Acute stressor exposure can facilitate innate immunity and suppress acquired immunity. The present study further characterized the potentiating effect of stress on innate immunity, interleukin-1β (IL-1β), and demonstrated that stress-induced potentiation of innate immunity may contribute to the stress-induced suppression of acquired immunity. The long-term effect of stress on IL-1β was measured by using an ex vivo approach. Sprague-Dawley rats were challenged with lipopolysaccharide (LPS) in vivo, and the IL-1β response was measured in vitro. Splenocytes, mesenteric lymphocytes, and peritoneal cavity cells had a dose- and time-dependent ex vivo IL-1β response to LPS. Rats that were exposed to inescapable shock (IS, 100 1.6 mA, 5-s tail shocks, 60-s intertrial interval) and challenged with a submaximal dose of LPS 4 days later had elevated IL-1β measured ex vivo. To test whether the acute stress-induced elevation in IL-1β contributes to the long-term suppression in acquired immunity, IL-1β receptors were blocked for 24 h after stress. Serum anti-keyhole limpet hemocyanin (KLH) immunoglobulin (Ig) was measured. In addition, the acute elevation (2 h post-IS) of splenic IL-1β in the absence of antigen was verified. Interleukin-1 receptor antagonist prevented IS-induced suppression in anti-KLH Ig. These data support the hypothesis that stress-induced increases in innate immunity (i.e., IL-1β) may contribute to stress-induced suppression in acquired immunity (i.e., anti-KLH Ig).
- interleukin-1 receptor antagonist
- keyhole limpet hemocyanin
acute stressor exposure modulates both innate and acquired immune function. It has been recently reported that stressor exposure can increase many measures of innate immunity. For example, the rate of benign (10, 17a) and infectious bacterial inflammation resolution (5), fever (17), macrophage/neutrophil nitric oxide (NO; Refs.5, 8, 21), proinflammatory cytokines (29, 32, 37,46, 50), acute-phase proteins (9, 18, 48), and complement activity (7, 17a) are all elevated after exposure to acute laboratory stressors. Exposure to acute stress can also suppress features of acquired immune function, for example, antigen-specific antibody (17, 20, 16, 25), T-cell proliferative (15, 19, 42), and cytotoxic T-lymphocyte (3) responses. Taken together, it is reasonable to suggest that exposure to acute stress can result in both, enhanced innate and suppressed acquired, immune function.
These seemingly opposing effects of stress on acquired and innate immunity may suggest that aspects of the facilitated innate immune responses are involved in the production of the inhibited acquired immune response. This direction of causation is suggested because the innate response typically lacks specificity and precedes the specific immune response in time. For example, exposure to 60 min of intermittent tail shock stress (IS) enhances nitric oxide (NO) production by macrophages for 2–4 days (21). NO is a key mediator of innate immune defense immunity (2, 13, 52) since it suppresses cellular proliferation of a broad class of pathogens. Thus stress-induced enhancement of NO production might well be adaptive. However, NO suppresses cellular proliferation nonspecifically, and so elevated NO at the time of antigen-specific lymphocyte proliferation could suppress that response and directly contribute to the stress-induced reduction in acquired immunity. Indeed, in vitro transfer of macrophages from a stressed rat into the lymphocyte culture from a nonstressed rat results in suppression of the T-cell proliferative response of the nonstressed rat (15). As expected from the present argument, the potentiated NO response also contributes to the positive consequences of the stress response. Our laboratory has recently reported that rats exposed to IS recover nearly 50% faster from Eschericia coli infection and that elevated NO at the inflammatory site contributes to this effect (5). Thus the combination of facilitated innate immunity and inhibited acquired immunity may result in a functional compensation, a possibility that has recently been suggested by Murray et al. (36).
