Sepsis does not impair tricarboxylic

acid cycle in the heart

RICHARD S. HOTCHKISS, SHENG-KWEI SONG, JEFFREY J. NEIL, ROBERT D. CHEN, JILL K. MANCHESTER, IRENE E. KARL, OLIVER H. LOWRY, AND JOSEPH J. H. ACKERMAN Departments of Anesthesiology, Chemistry, Pediatrics, Pharmacology, and Internal Medicine, Washington University School of Medicine, St. Louis 63110; and Department of Chemistry, Washington University, St. Louis, Missouri 63130

HOTCHKISS, NEIL, ROBERT KARL, OLIVER Sepsis does not

RICHARD S., SHENG-KWEI SONG, JEFFREY J. D. CHEN, JILL K. MANCHESTER, IRENE E. H. LOWRY, AND JOSEPH J. H. ACKERMAN. impair tricarboxylic acid cycle in the heart. Am.

J. Physiol. 260 (Cell Physiol. 29): C50-C57, 1991.-Sepsis has been reported to causemitochondrial dysfunction and inhibition of key enzymesthat regulate the tricarboxylic acid (TCA) cycle. We investigatedthe effect of sepsison high-energy phosphates, glycolytic and TCA cycle intermediates, and specific amino acidsthat are involved in regulating the size of the TCA cycle pool during changesin metabolicstate of the heart. Sepsis was induced in 12 female rats by the cecal ligation and perforation techniqueunder halothane anesthesia;sevencontrol rats underwent cecal manipulation without ligation. At 36-42 h postsurgery,the rats were reanesthetized,the chestwasopened, and the hearts were freeze-clamped.Perchloric acid extracts of the heartswere analyzed with fluorometric enzymatic methods and ‘lP nuclear magnetic resonancespectroscopy.There were no significant differencesin the levels of the TCA cycle intermediates or high-energy phosphatesbetween the septic and control rats. The major metabolic changeswere the 28% decreasein alanine and the 31% decreasein glutamate in the septic hearts comparedwith control (P < 0.05 and P < 0.005, respectively). Phosphocholine, a component of membrane phospholipids,was increasedby 91% in the septic hearts (P < 0.01). We conclude that sepsisdoesnot impair the TCA cycle or inducesignificant cellular ischemiain the heart. The increase in phosphocholinemay representsignificant cellular membrane disruption during sepsis. mitochondrial function; ischemia;cellular membrane;SpragueDawley rats

MYOCARDIAL CONTRACTILITY is a hallmark of sepsis and endotoxemia and is well documented in both human and animal studies (1, 2, 19, 25, 29). The myocardial depression is frequently profound and may be a cause of death in patients who are septic (25). Although it is well established that myocardial contractility is impaired in sepsis and endotoxemia, the etiology of the myocardial depression is not known. Metabolic abnormalities secondary to mitochondrial dysfunction have been proposed as potential factors in the decreased myocardial contractility (2, 4, 20, 32). The effect of sepsis and endotoxin on mitochondrial function in various organs remains extremely controversial (6,31, 32, 37). One reason for the different findings in various studies may relate to the different models that were DECREASED

