• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
Endocrinology. Author manuscript; available in PMC May 17, 2010.
Published in final edited form as:
PMCID: PMC2871297

Distinct, Time-Dependent Effects of Voluntary Exercise on Circadian and Ultradian Rhythms and Stress Responses of Free Corticosterone in the Rat Hippocampus


Previous work has shown that allowing rats to voluntarily exercise in a running wheel for 4 wk modifies the hypothalamic-pituitary-adrenal axis and behavioral coping responses to stress. To investigate whether long-term voluntary exercise would also affect the free, biologically active fraction of corticosterone in the brain, we conducted an in vivo microdialysis study in the hippocampus of rats. We monitored both the baseline circadian and ultradian patterns of corticosterone in hippocampus dialysates over the diurnal cycle and the responses to forced swim and novelty stress at different stages of exercise. Exercise for 1 d, 2 d, or 1 wk did not affect baseline circadian and ultradian pulse parameters or stress-induced hippocampal free corticosterone concentrations suggesting that acute or short-term periods of exercise do not affect baseline and stress-induced hormone levels. Baseline hormone parameters in 4 wk exercised rats, however, showed significantly increased pulse amplitudes (+108%) and mean free corticosterone levels (+42%) between 1500 and 2100 h but not between 0900 and 1500 h. Surprisingly, although our previous work showed substantial changes in stress-evoked plasma (total) corticosterone responses in long-term exercised animals, no differences in stress-induced hippocampal free hormone responses could be observed between exercised and sedentary animals. This lack of differences was not caused by compensatory changes in plasma corticosteroid-binding-globulin binding levels in exercising rats. Thus, long-term exercising rats show anticipatory increases in glucocorticoid output before the start of the active phase. These rats also reveal the putative existence of a containment mechanism preventing overexposure of the brain to glucocorticoid hormones.

Regular performance of exercise impacts on multiple systems in the body promoting fitness, health, and well-being. These systems include the motor, physiological, and neurobiological systems. For instance, exercise strengthens the bone structure and musculature (1), enhances autonomic balance (2, 3), and improves stress resilience through its effects on sleep architecture, cognition, and affective state (4-9). Although the beneficial effects of exercise are well documented, the mechanisms underlying how exercise impacts on these systems are still unclear.

We have suggested that the beneficial effects of exercise on stress resilience involve changes in the hypothalamic-pituitary-adrenocortical (HPA) axis, a neuroendocrine system implicated in protective and adaptive mechanisms evoked by stressful events (10-14). This notion was based on changes in HPA axis responses observed after subjecting long-term (4 wk) exercised animals to a stressful challenge. Exercised rats or mice show lower glucocorticoid responses than control, sedentary animals when exposed to a novel environment (11, 12, 14). This seems to reflect the exercised animals’ decreased anxiety-related behavior observed in the novel cage and in specific anxiety tests (6, 7, 11, 14). However, if animals are confronted with the stronger psychological and physical stimulus of forced swimming then the exercised animals, compared with the controls, respond with higher glucocorticoid levels (11, 12, 14). Given that the forced swim situation is a potentially life-threatening situation and involves increased metabolic demand, the enhanced hormone response is considered to be a better adaptive response (13, 15, 16). The exercised animals also showed stronger (glucocorticoid dependent) memory formation of the event (7). Thus, exercise benefits appropriate glucocorticoid responses to stressful events.

Exercised animals also show changes in the circadian variation of glucocorticoid hormones. They show significantly higher evening levels of plasma corticosterone than control animals, whereas early-morning or midday levels are not different (11). The circadian variation in glucocorticoid hormone is thought to be vital for maintaining normal physiology and well-being. Several diseases such as major depression are associated with alterations in the circadian rhythm of these hormones (17-21). Glucocorticoid hormones are secreted in a pulsatile fashion by the adrenal gland and the circadian variation in hormone levels is a reflection of the variation in pulse amplitude over the circadian cycle (22, 23). In plasma, these variations in hormone secretion are discernible as a pulsatile, ultradian rhythm forming a circadian rhythm over the day/night cycle (18, 24). Recently using intrahippocampal in vivo microdialysis, we demonstrated that the ultradian and circadian rhythms in glucocorticoid levels persist across the blood-brain barrier (25, 26). Due to the absence of glucocorticoid binding proteins [e.g. corticosteroid-binding globulin (CBG)] in the brain’s extracellular space, these glucocorticoid levels represent the free, biologically available fraction of the hormone. We also showed that stress evokes surges of free glucocorticoid levels in the brain, however, in the case of forced swim stress, peak levels lag 20 min behind the peak reached in total corticosterone levels in the circulation (25). Thus, in vivo microdialysis can determine changes in free glucocorticoid levels in target tissues with high time resolution, which is of great importance for the study of glucocorticoid physiology.

