Nonstressed rat model of acute endotoxemia that

0 downloads 0 Views 150KB Size Report
Brackett, D. J., C. F. Schaefer, P. Tompkins, L. Fagraeus, ... Ann. Math. Stat. 27: 251–271, 1956. 24. Spengler, R. N., S. W. Chensue, D. A. Giacherio, N. Blenk,.
Nonstressed rat model of acute endotoxemia that unmasks the endotoxin-induced TNF-a response DAVID W. A. BENO AND ROBERT E. KIMURA Section of Neonatology, Department of Pediatrics, Rush Children’s Hospital, Rush Presbyterian St. Luke’s Medical Center, Chicago, Illinois 60612 Beno, David W. A., and Robert E. Kimura. Nonstressed rat model of acute endotoxemia that unmasks the endotoxininduced TNF-a response. Am. J. Physiol. 276 (Heart Circ. Physiol. 45): H671–H678, 1999.—Previous investigators have demonstrated that the tumor necrosis factor-a (TNF-a) response to endotoxin is inhibited by exogenous corticosterone or catecholamines both in vitro and in vivo, whereas others have reported that surgical and nonsurgical stress increase the endogenous concentrations of these stress-induced hormones. We hypothesized that elevated endogenous stress hormones resultant from experimental protocols attenuated the endotoxin-induced TNF-a response. We used a chronically catheterized rat model to demonstrate that the endotoxin-induced TNF-a response is 10- to 50-fold greater in nonstressed (NS) rats compared with either surgical-stressed (SS, laparotomy) or nonsurgical-stressed (NSS, tail vein injection) models. Compared with the NS group, the SS and NSS groups demonstrated significantly lower mean peak TNF-a responses at 2 mg/kg and 6 µg/kg endotoxin [NS 111.8 6 6.5 ng/ml and 64.3 6 5.9 ng/ml, respectively, vs. SS 3.9 6 1.1 ng/ml (P , 0.01) and 1.3 6 0.5 ng/ml (P , 0.01) or NSS 5.2 6 3.2 ng/ml (P , 0.01) at 6 µg/kg]. Similarly, baseline concentrations of corticosterone and catecholamines were significantly lower in the NSS group [84.5 6 16.5 ng/ml and 199.8 6 26.2 pg/ml, respectively, vs. SS group 257.2 6 35.7 ng/ml (P , 0.01) and 467.5 6 52.2 pg/ml (P , 0.01) or NS group 168.6 6 14.4 ng/ml (P , 0.01) and 1,109.9 6 140.7 pg/ml (P , 0.01)]. These findings suggest that the surgical and nonsurgical stress inherent in experimental protocols increases baseline stress hormones, masking the endotoxininduced TNF-a response. Subsequent studies of endotoxic shock should control for the effects of protocol-induced stress and should measure and report baseline concentrations of corticosterone and catecholamines. catecholamines; corticosterone; stress; endogenous

has documented the complex interaction of the neuroendocrine and immune systems, an important consideration in research addressing septic shock (3, 21, 34). For both humans and animals, corticosterone and catecholamines have been shown to modulate the immune response in vivo and in vitro (7, 26, 33). Although these mechanisms are not completely defined, human data suggest that pretreatment with exogenous cortisol or catecholamine attenuates the endotoxin-induced tumor necrosis factor-a (TNF-a) response (2, 28). However, little research has focused on the release of endogenous corticosterone and catecholamines during experimental protocols for the study of

A SERIES OF STUDIES

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.

septic shock and whether these endogenous hormones have the capacity to attenuate the TNF-a response. The purpose of this study was to test the hypothesis that elevated baseline concentrations of endogenous corticosterone and catecholamines result from surgical and/or nonsurgical stressors in experimental protocols and that these elevated endogenous hormones attenuate the TNF-a response. We speculated that animals studied under nonstressed experimental conditions with physiological baseline concentrations of corticosterone and catecholamine would exhibit a significantly greater endotoxin-induced TNF-a response and that this response could be elicited with relatively low doses of endotoxin. MATERIALS AND METHODS

