The aim of this study was to elucidate the mechanism responsible for lymphopenia after exercise. Seven young healthy men volunteered for this study. Peripheral blood mononuclear cells (PBMC) were cultured with cortisol and analyzed for C-X-C motif chemokine receptor 4 (CXCR4) expression by flow cytometry. To determine the effects of exercise, subjects performed exhaustive cycling exercise. PBMC were cultured with plasma obtained before and after the cycling exercise. Alternatively, PBMC obtained before and after exercise were cultured without plasma or glucocorticoid to examine whether PBMC were primed in vivo for CXCR4 expression. We analyzed cortisol- or plasma-treated PBMC to determine their ability to migrate through membrane filters in response to stromal cell-derived factor 1α/CXCL12. Cortisol dose- and time-dependently augmented CXCR4 expression on T lymphocytes, with <6 h of treatment sufficient to augment CXCR4 on T lymphocytes. Postexercise plasma also augmented CXCR4 expression. Cortisol or postexercise plasma treatment markedly enhanced migration of T lymphocytes toward CXCL12. Augmentation of CXCR4 on T lymphocytes by cortisol or plasma was effectively blocked by the glucocorticoid receptor antagonist RU-486. Thus exercise-elicited endogenous cortisol effectively augments CXCR4 expression on T lymphocytes, which may account for lymphopenia after exercise.
- chemokine receptor
- T cell distribution
stress is thought to suppress the immune system and increase susceptibility to infections and cancer on one hand, whereas it is also thought to exacerbate inflammatory diseases like asthma and arthritis on the other. Glucocorticoid has been suggested to mediate the bidirectional modification of the immune reaction under stress, and modified lymphocyte trafficking by glucocorticoid seems to be one of the major causes of modified immune reaction. It is unknown, however, how glucocorticoid mobilizes lymphocytes in vivo.
Physical exercise has been accepted as a subset of physical stress, such as thermal and traumatic injury, surgery, acute myocardial infarction, or hemorrhagic shock. Because exercise is mostly harmless and controllable, it serves as a good experimental model for understanding the pathogenesis of stress-induced immune changes (22).
Exercise is well known to modify immune cell distribution (10, 23, 34). The number of circulating leukocytes changes during and after exercise (34). After a transient increase during physical exercise of relatively high intensity, the number of circulating T lymphocytes often declines below the initial level after exercise (10, 15, 34). Interestingly, although subjects whose plasma cortisol levels were elevated in response to exercise exhibited lymphopenia lasting for several hours after exercise, subjects whose plasma cortisol remained unchanged after the exercise did not exhibit lymphopenia after exercise (34). Administration of the synthetic glucocorticoid dexamethasone also induces transient lymphopenia (5). Thus glucocorticoid seems to play a major role in modified lymphocyte trafficking after exercise.
Chemokines, a family of 8- to 10-kDa chemotactic cytokines, provide signals that guide leukocytes to their appropriate location in the body through transmembrane receptors coupled to heterotrimeric GTP-binding proteins under inflammatory or physiological conditions (2, 31). Among various chemokine receptors, C-X-C motif chemokine receptor 4 (CXCR4) on CD4 T lymphocytes can be augmented by a synthetic glucocorticoid, dexamethasone (40). Although most chemokine receptors bind more than one ligand (3), CXCR4 is a unique receptor for stromal cell-derived factor 1α (SDF-1α)/CXCL12, which is constitutively expressed in lymph nodes, lung, liver, and bone marrow (26, 30, 36, 41). The interaction between CXCR4 and CXCL12 exhibits a highly potent chemoattraction in vivo (26, 30, 36, 41). Interestingly, the distinct pattern of breast cancer metastasis involving lymph nodes, lung, liver, and bone marrow strikingly coincides with the preferential expression of CXCL12 mRNA in these tissues (28). Stimulation of breast cancer cells expressing CXCR4 was shown to induce actin polymerization and pseudopodia formation (28). In vivo, neutralizing the interactions of CXCL12/CXCR4 significantly impairs metastasis of breast cancer cells to regional lymph nodes and lung (28). Administration of the human CXCR4 antagonist AMD-3100 resulted in marked increase of white blood cell counts in the bloodstream of human subjects (21). Thus the interaction between CXCR4 and CXCL12 may serve as a major determinant of lymphocyte distribution and breast cancer cells. Even though cytokines such as interleukin (IL)-2, IL-4, and transforming growth factor (TGF)-β also modify expression of cell surface CXCR4 (1, 24, 32, 40), the fact that synthetic glucocorticoid augments CXCR4 on CD4+ T lymphocytes led us to the idea that, under noninflammatory conditions, endogenous glucocorticoid, which could be induced above the level of circadian fluctuation in response to various stressors, including physical exercise, may have a potential role in altering the distribution of lymphocytes via augmentation of CXCR4.
