which were most affected by undernutrition. It is possible that exercise may have been of greater benefit in a model of more severe malnutrition-induced immunodepression. Recently Ueda and co-workers have also looked at interactive effects of malnutrition and exercise on immune functions. Running exercise significantly increased in vitro alveolar macrophage interleukin 1 (ILl) production and phagocytosis of SRBC in well-nourished but not protein-deficient rats (28). However, both these immune functions were higher in sedentary animals fed the low protein rather than the adequate diet which at least partly accounted for the lack of exercise-induced increases in the protein-deficient rats. The mechanism of the exercise effects probably will need to await an experimental design producing larger effects of both undernutrition and exercise on immune functions. It is unlikely that differences in protein and energy status in exercised versus non-exercised undernourished mice were responsible for immune function differences since weight loss, food intake, serum albumin level and carcass lipid content did not differ between these two groups of mice. Exercise significantly decreased serum corticosterone level in undernourished mice, suggesting an alteration in stress response. High levels of glucocorticoids are believed to be a causal agent in the immunodepression observed in many types of stress (8) and have been suggested as a principle agent in malnutritioninduced immune impairments (29). Also, high levels of glucocorticoids can induce intestinal ulceration and bleeding (30). It is possible that the lower serum corticosterone levels in undernourished, exercised mice contributed to the increases in some of their immune measures and prevented the intestinal bleeding. However, the lack of correlation between serum corticosterone level and any immune system parameter in the present study, and similar independence of corticosterone level and immune functions found in other models of undernutrition (Filteau, Kaido, O'Grady, Menzies and Hall, submitted for publication) suggest that serum levels of this hormone are not a major determinant of immune function in undernutrition. In addition, this lack of correlation undermines the idea that high serum corticosterone levels in well-nourished mice in the present work depressed immune functions and obscured immune differences between them and the

Vol.

51, NO. 8, 1992

Exercise and Immunity

in Undernutrition

573

undernourished mice. It might be instructive to look at bound versus free serum corticosterone or at lymphocyte glucocorticoid receptors and responsiveness in such conditions. Alterations in brain monoamine levels and turnover are also associated with the response to stress (8). Similarly to the present results, others have found increased hypothalamic norepinephrine but unchanged serotonin and dopamine in mice allowed voluntary exerise for 12 weeks (27). Endurance-trained rats had slightly higher hypothalamic levels of serotonin and 5-HIAA than sedentary rats but the response of both groups to the stress of an exhaustive exercise bout was similar (26). Stimulation of brain indoleamine metabolism is a frequent finding in studies of exercise (31) but was not apparent in the present study. Whereas the effects of early undernutrition on brain development in rodents has been studied, there exists little information on effects of undernutrition on brain monoamines of adult animals. Anorexia nervosa patients have been reported to have altered catecholamine metabolism (32,33) as do rats undernourished for several generations (34), but such changes were not seen in the present study. Brain uric acid levels increase markedly in certain pathological conditions, including brain ischemia (35) and cytomegalovirus infection (36) but we are unaware of previous reports concerning the effects of either undernutrition or exercise on brain uric acid levels. The biological significance of increased levels of uric acid in the dietary restricted groups as compared to nonrestricted groups is not clear, nor is the significance of the difference between the restricted-exercise groups. It is possible that these differences may reflect changes in adenosine metabolism. Body temperature is reported to affect certain immune functions (37,38) and lowering culture temperature can depress antibody responses in vitro (39). Since both undernutrition and exercise have well-known effects on body temperature, rectal temperature was measured at weekly intervals in mice before and after exercise. The large increases after exercise in the undernourished mice are consistent with the increases in certain immune functions in these animals. However, further research is necessary to determine any causal role of body temperature changes in the immune changes of undernutrition and exercise. In particular, different results might have been obtained if animals had been exercised on either the day of immunization with SRBC or the day of sacrifice. In conclusion, the present results suggest that further research into possible beneficial effects of exercise on immune functions in undemutrition is warranted. Such research may provide additional insight into fundamental aspects of neuroendocrine regulation of immune function and also into the mechanism of the maintenance of immune function and infectious disease resistance in anorexia nervosa patients. Such information could perhaps be applied advantageously to other conditions of stress and immunodepression, such as human immunodeficiency virus infection, in which beneficial effects of exercise have recently been documented (14). Acknowledgments The authors wish to thank Tiffany Omodio, Janelle Oliver, Sherry Rier-Bridges, and Julie Anderson for their excellent technical assistance and Cecilia Figueredo for manuscript preparation. This research was supported by NIH grant MH45646, a Medical Research Council of Canada postdoctoral fellowship awarded to SMF and The Cooper Institute for Advanced Studies in Medicine and the Humanities. Animals were housed at the James A. Haley Veterans Hospital, Tampa, FL. References 1. Y.G. LIU and S.Y. WANG, Immunol. Letters 1__4 117-120 (1987). 2. B.A. PETERS, M. SOTHMANN and W.B. WEHRENBERG, Life Sci. 45 2238-2245 (1989). 3. L. HOFFMAN-GOETZ, R. KEIR, R. THORNE and M.E. HOUSTON, Clin. Exp. Immunol. 66 551-557 (1986). 4. M.A. PAHLAVANI, T.H. CHEUNG, J.A. CHESKY and A. RICHARDSON, J. Appl. Physiol. 64 1997-2001 (1988). 5. J. RANDALL-SIMPSON and L. HOFFMAN-GOETZ, Canadian Federation of Biological Societies, 32rd meeting, p. 83, abstract, (1990). 6. L. HOFFMAN-GOETZ, R.J. THORNE and M.E. HOUSTON, Can. J. Phsyiol. Pharmacol. 66 1405-1419 (1988). 7. L. FITZGERALD, Immunol. Today 9 336-339 (1988).