Animals A total of 91 adult male Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA) weighing 200–250 g were used in this study. Rats were housed singly in standard cages and were fed chow and water ad libitum. The environment was temperature- and humidity-controlled, with lights on and off at 0630 and 1630, respectively. Operative Procedures Operative procedures were performed as previously described for our chronically catheterized rat model (17, 27). Briefly, the animals were anesthetized with intramuscular injections of 60 mg/kg ketamine and 5 mg/kg xylazine. Under an aseptic technique a 5-cm vertical midline abdominal skin incision was made from the subxiphoid process of the sternum to the suprapubic region and a 0.25-cm skin incision was made over the cervical vertebrae. Infusion sets (no. 4871, Abbott Laboratories, North Chicago, IL) were flushed with 0.9% saline solution containing 10 units of heparin/ml and were pulled through the skin opening over the vertebrae and into the abdominal incision. Then a 4.5-cm vertical midline incision was made through the abdominal wall. The infusion set tubes were introduced into the abdominal cavity through small punctures in the right abdominal wall. To prepare the abdominal aorta for catheterization, the gut was retracted onto sterile, saline-soaked gauze. The aortic catheter, consisting of an Insyte catheter tip (Becton Dickenson, Sandy, UT), Silastic tubing, and PE-60 tubing, in sequence, was introduced into the abdominal aorta over a 22-gauge Insyte needle. The Insyte tip of the catheter was advanced 0.5 cm into the aorta and secured with one drop of cyanoacrylate glue. The distal PE-60 tubing was inserted into its infusion tubing, and the line was flushed. This procedure was repeated for the inferior vena cava (IVC). The abdominal wall and skin were closed with 4-0 silk suture. The infusion sets exiting the cervical incision were sutured securely to the back of the rat with 2-0 silk suture and were glued postoperatively with silicon to form a single unit. Ampicillin (30 mg/kg)

0363-6135/99 $5.00 Copyright r 1999 the American Physiological Society

H671

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (178.171.097.009) on May 28, 2018. Copyright © 1999 American Physiological Society. All rights reserved.

H672

NONSTRESSED RAT MODEL OF ACUTE ENDOTOXEMIA

was injected into the IVC and aortic catheters. Catheters were flushed daily with saline. Reagents Endotoxin (Escherichia coli 0127:B8; Sigma Chemical, St. Louis, MO) was prepared in sterile saline, aliquoted, and stored at 280°C. Experimental Design Rats were given varying doses of endotoxin under three different experimental conditions: 1) nonstressed (NS), 2) surgical stressed (SS), and 3) nonsurgical stressed (NSS). This design allowed the identification of dose- and timedependent responses, while controlling for the effects of protocol-induced stressors on outcome measures. For each group, experiments were performed between 0900 and 1000 to control for circadian variation (10). NS group. This group of chronically catheterized rats experienced the operative procedures described and thereafter were maintained under nonstressed experimental conditions. These animals were considered nonstressed from an experimental perspective in that the environment was manipulated to ensure that animals were not subjected to further protocol-induced surgery, restraint, or pain. These 60 animals were assigned to one of four experimental groups, according to the endotoxin dose to be administered: 1) 2 mg/kg (n 5 7), 2) 6 µg/kg (n 5 37), 3) 10 ng/kg (n 5 8), or 4) control (saline diluent only, n 5 8). At a median of 7 days postsurgery (range 3–14 days), endotoxin doses were diluted in 0.5 ml of saline and infused through the IVC for 0.5 min. Then the IVC was flushed with 1 ml saline. SS group. The effect of surgical stress on outcome measures was studied for 16 rats who were injected with endotoxin immediately on completion of the described surgical procedures. The animals were assigned to receive one of three endotoxin doses: 1) 2 mg/kg (n 5 5), 2) 6 µg/kg (n 5 6), or 3) control (saline diluent only, n 5 5). Endotoxin was diluted and infused into the IVC as described for the NS group. NSS group. The effect of nonsurgical stress, including restraint, handling, and pain that are inherent in many other experimental protocols was studied for 15 rats on the 7th day after operative procedures were performed. These chronically catheterized rats received infusions into the tail vein of either 6 µg/kg of endotoxin (n 5 9) or saline (controls, n 5 6) using the following procedure. Each rat was secured in an individual restraining cage to allow tail access. The area of injection was swabbed with 70% ethanol and allowed to dry. A 25-g butterfly (no. 4573, Abbott Laboratories) was inserted into the tail vein, with placement confirmed by blood flow into the butterfly. Then blood samples were obtained from the aortic catheter. Endotoxin was infused into the tail vein, and the line was flushed with 1 ml of saline. The rat was removed from the restraining cage and placed into an individual cage for the duration of the experiment.

For 14 NS rats, weight gain and baseline concentrations of corticosterone, endotoxin, and TNF-a were measured daily between 0900 and 1000. Blood samples were cultured on tryptic soy, MacConkey’s, and blood agar plates. The measures of these 14 rats document that the NS group was studied under conditions of no or minimal stress, and no day-to-day variability existed in these measures over the 14-day study period. Statistical analysis. All results are expressed as means 6 SE. For kinetic studies, differences were compared by both one-way and repeated-measures ANOVA. Peak responses were compared with Student’s t-test or with post hoc comparison as appropriate. Type 1 error was set at 0.05. RESULTS