Long-term voluntary exercise impacts strongly on the circadian rhythm in (baseline) plasma glucocorticoid levels and on stress-induced plasma glucocorticoid responses in a stressor-specific manner. However, based on our recent findings, it seems that plasma glucocorticoid levels do not necessarily accurately reflect concentrations of free glucocorticoid hormone in a target tissue (25, 27). Therefore, to elucidate whether long-term voluntary exercise would have the same impact on baseline and stress-induced free corticosterone levels in the hippocampus as it did on plasma hormone levels, we conducted an in vivo microdialysis study on exercising and sedentary male Sprague Dawley rats. We determined changes in ultradian and circadian rhythms as well as changes in novelty- and forced swimming-evoked responses in hippocampal free corticosterone levels in rats that had voluntarily exercised for different time periods (i.e. 1 d to 4 wk). In addition to presenting the first detailed analysis of ultradian and circadian rhythms of free corticosterone levels in exercising rats we demonstrate that stress-induced responses in free corticosterone in the brain do not necessarily reflect those of hormone levels in the plasma.

Materials and Methods


Male Sprague Dawley rats (140–160 g) were singly housed under standard lighting (14 h light, 10 h dark cycle), humidity (50–60%), and temperature (21–22 C) conditions. Food and water were available ad libitum. After arrival rats were allowed to habituate to the housing conditions for 5 d. All procedures were in accordance with the Animals (Scientific Procedures) Act 1986 of the United Kingdom.

Voluntary exercise paradigm

The exercising group was allowed free access to a running wheel (diameter, 34 cm) in their home cages. Using a wheel-turning counting system, we monitored that all rats were running in the wheel. Rats ran mainly during the dark phase of the diurnal cycle. The housing of the sedentary (i.e. control) animals remained unchanged. The duration of wheel access depended on the study. In study 1 rats had access for 2 d, in study 2 for 1 wk, and in study 3 for 4 wk. The surgical and microdialysis procedures described below were identical for all three studies.


Rats were handled once per day (5 min per rat) starting 1 wk before surgery. Eight days before the microdialysis experiment, under isoflurane anesthesia, a guide cannula (Microbiotech AB, Stockholm, Sweden; MAB 6.14.IC) was implanted just entering the hippocampus at the dorsal site as described previously (25). After surgery, control rats were housed individually in plexiglas cages (length × width × height = 27 × 27 × 35 cm) with food and water ad libitum, whereas exercising animals were transferred into special plexiglas cages (length × width × height = 36 × 39 × 41 cm,) containing a compartment with a running wheel (diameter, 34 cm). Surgery had hardly any effects on running performance.

Microdialysis procedures

Microdialysis in freely behaving rats was conducted as before (25, 28-30). After 6 d of recovery, a microdialysis probe (polyethersulfone membrane; 15 kDa cutoff; length 4 mm; outer diameter 0.6 mm; Microbiotech AB; MAB 6.14.4) was inserted through the preimplanted guide cannula into the hippocampus under light isoflurane anesthesia, and the rats were connected to a liquid swivel and counterbalance arm system (Microbiotech) via a peg on their head. This system allows the animal to move freely. Microdialysis probes were perfused with sterile, pyrogen-free Ringer solution (Delta Pharma, Pfullingen, Germany; 147 mM NaCl, 4 mM KCl, 2.25 mM CaCl2) at a flow rate of 2 μl/min. Fluorethylenepolymer tubing with a dead volume of 1.2 μl per 100 mm length (Microbiotech) was used for all connections. Microdialysis samples were collected in cooled vials using automated refrigerated sample collectors (Microsampler 820; Univentor, Zejtun, Malta). The samples were stored at −80 C for corticosterone measurement (see below).

Experimental design

Study 1: first access to a running wheel (for up to 2 d)

Two days after insertion of the microdialysis probe into the hippocampus, sampling started at 0500 h and continued for 72 h. Microdialysis samples were collected in intervals of 10 min (between 0800 and 2200 h) or 30 min (between 2200 and 0800 h). On the first day of sampling, animals were left undisturbed in their home cage and had no access to a running wheel (i.e. the baseline day). On the second day at 0800 h, the animals were allowed free access for the first time to a running wheel by opening a compartment attached to their home cage containing a running wheel. Thereafter access was allowed until the end of the experiment (third day).

Study 2: short-term voluntary exercise (1 wk)

In this study we investigated changes in the circadian and ultradian rhythms as well as the stress responsiveness of hippocampal free corticosterone in 1-wk exercising and control rats. Approximately 8 h (at ~1800 h) after implantation of the guide cannula, some rats were given access (for the first time) to a running wheel, whereas other operated rats were not, i.e. the control animals. Six days later, in the early morning, the microdialysis probe was inserted after which the exercising rats were allowed to continue use of the running wheel. Two days after insertion of a microdialysis probe into the hippocampus, sampling started at 0500 h and continued for 48 h. Microdialysis samples were collected in intervals of 10 min (between 0800 and 2200 h) or 30 min (between 2200 and 0800 h). On the first day of sampling, animals were left undisturbed in their home cage. On the second day of sampling, at 1100 h, animals were exposed to forced swimming [15 min, 25 C water, as described previously (7, 25, 30, 31)]. After the procedure, rats were dried with a towel and returned to their home cage and left undisturbed for the remainder of the experiment while sampling continued. During the 48-h sampling period, the exercising animals had continuous access to a running wheel in their home cage except for the time of the stress procedure. To increase the time resolution of the free corticosterone response to stress, a sampling interval of 5 min was used between 1100 h (start of stressor) and 1300 h in the stressed animals.

