The Role of Endorphins and Vasopressin in Canine Endotoxin Shock’

JOURNAL OF SURGICAL RESEARCH 41,609-6 19 (1986)

JACK L. CRONENWETT, M.D., *,’ BETH S. BAVER-NEFF, B.S.,* ROGER J. GREKIN, M.D.,? AND JOHN N. SHEAGREN, M.D.?

Departments of *Surgery and TMedicine, the Veterans Administration Medical Center and the University of Michigan, Ann Arbor, Michigan 48104 Submitted for publication April 22, 1985 Chemical antagonists were used to assessthe role of @-endorphin and arginine-vasopressin (AVP) in canine endotoxin shock. Fifteen awake dogs were given Escherichia coli endotoxin IV. Within 5 min, CO decreased to 282, LV dP/dt to 46%, and MAP to 52% of baseline. Fifteen minutes after endotoxin, five do8s each received naloxone, AVP antagonist, or no treatment. Control (untreated) animals exhibited persistent cardiovascular depression, with CO 49%, LV dP/dt 691, and MAP 9 1% of baseline after 45 min. Naloxone improved CO to 69%, LV dP/dt to 94%, and MAP to 91% by 30 min after treatment. AVP blockade improved CO to 105%, LV dP/dt to 107%, and h4AP to 95% of baseline by 30 min after treatment, and caused significant tachycardia. Plasma cortisol and AVP increased markedly in all groups after endotoxin administration. AVP antagonist treatment increased mean survival from 1.4 to 4 days. These data suggest that abnormally elevated AVP contributes to cardiovascular depression in canine endotoxin shock and that AVP blockade is therapeutic in the animal model studied. o 1986

INTRODUCTION endotoxin or hemorrhagic shock, much more Endotoxin administration in dogs causes than occurs during water deprivation [28,44]. profound cardiovascular collapse, character- Furthermore, AVP has been shown to cause ized by systemic hypotension and low cardiac myocardial depression in dogs, both in vivo output [ 171. Although the mechanism of this and in an isolated left ventricle preparation response is not completely understood, a va- [29, 431. These observations suggested the riety of factors including hepatic venocon- possibility that excessively high AVP activity striction, reduced systemic venous return, cir- in endotoxin shock might contribute substan- culating myocardial depressant factor, and di- tially to cardiac depression. Our study was de- rect myocardial effects, have been implicated signed to test this hypothesis by treating en- [16, 18, 19, 25, 341. Opiate antagonism with dotoxin shocked dogs with a specific AVP an- naloxone has been found to increase survival tagonist and comparing the observed hemo- and improve cardiac performance in rats and dynamic changes and survival with untreated dogs during endotoxin shock [7,35,36]. Since control animals. Because of potential inter- opiates exert a centrally mediated myocardial actions between AVP and P-endorphin, a third depressant effect [24], these studies suggested group of naloxone-treated dogs was studied that endogenous opiates (endorphins), known to allow a comparison with AVP antagonist to be elevated in septic shock [5], might con- effects. tribute to hemodynamic deterioration. Plasma levels of arginine-vasopressin METHODS (AVP) also increase markedly in dogs during Fifteen mongrel dogs of both sexes (19-33 ’ This work was supported in part by the Veterans Ad- kg), free of dirofilaria and endoparasites, were ministration Medical Research Service, Grant HL18575 briefly anesthetized (thiamylal sodium) for of the NHLBI, and the NIH Student Research Program. * To whom reprint requests should be addressed: De- catheter placement 1 day prior to the actual partment of Surgery, Dartmouth-Hitchcock Medical experiment. Using sterile surgical technique, Center, Hanover, NH 03756. a cervical incision was used to insert a trans-