Experimental Conditions of NS Group Our data document that animals in the NS group were studied under conditions of no or minimal stress. At day 1 postsurgery these animals exhibited a 5.3 6 1.5% weight loss, but preoperative weight was reestablished by day 3. Mean daily weight gain for days 2–14 was comparable to that of control animals (7.2 6 2.8 vs. 7.8 6 1.2 g, P 5 0.98). An initial corticosterone surge (278 6 51 ng/ml) was noted in response to surgery, but mean daily corticosterone concentration for days 2–14 was 57 6 5 ng/ml (Fig. 1). A repeated-measures ANOVA demonstrated that mean corticosterone did not vary significantly from 2 to 14 days (P , 0.57). The rats were not bacteremic, and neither endotoxin nor TNF-a was noted in the daily blood and serum analyses (data not shown). Endotoxin-Induced TNF-a, Catecholamine, and Corticosterone Responses in NS Group Mean serum concentrations of TNF-a, catecholamines, and corticosterone revealed significant doseand time-dependent effects in response to endotoxin challenge (Fig. 2). For TNF-a a repeated-measures ANOVA revealed a significant effect for dose (P , 0.001) and time (P , 0.001). Peak TNF-a concentrations were noted at 90 min postendotoxin challenge in response to

Measures Mediators. The following mediators were measured by aortic blood sampling at baseline (0), and 30, 60, 90, 120, 180, and 240 min after endotoxin infusion: TNF-a (ng/ml), catecholamines (pg/ml), and corticosterone (ng/ml). TNF-a was measured by ELISA (Genzyme, Cambridge, MA). Catecholamines, a combination of epinephrine and norepinephrine, were measured by radioenzymatic assay (Amersham, Arlington Heights, IL) from plasma collected with EGTA (90 mg/ml) and glutathione (60 mg/ml). Corticosterone was measured by radioimmunoassay kit (ICN Biomedicals, Costa Mesa, CA).

Fig. 1. Daily corticosterone concentration for subgroup of 14 nonstressed (NS) rats from day 0 (immediately postsurgery) through day 14. Data are means 6 SE. These data demonstrate that NS rats were studied under conditions of no to low stress. Serum for corticosterone measurements was drawn each day between 0900 and 1000 to avoid circadian variation.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (178.171.097.009) on May 28, 2018. Copyright © 1999 American Physiological Society. All rights reserved.

NONSTRESSED RAT MODEL OF ACUTE ENDOTOXEMIA

H673

Fig. 2. Tumor necrosis factor-a (TNF-a; A), catecholamine (B), and corticosterone concentrations (C) for NS rats in response to 4 endotoxin doses: 2 mg/kg (n 5 7), 6 µg/kg (n 5 32), 10 ng/kg (n 5 8), or control (saline diluent only, n 5 8). Data are means 6 SE. Time 0 (baseline), immediately before endotoxin infusion.

2 mg/kg and 6 µg/kg, and at 60 min postendotoxin challenge in response to 10 ng/kg. No TNF-a was observed in the serum of NS rats in the control group. Peak catecholamine and corticosterone responses occurred, respectively, at 60 and 120 min postendotoxin challenge (Fig. 2). The catecholamine and corticosterone responses to 2 mg/kg of endotoxin were of significantly greater amplitude and duration than for 6 µg/kg (P , 0.001 and P , 0.005, respectively) and 10 ng/kg of endotoxin challenge (P , 0.001 and P , 0.001, respectively). Furthermore, catecholamine and corticosterone responses to 2 mg/kg of endotoxin had not returned to baseline by 240 min. Control rats displayed no changes in catecholamine or corticosterone concentrations, and control values were not significantly different from values for those given 10 ng/kg of endotoxin (P , 0.24). Endotoxin-Induced TNF-a Response Over Time To ensure that the TNF-a response was not affected by the number of days after surgery that experiments were conducted, we compared the peak TNF-a concentrations from 3 (61.0 6 14.8 ng/ml) to 14 (54.2 6 16.9 ng/ml) days postsurgery in 32 of the rats in the 6 µg/kg endotoxin challenge group. The median day of experimentation was 7 days (62.3 6 17.8 ng/ml). A repeatedmeasures ANOVA revealed no statistically significant time-dependent effect for peak TNF-a over this period (P , 0.65), indicating that the timing of the experiment did not influence the TNF-a response.

trations .200 ng/ml (mean 358.8 6 23.2 ng/ml), whereas their mean peak TNF-a concentration was only 3.9 6 2.8 ng/ml. In contrast, the remaining 32 rats demonstrated serum concentrations of corticosterone ,200 ng/ml (mean 84.5 6 16.5 ng/ml) and a mean peak TNF-a concentration of 63.6 6 7.6 ng/ml. Although the numbers were small, a Student’s t-test demonstrated significantly greater mean corticosterone (P , 0.001) and significantly lower mean peak TNF-a (P , 0.001) for these five animals compared with the remaining 32 rats. Mean daily weight gain for these five animals was not significantly different from the remaining 32 rats (6.1 6 3.6 vs. 7.2 6 2.8 g). Endotoxin-Induced TNF-a, Catecholamines, and Corticosterone Responses in SS Group For animals in the SS group a repeated-measures ANOVA revealed significant dose- and time-dependent TNF-a responses for 2 mg/kg (P , 0.01) and 6 µg/kg (P , 0.03) groups (Fig. 4). Dose- and time-dependent responses were noted for catecholamine but not for corticosterone in SS animals (Fig. 4). A striking differ-