Study 3: long-term voluntary exercise (4 wk)

Here we studied changes in the circadian and ultradian rhythms and the stress responsiveness of hippocampal free corticosterone in 4-wk exercising and control rats. After habituation, some rats were given access to a running wheel, whereas others were not, i.e. the control animals. Three weeks later, in the early morning, all rats were equipped with an intracranial guide cannula. Approximately 8 h after implantation (at ~1800 h), the exercising animals were given access again to their running wheels. Six days later, in the early morning, a microdialysis probe was inserted after which the exercising rats were allowed to continue use of the running wheel. Two days after insertion of the microdialysis probe into the hippocampus, sampling started at 0500 h and continued for 48 h. Microdialysis samples were collected in intervals of 10 min (between 0800 and 2200 h) or 30 min (between 2200 and 0800 h). On the first day of sampling, animals were left undisturbed in their home cage (i.e. baseline day). On the second day of sampling (i.e. stress day) at 1100 h animals were exposed to a novel environment [30 min in a novel cage (i.e. a clean cage containing no bedding material; light intensity raised to 450 Lux (normal light 90 Lux), as before (11, 14, 32, 33)] or submitted to forced swimming (15 min, 25 C water). After the procedure, rats were returned to their home cage and left undisturbed for the remainder of the experiment while sampling continued. During the 48-h sampling period, the exercising animals had continuous access to a running wheel in their home cage except for the time of the stress procedure. As before, a sampling interval of 5 min was used between 1100 h (start of stressor) and 1300 h in the stressed animals.

In a separate experiment, 4-wk exercising and control rats were either killed by decapitation under early morning baseline conditions or killed immediately after a 15-min forced swim session or after 30-min novelty exposure (n = 6 for all groups). Corticosterone levels in plasma prepared from trunk blood were determined by RIA (see below).

Measurement of free corticosterone in dialysates and total corticosterone in plasma

Dialysate corticosterone concentrations were measured using a commercially available RIA (MP Biomedicals, Irvine, CA) as described previously (25, 34). The detection limit was 0.00125 μg/dl; the inter- and intraassay variations were 16 and 14%, respectively. Dialysate corticosterone concentrations were not corrected for probe recovery (~25%).

Plasma corticosterone concentrations were measured using the same RIA as described previously (14, 31). The detection limit was 0.25 μg/dl (inter- and intraassay variation 7 and 4%).


After the experiment, animals were killed using an overdose of pentobarbital (Euthatal, 200 mg/kg body weight, ip). The brains were stored in 4% paraformaldehyde for later verification of probe placement (see Ref. 34). Only data from rats with correctly placed microdialysis probes were included in the analyses.

[3H]corticosterone binding assay for measurement of plasma CBG

After 4 wk of voluntary exercise, separate groups of exercising and control rats (n = 6) were killed under baseline conditions between 0900 and 1000 h or 1800 and 1900 h by quick anesthesia (<15 sec) in a glass jar containing isoflurane vapor followed immediately by decapitation. The trunk blood was collected in chilled EDTA containing tubes, centrifuged, and the obtained plasma stored at −80 C.

Plasma CBG levels were determined by [3H]corticosterone (specific activity, 70 Ci/mmol; NEN Life Science Products-PerkinElmer, Waltham, MA) binding, as described previously (11, 35). Briefly, binding of [3H]corticosterone in the absence (total binding) or presence of 1000-fold excess of nonradioactive corticosterone (nonspecific binding) was determined in 10-fold diluted plasma at a single greater than 95% saturating concentration of 60 nM in duplicates. After incubation (24 h, at 0–2 C) and separation of bound and free [3H]steroid by gel filtration, radioactivity was measured in a liquid scintillation counter. For the calculation of the number of plasma CBG binding sites, the specific activity of [3H]corticosterone was adjusted because of the presence of endogenous corticosterone [measured by RIA (see above)] in the plasma samples.

Data calculation and statistical analysis

The PULSAR algorithm (36) was used to calculate the pulse characteristics of free corticosterone levels measured on the baseline day as described by us previously (25). PULSAR analysis was initially performed over a 12-h period (0900–2100 h) because microdialysis samples had been collected with a high time resolution (10 min intervals) during this period, thus providing a sufficient number of data points for the analyses of ultradian rhythms. The following parameters were calculated (and expressed as group means ± SEM): pulse frequency (pulses/hour); mean pulse amplitude (mean height of pulses relative to a circadian rhythm baseline; micrograms per deciliter); mean pulse height (mean pulse height above zero; micrograms per deciliter); mean corticosterone concentration (micrograms per deciliter); area under the curve (AUC; arbitrary units). Furthermore, we analyzed in some experiments putative differences between pulse characteristics during the morning/early afternoon vs. the late afternoon/early night phase by splitting the 12-h period into two 6-h periods (0900–1500 and 1500–2100 h, respectively). Differences in parameters between these time periods were statistically assessed by paired Student’s t test and between experimental groups by unpaired Student’s t test. In one rat PULSAR analysis could not calculate reliable data; this animal was not included in the paired Student’s t tests.

