INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

STRENUOUS WHOLE BODY physical exercise at >70% of maximal oxygen uptake (29) and in the forms of long-distance running (11), treadmill (32), and cycle ergometry (46) has been shown to increase the level of circulating proinflammatory cytokines. More specifically, interleukin-6 (IL-6) has been consistently reported to increase (5, 9, 29, 31, 46), whereas the response of interleukin-1 (IL-1) is more variable, with some studies showing an increase (5, 29) and others reporting no change (9, 46). Both cytokines, especially IL-6, are powerful stimulants of the hypothalamic-pituitary-adrenal (HPA) axis in humans (8, 25, 26). The HPA axis and the sympathetic system are the peripheral limbs of the stress system of which the function is to maintain basal and stress-related homeostasis (8).

To our knowledge, the response of proinflammatory cytokines to strenuous exercise restricted to a specific muscle group has not been studied as yet. In particular, the respiratory muscles are a specific muscle group of which the function is pivotal for life (47). Inspiratory resistive breathing (IRB) represents a form of "exercise" for these muscles, which is clinically relevant because it is encountered in many disease states such as asthma and chronic obstructive pulmonary disease. When strenuous enough, IRB produces diaphragmatic fatigue (19, 37) and delayed diaphragmatic structural injury (50), phenomena that are not observed when the inspiratory resistance is moderate (19). Strenuous IRB also produces -endorphin through a largely unknown mechanism (38, 48). -Endorphin induces rapid shallow breathing, which is considered a protective strategy to prevent and/or postpone task failure, because the reduced tidal volume requires less pressure development by the respiratory muscles (38). However, the source of -endorphin remains elusive, and both central sites such as the HPA axis and peripheral sites such as the spinal cord and peripheral nerves have been implicated (33).

We hypothesized that strenuous IRB that could potentially lead to inspiratory muscle fatigue and task failure would cause increased production of the proinflammatory cytokines IL-1 and IL-6 and would lead to the production of -endorphin through stimulation of the HPA axis. This prospective study was done to test this hypothesis in normal human volunteers. The plasma ACTH, which is a proopiomelanocortin-derived peptide concurrently secreted with -endorphin from the HPA axis (13, 14), was used as a marker of the HPA axis activation (8, 25, 26).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Eleven healthy human volunteers (9 male, 2 female) with a mean age of 34 ± 5 yr (range 28-42), weighing 78 ± 13 kg and measuring 175 ± 9.5 cm in height, were studied. None of the subjects participated in competitive sporting activities or had febrile illness during the month before testing. The study protocol was approved by our institutional ethics committee, and all participants gave informed consent. All subjects refrained from exercising or any other strenuous activity for 24 h before testing. Testing was always performed at ~9 AM.

Protocol and measurements. Pulmonary function tests (PFT) and determination of the maximal static inspiratory pressure were performed in each subject on a separate day before the days of testing. Mouth pressure (P_m) was measured by a differential pressure transducer (Validyne, Northridge, CA) connected to a mouthpiece and was recorded on a polygraph recorder (Gould ES 1000, Gould Instruments, Cleveland, OH). Maximum inspiratory pressure (P_{m max}) was measured as the most negative mouth pressure sustained for at least 1 s during a maximum inspiratory effort from functional residual capacity against an occluded airway. Patients were in the sitting position, and each maneuver was repeated several times separated by 1 min until three reproducible measurements were recorded. The highest measured value was used for analysis. A small hole (1.5 cm diameter) in the mouthpiece prevented closure of the glottis. During the same session the subject was also accustomed to the environment and apparatus used for the protocol.

Resistive breathing runs. Each subject performed two resistive breathing runs at 35% and 75% of P_{m max} separated by at least 3 wk. The high load (75%) run was done first and was followed by a moderate load (35%) run of equal duration on the second occasion. These loads were chosen based on the fact that the high load was proven capable of producing inspiratory muscle fatigue and thus would represent strenuous IRB, whereas the moderate load has been shown to be sustainable indefinitely (12, 24, 36).