The purpose of the present study was to further characterize the long-term potentiating effect of stress on an additional measure of innate immunity [interleukin-1β (IL-1β)] and to determine whether stress-induced potentiation of innate immunity is causally related to stress-induced suppression of acquired immunity. To test this, we explored whether 1) exposure to IS primes or potentiates the IL-1β response to an important antigenic target on E. coli, lipopolysaccharide (LPS), 2) IS acutely increases IL-1 β in spleen and mesenteric lymph nodes, and 3) IS-induced enhancement of IL-1β contributes to the suppression of acquired immunity, specifically the antibody response against keyhole limpet hemocyanin (anti-KLH Ig). We developed an ex vivo approach to test whether stress produces a long-term potentiation of IL-1β production. Rats were challenged with LPS in vivo, but the IL-1β response was measured in vitro. This approach was necessary because testing the hypothesis that stress-induced potentiation of IL-1β contributes to the stress-induced suppression in anti-KLH Ig requires that the potentiated IL-1β response occur in immune tissues relevant to the antibody response, such as the spleen. The early effect of stress in the absence of LPS on tissue IL-1β concentrations in spleen and mesenteric lymph nodes was also measured. The effect of IS on circulating IL-1β has been previously reported (38). To test whether acutely elevated IL-1β is necessary for long-term stress-induced suppression of acquired immunity, IL-1β receptors were blocked after stress, and serum anti-KLH Ig was measured.
MATERIALS AND METHODS
Adult male Sprague-Dawley rats (4–10 per group; 300–350 g; Harlan Sprague Dawley, Indianapolis, IN) were individually housed in suspended metal wire cages (24.5 × 19 × 17.5 cm) with food and water available ad libitum. Colony conditions were maintained at 22°C on a 12:12-h light-dark cycle (lights on 0700–1900). Rats were given at least 2 wk to acclimate to the colony before experimental manipulation. Care and use of animals were in accordance with protocols approved by the University of Colorado Institutional Animal Care and Use Committee.
Time course study.
Before the impact of stress on IL-1β production could be tested, the time course of ex vivo IL-1β production needed to be determined to ensure an optimal response had been attained. Male Sprague-Dawley (4 mo) rats received intraperitoneal injections of either endotoxin- free saline (Abbott Laboratories, North Chicago, IL) or LPS (400 μg/kgE. coli endotoxin 0111:B4; lot number 86H18). Rats were killed 1.5, 3, 5, or 24 h after injection. Data from the saline-injected controls did not change across time, and so the data for this group were pooled for the final analyses.
Before the impact of stress on IL-1β production could be tested, a dose response of LPS was required to determine a dose that would result in a suboptimal response. Animals (6 per group) received intraperitoneal injections of 0.5 ml of 20, 200, or 400 μg/kg of LPS or endotoxin-free saline. Rats were killed 1.5 h after LPS challenge.
Stress LPS study.
Four days after stress termination, animals were injected intraperitoneally with LPS or saline and killed 1.5 h later. The 4-day interval was chosen because previous work from our laboratory has found that NO potentiation is still present 4 days after IS in KLH immunized rats (21).
Animals (6 per group) either remained in their cages as controls or were placed in Plexiglas tubes (23.4 × 7.0 cm) and subjected to a tail-shock stress procedure. The stress procedure consisted of 100 5-s, 1.6-mA inescapable tail shocks (IS), with an average intertrial interval of 60 s and the entire procedure lasting 110 min. All stress procedures occurred between 0900 and 1100 h. After stress termination, rats were returned to their home cages.
Splenic and Mesenteric Lymph Node IL-1β Concentrations 2 h After Stress
The acute effect of stress in the absence of LPS on tissue IL-1β concentrations in spleen and mesenteric lymph nodes was measured in rats (6 per group) 2 h after exposure to IS, as previously described. Rats were killed via decapitation. Spleen and mesenteric lymph nodes were removed and flash frozen in liquid nitrogen. Tissues were homogenized in homogenizing buffer (StressGen Biotechnologies, and 1 protease inhibitor cocktail tablet, Boehringer Mannheim, per 50 ml of extraction reagent). For each ∼0.5-cm3 piece of tissue, 1 ml of homogenizing buffer was used. Tissues were then dissociated with the use of sterile, modified, glass tissue homogenizers. The extract was transferred to a microcentrifuge tube and centrifuged at 21,000 g for 10 at 4°C. Supernatants were removed and stored at −20°C until time of assay. Presented IL-1β concentrations are corrected for total amount of protein (Bradford).
In experiments involving KLH antigen, rats received an intraperitoneal injection of 200 μg of soluble KLH (Calbiochem) in 0.5 ml of sterile endotoxin-free saline. Administration of KLH occurred immediately before IS or at an equivalent time for nonstressed control rats.