The effects of forced swimming and novelty stress on corticosterone levels were statistically analyzed (unpaired Student’s t test) after calculation of the maximum response and the AUC for each rat individually (between 1100 and 1300 h).

Plasma CBG concentrations were expressed as femtomoles per milligram protein and group means and SEM values were calculated. Statistical differences between exercised and control rat CBG levels in the morning and afternoon were tested by two-way ANOVA and, if significant, by post hoc Bonferroni testing.

The level of significance was P < 0.05.


Study 1: circadian and ultradian rhythm of hippocampal free corticosterone in rats having their first access (for 2 d) to a running wheel

The aim was to determine whether first access to a running wheel, and thus first bouts of running, in untrained rats would result in an enhanced HPA axis activity as a result of the acute metabolic demand due to the raised physical activity. Figure 1A shows the ultradian and circadian rhythm in hippocampal free corticosterone (means ± SEM, n = 5) in normal rats over the course of a complete day (0500 h until 0500 h, next day). The rats showed a clear circadian and ultradian pattern in hippocampal free corticosterone levels (Fig. 1A) that showed similar characteristics [see Table 1 published as supplemental data on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org; pulsar analysis (36) on 0900 to 2100 h data] to those previously reported by us in male and female Wistar rats (25, 27).

FIG. 1
Circadian and ultradian rhythm of hippocampal free corticosterone under control conditions and during the subsequent 2 d of voluntary exercise as provided by access to a running wheel. Hippocampal free corticosterone data (expressed as micrograms per ...
Hippocampal free corticosterone levels in rats after 1 wk of voluntary wheel running

At 0800 h on the next day, the separation wall between the rat’s living space and the running wheel was removed allowing the animal access to the wheel. The rats initially explored the newly opened compartment, few of the animals walked/ran some steps in the wheel, after which all animals resumed the normal activity for this time of the day, which is resting/sleeping. The rats used the running wheel during the night time of the 2 subsequent days, running 469 ± 126 m (mean ± SEM, n = 5) on the first day and 579 ± 116 m (mean ± SEM, n = 5) on the second day. However, no apparent changes in the circadian and ultradian rhythms of hippocampal free corticosterone were found as a result of these initial running periods (Fig. 1, B and C; and supplemental Table 1). Thus, acute voluntary exercise does not seem to influence free glucocorticoid levels in the hippocampus.

Study 2: baseline circadian and ultradian rhythm and forced swimming-induced changes in hippocampal free corticosterone in short-term (1 wk) voluntary exercising rats

Both 1-wk voluntary exercising and control rats showed a clear circadian and ultradian rhythm in hippocampal free corticosterone levels (Fig. 2A). Both groups of animals showed significantly higher pulse heights, mean free corticosterone levels and AUC values for the 1500–2100 h period than for the 0900–1500 h period (Table 1). However, there were no significant differences in pulsar parameters between exercising and control rats (Table 1). The free corticosterone levels in exercising rats during the early evening/night period appeared to be higher than those in the control animals, but the difference was not statistically significant.

FIG. 2
Baseline circadian and ultradian rhythm of hippocampal free corticosterone (A) and forced swimming-induced increases in hippocampal free corticosterone levels (B) in control and 1-wk voluntary exercising rats. Data are expressed as micrograms per deciliter ...

Forced swimming led in both 1-wk exercising and control rats to a surge in hippocampal free corticosterone levels peaking at approximately 1 h after the start of the stressful challenge (Fig. 2B). However, there were differences in neither peak levels nor the rise or fall phases of the response in free corticosterone between the exercising and control rats (Fig. 2B).

Study 3: baseline circadian and ultradian rhythm and stress-induced changes in hippocampal free corticosterone in long-term (4 wk) exercising rats

Our previous work on 4-wk voluntary exercising mice and rats showed profound changes in the circadian rhythm and stress-induced responses of plasma (total) corticosterone (11, 14). Exercised animals showed significantly higher circadian peak levels than the controls. Here we investigated whether such elevated circadian peak levels would also be discernible in hippocampal free corticosterone levels in 4-wk voluntary exercising rats. Figure 3, A and B, show representative ultradian and circadian time profiles of hippocampal free corticosterone in a control rat and an exercising rat that suggest greater late afternoon/evening pulses in the exercising animal than in the control animal. Also, the time profiles of 13 control and nine exercising rats (means ± SEM) clearly show that during the last 4 h before lights off, the pulse amplitudes of exercising animals reach higher levels than those of the control animals (Fig. 3C). This was confirmed by pulsar analysis (Table 2). Exercising rats showed significantly increased pulse amplitudes (+108%), pulse heights (+62%), mean free corticosterone levels (+42%), and AUC values (+41%) during the 1500–2100 h period compared with the control animals during this period. Significantly increased pulse heights, mean corticosterone levels, and AUC values were also seen in the exercising animals over the complete 0900–2100 h period (Table 2). No significant differences were seen during the 0900–1500 h period. Long-term voluntary exercise had no effects on the pulse frequency (Table 2). Thus, the exercising rats showed more pronounced ultradian and circadian rhythms in hippocampal free corticosterone.