609 0022-4804186 $1.50 Copyright 0 1986 by Academic FTes, Inc. All rights of reproduction in any form reserved. 610 JOURNAL OF SURGICAL RESEARCH: VOL. 41, NO. 6, DECEMBER 1986 jugular Swan-Ganz thermodilution catheter mg/kg/hr iv for 3 hr; five additional dogs re- in the pulmonary artery. A carotid artery ceived the A VP antagonist I-(B(mercapto+, catheter and a transcarotid left ventricular /3-cyclopentamethylenepropionic acid), 2- pigtail catheter were inserted through the same (O-methyl) tyrosine arginine-vasopressin incision. After the catheters were secured in (Bachem, Inc.), 10 pg/kg iv [22]. These doses place, the animals were allowed to recover were selected because of demonstrated efficacy from anesthesia and given free access to food in previous studies [36, 371. Both naloxone and water. The animal research protocol fol- and AVP antagonist were dissolved and ad- lowed in this study was approved by the Sub- ministered in saline, 2 ml/kg. All animals were committee for Animal Research, Ann Arbor monitored for 180 min after endotoxin ad- Veterans Administration Medical Center. ministration and received normal saline, 1 ml/ One day after catheter placement, fully kg/hr, during this period. After this observa- awake and normal-appearing animals were tion period, the dogs were returned to their placed in a modified Pavlov stand. Mean sys- cages and allowed free access to food and wa- temic arterial pressure (MAP) and left ven- ter, but were not treated with antibiotics or iv tricular (LV) pressure were continuously re- fluids. Catheters were removed under brief corded. Heart rate was measured with an am- thiamylal sodium anesthesia on the day fol- plifier triggered by the LV pressure pulse. The lowing the experiment. Survival was recorded maximum value during the cardiac cycle of for 7 days, at which time any surviving animal the first derivative of LV pressure (dP/dt max, was sacrificed. hereafter, dP/dt) was electronically derived Blood samples at baseline, 30 and 90 min from the LV pressure signal. Cardiac output after endotoxin administration were obtained. (CO) and temperature were intermittently re- White blood cell (WBC) count with differen- corded using the thermodilution Swan-Ganz tial, platelet count, and hematocrit were de- catheter. Hemodynamic variables were ana- termined. Ionized calcium was measured from lyzed and reported at 15-min intervals during heparinized blood immediately after sampling the 3-hr observation period, in addition to (Orion SS-20 Ionized Calcium Analyzer). To- measurements made 5 min after endotoxin tal hemolytic complement was assessed by the administration and 5 min after drug treatment. percentage lysis of sensitized sheep red blood Stroke volume (SV) was calculated as SV cells (RBC) by dog serum. Lysis of 60 X lo6 = CO/HR (ml/beat). Total peripheral resis- RBC by a 1:40 dilution of dog serum after 1 tance (TPR) was estimated as TPR = MAP/ hr incubation at 37°C was compared to stan- CO [(mm Hg/liter/min) X 80 = dyne * sec. dard (100%) lysis by distilled water using op- cme5], under the assumption that venous tical density at 540 nm. Plasma epinephrine, pressure contributed negligibly to the result. norepinephrine and cortisol were measured by Blood samples obtained through the carotid radioimmunoassay [23,32]. Plasma AVP was artery catheter were replaced with an equal measured by radioimmunoassay from samples volume of normal saline. extracted using Sep pak C, s cartridges (Waters After the awake dogs had stabilized for 30 Associates). Separation of free from bound min, baseline hemodynamic measurements AVP was achieved by adding rabbit -y-globulin and blood samples were obtained. Animals and polyethylene glycol [33]. Vasopressin an- were then given Escherichia coli endotoxin tagonist was tested for cross reactivity in the intravenously (0.75 mg/kg IV, Sigma No. L- AVP radioimmunoassay system. The antag- 3 124, serotype 0 127:BB, phenol extracted), an onist showed a 0.46% cross reactivity with LDrm dose at 48 hr in pilot studies. Fifteen AVP standard. minutes after endotoxin administration, Hematologic variables were compared treatment was initiated: five control dogs re- within treated groups at baseline, 30 and 90 ceived only normal saline, 2 ml/kg, iv; five min by paired Student’s t test, and between animals received naloxone, 2 mg/kg iv + 2 treatment groups by non-paired Student’s t CRONENWETT ET AL.: ENDORPHINS AND VASOPRESSIN 611

TABLE 1 HEMODYNAMIC PARAMETERS AT BASELINE

Treatment

Vasopressin Control Naloxone antagonist F value P value

Cardiac output (liter/min) 4.2 -t 0.4 3.8 + 0.8 2.1 k 0.3 2.5 0.13 MAP (mm I-k) 112+3 117+7 113k6 0.2 0.80 Heart rate (beats/min) llO-+ 11 90*9 111+14 1.0 0.41 LV dP/dt (mm Hg X lO’/sec) 3.1 kO.3 2.8 + 0.3 3.1 kO.3 0.2 0.79 Stroke volume W/beat) 4os4 44k9 26+4 2.1 0.16 TPR (dyne. sec. cme5) 2240 + 240 2800 + 480 352Ok400 2.6 0.11 Temperature (“C) 39.4 + 0.1 40.0 + 0.2 39.4 k 0.2 2.9 0.10

Note. X f SEA4 F, P values between groups by analysis of variance. Abbreviations: MAP, mean arterial pressure; LV dp/dt, left ventricular pressure-time derivative; TPR, total peripheral resistance. test. Hemodynamic parameters were com- within 5 min of treatment (T = 20 min). The pared between treatment groups and across difference between naloxone-treated and con- time by multiple analysis of variance. These trol animals (P = 0.08) gradually diminished, data were then expressed as the ratio to initial so that CO in both groups was 6570% of baseline measurement, to eliminate large in- baseline by T = 180 min. Animals treated with teranimal variations. AVP antagonist showed a rapid normalization of CO, within 10 min of treatment. CO re- RESULTS

Baseline hemodynamic parameters showed no statistically significant variance between treatment groups prior to endotoxin administration (Table 1). Apparent differences were due to large baseline variation among all dogs rather than systematic differences between drug treatment groups that were randomly assigned. Immediately after endotoxin admin- TREATMENT istration at T (time) = 0, CO fell to 28% of its D CONTROL initial level (Fig. 1). After this initial decrease, . -----NALOXONE there was a gradual improvement in control . WP ANTAGONIST animals, so that CO achieved 60% of its baseline level by 90 min and remained at this level during the subsequent 90 min observation pe- PIG. I. CO fell markedly after endotoxin (E), but re- riod. Naloxone-treated animals demonstrated turned to normal atIer drug treatment (T) with AVP an- a more rapid improvement in CO than control tagonist, significantly better than control (P < 0.01) or animals, with a return of CO to 80% of baseline naloxone (P -z 0.03) groups. 612 JOURNAL OF SURGICAL RESEARCH: VOL. 41, NO. 6, DECEMBER 1986