Relationship Between Endotoxin-Induced TNF-a Response and Baseline Corticosterone Concentration Five of the 37 rats from the NS group that were infused with 6 µg/kg endotoxin did not demonstrate a TNF-a response. To test the hypothesis that elevated baseline corticosterone concentrations in these animals had attenuated the TNF-a response, we plotted baseline corticosterone concentrations against peak serum TNF-a concentrations for the 37 NS rats (Fig. 3). The five rats that did not produce TNF-a in response to endotoxin challenge had serum corticosterone concen-

Fig. 3. Scatterplot of mean concentrations of baseline corticosterone and peak TNF-a for all 37 rats in 6 µg/kg endotoxin group. The 5 NS rats that demonstrated no TNF-a response had baseline corticosterone concentrations .200 ng/ml.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (178.171.097.009) on May 28, 2018. Copyright © 1999 American Physiological Society. All rights reserved.

H674

NONSTRESSED RAT MODEL OF ACUTE ENDOTOXEMIA

Fig. 4. TNF-a (A), catecholamine (B), and corticosterone concentrations (C) for surgical stressed (SS) rats in response to 3 endotoxin doses: 2 mg/kg (n 5 5), 6 µg/kg (n 5 6), or control (saline diluent only, n 5 5) and for 32 NS rats in response to 6 µg/kg endotoxin (data from Fig. 2). Data are means 6 SE. Time 0 (baseline for SS rats), immediately before endotoxin infusion but after laparotomy for catheter placement.

ence in peak TNF-a between NS and SS groups (Figs. 2 and 4) was observed. For the NS group peak TNF-a concentrations of 111 6 6.5 ng/ml and 64.3 6 5.9 ng/ml, for 2 mg/kg and 6 µg/kg endotoxin doses, respectively, were significantly higher (P , 0.001) than comparable values of 3.9 6 1.1 ng/ml and 1.3 6 0.5 ng/ml, respectively, for the SS group. Endotoxin-Induced TNF-a, Catecholamine, and Corticosterone Responses in NSS Group TNF-a, corticosterone, and catecholamine responses for the NSS group are depicted in Fig. 5, and compared with those for the NS group (Fig. 2). A repeatedmeasures ANOVA confirmed the difference in the TNF-a response (P , 0.001) for the NSS and NS groups. Effect of Associated Stress Hormones on TNF-a Response On the basis of post hoc analysis of the data presented in Fig. 3, we stratified the 37 NS rats that were

infused with 6 µg/kg endotoxin into two subgroups: 1) those with baseline corticosterone concentrations ,200 ng/ml and high peak TNF-a responses (NS, Lo-Cort; n 5 32) and 2) those with baseline corticosterone concentrations .200 ng/ml and low peak TNF-a responses (NS, Hi-Cort; n 5 5). Then we constructed two scatterplots (Fig. 6) to depict the relationships between peak TNF-a and baseline concentrations of corticosterone and catecholamines for 50 animals within the four subgroups: 1) NS, Lo-Cort, n 5 32; 2) NS, Hi-Cort, n 5 5; 3) SS, n 5 6; 4) NSS, n 5 9. These data reveal visual differences between the NS, Lo-Cort group and the remaining three subgroups. To confirm these groupspecific differences we performed separate one-way ANOVA for TNF-a, baseline corticosterone, and baseline catecholamines, using the four subgroups as independent variables. The ANOVA revealed statistically significant differences among the four groups (P , 0.001). All pairwise post hoc Scheffe´’s tests were significant between the NS,

Fig. 5. TNF-a (A), catecholamine (B), and corticosterone concentrations (C) for nonsurgical-stressed (NSS) rats in response to 2 endotoxin doses: 6 µg/kg of endotoxin (n 5 9) or saline (controls, n 5 6) and for 32 NS rats in response to 6 µg/kg endotoxin (data from Fig. 2). Data are means 6 SE. Time 0 (baseline for NSS rats), immediately before endotoxin infusion but after insertion of infusion needle into tail vein.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (178.171.097.009) on May 28, 2018. Copyright © 1999 American Physiological Society. All rights reserved.

NONSTRESSED RAT MODEL OF ACUTE ENDOTOXEMIA

H675

Fig. 6. Scatterplot of peak TNF-a compared with either baseline corticosterone (A) or baseline catecholamine (B) for 4 groups of rats in response to 6 µg/kg endotoxin: NS, Lo-Cort: baseline corticosterone ,200 ng/kg, n 5 32, data from Fig. 2; NS, Hi-Cort: baseline corticosterone .200 ng/kg, n 5 5, data from Fig. 2; SS: n 5 6; or NS: n 5 9. Peak TNF-a and baseline corticosterone and catecholamine were significantly different for the NS, Lo-Cort group than for the combination of the other 3 groups (Scheffe´’s post hoc, P , 0.001).