FIG. 3
Circadian and ultradian rhythm of hippocampal free corticosterone levels (expressed as micrograms per deciliter) under baseline conditions in 4-wk exercising rats (n = 9) and sedentary, control animals (n = 13). A, Representative example of a sedentary ...
Hippocampal free corticosterone levels in rats after 4 wk of voluntary wheel running

Previously we showed that 4 wk of voluntary exercise differentially impacts on forced swimming- and novelty stress-induced increases in plasma corticosterone concentrations (11, 14). Here we investigated whether these differential effects are also discernible in stress-induced free corticosterone responses in the hippocampus. Both forced swimming and novelty evoked rises in hippocampal free corticosterone levels in control and exercising rats whereby, as shown before (14), forced swimming led to higher glucocorticoid responses than novelty (Fig. 4). However, surprisingly, there were no significant differences in the effects of either stressor on free glucocorticoid levels between the exercising and the control animals (Fig. 4). The novelty-induced response in free corticosterone tended to be lower in the exercising rats, but this did not reach statistical significance. These results are in marked contrast to the observed changes in stress-induced responses in plasma corticosterone observed in long-term voluntary exercising rats and mice (11, 14).

FIG. 4
Effect of two different stressors on hippocampal free corticosterone levels in 4-wk exercising and control animals. A, Effect of forced swimming [1100–1115 h (gray column), 25 C water, n = 8]. B, Effect of novelty stress [1100–1130 h ( ...

At the time of these microdialysis experiments, we verified plasma (total) corticosterone under early morning baseline conditions and after forced swimming and novelty stress in 4-wk exercising and control rats. The obtained results were very similar to those published before [(14); baseline: control, 1.2 ± 0.1, exercise, 1.1 ± 0.1; forced swimming: control, 37.7 ± 3.2, exercise, 57.4 ± 4.6 (P < 0.05, Student’s t test); novelty: control, 14.3 ± 1.5, exercise, 6.6 ± 0.7 μg/dl (P < 0.05, Student’s t test; all groups, n = 6].

Plasma CBG binding capacity in control and long-term (4 wk) exercising rats

To check whether the absence of changes in stress-induced hippocampal free corticosterone responses of exercising rats may be due to altered circulating CBG levels in these animals, we determined CBG binding capacity in plasma of control and 4-wk exercising rats in the morning and evening just before lights off. There were no differences in plasma CBG levels between control and exercising animals either in the morning (control: 1360 ± 99; exercise: 1148 ± 83 fmol/mg protein) or in the evening [control: 1438 ± 112; exercise: 1374 ± 223 fmol/mg (all groups, n = 5)].


We showed that only long-term (4 wk) voluntary exercise resulted in changes in the ultradian and circadian rhythm of free corticosterone concentrations in the hippocampus. Furthermore, in contrast to previous observations in plasma, we found that in the extracellular space of the hippocampus stress-induced levels of (free) corticosterone in exercising rats were not different from those in control animals. We also found that this discrepancy was not due to changes in plasma CBG levels. These findings indicate that in exercising rats mechanisms act to regulate access of glucocorticoid to the target tissue, albeit distinctly under baseline and stress conditions.

The pulse parameter data generated by the pulsar algorithm on the basis of the hippocampal glucocorticoid levels of our control, male Sprague Dawley rats were similar to those generated from the free hormone levels in the hippocampus of male (and female) Wistar rats (25, 27). Thus, there seem to be no strain or sex differences at least between the Sprague Dawley and Wistar rat strains. Furthermore, pulse parameters and concentrations of free corticosterone seem to be similar throughout the brain (25).

Changes in the circadian and ultradian rhythm of hippocampal free corticosterone were only seen after 4 wk of exercise, but no alterations were seen at shorter periods of exercise. It appeared that the higher pulse amplitudes in exercising rats were mainly occurring during the 3- to 4-h period before lights off (at 1900 h; Fig. 3C). Given that rats start their daily running routine at the time when lights are switched off, the increases in glucocorticoid pulse amplitudes commencing several hours before lights off may be regarded as anticipatory increases in glucocorticoid availability before running possibly to support metabolism and other yet-to-be-identified processes. The elevated glucocorticoid output in 4-wk exercising rats specifically at the time before lights off corresponds well with our earlier observations in 4-wk voluntary exercising (wheel running) mice in which we found increased levels of plasma (total) corticosterone (in the absence of altered ACTH output) just before lights off and not at any other time of the day (11). Engeland and colleagues (37, 38) have shown that the circadian-driven increase in adrenocortical glucocorticoid secretion by the end of the light phase is mainly dependent on sympathoadrenomedullary activity rendering the adrenal cortex more sensitive to ACTH. We reported that 4-wk exercising rats and mice show an increased sympathoadrenomedullary capacity that accordingly may have boosted glucocorticoid secretion in these animals at this time (11, 14).

We found that acute or short-term voluntary exercise did not result in any significant changes in pulse parameters. Thus, the changes in glucocorticoid secretion did not develop acutely in exercising rats after having gained access to the running wheel, but this rather was a gradual process that required at least 1 wk of exercise to develop. Thus, physical activity per se does not activate the HPA axis, at least not the amount of physical activity displayed by our rats during the first 2 d.