mained greater than baseline throughout the experiment, being significantly greater than control (P < 0.0 1) or naloxone-treated (P < 0.03) dogs. LV dP/dt decreased to 46% of baseline immediately after endotoxin administration (Fig. 2). After treatment with naloxone and AVP antagonist, LV dP/dt returned to baseline level TREATMENT within 5- 10 min. These drug treatment groups q .._._____._.__-CONTROL were statistically comparable, but were both . -----NALOXONE different from control dogs, where LV dP/dt ‘3 . AVP ANTAGONIST returned gradually toward normal only after O.B! , , , , , , , , , , Tk 90 105120 135 IS0 165 110 120 min (naloxone > control, P < 0.01; AVP (Minutes) antagonist > control, P < 0.05). All dogs experienced tachycardia, beginning FIG. 3. Tachycardia developed after endotoxin (E) in all treatment (T) groups. AVP antagonist treated dogs de- approximately 20 min after endotoxin ad- veloped significantly higher HR than control or naloxone ministration (Fig. 3). AVP antagonist treat- dogs (both P < 0.05). ment was associated with the most rapid and profound tachycardia, beginning immediately after treatment and reaching a level 1.8 times tachycardia that developed in all groups. Nal- baseline after 60 min. Tachycardia in AVP oxone and AVP antagonist groups appeared antagonist treated dogs was significantly to have a higher SV than control animals, greater than control or naloxone-treated especially for the first 90 min after treatment, groups (P < 0.05). Although naloxone-treated but this was not statistically significant. dogs appeared to develop a more rapid heart After decreasing to 52% of baseline level rate than control animals, this was not statis- immediately after endotoxin administration, tically significant. After an initial marked de- MAP remained significantly depressed in crease, SV remained significantly depressed in control animals for the next 90 min, increasing all dogs throughout the 180 min observation gradually toward normal thereafter (Fig. 5). period (Fig. 4). This related primarily to Treatment with naloxone or AVP antagonist

TREATMENT

. AVP ANTAGONIST

TREATMENT

q ______.-.--.__.CONTROL

i . -----NALOXONE . A”P ANTAGONIST

90 105 00 135’ 150 165 1.30 Time (Minutes)

FIG. 2. LV dP/dt decreased 50% after endotoxin (E). FIG. 4. After endotoxin (E), SV fell markedly and re- Compared to control dogs, LV dP/dt returned rapidly to mained depressed. Drug treatment (T) with naloxone and normal after drug treatment (T) with naloxone (P < 0.05) AVP antagonist appeared to initially improve SV, although or AVP antagonist (P -c 0.01). this was not statistically significant. CRONENWETT ET AL.: ENDORPHINS AND VASOPRESSIN 613

baseline in both naloxone and control animals throughout the experiment. Dogs treated with AVP antagonist maintained their TPR below baseline, significantly different than control (P < 0.04) or naloxone (P < 0.02) groups. Tem- perature increased progressively throughout the 180 min observation period after endo-

TREATMENT toxin administration, being statistically com- m __-CONTROL .____.______parable in all groups (Fig. 7). . NALOXONE----- Hematologic variables showed minor, but . AVP ANTAGONIST occasionally statistically significant differences between the three treatment groups at baseline (Table 2). Since treatment groups were randomly assigned, these differences appeared to FIG. 5. MAP decreased rapidly after endotoxin (E), re- result from large variations among individual turning nearly to baseline level much more rapidly after naloxone (P <: 0.03) or AVP antagonist (P < 0.01) treat- dogs that were not eliminated by the small ment (T) than in control, untreated dogs. number of animals in each group. In all cases, these differences were small compared to the subsequent changes observed after experi- rapidly increased MAP to 90- 100% of baseline mental manipulation. After endotoxin ad- value. This improvement was statistically dif- ministration, WBC count decreased markedly ferent from control animals, where MAP re- (by 30 min) and remained depressed at the 90 turned to 90% of baseline only after 150 min min sample time. This decrease was compa- (naloxone > control, P < 0.03; AVP antago- rable in the three treatment groups, with leu- nist > control, P < 0.01). After endotoxin ad- kocyte counts decreasing to 15% of their base- ministration, TPR transiently increased, dur- line value. Differential counting revealed that ing the early, transient cardiovascular collapse the leukopenia was due to a reduction in het- (Fig. 6). TPR then decreased toward baseline, erophilic leukocytes (polymorphonuclear but remained elevated from 1.2 to 1.5 times forms), which decreased from 85% of the total leukocyte count at baseline to 38% of the leukocyte population 30 min after endotoxin ad- 2.5 1 TREATMENT q CONTROL_-.._____-----_ . NALOXONE----- . AVP ANTAGONIST

0: 15: JO 45 60 75 90 105 I20 135 150 16.5 190 FIG. 6. TPR increased transiently after endotoxin (E), Time (Minutes) and remained significantly higher in control (I’ < 0.04) and naloxone (P < 0.02) dogs, than in dogs treated (T) FIG. 7. Temperature increased gradually in ah dogs aher with AVP antagonist, where TPR was lower than baseline endotoxin (E), with no significant difference between after the initial transient increase. treatment (T) groups. 614 JOURNAL OF SURGICAL RESEARCH: VOL. 41, NO. 6, DECEMBER 1986