The aim of this study, therefore, was to examine whether endogenous glucocorticoid cortisol, elicited by exercise, may be involved in postexercise lymphopenia via augmentation of CXCR4 on circulating T lymphocytes.
MATERIALS AND METHODS
Chemokine and antibodies.
Recombinant human SDF-1α was purchased from R & D (Minneapolis, MN). Anti-human CD8 (SK1, IgG1) was purchased from BD (San Jose, CA). Anti-human CD3 (UCHT1, IgG1), CD4 (MT310, IgG1), CXCR4 (12G5, IgG2a), and isotype control (mouse IgG2a) antibodies were purchased from DAKO JAPAN (Kyoto, Japan). Anti-human CCR7.6B3 antibody was a generous gift from Dr. Hitoshi Hasegawa (Ehime University School of Medicine; see Ref. 20).
Seven healthy young men (age: 23.7 ± 0.8 yr) volunteered for the study. General characteristics of the subjects are presented in Table 1. Subjects were informed about the experimental procedures, potential risks, and discomfort of participating and gave signed informed consent. All of the experimental procedures were carried out under the approval of the ethical committee of Tohoku University.
Subjects refrained from any type of moderate or heavy exercise for at least 12 h before blood sampling. No food or drink other than water was allowed during 10 h before blood sampling. Blood samples were drawn from an antecubital vein between 8:00 and 10:00 AM with a sterile syringe containing 100 IU heparin (Novo Nordisk, Copenhagen, Denmark)/10 ml blood for either mononuclear cell preparation or plasma separation.
Peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood using a lymphoprep tube (NYCOMED, Oslo, Norway). PBMC were washed two times with PBS (Nissui, Tokyo, Japan) and resuspended in RPMI 1640 (Sigma, St. Louis, MO) containing 2 mM glutamine, 10% heat-inactivated FCS (Sigma), 10 mM HEPES (Wako, Tokyo, Japan), 100 U/ml penicillin (Sigma), and 100 μg/ml streptomycin (culture medium; Sigma).
Glucocorticoid treatment of PBMC.
PBMC were incubated in six-well flat-bottom culture plates in a culture medium with a cell density ranging from 2 × 106 to 4 × 106 cells/well, with or without cortisol (Sigma) for 24, 48, or 72 h at 37°C in a 5% CO2 humidified incubator. The concentration of cortisol in the culture ranged from 10−6 to 10−10 M. RU-486 (Sigma), a potent glucocorticoid receptor antagonist, was used at 10−6 M, as previously described (8). Considering the transient increase in plasma glucocorticoid level in vivo, we examined a shorter duration of glucocorticoid treatment in vitro. PBMC were incubated with or without 10−6 M cortisol in the same conditions as above except that they were washed with PBS for a total of three times at the 1st, the 3rd, and the 6th h of treatment, resuspended in fresh medium, and further incubated in a six-well plate up to a total incubation time of 72 h. We also examined the effect of 10−6 M RU-486 blockade.
Evaluation of maximal oxygen uptake.