c50

0363-6143/91

$1.50 Copyright

employed, i.e., endotoxin (l), early effects of sepsis (29), late effects of sepsis (38), isolated mitochondrial preparations (6), and cultured cells (15). The relevance of findings from many of these studies to the metabolic and physiological changes that actually occur in vivo during sepsis can be questioned. Undoubtedly, the septic syndrome results from the effects of a host of mediators (21, 25). Many of these mediators are recruited sequentially into the septic process as the infection progresses. Thus the particular period of time at which sepsis/endotoxemia is investigated may give different results (21). Recently investigators have reported that sepsis has direct effects on tricarboxylic acid cycle (TCA) function. Drapier and Hibbs (9) found that mouse macrophages activated in vivo by infection with Mycobacterium bovis strain Bacillus Calmette-Guerin (BCG) develop inhibition of aconitase, an important enzyme in the TCA cycle. These “activated” macrophages also are able to cause similar inhibition in the TCA cycle of various target cells and thus could potentially induce widespread metabolic abnormalities in other organs (15). Vary and associates (38) have reported decreased activity of pyruvate dehydrogenase, an enzyme that regulates entry of pyruvate into the TCA cycle, in tissues from rats with peritonitis. Burns et al. (2) reported that dichloroacetate, a stimulator of pyruvate dehydrogenase, was able to augment myocardial performance and restore the depressed ATP concentration in isolated perfused hearts from rats treated with endotoxin. Newsholme and colleagues (23) have proposed that impairment in the “second-half’ of the TCA cycle, i.e., the part of the cycle in which CYketoglutarate ultimately forms oxaloacetate, occurs in skeletal muscle during sepsis. The a-ketoglutarate leaves the TCA cycle in skeletal muscle via glutamate dehydrogenase and ultimately is converted to glutamine, which is the preferred substrate of macrophages and lymphocytes that are highly metabolically active during sepsis. Additional indirect evidence supports the concept of endotoxin and/or sepsis-induced mitochondrial abnormalities. Electron microscopic examination of cardiac muscle from endotoxin-treated rats has demonstrated mitochondrial swelling and disruption (4). One of the purposes of the present investigation was to examine the effect of sepsis on selected TCA cycle intermediates in the heart. The levels of selected metabolites will be indicative of TCA cycle function and may point to po-

0 1991 the American

Physiological

Society

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (196.011.235.237) on August 2, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

EFFECT

OF

SEPSIS

ON

TRICARBOXYLIC

tential sites of metabolic control or inhibition (13). A second major purpose of this investigation was to determine whether myocardial ischemia occurs during sepsis. Although global myocardial ischemia is uncommon during sepsis and does not occur until significant arterial hypotension exists (5, 7), a growing number of investigators argue that localized cellular ischemia does occur because of either defective coronary autoregulation (30) or capillary obstruction due to microvascular thrombosis and/or capillary luminal swelling (10, 28). Indeed, many investigators believe that cellular ischemia is a major factor in the development of multiple organ failure in the sepsis syndrome (28). Analysis of selected metabolites including the high-energy phosphates will help determine whether significant cellular ischemia is occurring in the heart during the septic process prior to agonal events. The cecal ligation and perforation technique of inducing sepsis was selected as an appropriate model because a variety of gram-positive and gram-negative bacteria are routinely cultured from the blood of the septic rats (16,39) and the septic process follows a prolonged course over several days, which reflects the clinical condition as it occurs in vivo. The intent of the experiment was to examine metabolic effects at a point when the septic process was well established, 36-42 h postsurgery. If no evidence of TCA cycle dysfunction or myocardial ischemia is detected at this time, it would be unlikely that TCA cycle inhibition or ischemia are major pathophysiological abnormalities in sepsis. MATERIALS

AND

METHODS

Septic model. The cecal ligation and perforation technique was used to induce sepsis in the rats (39). Briefly, female Sprague-Dawley rats (175250 g) were anesthetized with halothane (4% induction, 1.5% maintenance) and oxygen. In the septic rats, the cecum was ligated and punctured twice with a 19-gauge needle. The control rats had exteriorization and manipulation of the cecum without ligation or perforation. The abdominal wall was closed in two layers, and the animals received 10 ml of 0.9% saline solution subcutaneously. After surgery, the animals were housed in metabolic cages and kept fasted but allowed free access to water. (The septic rats eat little to no food, and therefore it is appropriate to keep both groups fasted.) Approximately 36-42 h after surgery, 12 surviving septic rats (there is an -25% mortality in the rats at the 36-h time period) and 7 control rats were anesthetized with halothane. The chest was rapidly opened and the heart frozen in situ with copper tongs precooled in liquid nitrogen. The period of time required to appropriately anesthetize the rats was l-l.5 min. The chest could be opened and the heart freeze-clamped in -5-10 s. At the conclusion of each experiment, the abdominal cavity of the rats was examined. Only data from “septic rats,” defined as those whose abdomen confirmed the presence of foul-smelling, purulent abdominal fluid and a gray-black cecum, were included in the septic group (12 rats). Abdominal examination of control rats revealed minimal peritoneal fluid, and the bowel wall had the normal pink color of healthy tissue.