TABLE 2 HEMATOLOGIC PARAMETERS

Time Treatment group Baseline 30 min 90 min

WBC C 14.2 f 1.9 t 1.3kO.3 t 1.4kO.3 (cells/mm3) N 22.5 f 2.7 t 4.4 f 0.8 t 2.5 f 0.5 V 14.3 + 1.7 t 1.9kO.6 t 3.0 f 1.2 PMN C 84+2 t28+8 t26?11 (%I N 83?4 t45*11 t40k8 V 87?3 t40* 13 t54+8 Hematocrit C 39a 1 t44-c1 iSOk (Vol %) N 44+2 *49*3 ?51&2 V 48 f 3 t56f3 t58+2 Platelets C 199+ 17 t72+11 t99* 15 (cells/mm3) N 95215 t46flO t46+8 V 135 + 16 t49+ 17 t 53 -+ 12 Ionized calcium C 2.53 f 0.08 t 2.46 -+ 0.10 t 2.30 f 0.13 (mEq/liter) N 2.5 1 f 0.05 2.48 -+ 0.05 2.29aO.11 V 2.47 k 0.05 t 2.3 1 2 0.05 t 2.24 k 0.07 Epinephrine C 100+23 * 8364 k 3910 t4519+1256 (t-MN N 263 + 53 * 620712644 t4986k 1891 V 283 f 59 * 12332 f 5031 t 10174+3046 Norepinephrine C 133*9 t 1497f519 t 1216f206 bx/dl) N 224 f 46 * 1246f608 18182822 V 212 f 36 t 1645 + 272 t 1803 +271 Hemolytic complement C 64f 12 t 26 f 10 16?8 (% RBC lysis) N 64f 13 *44*15 402 12 V 39+ 14 21+9 23+ 11 Cortisol C 7.8 k 0.9 t 15.1 k2.1 t 16.4 -t 1.3 (/e/dl) N 6.2 + 1.6 t 12.0 + 2.8 t 14.21 1.1 V 11.9k2.2 t 19.7k2.1 t 25.3 zk2.9 Vasopressin C 5.5 f 0.3 t 140+40 t5822 Wml) N 8.5 XL 1.7 t 103+ 15 68f 12 V 13.2f5.1 t507+49 ?232f24

Note. X + SEM. Within group statistics (paired Student’s t test): * P < 0.10, t P < 0.05, t P < 0.01, comparison of baseline vs 30- and 90-min samples. Between group statistics (non-paired Student’s t test): baseline: WBC, N > C,V, P < 0.04; hematocrit, V > C, P < 0.03; platelets, C > N,V, P < 0.03; epinephrine, N,V > C, P < 0.03. 30 min: ionized calcium, C,N > V, P < 0.05; vasopressin, V > C,N, P < 0.01. 90 min: PMN, V > C, P = 0.06;cortisol, V > C,N, P < 0.01; vasopressin, V > C,N, P < 0.01. Abbreviations: C, control, N, naloxone; V, vasopressin antagonist; WBC, white blood cells; PMN, polymorphonuclear leukocytes; RBC, red blood cells.

ministration (Table 2). Lymphocytic and groups had decreased to 39% of baseline, dif- mononuclear cells were preserved. Peripheral ferences that were statistically significant and platelet count also decreased significantly in that persisted for 90 min. Hematocrit in- all treatment groups after endotoxin admin- creased progressively in all dogs during the istration. After 30 min, platelet counts in all study, reaching levels of 123% of baseline CRONENWETT ET AL.: ENDORPHINS AND VASOPRESSIN 615 by 90 min after endotoxin administration decreasing to 10 times baseline by 90 min. Al- (Table 2). though this 30 min AVP value appeared to be During this study, ionized calcium de- less than that seen in control animals, this was creased slightly in all dogs, to 9 1% of baseline not statistically significant (Table 2). AVP an- by 90 min after endotoxin administration. tagonist treated dogs showed a much greater This decrease was statistically significant only (6 l-fold) increase in plasma AVP at 30 min for control and AVP antagonist animals. Dogs and a 28-fold increase at 90 min, statistically treated with AVP antagonist showed a more greater than both control and naloxone-treated rapid decrease in ionized calcium, significantly dogs (P < 0.01). Due to the cross-reactivity of different from control and naloxone animals AVP with the AVP antagonist, however, it is at 30 min (P c 0.05, Table 2). Epinephrine estimated that this difference represented the and norepinephrine showed considerable amount of antagonist administered. Assuming variance between individual dogs, but both an initial volume of distribution for AVP an- catecholamines increased profoundly in all tagonist of 67 ml/kg [22], the 10 /*g/kg dose treatment groups by 30 min after endotoxin of antagonist would have resulted in a mea- administration, being statistically equivalent surable 685 pg/ml AVP activity, based on the in the three treatment groups (Table 2). 0.46% cross-reactivity. Assuming that the vol- Total hemolytic complement was activated ume of distribution increased and noting the in all treatment groups, as indicated by a de- antagonist half-life of 3-4 hr [22], the antag- crease in percent lysis of sensitized sheep red onist activity measured 15 and 75 min after blood cells at 30 and 90 min after endotoxin administration is sufficient to explain the dif- administration (Table 2). There appeared to ference in AVP activity measured between be more complement activation in control control or naloxone dogs and the AVP antag- animals than in naloxone or AVP antagonist onist group. treated dogs. At 90 min, available hemolytic Although there was no statistically signifi- complement had decreased to 25% of its base- cant difference in survival between treatment line level in control dogs, compared with de- groups, there was a tendency for improved creases to only 63 and 59% of baseline in‘nal- survival among AVP antagonist treated dogs oxone- and AVP antagonist-treated dogs. Due (Fig. 8). Mean survival was 1.4 days in control, to the large variance among all animals, how- 2.2 days in naloxone, and 4.0 days in AVP ever, these changes were not statistically sig- antagonist groups. Median survival of 1,2, and nificant. Plasma cortisol increased after endotoxin administration in all treatment groups, increasing 1.8 times baseline by 30 min, and to more than twice baseline by 90 min. Although AVP antagonist-treated dogs had significantly greater cortisol levels by 90 min than both naloxone (P < 0.01) and control (P < 0.03) dogs, the percentage change compared to baseline was comparable in these three groups (increases of 210% in control, 230% in naloxone and 2 12% in AVP antagonist groups). Plasma AVP increased markedly after endotoxin administration. By 30 min, AVP in control animals had increased 27 times baseline, with a persistent elevation of 11 times FIG. 8. Survival after E. coli endotoxin shock (0.75 mg/ baseline at 90 min. Dogs treated with naloxone kg, LDloo at 48 hr) in dogs treated with naloxone, AVP showed a 1Cfold increase in AVP by 30 min, antagonist, or saline (control). 616 JOURNAL OF SURGICAL RESEARCH: VOL. 41, NO. 6, DECEMBER 1986