Lo-Cort group and each of the three subgroups separately (P , 0.001), with the exception of baseline catecholamine for NS, Lo-Cort and NS, Hi-Cort. Using separate post hoc Scheffe´’s comparisons, we tested the hypothesis that the NS, Lo-Cort group was different from the remaining three groups that were characterized by elevated stress hormones and lower TNF-a. Resulting mean values for the NS, Lo-Cort group were significantly different from mean values for the combination of the three ‘‘stressed’’ groups; peak TNF-a (64.3 6 5.9 vs. 4.0 6 1.5 ng/ml, P , 0.001), corticosterone (84.5 6 16.5 vs. 254.6 6 27.3 ng/ml, P , 0.001), and catecholamine (198.8 6 26.2 vs. 940.8 6 290.9 pg/ml, P , 0.001)(23). DISCUSSION

These data are the first to demonstrate that rats with low baseline concentrations of corticosterone and catecholamine possess the ability to mount a TNF-a response to relatively low doses of endotoxin. The peak TNF-a concentrations for animals in the NS group were greater than previously described at endotoxin doses equivalent to from 1 to 1/1 3 106 those used in previous studies (13, 15, 18, 22, 31). In contrast, rats that were studied under surgically or nonsurgically induced stress demonstrated elevated baseline corticosterone and catecholamine concentrations and diminished TNF-a responses for all doses examined. Although our findings confirm the hypotheses of others that experimental conditions associated with highstress hormones affect physiological responses to endotoxin, our study is the first to experimentally manipulate protocol-induced stress while examining the effect on baseline corticosterone, catecholamine, and peak TNF-a (6, 8, 14, 32). Several methodological controls ensured that animals in our NS group were maintained and studied under experimental conditions that were free of surgically and nonsurgically induced stress. During the 3–14 days postsurgery when experiments were conducted, animals demonstrated a mean daily weight gain comparable to control rats and did not have bacterial infection or endotoxin in the blood. Similarly, no difference over the 3–14 days was noted in the mean baseline corticosterone or the peak TNF-a response. In

contrast, our SS and NSS groups were studied under conditions of laboratory stress that are standard in experimental protocols used by other investigators. Whereas both the SS and NSS groups demonstrated an attenuated TNF-a response to endotoxin, baseline stress hormones differed for the two groups. Our SS group displayed elevated baseline corticosterone, suggesting stress from surgery, whereas the NSS group displayed elevated baseline catecholamine suggesting stress from handling and pain. We observed that five chronically catheterized rats (NS, Hi-Cort) demonstrated baseline serum corticosterone concentrations .200 ng/ml. These rats showed no overt signs of stress and demonstrated weight gain that was not statistically different from other rats in the NS group. Although the reasons for these elevated baseline corticosterone values are unknown, we suspect that these animals experienced nonbacterial surgical complications. The fact that these rats appeared to be NS, yet demonstrated elevated baseline corticosterone, underscores the importance of measuring and reporting stress hormones for all animals in studies of endotoxic shock. Although experimental models with indwelling catheters for endotoxin infusion and blood sampling have been reported, these protocols incorporate significantly greater endotoxin doses to elicit the TNF-a response than doses required by animals in our NS group (11, 15, 18, 22, 31). Previous investigators administered $1.0 mg/kg of endotoxin to achieve a mean peak TNF-a response that was only 1⁄3 to 1⁄10 of values for our NS animals that received this endotoxin dose. Our findings are the first to reveal a statistically significant TNF-a response to 10 ng/kg of endotoxin. Feuerstein et al. (11) described a minor TNF-a response to 100 ng/kg of endotoxin, and Givalois et al. (13) used 5 µg/kg of endotoxin administered intraarterially to achieve a mean peak TNF-a concentration equivalent to one-half that reported in our NS animals. These different findings can be explained by the fact that previous protocols did not permit sufficient postsurgical recovery, with resultant elevations in baseline stress hormones at the time of endotoxin challenge. Indexes of stress, such as weight loss, baseline corticosterone, and catecholamines were not reported in these studies.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (178.171.097.009) on May 28, 2018. Copyright © 1999 American Physiological Society. All rights reserved.