ACID

CYCLE

OF

HEART

c51

[Blood cultures performed on a previous group of septic and control rats were positive in the septic but not in the control rats for gram-positive and gram-negative bacteria (16) .] Whole blood was obtained from a separate group of seven septic and seven control rats, which were identically prepared as previously described. The rats were anesthetized with halothane and blood withdrawn from the abdominal aorta into a heparinized syringe. Blood was centrifuged and the plasma obtained. Plasma was then analyzed for various amino acids and other metabolites. A portion of the whole blood sample was also used for determination of hematocrit with the use of a microcentrifuge. Additionally, in another group of septic and control rats, hearts were removed, weighed, dried until reaching constant weight in a 100°C oven, and dry-towet weight ratios were determined. Perchloric acid extraction. The hearts were stored at -80°C until perchloric acid (PCA) extraction. Samples weighing an average of 0.36 g were extracted as previously described (18) except for the inclusion of EDTA to give a final concentration of 5.5 mM. This was to sequester paramagnetic ions that tend to broaden 31P nuclear magnetic resonance (NMR) signals. After centrifugation and neutralization, the supernatant from the samples was lyophilized. After lyophilization, each sample was dissolved in 1 ml of deuterium oxide (DzO), required as a field-lock-signal solvent for NMR experiments, and divided into two portions. One portion was examined by 31P-NMR spectroscopy, and the other portion was analyzed for metabolites via enzymatic methods. NMR measurements. 31P-NMR spectroscopy was employed because of its capability to provide simultaneous species resolution and concentration analysis of a number of high-energy phosphates and phosphorylated metabolites. It can readily provide a quantitative concentration analysis on certain low-molecular-weight solutionstate (i.e., highly mobile) metabolites present in millimolar concentration and qualitative information on other metabolites that are present at lower concentrations. Therefore, it can serve not only as an analytical capability but also as a useful screening device for a number of intermediates, thereby directing the investigator to pursue other relevant areas. 31P-NMR spectra (202.34 MHz) were acquired at 4°C using a Varian VXR500 spectrometer. The D20 signal was used for adjustments of field homogeneity and to maintain a frequency lock. Proton-decoupled spectra were acquired under fully relaxed, i.e., quantitative conditions as follows: acquisition time, 1.5 s; pulse-repetition period, 10.0 s; pulse width (go”), 11.0 PCS;sweep width, 9,756 Hz. Phenylphosfonic acid was added as an internal concentration reference. Most peak assignments were determined by both characteristic appearance and chemical shift relative to phosphocreatine (PCr). Peak assignments in the phosphomonoester region were also determined by graded additions of phosphocholine, phosphoethanolamine, 2,3diphosphoglycerate (2,3-DPG), adenosine monophosphate, and glucose 6-phosphate (G-6-P) to a heart extract and examining for peak overlap/position. Metabolite assays in heart. All but one of the analytical methods has been described (18). In each case an enzy-

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (196.011.235.237) on August 2, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