3 days, respectively, for these treatment groups return. These stimuli increase plasma AVP also suggested an improved survival after AVP concentration in excess of those normally re- antagonist treatment, and intermediate sur- quired to achieve osmotic or fluid balance vival after naloxone treatment. The only long- during water deprivation [44]. Such observa- term survivor among these dogs was in the tions have led to increased recognition of a AVP antagonist group, and this animal ap- role for AVP in cardiovascular control [29]. peared healthy when sacrificed at 7 days. Dogs In fact, AVP blockade in dogs subjected to treated with AVP antagonist were subjectively mild hemorrhage results in significantly greater improved immediately after drug treatment hypotension, suggesting that AVP has a sup- when compared with both control and nal- portive cardiovascular effect during mild oxone dogs. AVP antagonist treated dogs were hemorrhage. Data from our experiment, how- less lethargic and had fewer symptoms of gas- ever, suggest that in canine endotoxin shock, trointestinal emptying than did control or AVP secretion has gone awry; that is, excessive naloxone-treated animals. concentrations of AVP resulted in detrimental cardiovascular effects. DISCUSSION The AVP antagonist used in our study has been shown to predominantly block pressor In this study, E. coli endotoxin administered rather than antidiuretic effects of AVP [22]. It to awake dogs caused a rapid, 25-fold increase is therefore a useful agent to probe the poten- in plasma AVP levels. Wilson et al. have pre- tial cardiovascular effects of AVP in shock. To viously reported increased AVP levels in dogs our knowledge, this report is the first docu- and baboons after endotoxin or live E. coli mentation that AVP blockade after endotoxin shock. In their study, the rapid increase in administration significantly improves cardiac AVP appeared to be stimulated by the effect performance in awake dogs. In fact, AVP an- of decreased venous return on left atrial stretch tagonist administered 15 min after endotoxin receptors, since MAP had not yet decreased actually normalized CO and MAP by im- sufficiently to involve baroreceptors [44]. proving LV dP/dt and increasing HR. Tachy- These investigators found peak AVP concen- cardia seen after AVP antagonist may repre- trations 30 min after endotoxin administration sent an increased cardiac responsiveness to that were equal to or greater than those found circulating epinephrine, shown to be signifi- in our study. At such high concentrations, cantly elevated in these dogs. Archer et al. have AVP is known to have cardiac depressant ef- shown that perfusion of an isolated canine fects [ 111. Even modest doses of AVP reduce heart with endotoxin significantly reduces the CO and increase TPR in normal, conscious normal cardiac response to epinephrine [I]. dogs [29]. In an isolated canine left ventricle Furthermore, Wilson et al. have shown that preparation, Wilson et al. showed that AVP epinephrine can reverse myocardial depression infusion has direct myocardial depressant ef- caused by AVP in the isolated canine left ven- fects, decreasing contractility despite relatively tricle [43]. normal coronary blood flow [43]. Clinically, Since the initial reports of Faden and Hol- myocardial dysfunction with reduced cardiac aday, naloxone treatment of endotoxin and output is a side effect of AVP infusion used to septic shock has received considerable exper- treat gastrointestinal hemorrhage [38]. imental and clinical support [7, 8, 311. In our The above observations promote the hy- study, hemodynamic improvements seen after pothesis that excessively high AVP levels in naloxone treatment agree with reports of the septic or endotoxin shock might actually have salutary effects of opiate antagonism in these detrimental cardiovascular effects. In this set- studies [35, 361. Although /3-endorphin was ting, AVP is secreted in response to barore- not directly measured in our experiment, it ceptor or left atria1 stretch receptor stimuli would predictably have increased, based on provoked by hypotension or decreased venous the measured increases in plasma cortisol, CRONENWETT ET AL.: ENDORPHINS AND VASOPRESSIN 617