H676

NONSTRESSED RAT MODEL OF ACUTE ENDOTOXEMIA

Similarly, the fact that our experimental protocol permitted the manipulation of NS, SS, and NSS conditions challenges previous hypotheses of a speciesspecific inconsistency in endotoxin doses for eliciting the TNF-a response. Whereas previous work has suggested that some animal models, such as rats, require larger (.1 mg/kg) doses than other models, such as humans (4 ng/kg), none of these protocols controlled for conditions of handling, restraint, pain, and surgical stress. Our data reveal significantly different mean peak TNF-a for the NS, Lo-Cort group when compared with the combination of the remaining three stressed groups, for the same endotoxin dosage. This finding suggests a within-species variability in the TNF-a response, which is dependent on the concentrations of baseline stress hormones. The hypothesis linking elevated endogenous baseline stress hormones to the TNF-a response is supported by findings that the study of rats in stressed conditions, such as surgery and handling, influences physiological responses. Brackett et al. (6) demonstrated that anesthetizing rats induces norepinephrine, an inhibitor of the endotoxin-induced TNF-a response, whereas Zellweger et al. (32) have shown significant impairment of cell-mediated immunity for at least 24 h postlaparotomy. Bagby et al. (1) demonstrated that exerciseinduced stress before endotoxin challenge suppresses the TNF-a response. Furthermore, several chronically catheterized rat models, similar to our NS animals, have been developed to examine the physiological and hormonal response to endotoxin and bacterial infections (8, 9, 12, 13, 16). These investigators have described significant biochemical and physiological differences between chronic and acute catheterization models and have emphasized that experimental protocols for the study of septic shock and hormone function should incorporate a conscious animal model and the avoidance of anesthesia. These experimental conditions are essential for maintaining physiological concentrations of corticosterone and catecholamines that characterized our NS group. A related body of research has linked exogenous corticosterone and catecholamines to suppression of the immune defense mechanisms (20, 26). In vitro, treatment with epinephrine or corticosterone inhibits macrophage activation and the secretion of TNF-a postendotoxin challenge (24). Exogenous epinephrine and corticosterone also inhibit the in vivo TNF-a response to endotoxin when given before or at the time of endotoxin challenge in humans and rats (2, 29, 30). In one study, TNF-a for rats that had undergone adrenalectomy or hypophysectomy remained elevated following endotoxin challenge (34). Under NS, physiological conditions in animals and humans, endotoxin-induced TNF-a, corticosterone, and catecholamines have been shown to reach peak concentrations in a predictable pattern (7). This pattern, in which glucocorticoids peak immediately after TNF-a, but catecholamines peak before TNF-a, is similar to the

time- and dose-pattern for these hormones in our NS group. In contrast, the animals with elevated baseline corticosterone (NS, Hi-Cort; SS; NSS groups) demonstrated no time-dependent corticosterone responses. Our findings explain those of previous investigators who suggested that corticosterone and catecholamines may be involved in the counterregulation of the TNF-a response under both stressed and normal, NS conditions. Our finding that corticosterone inhibits the TNF-a response is supported by related studies of septic shock in both humans and animals. Several investigators have reported that exogenous glucocorticoids are efficacious during the early phases of septic shock but that they do not increase survival in humans when administered during severe late septic shock (5, 19, 25). These therapeutic attempts most likely failed because of the transient nature of the rise in serum TNF-a. TNF-a orchestrates the immunological response following its very early appearance after endotoxin challenge. Removal of TNF-a from experimental models of septic shock minimizes the shock response (4). However, in patients presenting with septic shock, the acute phase of the disease has passed and the majority of patients do not possess detectable serum concentrations of TNF-a (5). The data from our study are consistent with those of previous researchers in suggesting that glucocorticoid concentration at the time of endotoxin challenge regulates the TNF-a response. These findings imply that therapeutic attempts to limit septic shock with glucocorticoids must concentrate on the early stages of the disease process. Catecholamines may also regulate the TNF-a response since our SS and NSS animals demonstrated a dose- and time-dependent catecholamine response. However, the response for these groups was slower and reflected a lower catecholamine concentration compared with the NS group. Furthermore, the NSS group had a significantly higher (P , 0.05) peak TNF-a response to 6 µg/kg of endotoxin compared with the SS group. This time-dependent response occurred in the NSS group despite the markedly elevated increases in baseline catecholamine. The observed differences in catecholamine concentrations for the three groups suggest that in addition to the baseline concentration subsequent increases in catecholamines further attenuate the TNF-a response to endotoxin. Previous studies have raised the possibility that prior exposure to exogenous stress hormones may enhance the observed TNF-a response to endotoxin. Barber et al. (2) showed that, in humans, antecedent treatment of cortisol enhanced the endotoxin-induced (4 ng/kg) TNF-a response from 12 to 144 h postadministration but inhibited the response at 6 h postadministration. Van der Poll and Lowry (29) reported similar results with epinephrine. We controlled for this possibility in the current study by examining the TNF-a response to 6 µg/kg for 3–14 days postsurgery during the time that all experiments were conducted. Al-

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (178.171.097.009) on May 28, 2018. Copyright © 1999 American Physiological Society. All rights reserved.