C52

EFFECT

OF SEPSIS ON TRICARBOXYLIC

matic reaction or series of reactions result in oxidation or reduction of a pyridine nucleotide. ATP, phosphocreatine, and glucose were measured directly by the fluorescence of NADPH generated; glutamate, lactate, and NAD+ were measured by the fluorescence of NADH generated; and aspartate and alanine were measured by the fall in fluorescence when NADH is reduced to NAD+. The rest of the metabolites were measured indirectly with the use of enzymatic cycling (18) to amplify (as much as 5,OOO-fold) the pyridine nucleotide that had been formed: NAD+ in the case of citrate, pyruvate, and a-ketoglutarate, NADH in the case of malate, and NADPH in the case of G-6-P. When NAD+ was the indicator, pyridine nucleotide formed. The preformed tissue NAD+ was first removed by heating with alkali. The unpublished fluorescence method for NAD+ itself consisted simply of reducing it to NADH with G-6-P and G-6-P dehydrogenase from Leuconostoc mesenteroides. In 50 mM tris( hydroxymethyl)aminomethane (Tris)HCl buffer, pH 8.1, with 1 mM G-6-P and 1 pg/ml of enzyme, the reaction is completed in Cl0 min. Plasma metabolite assays. Perchloric acid extracts of arterial plasma samples were made as previously described (see MATERIALS AND METHODS). The analytical methods for the plasma metabolites were similar to the tissue metabolites (18) with the exception that enzymatic cycling was not necessary because of the increased concentration of the plasma metabolites. Determination of cardiac hemodynamics and myocardial blood flow. To determine the effect of sepsis on

cardiac function, a separate group of rats was surgically prepared in an identical manner to that previously described. The rats were anesthetized with halothane, and polyethylene cannulas (PE-50, Fisher Scientific, Pittsburgh, PA) were placed in the femoral artery and femoral vein of one hindlimb and in the tail artery. The right carotid artery was similarly cannulated, and the catheter was advanced into the left cardiac ventricle by observing the characteristic decrease in diastolic pressure as the aortic valve was transversed. Position of the ventricular catheter was confirmed after the rat was killed at the end of the experiment. Temperature of the rats was monitored via a rectal thermistor, and the temperature was maintained at 37-38°C by use of a heating pad. The rats were maintained on l--1.5% halothane anesthesia in O2 delivered via a nose cone throughout the experiment. Dry microspheres (Du Pont-New England Nuclear, Boston, MA) (15 t 0.8 pm) labeled with either 46Sc or lo3Ru were injected into the left cardiac ventricle for determination of cardiac output (CO) and organ blood flow in 10 septic and 7 control rats. Each rat had two independent measurements, i.e., one each with 46Sc and lo3Ru, randomly selected. The microspheres were suspended in a 50% dextrose solution in saline containing 0.05% polyoxyethylene sorbitan monooleate (Tween 80), which was dispersed by ultrasonication for 5 min immediately prior to infusion. A O.l-ml suspension of microspheres was injected into a 36 cm length of coiled PE-50 tubing, and the number of microspheres was determined by gamma counting. Between 100,000 and 140,000 microspheres were injected into the left cardiac ventricle in a volume of 0.1 ml over

ACID CYCLE OF HEART

15 s. Withdrawal of the reference blood sample of -1 ml from the femoral arterial catheter into a preweighed syringe by a second infusion pump was begun 10 s before microsphere injection and continued for -45 s after injection was complete. Microspheres are cleared from the circulation within this time period (unpublished observation). An infusion-withdrawal pump (Harvard Apparatus, South Natick, MA, model 944) was used to coordinate withdrawal of the blood sample from the femoral artery and simultaneous infusion of 1 ml of heparinized whole blood (obtained from a nonoperated littermate) into the femoral vein. The arterial blood pressure and heart rate were monitored via the tail artery during the experiment, and there were no alterations in systolic or diastolic arterial blood pressure or heart rate. The reference sample of blood was weighed, and the actual volume of blood was determined by dividing the weight of the sample by the density of blood (1.05 g/ml). The exact rate of withdrawal was determined using the volume of blood in the syringe and the time required for withdrawal. The reference blood sample was mixed with 1.0 ml of a 20% gelatin solution, and its radioactivity was determined. At the conclusion of the experiment, death of the animals was produced by pentobarbital sodium injection, and the various organs were removed, weighed, and counted for radioactivity on a two-channel Tracer 1191 gamma counter. Cardiac output and myocardial flow were determined using the methods of Heymann and associates (14). StatisticaL analysis. The significance of difference test between two samples of data was carried out by using RS/l system from BBN Software. The paired t test was performed on the data measured from the same sample, i.e., testing the ATP and PCr concentrations by 31PNMR and fluorometric enzymatic analysis. For unpaired data, an F test was first employed to test for the equality of variance. The pooled variance t test was employed for data with equal variances. The unpooled variance t test was employed for data with unequal variances. Statistical significance was accepted at the 95% confidence limit. The data are presented as the means t SE. RESULTS