A blood sample for antibody assessment was quickly taken (within 2 min of touching cage) by gently wrapping the rat in a small towel and lightly restraining it using a Velcro strapping apparatus. A small nick was made in the exposed tail with a no. 15 scalpel, and a blood sample (300 μl) was quickly milked from the tail vein. This procedure was conducted at 0900 on days 5, 7, 10, and 14 post-KLH immunization. Anti-KLH IgM levels were measured in samples taken days 5, 7,10, and 14. Anti-KLH IgG levels were measured only in samples taken on days 7, 10, and14.
Anti-KLH IgM assessment.
An ELISA was used for antibody assessment. Microtiter plates (96-well, Immulon-4, Dynex) were coated with 0.5 mg/ml dialyzed KLH for 3 days at 4°C. Plates were then washed and blocked with 5% bovine serum albumin (Sigma Chemical) overnight at 4°C. Serum samples were diluted (IgM 1:400, IgG 1:3,000) in phosphate- buffered saline (PBS) containing 0.05% Tween 20 (Sigma Chemical). A single serial dilution (1:2) of these concentrations was performed. These dilutions ensured that the sample concentration fell within the linear range of the plate reader. Microtiter plates were incubated for 3 h at 37°C and then washed with PBS-Tween 20. Secondary antibody, alkaline phosphatase-conjugated goat anti-rat IgM (1:5,500 dilution, Cappel) or alkaline phosphatase conjugated goat anti-rat IgG (1:2,000, Cappel) was added to each well for 60 min at 37°C. Plates were again washed three times before addition of p-nitrophenyl phosphate substrate (Sigma Chemical). Plates were incubated at room temperature in the dark until plate positive control wells registered an optical density of ∼1.0 at 405 nm on a Dynatech plate reader. Results are presented as a percent change from nonstressed, drug- matched controls.
Rats were injected subcutaneously with interleukin-1 receptor antagonist (IL-1ra; 100 mg · kg−1 · ml−1 in CSE buffer, Amgen) or CSE buffer (vehicle) every 4 h for 24 h after IS + KLH or no stress + KLH. This dosing regimen of IL-1ra was chosen because we have previously reported it to block stress-induced NO elevations measured 4 days after IS termination (34).
Cell Culture Procedures
Animals were briefly anesthetized by exposure to ether and killed by cervical dislocation. Peritoneal cells were removed by lavage. Cold dissection medium (30.0 ml of Iscove's medium with 1% penicillin-streptomycin) was placed into the peritoneal cavity, the abdomen was briefly massaged, and the fluid was removed (20 ml). The medium was centrifuged, and the cells were resuspended to 2.0 × 106 cells/ml in culture medium (Iscove's medium containing 10% fetal bovine serum with 1% penicillin-streptomycin and 2 μM l-glutamine; all media reagents from GIBCO). Cells were counted using a Coulter particle counter. Mesenteric lymph node tissue and spleen were aseptically dissected and then placed in dissection medium over ice. Lymph nodes and spleens were dissociated using sterile, modified, glass tissue homogenizers. Lymph node and spleen cells were resuspended in culture medium at 10.0 × 106 cells/ml. One milliliter of each cell suspension was plated into a well of a 24-well flat-bottom culture plate (Falcon). Cells were cultured without additional stimulation at 37°C with 5% CO2. After 48 h, culture supernatant was collected and stored at −20°C until time of assay.
After ether anesthetization, but before cell or organ collection, blood was collected by cardiac puncture with a 23-gauge needle. Blood was allowed to clot for 30 min over ice before it was centrifuged. Serum was then collected and stored at −20°C until assayed.
IL-1β was measured from culture supernatants, tissue supernatants, and serum with the use of a rat-specific ELISA kit (R&D Systems). Culture samples were assayed according to manufacturer's instructions. Tissue supernatants were assayed at optimal concentrations (spleen, 1:40; lymph node, 1:10). Serum was assayed undiluted.
A one-factor ANOVA was used to determine statistical differences for the LPS time course and dose-response study. Post hoc pairwise comparisons were performed using Fisher's least significant difference. For the study involving stress, a stress (home cage control vs. IS) by drug (saline vs. LPS) two-factor ANOVA was used. For the study involving IL-1ra and anti-KLH Ig, two-factor, repeated-measures ANOVAs were performed on the change from control anti-KLH Ig data. Statistical analyses were performed by using SuperAnova with a statistical difference accepted at α = 0.05.