To determine the exercise intensity of an acute exercise load, we evaluated the peak exercise capacity by measuring oxygen uptake during stepwise incremental exercise on a bicycle ergometer. Subjects were told to pedal at a constant rate of 60 ± 5 rpm. The initial workload was set at 60 W. The workload was increased thereafter at 15 W/min stepwise until either the subject could not maintain the required pedaling rate (60 rpm) or the subject reached exhaustion. Heart rate and rating of perceived exertion were monitored throughout the exercise. We measured oxygen uptake every 30 s during the exercise test using an AeroMonitor 280 (Minato Medical Science, Osaka, Japan).
Acute exercise load and plasma sampling schedule.
Subjects performed bicycle exercise at the intensity of 70% of each subject's maximal oxygen uptake for 90 min. Subjects started exercise after 8:00 AM and before 9:30 AM. Peripheral blood samples were obtained immediately before and after the exercise. Heparinized peripheral blood samples were centrifuged for 10 min at 3,000 rpm at 4°C. Separated plasma was stored in aliquots at −80°C until cortisol was measured or samples were used in cell cultures.
Measurement of plasma cortisol.
Plasma cortisol concentration was measured by RIA (Immunetech, Marseille, France), according to the manufacturer's instructions.
Plasma treatment of PBMC.
To examine the effect of endogenous glucocorticoid induced by acute exercise on PBMC, we incubated PBMC in a 24-well flat-bottom plate in culture medium at a cell density ranging from 1 × 106 to 2 × 106 cells/well, with plasma obtained either before or after the exercise. The cells were incubated with pre- or postexercise plasma at 25% of culture volume for 72 h at 37°C in a 5% CO2 humidified incubator.
In vivo priming effect of exercise on CXCR4 expression on PBMC.
Blood samples were collected before and after exercise. PBMC were incubated with culture medium, and cultured cells were stained and analyzed by four-color flow cytometry.
Flow cytometry analysis.
Cultured PBMC were washed three times with PBS before staining procedures. For blocking of IgG Fc receptors, freshly isolated or cultured PBMC were incubated with 4 μl mouse serum (Sigma) per 4 × 105 cells for 30 min at 4°C. Cells were then washed two times in ice-cold PBS and stained with the following combination of antibodies: allophycocyanin (APC)-labeled anti-CD3 monoclonal antibody (MAb), peridinin chlorophyll protein (PerCP)-labeled anti-CD8 MAb, and phycoerythrin (PE)-labeled anti-CXCR4 MAb. After 30 min of incubation with MAb at 4°C, PBMC were washed two times in ice-cold PBS and analyzed by FACSCalibur (BD). Because multiparameter staining of circulating T lymphocytes used in this study showed that the vast majority of CD8− cells were indeed CD4+ (data not shown), CD8 MAb alone could effectively identify CD4+ and CD8+ subsets. Therefore, we considered the CD8− subset to be equivalent to the CD4+ subset in these experiments. We evaluated the level of chemokine receptor expression on lymphocytes using mean fluorescence intensity (MFI).
We assessed the migration of PBMC in a 24-well trans-well culture insert with a 3-μm pore size (BD). The lower well was filled with culture medium with or without a supplement of human SDF-1α. PBMC (5 × 105) suspended in culture medium were added to the upper wells and incubated at 37°C in a 5% CO2 incubator for 120 min. After incubation, the inserted membrane was removed, and migrated cells were stained with PE-conjugated anti-CD4 MAb, PerCP-conjugated anti-CD8 MAb, and APC-conjugated anti-CD3 MAb. CD4+ and CD8+ T lymphocytes were then analyzed by FACSCalibur (BD).
We examined the statistical difference with one-way ANOVA, repeated-measures ANOVA, simple linear regression analysis, or Student's t-test, as appropriate. All statistical analyses were performed with StatView software (SAS Institute, Cary, NC). Mean values are described along with SE. P < 0.05 was regarded as statistically significant.