Physical examination and physiological parameters of the septic and control rats. The septic rats’ appearance

was characterized

by piloerection,

exudates around the

1. Physiological and myocardial blood flow variables in control and septic rats TABLE

Septic

n

Control

n

Heart rate, beats/min 393k25.8 20 321.4*13.1* 14 Mean arterial pressure, 84.8t3.7 20 78.5t2.5 14 mmHg 21.0k1.7 20 26.Otl.O* 14 Cardiac index, ml. min-’ 100 g body wt-’ 0.56t0.008 20 0.82kO.O04* 14 Stroke volume index, ml Left ventricular blood 3.11kO.37 20 3.49k0.62 12 flow, ml min-’ g-’ Values are means t SE; n, no. of measurements. Ten septic and seven control rats were used in microsphere experiments. Each rat had 2 separate determinations performed using 2 radiolabeled tracers, 46Sc and lo3Ru . * P < 0 . 02 . l

l

l

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (196.011.235.237) on August 2, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

EFFECT

OF SEPSIS ON TRICARBOXYLIC

ACID CYCLE OF HEART

2,3 DPG PC I

UK I’ 1

Ref

G6P tA

Pi X ATP

/x.AJ~~“L

Ai

PC A

PC 2,3 DPG ;

f/HAD7

-15 PPM nose and eyes, diarrhea, and tachypnea. The septic rats usually exhibited little spontaneous movement within their cages. The sham-operated control rats appeared grossly normal and were active within the cages. Examination of the abdominal cavity at the end of the experiments in the septic rats confirmed the presence of copious amounts of foul-smelling, purulent peritoneal fluid. Portions of the bowel were adherent and covered with fibrinous exudate. The segment of the cecum that had been ligated and punctured was grossly dilated and gray-black in color. In the control rats, there was minimal peritoneal fluid, and the bowel was pink in color. In the group of septic and control rats that underwent vascular cannulation for microsphere blood flow determinations, various hemodynamic parameters were recorded (Table 1). The most striking differences found between the septic and control rats were the 22% increase in heart rate (P < O.Ol), the 32% decrease in stroke volume index (P < 0.02), and the 19% decrease in cardiac index (P < 0.02) (Table 1). There were no significant differences in mean arterial pressure (P > 0.10) or left ventricular blood flow (P > 0.05). The dry weight-to-wet weight ratio of the hearts was 0.238 t 0.007 and 0.233 t 0.001 for the septic and control rats, respectively, and is not statistically different (P >

0.05).

I

31P-NMR. A typical

31P-NMR spectrum is displayed in Fig. 1 and reveals the characteristic resonances of the various high-energy phosphates, inorganic phosphate (Pi), and other phosphate-containing compounds. The region of the spectrum between 6.8 and 7.2 ppm was not sufficiently resolved to permit unequivocal peak assignment, but the 2 phosphate of 2,3-DPG and phosphoethanolamine were present in this area. There were no statistical differences in the levels of ATP or PCr in the hearts of the septic or control rats (Table 2). Comparison of the NMR determination of [ATP] and [PC,] to the enzymatic method revealed good agreement between the two techniques (correlation coefficient = 0.82).