since P-endorphin and ACTH are known to eters were measured. As expected, endotoxin be secreted concomitantly from the same pre- induced marked leukopenia and thrombocy- cursor [ 15, 271. Experimental evidence of a topenia, associated with complement activa- close interrelationship between P-endorphin tion. These effects are well documented in this and vasopressin secretion has prompted spec- and other endotoxin shock models [30], and ulation that vasopressin, in fact, might mediate were largely unaffected by the drug treatment effects attributed to &endorphin [ 131. In rab- used in the study. Although not statistically bits, both IV and third ventricular injection significant, there appeared to be less comple- of /3-endorphin stimulated a rapid increase in ment activation and less depression of poly- plasma AVP concentration [9,41]. Naloxone morphonuclear leukocytes in both naloxone- did not block this increased AVP secretion in and AVP antagonist-treated dogs. This trend rabbits, suggesting that a naloxone indepen- suggests that improved cardiovascular perfor- dent opiate receptor is responsible [9]. In man, mance might reduce complement activation, however, naloxone has been shown to decrease which is otherwise aggravated by reduced cap- plasma AVP concentration and prevent AVP illary blood flow. increases normally caused by orthostatic hy- Reduced ionized calcium in baboon septic potension [26]. Thus, data exist to indicate that shock has been reported by Trunkey et al., fi-endorphin and naloxone may have different who concluded that calcium reduction con- effects on AVP secretion in different species tributes to decreased myocardial contractility and after different stimuli [ 14, 20, 2 11. [40]. Although reduced ionized calcium is as- Based on the above studies, a complex in- sociated with septic shock, in some cases this teraction appears to exist between AVP and reduction might indicate a favorable move- &endorphin. Recent experiments suggest that ment of calcium into cells, essential for the endogenous opiates may, in fact, promote action of certain hormones, such as catechol- AVP secretion by inhibiting oxytocin release amines [ lo]. The early decrease in serum ion- [2, 391. Furthermore, Gillies et al. reported ized calcium seen in AVP antagonist-treated that AVP may act as a corticotrophin releasing dogs in our study was associated with positive factor in the rat and, therefore, may be capable hemodynamic results, and thus, might reflect of promoting endorphin secretion [ 121. Based favorable cellular calcium uptake and utili- on this potential interaction of AVP and fi- zation. endorphin, we were stimulated to measure Although there appears to be ample support cortisol and AVP concentration in both nal- for a direct cardiac mechanism to explain fa- oxone- and AVP antagonist-treated dogs. Our vorable cardiovascular effects produced by data indicate that the hemodynamic effects AVP blockade in this study, other explanations seen after naloxone treatment were not me- are possible. In dogs, endotoxin produces diated by AVP inhibition, since AVP was marked mesenteric vasoconstriction, as op- markedly elevated despite naloxone treatment, posed to primates, where slight vasodilation Conversely, increased cortisol in dogs receiving may occur [4]. In fact, sequelae of mesenteric AVP antagonist indicates that AVP blockade &hernia are a prominant cause of mortality did not achieve its hemodynamic effects by in the canine model. Despite improved early reducing /3-endorphin secretion. survival of dogs in endotoxin shock after nal- Despite extensive investigation, the under- oxone treatment, significant mesenteric isch- lying mechanisms for cardiac depression and emia persists, prompting Reynolds et al. to hypotension in endotoxin shock are unclear. suggest that severe intestinal ischemic damage Although AVP appears to have a significant might abrogate the early beneficial effects of role in the current model, its interaction with naloxone [ 361. High plasma AVP levels found other vasoactive factors is undoubtedly com- in our study suggest a potential role for AVP plex. In an attempt to further characterize our in promoting mesenteric vasoconstriction. model, a variety of other hematologic param- Relief of this vasoconstriction by AVP block- 618 JOURNAL OF SURGICAL RESEARCH: VOL. 41, NO. 6, DECEMBER 1986