Investigation of the effects of novelty and forced swimming on hippocampal free corticosterone levels in exercising and control rats delivered some very surprising findings. Although our previous work repeatedly demonstrated that 4-wk exercising animals show a substantially lower plasma corticosterone response to novelty (11, 12, 14; present study), we found in the present study in 4-wk exercising rats no significant changes in novelty-induced hippocampal free corticosterone responses. It is unlikely that we ‘missed’ an effect as we collected samples at 5-min intervals. Furthermore, plasma CBG concentrations did not differ between control and exercising rats, thus providing no explanation for the discrepancy between plasma total and hippocampal free corticosterone responses. A factor likely to contribute to the lack of a significant difference between the experimental groups is the relatively high variability in the free corticosterone data, especially at the peak of the stress response. An additional factor contributing to the variability in the free corticosterone values are the relatively small plasma corticosterone responses to novelty stress (39-41), which in the presence of CBG with its considerable buffering capacity at these levels of plasma corticosterone will reduce the resultant free corticosterone responses. Finally, fundamental changes in blood-brain barrier function in exercising rats affecting the access of corticosterone to the brain cannot be excluded, although there is no evidence for such a physiological response.

Exposing 1- and 4-wk exercising and control rats to forced swimming produced no differences in hippocampal free corticosterone responses. This is remarkable, given that this stressor evokes almost double the plasma corticosterone response in 4-wk exercising rats and mice than sedentary animals (11, 12, 14; present study). No differences in free corticosterone responses were observed over the whole time course and, as mentioned, there were no exercise-induced changes in plasma CBG levels. Previously we reported that 4-wk exercising rats and mice show similar plasma ACTH responses to forced swimming as control animals, and we hypothesized that the enhanced plasma corticosterone responses were due to enhanced sympathoadrenomedullary responses in these animals (11, 12, 14), increasing adrenocortical sensitivity to ACTH (38). Nevertheless, apparently despite the enhanced sympathoadrenomedullary activity leading to higher plasma corticosterone levels in the exercising animals, the levels of free corticosterone attained in the hippocampus were not different from those in control animals. Although the mechanism(s) restraining access of corticosterone to the brain of exercising rats is unknown, these observations may have considerable implications. Although the hippocampal free corticosterone responses after forced swimming were similar in exercising and control rats, evidently these responses are shaped by distinct physiological mechanisms in the two experimental groups.

Apparently these physiological mechanisms are regulated to ensure that despite the greatly elevated circulating levels of corticosterone in the blood of exercising rats, the brains of these animals are exposed to normal levels of glucocorticoid hormone. In these animals, it seems that a glucocorticoid access-to-brain restraining mechanism is counteracting the enhanced glucocorticoid-secretion promoting mechanism of an enhanced sympathoadrenomedullary system (a result of exercise-associated increased physical demand). This implies that in addition to hypothalamic, pituitary, and (sympatho-)adrenal mechanisms, additional mechanisms are involved in fine-tuning glucocorticoid synthesis and access to the brain. Such mechanisms may include blood-brain barrier-associated processes and 11ß-hydroxysteroid dehydrogenase, an enzyme well known to be involved in modulating glucocorticoid availability in glucocorticoid target tissues (42, 43) but not CBG as shown in the present study. Changes in 11ß-hydroxysteroid dehydrogenase expression and activity have been studied in several tissues but not brain tissue of voluntary exercising hamsters (44). Future research should provide insight into these mechanisms as well as determine whether these are brain-specific mechanisms. Often conclusions regarding glucocorticoid function in basic and clinical studies are drawn on the basis of measured plasma glucocorticoid concentrations. Our observations, however, point out the weakness in interpreting the physiological and functional importance of changes in plasma glucocorticoid levels without knowing the free hormone levels in the tissue compartment of interest.

HPA axis regulation including control of glucocorticoid function is a highly dynamic process. Based on our results, it appears that access of corticosterone to the brain is differentially controlled under baseline and stress conditions. Physiologically, attenuation of glucocorticoid access to the brain in exercising rats may be a homeostatic, protective mechanism to prevent their brains from exposure to excessive levels of glucocorticoids and/or to ensure that after stress a set (stressor specific?) level of glucocorticoid hormone is attained in the tissue. Although forced swimming evoked similar free corticosterone responses, exercising rats show enhanced glucocorticoid-dependent, hippocampal epigenetic, gene expression, and behavioral responses (7), possibly as a result of the increased glucocorticoid receptor expression in the hippocampus of these animals (14).

Supplementary Material

Supplemental data Table 1


We thank Drs. Yalini Chandramohan and Alison Leggett-Overbury for their technical assistance.

This work was supported by the Wellcome Trust (Grant 082628), United Kingdom.


Area under the curve
corticosteroid-binding globulin


Disclosure Summary: The authors have nothing to disclose.