NONSTRESSED RAT MODEL OF ACUTE ENDOTOXEMIA

though our data confirm that the NS group was not affected by surgery-induced stress during this time, we did not examine rats 1 or 2 days postsurgery because they had not returned to preoperative weight and because we have previously reported intestinal blood flow and metabolism during this time (27). The possibility exists in our NS group that the enhanced TNF-a response to endotoxin persists beyond 14 days, but the inability to maintain consistently low corticosterone concentrations has limited our investigation of this hypothesis. In summary, we believe that this is the first study to demonstrate the relationship between baseline stress hormones and the endotoxin-induced TNF-a response when the conditions of NS, SS, and NSS are manipulated experimentally within the same animal model. Our findings underscore the importance of measuring and reporting baseline concentrations of corticosterone and catecholamines in subsequent studies of endotoxin shock. We gratefully acknowledge Drs. Michael Uhing, Luven Tujero, and Yong Chen for assistance in surgery, experiments, and assays. We also extend gratitude to Paula P. Meier, for a critical review of the manuscript. Address for reprint requests: D. W. A. Beno, Section of Neonatology, Dept. of Pediatrics, MU 622, Rush Children’s Hospital, Rush Presbyterian St. Luke’s Medical Center, 1653 W. Congress, Chicago, IL 60612. Received 30 March 1998; accepted in final form 21 October 1998. REFERENCES 1. Bagby, G. J., D. E. Sawaya, L. D. Crouch, and R. E. Shepherd. Prior exercise suppresses the plasma tumor necrosis factor response to bacterial lipopolysaccharide. J. Appl. Physiol. 77: 1542–1547, 1994. 2. Barber, A. E., S. M. Coyle, M. A. Marano, E. Fischer, S. E. Calvano, Y. Fong, L. L. Moldawer, and S. F. Lowry. Glucocorticoid therapy alters hormonal and cytokine responses to endotoxin in man. J. Immunol. 150: 1999–2006, 1993. 3. Bateman, A., A. Singh, T. Kral, and S. Solomon. The immunehypothalamic-pituitary-adrenal axis. Endocr. Rev. 10: 92–112, 1989. 4. Beutler, B., I. W. Milsark, and A. C. Cerami. Passive immunization against cachectin/tumor necrosis factor protects mice from lethal effect of endotoxin. Science 229: 869–871, 1985. 5. Bone, R. C., C. J. Fisher, Jr., T. P. Clemmer, G. J. Slotman, C. A. Metz, and R. A. Balk. A controlled clinical trial of high-dose methylprednisolone in the treatment of severe sepsis and septic shock. N. Engl. J. Med. 317: 653–658, 1987. 6. Brackett, D. J., C. F. Schaefer, P. Tompkins, L. Fagraeus, L. J. Peters, and M. F. Wilson. Evaluation of cardiac output, total peripheral vascular resistance, and plasma concentrations of vasopressin in the conscious, unrestrained rat during endotoxemia. Circ. Shock 17: 273–284, 1985. 7. Butler, L. D., N. K. Layman, P. E. Riedl, R. L. Cain, J. Shellhaas, G. F. Evans, and S. H. Zuckerman. Neuroendocrine regulation of in vivo cytokine production and effects: I. In vivo regulatory networks involving the neuroendocrine system, interleukin-1 and tumor necrosis factor-alpha. J. Neuroimmunol. 24: 143–153, 1989. 8. Carlson, D. E. Adrenocorticotropin correlates strongly with endotoxemia after intravenous but not after intraperitoneal inoculations of E. coli. Shock 7: 65–69, 1997. 9. Ciancio, M. J., J. Hunt, S. B. Jones, and J. P. Filkins. Comparative and interactive in vivo effects of tumor necrosis factor alpha and endotoxin. Circ. Shock 33: 108–120, 1991. 10. Dunn, J., L. Scheving, and P. Millet. Circadian variation in stress-evoked increases in plasma corticosterone. Am. J. Physiol. 223: 402–406, 1972.