c53

FIG. 1. Proton-decoupled 31P-NMR spectrum (202.34 MHz) of perchloric acid extract of septic rat heart. Exponential time domain apodization resulting in a ~-HZ frequency domain line broadening was applied to accumulated free induction decays. Chemical shifts are assigned relative to phosphocreatine (PCr) at 0 ppm. Spectrum was obtained at 4°C; acquisition time, 1.5 s; pulse repetition period, 10.0 s; pulse width (go”), 11 .O ps; sweep width, 9,756 Hz; number of transients, 648. G-6-P, glucose-6-phosphate; Ref, reference peak of phenyl phosphonic acid* arrow identifies 3 phosphate of 2 3-DPG; PC phosphocholine* UK unknown group of peaks in which resol lution would not enable discrimination. This region corresponds to 2 phosphates of 2,3-DPG and other metabolites including phosphoethanolamine, ribose 5 phosphate, adenosine 5’-monophosphate, and inosine 5’-monophosphate.

There were also no statistical differences in the concentrations of G-6-P, 2,3-DPG, or Pi (Table 3). The most marked finding was a 91% increase in phosphocholine in the septic rat hearts (P < 0.01). The mean values of G6-P determined by NMR and enzymatic methods were comparable, but the correlation coefficient was only 0.62. Cardiac met&o&es. There were no significant differences between the cardiac levels of glucose, G-6-P, pyruvate, lactate, or NAD+ in septic and control rats (Table 4). Similarly, the TCA cycle intermediates citrate, cyketoglutarate, and malate were not different in septic versus control rat hearts. The only statistically significant differences were the 28% decrease in alanine and the 31% decrease in glutamate (P < 0.05 and P < 0.005, respectively) in the septic rat hearts compared with the control. Although aspartate tended to be lower in the septic rats, the difference was not statistically significant (P > 0.10). Plasma metchokes and hematocrits. In marked

contrast to the myocardium, plasma lactate and pyruvate levels in septic animals were increased more than twofold (Table 4). It should be noted, however, that the lactateto-pyruvate ratios were not significantly increased (12.1 and 10.6 in septic and control plasma, respectively; see DISCUSSION). Again, in contrast to the myocardium, plasma alanine and glutamate were not significantly decreased by sepsis. Glucose levels were marginally (17%) reduced (P < 0.05). The hematocrits of the septic and control rats were 39.6 t 0.8 and 42.0 t 0.9, respectively, and were not statistically different (P > 0.05). DISCUSSION

The major aim of this investigation was to determine whether mitochondrial abnormalities in the TCA cycle were present in the heart during sepsis. Analyses of TCA cycle intermediates and specific amino acids that help regulate the total pool size of the TCA cycle are indicative of mitochondrial function and may identify potential

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (196.011.235.237) on August 2, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

c54

EFFECT

OF

SEPSIS

ON

TRICARBOXYLIC

ACID

CYCLE

OF

HEART

2. Effect of sepsison high-energy phosphates in freeze-clamped hearts determined by direct assay and by 31P-NMR spectroscopy

TABLE

ATP D

n

PCr NMR

n

n

D

Septic Control

3.72kO.12 12 3.79t0.88 11 5.14kO.24 3.41t0.12 7 3.13kO.47 6 4.57zk0.23 Values are means t SE in mmol/kg wet wt; n, no. of measurements. ATP, NMR, nuclear magnetic resonance. TABLE 3. 31P-NMR analysis of selected phosphorus metabolites Septic

Control

Confidence Limits

0.25t0.03 (12) 0.241kO.03 (6) P > 0.85 0.43kO.05 (12) 0.31t0.05 (6) P > 0.15 0.21t0.02 (12) 0.11&0.02 (6) P c 0.01 2.72zko.29 (12) 1.88t0.16 (6) P > 0.07 Pi Values are means t SE in mmol/kg wet wt; no. of measurements in parentheses. G-6-P, glucose 6-phosphate; 2,3-DPG, 2 phosphorus of 2,3-diphosphoglycerate; PC, phosphocholine. G-6-P