1. Allen DL, Harrison BC, Maass A, Bell ML, Byrnes WC, Leinwand LA. Cardiac and skeletal muscle adaptations to voluntary wheel running in the mouse. J Appl Physiol. 2001;90:1900–1908. [PubMed]
2. Kramer JM, Plowey ED, Beatty JA, Little HR, Waldrop TG. Hypothalamus, hypertension, and exercise. Brain Res Bull. 2000;53:77–85. [PubMed]
3. Gielen S, Schuler G, Hambrecht R. Exercise training in coronary artery disease and coronary vasomotion. Circulation. 2001;103:e1–e6. [PubMed]
4. Lancel M, Droste SK, Sommer S, Reul JMHM. Influence of regular voluntary exercise on spontaneous and social stress-affected sleep in mice. Eur J Neurosci. 2003;17:2171–2179. [PubMed]
5. van Praag H, Christie BR, Sejnowski TJ, Gage FH. Running enhances neurogenesis, learning and long-term potentiation in mice. Proc Natl Acad Sci USA. 1999;96:13427–13431. [PMC free article] [PubMed]
6. Binder E, Droste SK, Ohl F, Reul JMHM. Regular voluntary exercise reduces anxiety-related behaviour and impulsiveness in mice. Behav Brain Res. 2004;155:197–206. [PubMed]
7. Collins A, Hill LE, Chandramohan Y, Whitcomb D, Droste SK, Reul JMHM. Exercise improves cognitive responses to psychological stress through enhancement of epigenetic mechanisms and gene expression in the dentate gyrus. PLoS One. 2009;4:e4330. [PMC free article] [PubMed]
8. Steptoe A, Edwards S, Moses J, Mathews A. The effects of exercise training on mood and perceived coping ability in anxious adults from the general population. J Psychosom Res. 1989;33:537–547. [PubMed]
9. Salmon P. Effects of physical exercise on anxiety, depression, and sensitivity to stress: a unifying theory. Clin Psychol Rev. 2001;21:33–61. [PubMed]
10. Steckler T, Holsboer F, Reul JMHM. Glucocorticoids and depression. In: Holly JMP, editor. Bailliére’s best practice, research. 13 ed Bailliére Tindall; London: 1999. pp. 597–614. [PubMed]
11. Droste SK, Gesing A, Ulbricht S, Müller MB, Linthorst ACE, Reul JMHM. Effects of long-term voluntary exercise on the mouse hypothalamic-pituitary-adrenocortical axis. Endocrinology. 2003;144:3012–3023. [PubMed]
12. Droste SK, Schweizer MC, Ulbricht S, Reul JMHM. Long-term voluntary exercise and the mouse hypothalamic-pituitary-adrenocortical axis: impact of concurrent treatment with the antidepressant drug tianeptine. J Neuroendocrinol. 2006;18:915–925. [PubMed]
13. Reul JMHM, Droste SK. The hypothalamic-pituitary-adrenal axis as a dynamically organized system: lessons from exercising mice. In: Steckler T, Kalin NH, Reul JMHM, editors. Handbook of stress and the brain, part 1: the neurobiology of stress. Elsevier; Amsterdam: 2005. pp. 95–112.
14. Droste SK, Chandramohan Y, Reul JMHM. Voluntary exercise impacts on the rat hypothalamic-pituitary-adrenal axis mainly at the adrenal level. Neuroendocrinology. 2007;86:26–37. [PubMed]
15. Sapolsky RM, Romero LM, Munck AU. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocr Rev. 2000;21:55–89. [PubMed]
16. De Kloet ER, Vreugdenhil E, Oitzl MS, Joëls M. Brain corticosteroid receptor balance in health and disease. Endocr Rev. 1998;19:269–301. [PubMed]
17. Carroll BJ, Curtis GC, Mendels J. Neuroendocrine regulation in depression. Arch Gen Psychiatry. 1976;33:1039–1044. [PubMed]
18. Linkowski P, Mendlewicz J, Leclercq R, Brasseur M, Hubain P, Golstein J, Copinschi G, Van Cauter E. The 24-hour profile of adrenocorticotropin and cortisol in major depressive illness. J Clin Endocrinol Metab. 1985;61:429–438. [PubMed]
19. Seckl JR, Campbell JC, Edwards CRW, Christie JE, Whalley LJ, Goodwin GM, Fink G. Diurnal variation of plasma corticosterone in depression. Psychoneuroendocrinology. 1990;15:485–488. [PubMed]
20. Schlechte JA, Sherman B, Pfohl B. A comparison of adrenal cortical function in patients with depressive illness and cushing’s disease. Horm Res. 1986;23:1–8. [PubMed]
21. Heesen C, Gold SM, Huitinga I, Reul JMHM. Stress and hypothalamic-pituitary-adrenal axis function in experimental autoimmune encephalomyelitis and multiple sclerosis—a review. Psychoneuroendocrinology. 2007;32:604–618. [PubMed]
22. Jasper MS, Engeland WC. Synchronous ultradian rhythms in adrenocortical secretion detected by microdialysis in awake rats. Am J Physiol. 1991;261:R1257–R1268. [PubMed]
23. Jasper MS, Walworth JT, Engeland WC. Pulsatile corticosterone secretion as measured by intraadrenal microdialysis. In: Levine JE, editor. Pulsatility in neuroendocrine systems. Academic Press; San Diego: 1994. pp. 247–264.