ade has the potential to reduce this contribu- 2. Bicknell, R. J., and Leng, G. Endogenous opiates reg- tion to mortality in canine endotoxin shock. ulate oxytocin but not vasopressin secretion from the Furthermore, myocardial depressant factors, neurohypophysis. Nature (London) 298: 161, 1982. 3. Brackett, D. J., Schaefer, C. F., and Wilson, M. F. suggested by Lefer et al. to originate in isch- The role of vasopressin in the maintenance of cardio- emit intestine during shock [25], might be re- vascular function during early endotoxin shock. Adv. duced. If this mechanism proves to be impor- Shock Rex 9: 147, 1983. tant for the favorable effects of AVP blockade 4. Brobmann, G. F., Ulano, H. B., Hinshaw, L. B., and Jacobson, E. D. Mesenteric vascular responses to en- in canine endotoxin shock, considerable cau- dotoxin in the monkey and dog. Amer. J. Physiol. tion should be exercised in generalizing these 219: 1464, 1970. results to primates, where significant splanch- 5. Carr, D. B., Bergland, R., Kasting, N., Arnold, M., nit vasoconstriction does not occur during and Martin, J. B. Endotoxin-stimulated opioid peptide endotoxin shock. secretion: Two secretory pools and feedback control in vivo. Science 217: 845, 1982. It would appear teleologically unsound for 6. Cowley, A. W., Switzer, S. J., Jr., and Guinn, M. M. elevated AVP in endotoxin shock to precipi- Evidence and quantification of the vasopressin arterial tate myocardial depression. Early cardiovas- pressure control system in the dog. Circ. Rex 46: 58, cular collapse during endotoxin shock in rats 1980. and dogs, however, is primarily due to hepatic 7. Faden, A. I., and Holaday, J. W. Experimental endotoxin shock: The pathophysiologic function of en- venous pooling that probably stimulates early dorphins and treatment with opiate antagonists. J. AVP release [19, 42, 441. As hepatic venous Infect. Dis. 142: 229, 1980. congestion resolves, this initially favorable 8. Feuerstein, G., Chiueh, C. C., and Kopin, I. J. Effect AVP response might become excessive. It is of naloxone on the cardiovascular and sympathetic known that AVP depresses renin secretion, an response to hypovolemic hypotension in the rat. Eur. alternative pressor mechanism with more J. Pharmacol. 75: 65, 198 1. 9. Firemark, H. M., and Weitzman, R. E. Effectsof j3- beneficial cardiac effects [28]. AVP also re- endorphin, morphine and naloxone on arginine va- duces cardiac sensitivity to epinephrine [43]. sopressin secretion and the electroencephalogram. Thus, although AVP may be beneficial in ca- Neuroscience 4: 1895, 1979. nine or rat endotoxin shock during early 10. Gann, D. In (Trunkey, D., Carpenter, M. A., and splanchnic venous pooling, continuation of Holcroft, J., Eds.), Ionized calcium and magnesium: The effect of septic shock in the baboon. J. Trauma this response could be detrimental. If this hy- 18: 170, 1978. pothesis is correct, we would predict that AVP Il. Gelman, S., and Ernst, E. E. Nitroprusside prevents blockade prior to endotoxin administration adverse hemodynamic effects of vasopressin. Arch. would have detrimental effects in rats or dogs. Surg. 113: 1465, 1978. 12. Gillies, G., Van Wimersma Greidanus, T. B., and In fact, Brattleboro rats that congenitally lack Lowry, P. J. Characterization of rat stalk median em- AVP demonstrate significantly worse hypo- inence vasopressin and its involvement in adreno- tension after endotoxin administration than corticotropin release. Endocrinology 103: 528, 1978. Sprague-Dawley rats with AVP present [3]. 13. Gold, P. W., Goodwin, F. K., and Reus, V. I. Vaso- Additional experiments are required to clarify pressin in affective illness. Lance1 1: 1233, 1978. 14. Grossman, A., Besser, G. M., Milles, J. J., and Baylis, this hypothesis. P. H. Inhibition of vasopressin release in man by an opiate peptide. Lancet 2: 1108, 1980. ACKNOWLEDGMENTS 15. Guillemin, R., Vargo, T., Rossier, J., et al. &Endor- Gary Robertson, M.D. provided antiserum for the va- phin and adrenocorticotropin are secreted concomi- sopressin radioimmunoassay. Andrew J. Zweifler, M.D. tantly by the pituitary gland. Science 197: 1367, 1977. performed the catecholamine radioimmunoassays. Gerd 16. Hess, M. L., Soulsby, M. E., Davis, J. A., and Brings, 0. Till, M.D. performed the hemolytic complement assays. F. N. The influence of venous return on cardiac me- Morton Brown, Ph.D. provided statistical consultation. chanical and sarcoplasmic reticulum function during Endo Laboratories supplied the naloxone. endotoxemia. Circ. Shock 4: 143, 1971. 17. Hinshaw, L. B., Archer, L. T., Greenfield, L. J., and REFERENCES Geuenter, C. A. Effects of endotoxin on myocardial hemodynamics, performance, and metabolism. Amer. 1. Archer, L. T., Black, M. R., and Hinshaw, L. B. Myo- J. Physiol. 221: 504, 1971. cardial failure with altered response to adrenaline in 18. Hinshaw, L. B., Archer, B. L., Spitzer, J. J., Black, endotoxin shock. Brit. J Pharmacol. 54: 145, 1975. M. R., Peyton, M. D., and Greenfield, L. J. Effectsof CRONENWETT ET AL.: ENDGRPHINS AND VASOPRESSIN 619