Time Course of LPS Stimulation
As described above, animals received a high dose of LPS (400 μg/kg) and were killed either 1.5, 3, 5, or 24 h later. Splenocytes, mesenteric lymph node cells, and peritoneal cells were then placed in culture for 48 h with no additional stimulation. Figure 1 shows IL-1β measured in supernatants from these cultures. A significant group difference was detected [F(4,17) = 462.0,P < 0.01]. IL-1β production in splenocytes of rats treated with LPS peaked after 1.5 h of stimulation and remained significantly greater than vehicle control through the 5-h measurement point (Fig. 1 A, P < 0.01). A progressive decrease in IL-1β production was measured at each time point and by 24 h of in vivo LPS stimulation IL-1β production returned to control levels (P = 0.17).
Similar to splenocytes, mesenteric lymph node cells also exhibited significant group differences [F(4,17) = 19.8, P < 0.01] in IL-1β production, with peak levels also measured from cells with 1.5 h after in vivo LPS stimulation (Fig. 1 B). Post hoc tests confirmed that IL-1β levels were higher than saline control for the 1.5, 3, and 5 h groups (P < 0.01) but returned to baseline after 24 h (P = 0.70). Likewise, a significant group difference in IL-1β from LPS stimulated peritoneal cultures was present [Fig. 1 C, F(4,17) = 12.5, P < 0.01]. IL-1β production also peaked in cells stimulated for 1.5 h with LPS but approached control levels after 5 h (P = 0.07), and returned to vehicle control levels at 24 h (P = 0.99).
Dose Response of LPS Stimulation
Results from the first study indicated that the peak IL-1β response measured in vitro occurred 1.5 h after in vivo LPS stimulation. Therefore, this time was chosen to optimize dose-response measures. IL-1β from cells taken from spleen, mesenteric lymph node, and peritoneal lavage in response to various in vivo LPS doses is presented in Fig. 2,A–C. Splenocyte production of IL-1β indicated a significant group effect [F(3,20) = 97.8, P < 0.01]. All doses of LPS elevated IL-1β production over vehicle control levels (P < 0.01). Additionally, animals given the higher doses of LPS (200 or 400 μg/kg) produced significantly more IL-1β than did animals given the 20 μg/kg dose (P < 0.01). A significant group effect was also detected for IL-1β production in cultured mesenteric lymphocytes [Fig. 2 B, F(3,20) = 7.14,P < 0.01]. Post hoc analyses confirmed that both the 200 and 400 μg/kg doses caused an elevation in IL-1β over vehicle control (P < 0.01). The 20 μg/kg dose did not significantly elevate IL-1β in lymph node cells (P = 0.38). Peritoneal cells stimulated in vivo with LPS for 1.5 h produced a dose-dependent increase in IL-1β [Fig. 2 C,F(3,20) = 18.6, P < 0.01]. Similar to the spleen, all doses of LPS used significantly elevated IL-1β in peritoneal cells over vehicle control (P < 0.05).
Stress and LPS Stimulation
To study the effects of tail shock stress on the response to LPS, a suboptimal dose of LPS was used to prevent a ceiling effect on IL-1β production. Therefore, the 20 μg/kg dose of LPS was chosen for this study. Additionally, prior work from our laboratory has shown that the presence of antigen at the time of tail shock is necessary to produce elevated NO responses 4 days after stressor termination (A. Moraska, unpublished results). The experimental design is depicted in Fig. 3. Rats were injected, therefore, with 200 μg ip of soluble KLH antigen immediately before tail shock. Four days after stressor termination, rats were injected with 20 μg/kg of LPS and killed 1.5 h later, and cells were removed and cultured for an additional 48 h. IL-1β levels from splenocyte culture supernatants (Fig. 4 A) were increased by stress [F(1,20) = 4.92,P < 0.05] and LPS [F(1,20) = 43.1, P < 0.01]. In addition, there was a significant stress × LPS interaction [F(1,20) = 4.47,P < 0.05], indicating that stress potentiated the effects of LPS. Mesenteric lymph node cultures (Fig. 4 B) showed a different pattern. There was a significant main effect for LPS administration [F(1,20) = 6.28,P < 0.02], but no stress × LPS interaction [F(1,20) = 1.26, P = 0.98]. The peritoneal cultures (Fig. 4 C) yielded data similar to the mesenteric lymph nodes in that there was a main effect of stress [F(1,20) = 26.0,P < 0.01] but no interaction [F(1,20) = 0.69, P = 0.42]. Interestingly, IL-1β concentrations in the supernatants from cultured cells were not elevated by stress alone.