2,3-DPG PC

sites of metabolic control or inhibition (13, 26, 36). Our results establish that sepsis does not cause an impairment in the TCA cycle in the heart. The fact that the concentrations of citrate, ar-ketoglutarate, and malate were not significantly different in control and septic hearts strongly suggests that the TCA cycle of the septic hearts was functionally normal under the given conditions of this study. The TCA cycle is only one part of complex metabolic processes that occur within mitochondria, and the present study does not definitively rule out other abnormalities that may be present in mitochondrial function during sepsis. It would seem unlikely, however, that significant defects could be present in the major energy-generating mitochondrial pathways, i.e., oxidative phosphorylation, during sepsis given the fact that both [ATP] and [PCr] were normal. We determined the effect of sepsis on myocardial metabolism at a point when the septic process was well established, but prior to development of agonal changes. At the 36. to 42-h time period of the present study, the septic rats were obviously quite ill and exhibited the characteristic features of sepsis previously detailed. Approximately 25% of the initial group of rats that had undergone cecal ligation and perforation had died during

PCr/ATP NMR

n

12 5.09H.18 7 4.18t0.88 adenosine triphosphate;

D

NMR

11 1.37t0.11 6 1.32t0.18 PCr, phosphocreatine;

1.35t0.11 1.33kO.15 D, direct assay;

this period of time. The severity of the septic condition was confirmed also on abdominal examination, which demonstrated gross peritonitis. The marked increase in plasma lactate and pyruvate is consistent with the systemic nature of the illness. All efforts were made to minimize potential confounding effects of anesthesia on tissue metabolites by keeping the duration of anesthesia as short as possible. Thus we believe that the present findings are most likely to reflect the effect of sepsis on myocardial intermediary metabolism under conditions that are likely to exist in vivo prior to agonal metabolic changes. It is possible that if myocardial work had been increased, for example, by using an isolated perfused working heart model, some limitation in mitochondrial function by sepsis may have been uncovered. However, the relevance of findings from such a study to the actual clinical condition of sepsis would be questionable. Our second major purpose was to determine the importance of ischemia-hypoxia in the cardiac failure of sepsis. The issue of ischemia is of utmost importance because of its implication in the therapy of this disorder. Although global myocardial ischemia is not believed to occur during sepsis until profound systemic arterial hypotension exists (5,7), uncoupling of coronary flow from regulation by cellular metabolism has been documented (5, 30) and may result in localized ischemic regions. Diffuse capillary thrombosis, endothelium swelling, and intravascular inflammation also occur during sepsis, and many investigators believe that these processes result in organ ischemia and multiple systems organ failure (10, 28). The characteristic changes in TCA cycle intermediates and key amino acids that occur during ischemiahypoxia have been well characterized (22, 27) and may be a more sensitive indicator of ongoing “relative” and intermittent ischemia than determination of global net lactate extraction (5, 11) or high-energy phosphate con-

4. Effect of sepsison TCA cycle intermediates and other selected metabolites in freeze-clamped hearts and arterial plasma

TABLE

n

Glucose

G-6-P

12

1.05 to.03 1.01 20.03

0.268 to.030 0.291 to.017

Pyruvate

Lactate Cardiac

Septic Control

7

Septic

7

0.233 to.010

0.246 kO.009

Citrate metabolites,

1.43 to.09 1.55 to.14 Plasma

KG mmollkg

0.242 to.017 0.291 to.024 metabolites,

0.118 t0.008 0.113 to.009

Malate

Alanine

Glutamate

Aspartate

NAD’

wet wt

0.200 kO.014 0.195 to.012

0.327* kO.030 0.452 to.035

2.89-f

kO.11 4.19 20.06

1.24 zko.13 1.41 to.12

0.566 to.018 0.587 to.018

pmol/l

124.6*$ 170.4-f 2,211.1-f 363.6" 78.5 t6.9 t17.0 t259.3 k55.4 k11.9 Control 7 149.51 81.2 919.3 378.8 78.0 k8.9 t5.2 t57.0 k17.4 t4.5 Values are means t SE; n, no. of measurements. G-6-P, glucose 6-phosphate; KG, a-ketoglutarate. * P c 0.05, t P c 0.005. $ Values in mg/dl.