574

Exercise

and Immunity

in Undernutrition

Vol.

51, No. 8, 1992

8. D.N. KHANSARI, A. J. MURGO and R. E. FAITH, Immunol. Today 1_!_1 170-175 (1990). 9. A. FERRY, B.L. WEILL, and M. RIEV, J. Appl. Physiol. 69 1912-1915 (1990). 10. M.A. FIATARONE, J.E. MORLEY, E.T. BLOOM, D. BENTON, T. MAKINODAN and G.F. SOLOMON, J. Lab. Clin. Med. 112 543-552 (1988). 11. W.P. MORGAN, Med. Sci. Sports Exercise 17 94-100 (1985). 12. V.E. BARACOS, Can. J. Physiol. Pharmacol. 67 864-870 (1989). 13. D.M. CRIST, L.T. MACKINNON, R.F. THOMPSON, H.A. ATTERBOM and P.A. EGAN, Gerontology 34 66-71 (1989). 14. A. LAPERRIERE, M. H. ANTONI, N. SCHNEIDERMAN, G. IRONSON, N. KLIMAS, P. CARALIS, and M.A. FLETCHER, Biofeedback and Self-Regulation (in press). 15. A.M. TOMKINS, Proc. Nutr. Soc. 4___~S279-294 (1986). 16. C.A. ARMSTRONG-ESTHER, A.H. CRISP, J.H. LACEY and T.N. BRYANT, Postgrad. Med. J. 54 395-399 (1978). 17. J.A. GOLLA, L.A. LARSON, C.F. ANDERSON, A.R.LUCAS, W.R. WILSON and T.B. TOMASI, JR., Am. J. Clin. Nutr. 3.3 2656-2762 (1981). 18. M.J. PERTSCHUK, L.O. CROSBY, L. BAROT and J.L. MULLEN, Am. J. Clin. Nutr. 35 968-972 (1982). 19. P.S. DOWD, J. KELLEHER, B.E. WALKER and P.J. GUILLOU, Clin. Nutr. 2 79-83 (1983). 20. W. NELSON, J. Nutr. 118 284-289 (1988). 21. E.G. BLIGH and W.J. DYER, Can. J. Biochem. Physiol. 3__2_7911-917 (1959). 22. SAS INSTITUTE INC. SAS User's Guide: Statistics 5th ed. SAS Institute, Cary, NC (1985). 23. R.L. GROSS and P.M. NEWBERNE, Physiol. Rev. 6(___t)188-301 (1980). 24. C. GOMEZ-SANCHEZ, B.A. MURRY, D.C. KERN and N.M. KAPLAN, Endocrinol. 96 796-798 (1975). 24. V. RILEY, Science 212 1100-1109 (1981). 25. E. BLOMSTRAND, D. PERRETT, M. PARRY-BILLINGS and E.A. NEWSHOLME, Acta Physiol. Scand. 135 473-481 (1989). 27. T. SAMORAJSKI, C. ROLSTEN, A. PRZYKORSKA and C.M. DAVIS, Exp. Gerontol. 22 411-420 (1987). 28. N. UEDA, J. KAYASHITA, S. MORIGUCHI and Y. KISHINO, Nutr. Res. 1_._0.0428-437 (1990). 29. R.R. WATSON, Nutrition, Disease Resistance and Immune Function, R.R. Watson (ed), 273-283, Marcel Dekker, New York and Basel (1984). 30. British Medical Association, British National Formulary, #15, p. 243, British Medical Association, London (1988). 31. F. CHAOULOFF, Acta Physiol. Stand. 137 1-13 (1989). 32. T. SATO, N. IGARASHI, K. MIYAGAWA, T. NAKAJIMA and K. KATAYAMA, Endocrinol. Japan. 35 295-301 (1988). 33 W.H. KAYE, D.C. JIMERSON, C.R. LAKE and M.H. EBERT, Psych. Res. 14. 333-342 (1985). 34. C. ANNAMMA and T. DESIRAJU, Biogenic Amines 5 323-337 (1988). 35. H. KANEMITSU, A. TAMURA, T. KIRINO, H. OKA, K. SANO, T. IWAMOTO, M. YOSHIURA and K. IRIYAMA, Brain Res. 499 367-370 (1989). 36. B.E. BOYES, D.G. WALKER, E.G. McGEER and J.R. O'KUSKY, J. Neurochem. 53 1719-1723 (1989). 37. A.J. NEVILLE and D.N. SAUDER, Lymphokine Res. 7 201-206 (1988). 38. M . M . PARK, N. B. HORNBACK, S. ENDRES AND C. A. DINARELLO, Lymphokine Res. 9 213-223 (1990). 39. M.C. WANG-YANG, T.M. BUTTKE, N.W. MILLER and L.W. CLEM, Cell. Immunol. 126 354-366 (1990).