The pattern for the IL-1β response in the serum followed that for splenocyte cultures (Fig. 5). LPS significantly elevated serum IL-1β [F(1,18) = 20.0, P < 0.01] in both stress and nonstressed rats, but the elevation in LPS-stimulated serum IL-1β was significantly greater in rats that were stressed [F(1,18) = 4.23,P < 0.05].
Splenic and mesenteric lymph node IL-1β concentrations 2 h after stress. Splenic [Fig. 6,F(1,9) = 6.5, P < 0.05], but not mesenteric lymph node [Fig. 6,F(1,9) = 1.9, P = 0.2], tissue concentration of IL-1β is elevated 2 h after IS termination. The short-term increase occurred in the absence of LPS.
Stress and IL-1ra
Our laboratory has previously reported that exposure to IS reliably increases in circulating IL-1β at 0 h and 2 h, but not 24 h, after IS termination (38). As shown in Fig.6, IS also produces an increase in splenic IL-1β that is detectable 2 h after IS. To test the potential effect of this early stress-induced increase in IL-1β on stress-induced suppression of anti-KLH Ig, rats were injected with IL-1ra or vehicle every 4 h for 24 h after IS or no stress. The experimental protocol is shown in Fig. 7. Data are presented as a percent change from not stressed, drug-matched control. IL-1ra reduced the suppressive effect of IS on anti-KLH IgM (Fig.8 A) and anti-KLH IgG (Fig.8 B). A 2 (IL-1ra vs. vehicle) × 3 or 4 (days) repeated-measure ANOVA revealed a trend for a reduction in the stress-induced suppression of anti-KLH IgM [F(3,60) = 2.5, P = 0.06] and reliable reduction in the suppression of anti-KLH IgG [F(2,40) = 6.0, P < 0.01] across time.
LPS in vivo resulted in a large and sustained IL-1β response that was detectable in vitro for at least 5 h after challenge. This response was dose dependent, with 20 μg/kg resulting in a submaximal response. The IL-1β response to in vivo LPS was produced by splenocytes, mesenteric lymphocytes, and peritoneal cavity cells. However, only splenocytes had a potentiated IL-1β response to LPS challenge given 4 days after IS termination.
It has been previously reported (28), that serum IL-1β responses to LPS challenge are also increased 4 days after IS. In these studies, the long-lasting potentiated response only occurred if the LPS challenge was given in vivo. If the cells were removed from the animal after stress, and stimulated with LPS in vitro, no stress-induced potentiation was found. In the present studies, an ex vivo approach was used. This approach was advantageous because 1) the cells of the immune system remained in the hormonal milieu when responding to LPS and 2) the dissection of specific cell population responses was possible. The present data suggest, therefore, that the spleen may be an important source of the potentiated IL-1β concentration detected in the serum. This idea is further supported by the fact that the concentration of IL-1β in the spleen is 10 times that detected in the blood and could be primary source of the cytokine detected in the serum.
Exposure to IS acutely (0–2 h) increases circulating IL-1β (37, 38), splenic IL-1β (Fig. 6) and NO production (5, 21). Our laboratory has previously reported that administration of IL-1ra after IS prevents the stress-induced potentiation of NO 4 days after stressor exposure in KLH immunized rats (34). Thus, in the present study, we tested whether IL-1ra would also prevent stress-induced suppression of acquired immunity as measured by serum anti-KLH Ig. The results were that administration of IL-1ra did indeed reduce the immunosuppressive effect of IS on anti-KLH IgM and anti-KLH IgG. These data support the hypothesis that stress-induced increases in innate immunity (i.e., IL-1β) contribute to stress-induced suppression in acquired immunity (i.e., anti-KLH Ig). They also extend the previous observations of the effect of IS on products released by cells of the innate immune system to also include IL-1β.