H677

11. Feuerstein, G., J. M. Hallenbeck, B. Vanatta, R. Rabinovici, P. Y. Perera, and S. N. Vogel. Effect of gram-negative endotoxin on levels of serum corticosterone, TNF alpha, circulating blood cells, and the survival of rats. Circ. Shock 30: 265–278, 1990. 12. Fish, R. E., and J. A. Spitzer. Continuous infusion of endotoxin from an osmotic pump in the conscious, unrestrained rat: a unique model of chronic endotoxemia. Circ. Shock 12: 135–149, 1984. 13. Givalois, L., J. Dornand, M. Mekaouche, M. D. Solier, A. F. Bristow, G. Ixart, P. Siaud, I. Assenmacher, and G. Barbanel. Temporal cascade of plasma level surges in ACTH, corticosterone, and cytokines in endotoxin-challenged rats. Am. J. Physiol. 267 (Regulatory Integrative Comp. Physiol. 36): R164– R170, 1994. 14. Hensler, T., H. Hecker, K. Heeg, C. D. Heidecke, H. Bartels, W. Barthlen, H. Wagner, J. R. Siewert, and B. Holzmann. Distinct mechanisms of immunosuppression as a consequence of major surgery. Infect. Immun. 65: 2283–2291, 1997. 15. Jiang, J. X., Y. F. Diao, K. L. Tian, H. S. Chen, P. F. Zhu, and Z. G. Wang. Effect of hemorrhagic shock on endotoxin-inducing TNF production and intra-tissue lipopolysaccharide-binding protein mRNA expression and their relationship. Shock 7: 206–210, 1997. 16. Jones, S. B., and F. D. Romano. Dose- and time-dependent changes in plasma catecholamines in response to endotoxin in conscious rats. Circ. Shock 28: 59–68, 1989. 17. Kimura, R. E., T. R. LaPine, and W. M. Gooch. Portal venous and aortic glucose and lactate changes in a chronically catheterized rat. Pediatr. Res. 23: 235–240, 1988. 18. Kotanidou, A., A. M. Choi, R. A. Winchurch, L. Otterbein, and H. E. Fessler. Urethan anesthesia protects rats against lethal endotoxemia and reduces TNF-alpha release. J. Appl. Physiol. 81: 2305–2311, 1996. 19. Luce, J. M., A. B. Montgomery, J. D. Marks, J. Turner, C. A. Metz, and J. F. Murray. Ineffectiveness of high-dose methylprednisolone in preventing parenchymal lung injury and improving mortality in patients with septic shock. Am. Rev. Respir. Dis. 138: 62–68, 1988. 20. Martich, G. D., A. J. Boujoukos, and A. F. Suffredini. Response of man to endotoxin. Immunobiology 187: 403–416, 1993. 21. Michie, H. R., K. R. Manogue, D. R. Spriggs, A. Revhaug, S. O. Dwyer, C. A. Dinarello, A. Cerami, S. M. Wolff, and D. W. Wilmore. Detection of circulating tumor necrosis factor after endotoxin administration. N. Engl. J. Med. 318: 1481– 1486, 1988. 22. Ribeiro, S. P., J. Villar, G. P. Downey, J. D. Edelson, and A. S. Slutsky. Effects of the stress response in septic rats and LPS-stimulated alveolar macrophages: evidence for TNF-alpha posttranslational regulation. Am. J. Respir. Crit. Care Med. 154: 1843–1850, 1996. 23. Scheffe´, H. Alternative models for the analysis of variance. Ann. Math. Stat. 27: 251–271, 1956. 24. Spengler, R. N., S. W. Chensue, D. A. Giacherio, N. Blenk, and S. L. Kunkel. Endogenous norepinephrine regulates tumor necrosis factor-alpha production from macrophages in vitro. J. Immunol. 152: 3024–3031, 1994. 25. Sprung, C. L., P. V. Caralis, E. H. Marcial, M. Pierce, M. A. Gelbard, W. M. Long, R. C. Duncan, M. D. Tendler, and M. Karpf. The effects of high-dose corticosteroids in patients with septic shock. A prospective, controlled study. N. Engl. J. Med. 311: 1137–1143, 1984. 26. Tracey, K. J., and A. Cerami. Tumor necrosis factor: an updated review of its biology. Crit. Care Med. 21, Suppl 10: S415–S422, 1993. 27. Uhing, M. R., and R. E. Kimura. The effect of surgical bowel manipulation and anesthesia on intestinal glucose absorption in rats. J. Clin. Invest. 95: 2790–2798, 1995. 28. Van der Poll, T., S. M. Coyle, K. Barbosa, C. C. Braxton, and S. F. Lowry. Epinephrine inhibits tumor necrosis factor-alpha

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (178.171.097.009) on May 28, 2018. Copyright © 1999 American Physiological Society. All rights reserved.

H678

NONSTRESSED RAT MODEL OF ACUTE ENDOTOXEMIA

and potentiates interleukin 10 production during human endotoxemia. J. Clin. Invest. 97: 713–719, 1996. 29. Van der Poll, T., and S. F. Lowry. Tumor necrosis factor in sepsis: mediator of multiple organ failure or essential part of host defense? Shock 3: 1–12, 1995. 30. Waage, A. Production and clearance of tumor necrosis factor in rats exposed to endotoxin and dexamethasone. Clin. Immunol. Immunopathol. 45: 348–355, 1987. 31. Xie, J., K. O. Joseph, G. J. Bagby, T. D. Giles, and S. S. Greenberg. Dissociation of TNF-alpha from endotoxin-induced nitric oxide and acute-phase hypotension. Am. J. Physiol. 273 (Heart Circ. Physiol. 42): H164–H174, 1997.

32. Zellweger, R., A. Ayala, X. L. Zhu, M. H. Morrison, and I. H. Chaudry. Effect of surgical trauma on splenocyte and peritoneal macrophage immune function. J. Trauma 39: 645–650, 1995. 33. Zuckerman, S. H., and A. M. Bendele. Regulation of serum tumor necrosis factor in glucocorticoid-sensitive and resistant rodent endotoxin shock models. Infect. Immun. 57: 3009–3013, 1989. 34. Zuckerman, S. H., J. Shellhaas, and L. D. Butler. Differential regulation of lipopolysaccharide-induced interleukin 1 and tumor necrosis factor synthesis: effects of endogenous and exogenous glucocorticoids and the role of the pituitary-adrenal axis. Eur. J. Immunol. 19: 301–305, 1989.

Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (178.171.097.009) on May 28, 2018. Copyright © 1999 American Physiological Society. All rights reserved.