High-load run. An intravenous catheter for blood sampling was placed in one forearm vein, and a baseline sample was obtained (point R). Afterwards, the P_{m max} was measured as described and displayed to the subject on an oscilloscope (Tektronik 2213, Beaverton, OR). The subjects were then instructed to breathe through an inspiratory resistive load while maintaining 75% of the P_{m max}. The resistive load consisted of an alinear resistance adjusted at the start of each experiment to help each subject reach the required P_m. The resistance device was a tube with an adjustable orifice to alter inspiratory resistance. The expiratory line was not loaded. The subjects were instructed to maintain a constant P_m throughout inspiration as reflected by a square-wave pattern on the oscilloscope. Otherwise, they were allowed to choose their own breathing pattern with no special instructions as to how to achieve the target P_m. The subjects were encouraged to maintain the target P_m and to endure the test to their limit. When the subjects were unable to generate the target P_m for five consecutive inspiratory efforts, we assumed that inspiratory muscle fatigue might have been produced. At this time a second blood sample was drawn (point F). The patients continued to breathe through the resistance as hard as possible for another 15 min, at which point the resistive breathing run ended and a final blood sample was collected (point E). The time interval from the beginning of the run to points F and E was recorded for each subject.

Moderate-load run. Each subject performed a second resistive breathing run at least 3 wk after the first, at 35% of P_{m max}, in which the procedure of the first run was repeated. Because all subjects were able to sustain this load, the time course of the initial high-load run was used to set the time of blood sampling and the total duration of resistive breathing, i.e., the second and third blood samples (points F and E) were taken at the same time interval from the beginning of the run as during the first high-load run. Thus the total duration of both resistive breathing runs was the same.

Blood samples. Blood was drawn into sterile syringes and transferred to precooled sterile EDTA tubes. Samples were immediately spun in a refrigerated centrifuge to separate plasma from cells and thus avoid ex vivo cytokine secretion (7) and were next placed in polysterene tubes and stored at 70 °C until assayed.

Assays. Plasma levels of IL-1 and IL-6 were measured with commercially available ELISA kits (Quantikine, R&D Systems, Minneapolis, MN). All assays were performed in duplicate, and the intra- and interassay coefficients of variation were <10% in all cases.

The plasma concentrations of ACTH and -endorphin were measured in duplicate by RIA using commercial kits (Diagnostic Product, Los Angeles, CA, and Nichols Institute, San Juan Capistrano, CA, respectively). The intra- and interassay coefficients of variation were <10%.

Statistics. Values are expressed as means ± SE. The data were analyzed by two-way ANOVA followed by least squares difference test for post hoc comparisons. To evaluate correlations, the Pearson product moment test was performed. A P value <0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All subjects had PFTs within normal limits. During the high-load run, the average time required to reach points F and E was 39 ± 9 and 54 ± 10 min, respectively. The moderate-load run was of equal duration by study design.

High-load run. At rest, P_{m max} averaged 140 ± 7 cmH₂O (range 96-172) and fell to 95 ± 4 cmH₂O at the end of the run (point E; P < 0.01).

Circulating cytokines. Figure 1, A and B, illustrates the changes in plasma IL-1 and IL-6 concentrations during the strenuous inspiratory resistive loading experiment. Severe IRB significantly increased the plasma concentration of IL-1 (P < 0.01) and IL-6 (P < 0.01). IL-1 increased from 0.83 ± 0.12 pg/ml at rest (point R) to 1.88 ± 0.53 pg/ml at point F (P = 0.07) and to 4.06 ± 1.27 pg/ml at the end of resistive breathing (point E; P < 0.01, point E vs. F). IL-6 increased from 5.30 ± 1.02 pg/ml at rest (point R) to 10.33 ± 2.14 pg/ml at point F (P = 0.02) and had a small but not significant further increase to 11.66 ± 2.29 pg/ml at the end of the resistive breathing (point E).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Mean plasma level of interleukin-1 (IL-1; A), interleukin-6 (IL-6; B), -endorphin (C), and ACTH (D) at rest (point R), at point at which subjects could not generate target maximum inspiratory pressure (point F), and at end of resistive breathing (point E). Vertical bars are ±SE. * Statistically significant difference (P < 0.05) from point R; # statistically significant difference (P < 0.05) from point F; + statistically significant difference (P < 0.01) from the moderate-load run. Closed symbols, high-load run; open symbols, moderate-load run.