Cortisol-augmented CXCR4 expression on CD4, CD8 T lymphocytes.
CXCR4 expression on CD4 and CD8 T lymphocytes when incubated with 10−6 M cortisol was augmented in a time-dependent manner, peaking at 72 h (P < 0.01; Fig. 1). The augmentation of CXCR4 was detectable as early as 24 h after the start of incubation. The level of CXCR4 augmentation by cortisol for CD4 and CD8 T lymphocytes was 5.98-fold (ranging from 4.15 to 7.60) and 4.66-fold (ranging from 3.42 to 6.31) of the control MFI, respectively.
Dose of cortisol comparable to physiological concentration was able to augment CXCR4 expression on CD4 and CD8 T lymphocytes.
CXCR4 expression on CD4 and CD8 T lymphocytes was augmented by cortisol in a dose-dependent manner. Even a dose of cortisol as low as 1 × 10−7 M, which is comparable to the lower limit of physiological plasma concentration, was able to augment CXCR4 expression on CD4 and CD8 T lymphocytes (P < 0.01). Glucocorticoid receptor antagonist RU-486 could reverse CXCR4 augmentation, indicating that a classical glucocorticoid receptor pathway is necessary to augment CXCR4 (Fig. 2).
Short-term exposure to cortisol was sufficient to augment CXCR4 expression on CD4 and CD8 T lymphocytes.
It is well documented that plasma cortisol elevation after acute stress is transient, lasting for no more than several hours. Thus we examined whether shorter exposure to cortisol in vitro could induce CXCR4 on CD4 and CD8 T lymphocytes. Less than 6 h of cortisol treatment was sufficient to induce CXCR4 after 72 h of incubation on CD4 (P < 0.05) and CD8 (P < 0.01) T lymphocytes, which again was inhibited by RU-486 (Fig. 3).
Change in plasma cortisol concentration after exercise.
Plasma cortisol was elevated in six out of seven subjects (Table 2) after exercise. The average increase of plasma cortisol in the responders was 38%, ranging from 6 to 98%. The nonresponder subject was left out of the plasma cortisol stimulation studies.
Cortisol in postexercise plasma is responsible for CXCR4 induction on CD4 and CD8 T lymphocytes.
To examine whether exercise-induced cortisol in plasma could augment CXCR4 expression, we cocultured PBMC for 72 h with plasma obtained before and immediately after 90 min of exercise. Incubation with postexercise plasma from the subjects with increased cortisol after exercise significantly augmented CXCR4 expression on CD4 and CD8 T lymphocytes by 1.57-fold and 1.50-fold, respectively (P < 0.05), when compared with CXCR4 expression on PBMC incubated with control plasma obtained before exercise. RU-486 effectively blocked the postexercise plasma augmentation of CXCR4 by 52.1% (CD4) and 55.2% (CD8; Fig. 4). We found a high correlation between the percent increase in plasma cortisol concentration postexercise and the magnitude of CXCR4 augmentation on CD4 T lymphocytes (r = 0.93, P < 0.01). Interestingly, there was no correlation between the percent increase in plasma cortisol concentration postexercise and the magnitude of CXCR4 augmentation on CD8 T lymphocytes, which express less CXCR4 compared with CD4 T lymphocytes (r = 0.53, P = 0.22; Fig. 5).
PBMC obtained postexercise are primed for CXCR4 expression.
We tested whether PBMC obtained postexercise, which have definitely been exposed to elevated cortisol, are primed for CXCR4 expression. We simply incubated PBMC obtained at various time points after exercise for 72 h and analyzed them for the expression of CXCR4. CXCR4 expression on CD4 and CD8 T lymphocytes obtained after exercise was augmented compared with cells obtained before exercise (P < 0.05; Fig. 6).
Cortisol and postexercise plasma treatment enhanced migration of T lymphocytes toward CXCL12.