Because the innate immune response is the arm of the immune response that is responsible for the “first line of defense” against infection, stress-induced priming of this response could be adaptive. That is, it is reasonable to suggest that an acutely stressed organism would be better able to mount a greater or faster response to antigenic challenge. It has been previously reported, for example, that exposure to acute stress potentiates responses of cells (neutrophils and macrophages) classically associated with innate immunity, and facilitates responses to antigens (E. coli and LPS) classically cleared by innate immune responses (5, 8, 10, 13, 14,24, 33, 39, 41, 43-45, 49). The potentiating effect of acute stressor exposure (defined as a single exposure, <2 h in duration) on measures of innate immunity is pervasive in the literature. In fact, one is hard pressed to find exceptions to this rule. In the literature, if stress is reported to suppress measures of innate immunity, it commonly occurs after repeated stress sessions or a single stress session greater than 2 h in duration (30, 31,53). Thus, if a stressor is chronic, the potentiating effect of stress on innate immunity can be lost.
Although the majority of the literature supports the suppressive effect of acute stress on acquired immunity, there are examples whereby acute stressor exposure potentiates measures of immunity involving T cells and nonbacterial antigen (11, 35). Interestingly, in these other studies, the antigenic challenge was administered to the skin or involved a relatively mild stressor. Perhaps, the potentiating effect of acute stress on acquired immunity occurred because the skin is often considered to be the most important tissue associated with innate immunity (27) or that moderate-intensity acute stressors tend to produce immunopotentiation. More work needs to be done before conclusions can be made.
The mechanism(s) responsible for the relatively long-lasting (up to 4 days after IS) potentiation of LPS-induced IL-1β production or NO are not known. It seems unlikely that IS-induced changes in migration of leukocytes could play a role in these effects since such changes are quite short lived, usually disappearing within 24 h (17a, 19, 22). The data presented here suggest that IL-1β elevations produced during and immediately after stress may play a role. IL-1β is known to stimulate NO release (12). Our laboratory has previously reported (34) that IL-1ra administered every 4 h for 24 h after stress prevents the long-term (2–4 days after IS) increase in NO found in rats exposed to KLH + IS. The potentiated NO response is likely to be involved in stress-induced suppression of anti-KLH T-cell proliferation, leading to suppression in anti-KLH Ig (21, 15). The present study demonstrated that IL-1ra administered every 4 h for 24 h after stress prevents anti-KLH Ig suppression, just as it prevents stress-induced enhancement of NO production. Thus IL-1β released during stress may be involved in priming subsequent antigen-stimulated IL-1β and NO responses. One potential scenario is that exposure to IS increases circulating IL-1β (38) and splenic IL-1β (Fig. 6), leading to increased IL-1β receptor (IL-1βr) expression on macrophages and neutrophils (4). Upregulation of IL-1βr can persist for several days, and so when the stressed organisms is exposed to antigen (LPS or KLH), the cells of the innate immune system have an exaggerated IL-1β and NO response. In fact there is evidence that exposure to a single session of restraint stress can increase IL-1βr expression in the pituitary (1, 26). The fact that both the IS-induced elevation in IL-1β and the potentiated LPS-stimulated NO response was more robust in spleen compared with mesenteric lymph node lends support to this idea.
There is also evidence supporting a role of stress-reactive hormones and peptides, such as substance P (51), catecholamines (47), and moderate levels of glucocorticoids (40), in potentiating neutrophil and macrophage function. The results of the present studies, however, do not address whether these hormones and peptides are involved in the mediation of IS-induced potentiation of innate immune function.
The finding that exposure to acute stress produced a long-lasting potentiation in IL-1β responses to LPS, which may play an important role in decreasing KLH-specific Ig, suggest that the immunological response to acute stress is a “double-edged sword.” Acute stressor exposure, and in some circumstances chronic stress (36), can increase features of the innate immune response, and this improved innate response can be adaptive. If, however, the organism is challenged with a pathogen or via a route that requires the proliferation of T and B cells and the generation of an acquired immune response, then the stress-induced enhancement of innate immunity could be detrimental to host defense.
Address for reprint requests and other correspondence: M. Fleshner, Dept. of KAPH; Center for Neuroscience, Campus Box 354, Univ. of Colorado at Boulder, Boulder, CO 80309-0354 (E-mail:).
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 © 2002 the American Physiological Society