ACTH and -endorphin. Strenuous IRB was also associated with increased plasma levels of ACTH (P < 0.01) and -endorphin (P < 0.01) (Fig. 1, C and D). However, the pattern of increase was different from that of cytokines. The plasma ACTH and -endorphin levels were only significantly elevated at point E, whereas their levels at point F were not different from those measured at point R (Fig. 1).

Correlations. The plasma levels of ACTH and -endorphin were strongly and positively correlated both at points F (r = 0.97, P < 0.01) and E (r = 0.95, P < 0.01; Fig. 2), as were their respective increases from point F to E (r = 0.95, P < 0.01; Fig. 3). There was a strong positive correlation between IL-6 level at point F and ACTH and -endorphin levels at point E (r = 0.94 and 0.88 respectively, P < 0.001 in both cases; Fig. 4). The correlation between the plasma levels of IL-6 and ACTH (r = 0.80, P = 0.03) and -endorphin (r = 0.85, P = 0.01) at the end of resistive breathing was also statistically significant but weaker. There was also a strong positive correlation between the change in IL-6 level from rest to point F and the corresponding changes in either ACTH or -endorphin level from point F to E (r = 0.91, P < 0.01 and r = 0.92, P < 0.01, respectively; Fig. 5). No significant correlation was observed between IL-6 at point F and either ACTH or -endorphin level measured at point F or between IL-1 and either ACTH or -endorphin level at any of the test points.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Correlations between plasma level of -endorphin and ACTH at points F and E during high-load run. For definitions of points F and E, see Fig. 1 legend.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Correlation between respective increases in plasma levels of -endorphin and ACTH from point F to E during high-load run. For definitions of points F and E, see Fig. 1 legend.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Correlations between plasma level of -endorphin and ACTH at point E and IL-6 level 15 min before, i.e., at point F during high-load run. For definitions of points F and E, see Fig. 1 legend.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Correlations between increase in IL-6 from point R to F and corresponding increases in -endorphin and ACTH levels between points F and E during high-load run. For definitions of points R, F, and E, see Fig. 1 legend.

Moderate-load run. At rest, P_{m max} averaged 138 ± 8 cmH₂O and remained unchanged throughout the run (P_{m max} = 137 ± 7 cmH₂O at point E).

Resting levels of plasma IL-1, IL-6, ACTH, and -endorphin during moderate IRB were similar to those measured at rest before the severe IRB experiment (Fig. 1). At points F and E, plasma levels of cytokines, ACTH, and -endorphins during the moderate IRB remained similar to those measured at rest (point R; Fig. 1).

Moreover, comparison between the two resistive loading experiments at point F revealed that the IL-6 level was significantly higher in the severe loading experiment (10.33 ± 2.14 vs. 4.16 ± 1.03 pg/ml, P < 0.01; Fig. 1). The IL-1 level was also higher, although the difference failed to reach statistical significance (1.88 ± 0.53 vs. 0.86 ± 0.13 pg/ml, P = 0.08; Fig. 1). No difference was observed between either the ACTH or the -endorphin level between the two loading experiments (Fig. 1).

At point E, the cytokine and hormonal levels at the end of the high-load run were significantly higher than the respective values at the end of the moderate-load run (Fig. 1).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main finding of our study is that severe but not moderate IRB leads to a significant rise in plasma level of IL-1, IL-6, ACTH, and -endorphin. The strong relationships between the rise in the -endorphin and ACTH and the preceding increase in circulating IL-6 suggest that proinflammatory cytokines and especially IL-6 are responsible for the activation of the HPA axis, leading eventually to an increase in plasma -endorphin and ACTH.