24. Windle RJ, Wood SA, Shanks N, Lightman SL, Ingram CD. Ultradian rhythm of basal corticosterone release in the female rat-dynamic interaction with the response to acute stress. Endocrinology. 1998;139:443–450. [PubMed]
25. Droste SK, de Groote L, Atkinson HC, Lightman SL, Reul JMHM, Linthorst ACE. Corticosterone levels in the brain show a distinct ultradian rhythm but a delayed response to forced swim stress. Endocrinology. 2008;149:3244–3253. [PubMed]
26. Linthorst ACE, Reul JMHM. Stress and the brain: solving the puzzle using microdialysis. Pharmacol Biochem Behav. 2008;90:163–173. [PubMed]
27. Droste SK, de Groote L, Lightman SL, Reul JMHM, Linthorst ACE. The ultradian and circadian rhythms of free corticosterone in the brain are not affected by gender: an in vivo microdialysis study in Wistar rats. J Neuroendocrinol. 2009;21:132–140. [PubMed]
28. Linthorst ACE, Flachskamm C, Müller-Preuss P, Holsboer F, Reul JMHM. Effect of bacterial endotoxin and interleukin-1β on hippocampal serotonergic neurotransmission, behavioral activity, and free corticosterone levels: an in vivo microdialysis study. J Neurosci. 1995;15:2920–2934. [PubMed]
29. Linthorst AC, Reul JMHM. Brain neurotransmission during peripheral inflammation. Ann NY Acad Sci. 1998;840:139–152. [PubMed]
30. Linthorst AC, Peñalva RG, Flachskamm C, Holsboer F, Reul JMHM. Forced swim stress activates rat hippocampal serotonergic neurotransmission involving a corticotropin-releasing hormone receptor-dependent mechanism. Eur J Neurosci. 2002;16:2441–2452. [PubMed]
31. Bilang-Bleuel A, De Carli S, Rech J, Holsboer F, Reul JMHM. Forced swimming evokes a biphasic response in CREB phosphorylation in extrahypothalamic limbic and neocortical brain structures. Eur J Neurosci. 2002;15:1048–1060. [PubMed]
32. de Groote L, Linthorst ACE. Exposure to novelty and forced swimming evoke stressor-dependent changes in extracellular GABA in the rat hippocampus. Neuroscience. 2007;148:794–805. [PubMed]
33. Chandramohan Y, Droste SK, Reul JMHM. Novelty stress induces phospho-acetylation of histone H3 in rat dentate gyrus granule neurons through coincident signalling via the N-methyl-D-aspartate receptor and the glucocorticoid receptor: relevance for c-fos induction. J Neurochem. 2007;101:815–828. [PubMed]
34. Linthorst ACE, Flachskamm C, Holsboer F, Reul JMHM. Local administration of recombinant human interleukin-1β in the rat hippocampus increases serotonergic neurotransmission, hypothalamic-pituitary-adrenocortical axis activity, and body temperature. Endocrinology. 1994;135:520–532. [PubMed]
35. Stefferl A, Storch MK, Linington C, Stadelmann C, Lassmann H, Pohl T, Holsboer F, Tilders FJ, Reul JMHM. Disease progression in chronic relapsing experimental allergic encephalomyelitis is associated with reduced inflammation-driven production of corticosterone. Endocrinology. 2001;142:3616–3624. [PubMed]
36. Merriam GR, Wachter KW. Algorithms for the study of episodic hormone secretion. Am J Physiol. 1982;243:E310–E318. [PubMed]
37. Jasper MS, Engeland WC. Splanchnic neural activity modulates ultradian and circadian rhythms in adrenocortical secretion in awake rats. Neuroendocrinology. 1994;59:97–109. [PubMed]
38. Engeland WC, Arnhold MM. Neural circuitry in the regulation of adrenal corticosterone rhythmicity. Endocrine. 2005;28:325–332. [PubMed]
39. Reul JMHM, Stec I, Söder M, Holsboer F. Chronic treatment of rats with the antidepressant amitriptyline attenuates the activity of the hypothalamic-pituitary-adrenocortical system. Endocrinology. 1993;133:312–320. [PubMed]
40. Reul JMHM, Labeur MS, Grigoriadis DE, De Souza EB, Holsboer F. Hypothalamic-pituitary-adrenocortical axis changes in the rat after long-term treatment with the reversible monoamine oxidase-A inhibitor moclobemide. Neuroendocrinology. 1994;60:509–519. [PubMed]
41. Gesing A, Bilang-Bleuel A, Droste SK, Linthorst ACE, Holsboer F, Reul JMHM. Psychological stress increases hippocampal mineralocorticoid receptor levels: involvement of corticotropin-releasing hormone. J Neurosci. 2001;21:4822–4829. [PubMed]
42. Seckl JR, Yau J, Holmes M. 11β-Hydroxysteroid dehydrogenases: a novel control of glucocorticoid action in the brain. Endocr Res. 2002;28:701–707. [PubMed]
43. Holmes MC, Seckl JR. The role of 11β-Hydroxysteroid dehydrogenases in the brain. Mol Cell Endocrinol. 2006;248:9–14. [PubMed]
44. Coutinho AE, Campbell JE, Fediuc S, Riddell MC. Effect of voluntary exercise on peripheral tissue glucocorticoid receptor content and the expression and activity of 11β-HSD1 in the Syrian hamster. J Appl Physiol. 2006;100:1483–1488. [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...