To determine whether cortisol-induced CXCR4 expression was functional, we measured the migration activity of cortisol- or plasma-treated lymphocytes. After 72 h of incubation in the presence of cortisol, CD4 (P < 0.01) and CD8 (P < 0.05) T lymphocytes migrated in a dose-dependent fashion in response to CXCL12 (Fig. 7). The migration activity of postexercise plasma-treated PBMC in response to CXCL12 was apparently larger than that of PBMC treated with control plasma (P < 0.05;Fig. 8).
In this study, we have clearly shown for the first time that cortisol is as potent as dexamethasone in augmenting CXCR4 expression in vitro at a range of doses comparable to physiological concentration. Our observation strongly suggests that endogenous glucocorticoid induced by exercise or other stressors may modify lymphocyte trafficking by upregulating the level of chemokine receptor CXCR4 on lymphocytes. Regulation of CXCR4 level by endogenous glucocorticoid may therefore be the potential mechanism responsible for postexercise lymphopenia and for dexamethasone-induced lymphopenia. Because elevation of plasma cortisol is not solely induced by exercise but also by various stressors, our observation may be one of the common mechanisms by which stressors modify immune responses.
Modified lymphocyte trafficking dependent on CXCR4 expression may lead to an altered immune response, such as enhancement in delayed-type hypersensitivity (DTH), as shown by Dhabhar and McEwen (11–13). They showed an extensive relocation of lymphocytes to skin in rats after restraint stress, which consequently led to enhanced DTH to keyhole limpet hemocyanin antigen. In their experiment, deployment of lymphocytes to skin and enhanced DTH was abrogated by adrenalectomy, showing the involvement of corticoid hormone. Although skin is not a tissue abundant in CXCL12 transcripts, it distinctly expresses CXCL12 (28, 37), the unique ligand for CXCR4. Therefore, it is possible that the immune response in tissues abundant in CXCL12 may be affected by acute stress as well as in the skin tissue.
Augmentation of CXCR4 reached a plateau after 72 h of incubation with cortisol in vitro. Increased plasma cortisol levels resulting from hypothalamo-pituitary-adrenal axis (HPA axis) activation in response to acute exercise, however, is reported to last no longer than several hours (16, 17, 34). The apparent time of exposure to an elevated cortisol level for circulating lymphocytes can hardly be beyond several hours under physiological conditions. We could, accordingly, detect augmentation of CXCR4 on T lymphocytes by cortisol as soon as 1 h after exposure to a physiological concentration of cortisol, with the level becoming statistically significant at 6 h of exposure. Thus it is possible that cortisol elicited in response to exercise may be sufficient to augment CXCR4 in vivo.
The augmentation of CXCR4 on postexercise plasma-stimulated T lymphocytes was less than twofold and seemed not as big as that in cortisol-treated lymphocytes. Considering the expected concentration of cortisol ranging from 10−7 to 10−8 M in the culture medium as calculated from plasma cortisol concentration, however, the result was in line with the dose response of CXCR4 augmentation by cortisol.
Although the exercise protocol was rather severe with the exercise intensity of 70% maximal oxygen uptake and the duration of 90 min, and plasma cortisol was elevated in six out of seven subjects after the exercise, the level of increase was <30% in four subjects (subjects C-F). All four subjects and the subject who exhibited decreased cortisol level after exercise belonged to a triathlon team and were undergoing regular training. They could well be accustomed to the type of exercise in this study, resulting in the least changes of plasma cortisol in response to exercise. Untrained subjects may have exhibited more enhanced elevation in cortisol, with milder condition. However, even though plasma cortisol elevation does not necessarily depend on exercise intensity, with lower-intensity or shorter-duration exercise a rise in plasma cortisol level is not likely. Thus our observation may be limited to the type, intensity, and duration of exercise in which plasma cortisol is elevated. Interestingly, there was a high correlation between the percent increase in plasma cortisol concentration postexercise and the magnitude of CXCR4 augmentation on CD4 T lymphocytes (CD4: r = 0.93, P < 0.01; CD8: r = 0.53, P = 0.22). Expression of CXCR4 on circulating CD4 T lymphocytes may therefore be sensitive to changes in plasma concentration of cortisol. Alteration in the level of CXCR4 expression may directly affect the distribution of CD4 T lymphocytes in the body.