coronary hypotension and endotoxin on myocardial Mitch, W. E. Pressor effectof naloxone in septic shock. performance. Amer. J. Physiol. 227: 1051, 1974. Lancet 1: 529, 1981. 19. Hinshaw, L. B., Gilbert, R. P., Kuida, H., and 32. Peuler, J. D., and Johnson, G. A. Simultaneous single Visscher, M. B. Peripheral resistance changes and isotope radioenzymatic assayof plasma norepineph- blood pooling after endotoxin in eviscerated dogs. rine, epinephrine and dopamine. Life Sci. 21: 625, Amer. J. Physiol. 195: 631, 1958. 1977. 20. Knepel, W., Nutto, D., Anhut, H., and Hertting, G. 33. Pierce, E. T., Grekin, R. J., and Mouw, D. R. Efferent Vasopressin and fl-endorpbin release after osmotic and role of ADH in CNS-induced natriuresis. Amer. J. non-osmotic stimuli: Effect of naloxone and dexa- Physiol. 246: F32, 1984. methasone. Eur. J. Pharmacol. 77: 299, 1982. 34. Raffa, J., and Trunkey, D. D. Myocardial depression 2 1. Knepel, W., Nutto, D., and Hertting, G. Evidence for in sepsis. J. Trauma 18: 617, 1978. inhibition by /I-endorphin of vasopressin release dur- 35. Raymond, R. M., Harkema, J. M., Satoffs, W. V., ing foot shock-induced stress in the rat. Neuroendo- and Emerson, T. E., Jr. Effects of naloxone therapy crinology 34: 353, 1982. on hemodynamics and metabolism following a su- 22. Kruszynski, M., Lammek, B., Manning, M., Seto, J., perlethal dosage of Escherichia coli endotoxin in dogs. Haldar, J., and Sawyer, W. H. [1-fi-Mercapto+, &- Surg. Gynecol. Obstet. 152: 159, 198 1. cyclopentamethylenepropionic acid), 2-(G-methyl) 36. Reynolds, D. G., Gurll, N. J., Vargish, T., Lechner, tyrosine] arginine-vasopressin and [ l-(&Mercapto-0, R. B., Faden, A. I., and Holaday, J. W. Blockade of @cyclopcntamethylenepropionic acid)] arginine-va- opiate receptors with naloxone improves survival and sopressin, two highly potent antagonists of the vaso- cardiac performance in canine endotoxic shock. Circ. pressor response to arginine-vasopressin. J. Med. Shock 7: 39, 1980. Chem. 23: 364, 1980. 37. Schwartz, J., and Reid, I. Effect of vasopressin block- 23. Kumar, M. S., Safa, A. M., and Deodhar, S. D. Ra- ade on blood pressure regulation during hemorrhage dioimmunoassay for serum cortisol with 1251-labelled in conscious dogs. Endocrinology 109: 1778, 198 1. ligand: Comparison of three methods. Clin. Chem. 38. Sherman, L. M., Shenoy, S. S., and Cerra, F. B. Se- 22: 1845, 1976. lective intraarterial vasopressin: Clinical efficacyand 24. Laubie, M., Schmitt, H., Vincent, M., and Remond, complications. Ann. Surg. 189: 298, 1979. G. Central cardiovascular effectsof morphinomimetic 39. Summy-Long, J., Miller, D., Rosella-Dampman, L., peptides in dogs. Eur. J. Pharmacol. 46: 67, 1977. Hartman, R., and Emmett, S. Gpioid peptides mediate 25. Lefer, A. M. Properties of cardioinhibitory factors the differential release of vasopressin and oxytocin produced in shock. Fed. Proc. 31: 2734, 1978. from the hypothalamoneurohypophysial system.Fed. 26. Lightman, S. L., and Forsling, M. L. Evidence for Proc. 43: 806, 1984. endogenous opioid control of vasopressin release in 40. Trunkey, D., Carpenter, M. A., and Holcroft, J. Ion- man. J. Clin. Endocrinol. Metab. 50: 569, 1980. ized calcium and magnesium: The effect of septic 27. Mains, R. E., Eipper, B. A., and Ling, N. Common shock in the baboon. J. Trauma 18: 166, 1978. precursor to corticotropins and endorphins. Proc. Natl. 41. Weitzman, R. E., Fisher, D. A., Minick, S., Ling, N., Acad. Sci. USA 74: 3014, 1977. and Guillemin, R. &Endorphin stimulates secretion 28. Malayan, S. A., Ramsay, D. J., Keil, L. C., and Reid, of arginine vasopressin in vivo. Endocrinology 101: I. A. Effects of increases in plasma vasopressin con- 1634, 1977. centration on plasma renin activity, blood pressure, 42. White, F. N., Ross, G., Barajas, L., and Jacobson, heart rate, and plasma corticosteroid concentration E. D. Hemodynamics of endotoxin shock in the rat in conscious dogs. Endocrinology 107: 1899, 1980. and the effectsof phenoxybenzamine. Proc. Sot. Exp. 29. Montani, J. P., Liard, J. F., Schoun, J., and Mohrin, Biol. Med. 122: 1025, 1966. J. Hemodynamic effectsof exogenous and endogenous 43. Wilson, M. F., Bracket& D. J., Archer, L. T., and Hin- vasopressin at low plasma concentrations in conscious shaw, L. B. Mechanisms of impaired cardiac function dogs. Circ. Res. 47: 346, 1980. by vasopressin. Ann. Surg. 191: 494, 1980. 30. Morrison, D. C., and Ulevitch, R. J. The effects of 44. Wilson, M. F., Bracket& D. J., Tompkins, P., Ben- bacterial endotoxins on host mediation systems: A jamin, B., Archer, L. T., and Hinshaw, L. B. Elevated review. Amer. J. Pathol. 93: 527, 1978. plasma vasopressin concentrations during endotoxin 31. Peters, W. P., Johnson, M. W., Friedman, P. A., and and E. coli shock. Adv. Shock Res. 6: 15, 198 1.