We employed two different resistive loads (first 75%, then 35% of P_{m max}) separated by at least 3 wk to allow respiratory muscles to recover from a possible loading-related muscle fiber injury (11, 19, 50). The moderate IRB served as control for the strenuous IRB, because our hypothesis was based on the observation that only strenuous whole body exercise produces proinflammatory cytokines (5, 9, 29, 31, 44, 46).

Cytokine response to IRB. Previous studies indicate that circulating levels of proinflammatory cytokines increase significantly in response to strenuous but not mild whole body exercise (5, 9, 29, 31, 46). Our study indicates for the first time that proinflammatory cytokines increase in response to strenuous IRB. The increase in IL-6 is ~25% of that observed after strenuous whole body exercise (30), which is in line with the fact that the respiratory muscles comprise ~15% of total muscular mass (35). Interestingly, IL-1 and IL-6 were induced 39 ± 9 min after the onset of resistive breathing, a time course analogous to that reported during whole body exercise (31, 46). In fact, IL-6 may increase as early as 15 min after the start of exercise (31), indicating that the induction of cytokines is quite rapid.

The etiology of cytokine increase during strenuous exercise remains unclear. Theoretically, it could have resulted from shifts of cytokine-rich body fluids, decreased clearance, or increased secretion. Our experiment was not designed to address this issue, and we can neither prove nor refute any of these possibilities. However, some speculations can be made. Increased secretion could be partly or mainly responsible for the observed increase, although the source of cytokine release is elusive. Blood mononuclear cells have been excluded as a source of IL-6, but their involvement in IL-1 production is controversial with both suggestive (30) and negative (46) reports. Sympathetic activation acting in a paracrine and endocrine fashion on immune organs is another candidate (31). The third possible source is the intensely working respiratory muscles. Indeed, a significant increase in IL-6 mRNA expression was detected in homogenates of lower limb muscle biopsies obtained from normal humans after marathon running, coinciding with the rise in circulating IL-6 levels (30). Although the stimuli for the production of the cytokines are not known, reactive oxygen species (ROS) produced during both strenuous whole body exercise (42) and fatiguing resistive breathing within the respiratory muscles (2, 3) could probably be responsible. ROS may induce cytokine production through both transcriptional and posttranscriptional mechanisms (4, 23, 27, 42), although the time might be relatively short for de novo protein synthesis, and consequently the latter or even posttranslational mechanisms are more likely. Accordingly, pretreatment with vitamin E (a well-established antioxidant) clearly decreased the IL-1- and IL-6-producing capacity of humans subjected to exercise (6). As to the cellular source of cytokine production within the muscle, the vascular endothelial cells are probable candidates, because they can produce both IL-1 and IL-6 (20), especially on suffering attack by ROS (1). Alternatively, the respiratory muscle myocytes may be the cytokine-producing cells (30, 43). Clearly, more studies are needed to elucidate the source and stimuli of cytokine increase during strenuous exercise.

Stimulation of HPA axis. Strenuous IRB resulted in increased levels of circulating -endorphin and ACTH. The increase in -endorphin has been previously reported (33, 38, 48). However, the source of -endorphin is not clear, and both central and peripheral sites have been implicated (33, 38). Our study extends these observations and provides the first evidence that strenuous IRB causes elevation of the ACTH level. The strong correlations between these two hormones, both at points F and E (Fig. 2), as well as their simultaneous increase (evidenced by the strong correlation found between their respective increases from point F to E; Fig. 3) suggest a common source for both of them. This is very likely, given the fact that both are derived from posttranslational modification of the same molecule, proopiomelanocortin (22), and are concomitantly secreted by the pituitary gland (13, 14). Because ACTH originates exclusively from the pituitary gland in healthy humans, it stands to reason that both molecules were coreleased secondary to activation of the HPA axis.