Catecholamines elicited during exercise or psychological stress may potentially modify leukocyte distribution as well as glucocorticoids. Human lymphocytes, in particular T lymphocytes, carry β-adrenergic receptors (25, 33, 39). Administration of epinephrine and norepinephrine in vivo transiently increases circulating T lymphocyte numbers (9, 14). Furthermore, treatment with β-adrenoceptor antagonists inhibits the increase in circulating lymphocytes induced by exercise or psychological stress (4, 29). In a rat study, low-dose epinephrine administration significantly enhanced skin DTH response and produced a significant increase in the number of T lymphocytes in the lymph nodes draining the site of DTH reaction (13). In our preliminary experiment, catecholamines in their physiological concentration failed to augment CXCR4 expression on T lymphocytes in vitro. The mechanism by which catecholamines relocate blood leukocytes must be further examined.
Although RU-486 effectively blocked the postexercise plasma augmentation of CXCR4 expression, its blockade by RU-486 was incomplete. Cytokines such as IL-2, IL-4, and TGF-β, known to be potent inducers of CXCR4, are also candidates to be investigated for the involvement in postexercise augmentation of CXCR4 (7, 24, 32). Interferon (IFN)-γ, on the other hand, has been shown to downregulate CXCR4 expression (1, 31, 32). Various types of cytokines, such as IL-6 or tumor necrosis factor-α, are known to be present in the bloodstream after exercise (6, 38). IL-2, IL-4, TGF-β, and IFN-γ, however, are least likely to be induced during or after acute exercise (18, 19, 27, 35). Thus it is reasonable to suppose that these cytokines may not affect CXCR4 expression.
CXCL12, BLC/CXCL13, ELC/CCL19, and SLC/CCL21 are called homeostatic chemokines, and they are important for the regulation of the homing and traffic of T lymphocytes in the secondary lymphoid tissues. The corresponding receptors are CXCR4, CXCR5, and CCR7 for the two latter chemokines, respectively. In our preliminary experiments, cortisol at a physiological concentration failed to augment CXCR5 and CCR7 expression on T lymphocytes in vitro. Thus T lymphocyte distribution may be controlled with cortisol via the regulation of CXCR4 expression under stressed conditions.
Interestingly, CXCL12 was shown to reduce CXCR4 expression on T cells (1, 31, 32). Downregulation of CXCR4 by its unique ligand may serve as a negative feedback mechanism, ensuring the termination of T cell accumulation. If the expression of CXCL12 in any organs is considerably modified by exercise or stress, or if the circulating level of CXCL12 changes after exercise, it may largely affect the distribution of T cells. Because we lack methods to quantify the tissue level of CXCL12 in healthy humans, animal study is necessary to further understand the regulation of lymphocyte tracking.
It must also be noted that circulating T lymphocytes express a low level of CXCR4 without HPA axis activation. Considering the fact that enhanced CXCR4 expression leads to enhanced migration, there may be a certain equilibrium between the circulating pool and the lodged pool of T lymphocytes expressing CXCR4. Thus we suggest that investigation of physiological roles and regulation of leukocyte trafficking is one of the major keys to understanding the influence of stress on the immune response.
In this study, RU-486 independently augmented CXCR4 expression. RU-486 is known as a potent competitive antagonist of steroid hormones for their receptors, but it may also have partial agonistic action on glucocorticoid receptors.
In conclusion, the present study demonstrates that endogenous cortisol augments CXCR4 expression on T lymphocytes, which may account for the transient mobilization of T lymphocytes under stress such as physical exercise.
We are grateful to Dr. Hitoshi Hasegawa (Ehime University School of Medicine), who kindly provided the anti-human CCR7.6B3 antibody.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society