The cause of the IRB-induced HPA axis activation is not known, and our experimental design does not allow us to sort it out. However, there are two alternative, not mutually exclusive possible explanations. First, it may have resulted secondary to the increased IL-1 and IL-6 induced by strenuous IRB. Previous studies indicate that both IL-1 and IL-6 are potent stimulants of the HPA axis (8, 23, 24), exhibiting significant synergism (8, 28, 32). Both IL-1 and IL-6 stimulate hypothalamic corticotropin releasing hormone secretion from parvicellular neurons located in the paraventricullar nuclei (8, 25), which are the major ACTH and -endorphin secretagogues by the pituitary corticotroph. In fact, even suboptimal amounts of either recombinant IL-1 or IL-6 that fail to stimulate the HPA axis synergistically stimulate the release of ACTH when they are coadministered (28). Whereas IL-6 plays a fundamental role in the stimulation of the HPA axis, the participation and interaction of IL-1 appears necessary for the full effects of IL-6 on the axis (32). However, the increase in plasma IL-6 concentration achieved is relatively modest, smaller than the minimum IL-6 concentration required to stimulate the HPA axis in experiments of recombinant IL-6 administration (25, 26). The extent to which this could be surpassed by the synergistic interaction with the IL-1 is not known.

Second, an alternative/complementary mechanism accounting for the increase of -endorphin and ACTH is the stimulation of small afferent nerve fibers (types III and IV) in the respiratory muscles. These project to various levels of the central nervous system (17) and are stimulated by fatigue-induced metabolic changes such as acidosis (15-18). Their influence on breathing pattern and respiratory muscle recruitment has been investigated in conscious goats. In these animals, strenuous resistive breathing is associated with a biphasic electromyographic response in the diaphragm (33), consisting of an initial and immediate increase (facilitation) followed by a partial decrease (inhibition). Both responses have been attributed to small afferent fiber activation, initially causing facilitation and, in a time- and/or intensity-dependent manner, partial inhibition through the elaboration of -endorphin (33). However, dichloroacetate, which prevented the intramuscular acidosis, altered this biphasic response by blunting the early facilitation, whereas the latter electromyographic response in the diaphragm was not affected (34). It is therefore likely that small afferent stimulation was involved in the early excitatory response, whereas another factor produced in a time- and intensity-dependent manner was responsible for the later partial inhibition through the elaboration of -endorphin. The results of the present study suggest that this factor may be the cytokines and especially IL-6. The time course of the production of -endorphin and ACTH also negates a major exclusive role of small afferent activation, because this would be expected to lead to an earlier response, whereas the hormonal response we observed was late. However, small afferents could interact synergistically with the cytokines to stimulate the HPA axis. Alternatively/complementary, the cytokines might directly stimulate small afferent fibers to induce -endorphin and ACTH elaboration from the HPA axis. In fact, global depletion of small afferent fibers (by repeated systemic treatment with capsaicin) inhibits the plasma ACTH response to intravenous IL-1 (25, 49). Activation of small afferent fibers originating from the respiratory muscles rather than global (whole body) small afferent fiber activation likely stimulated the HPA axis, because the local within muscle concentration of cytokines is expected to be higher (45). This would be analogous to the role of small vagal afferent activation in eliciting hypothalamic and HPA axis responses after intraperitoneal injection of IL-1 in animals, which are abolished after subdiaphragmatic transection of the vagus (26-28).

The notion that increased levels of -endorphin and ACTH induced by strenuous IRB may be secondary to the elevation of IL-6 and IL-1 concentration is supported by four observations: 1) the time course, with cytokine elevation appearing first at point F followed by the increase in -endorphin and ACTH levels at point E (Fig. 1); 2) the strong correlation between the IL-6 at point F and both -endorphin and ACTH at point E (Fig. 4); 3) the strong correlation between the increase in IL-6 (from point R to F) and the increase both in -endorphin and ACTH (from point F to E; Fig. 5); and 4) the fact that moderate IRB did not change either -endorphin or ACTH, which is in line with the lack of production of proinflammatory cytokines and supports the key role played by the cytokines in the stimulation of the HPA axis induced by strenuous IRB. Taken together, all the above may suggest, but certainly do not prove, a cause and effect relationship.

Perspectives

In conclusion, in healthy humans, strenuous IRB (contrary to moderate) increases the level of the proinflammatory cytokines IL-1 and IL-6 and stimulates the HPA axis, probably secondary to the increased cytokine (especially IL-6) production.