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (196.011.235.237) on August 2, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

EFFECT

OF

SEPSIS

ON

TRICARBOXYLIC

centrations (27). Also, studies on ischemic myocardium have documented large changes in the concentrations of certain amino acids and TCA cycle intermediates prior to any significant decrease in ATP (27). This investigation confirms that sepsis does not cause myocardial cellular ischemia. Studies on rabbit (35) and rat (27) myocardium have demonstrated that alanine is an end product of anaerobic metabolism and is markedly increased during ischemia. Coincident with the increase in alanine are decreases in aspartate and glutamate (27, 35). One of the most thorough studies on the effect of ischemia on amino acids and TCA cycle intermediates in the heart was reported by Peuhkurinen and colleagues (27). They demonstrated the coupling between alanine formation and the almost stoichiometric consumption of aspartate. These investigators have shown that the [alanine] -to- [ aspartate] ratio is a sensitive reflection of cellular ischemia-hypoxia and rapidly increased from 0.34 during normoxia to 1.5 during ischemia. Our values for the [alanine] -to-[aspartate] ratio in the control and septic rat hearts were 0.32 and 0.26, respectively. The control value of 0.32 is almost identical to the 0.34 of the normoxic rat hearts obtained by Peuhkurinen et al. (27). Because the ratio for the septic rat heart tends to be slightly decreased rather than increased in comparison to control, ischemia is unlikely. Analysis of the TCA cycle intermediates provides further evidence against ischemia. There were no significant differences in cyketoglutarate and malate in the septic and control rats, whereas ischemia induces a rapid decrease in a-ketoglutarate and an increase in malate (27). Finally, the highenergy phosphates ATP and PCr were not statistically different in the septic vs. the control rats (P > O.Ol), which is further supportive evidence against significant ischemia. The high-energy phosphate results in this study are in contrast to those of Burns et al. (2), who reported decreased levels of ATP in rat hearts using an endotoxin model. However, our findings are in agreement with McDonough et al. (19), who reported normal levels of ATP and PCr in rat heart using a peritonitis model. In summary, although defective coronary autoregulation and/or capillary endothelial swelling may occur during sepsis, these processes do not result in cellular ischemia in the heart. The levels of the amino acids alanine and glutamate were significantly lower in the septic rat hearts compared to controls. Potential mechanisms for these changes in amino acids during sepsis could be differences in myocardial substrate utilization. Changes in the concentrations of the available substrates to the heart lead to changes in the concentration of the pools feeding into the citric acid cycle (33, 36). The particular substrate utilized depends on its plasma concentration, the work load of the heart, the hormonal milieu, and dietary state (36). Many of these factors are different in the septic versus the control hearts. The plasma lactate levels were twofold higher in the septic rats. Drake and colleagues (8) have shown that lactate is the preferred energy substrate in the normal dog heart and that once the lactate concentration becomes greater than 4 mM, it constitutes the majority of the fuel utilized by the heart even in the presence of elevated glucose and free fatty acids. Also,

ACID

CYCLE

OF

HEART

c55

recent 13C-NMR spectroscopy studies performed on isolated perfused hearts demonstrated that the concentrations of intracellular metabolites glutamate, aspartate, and alanine were remarkably different depending upon whether pyruvate or lactate was used in the perfusion medium (33). Alanine and aspartate were detected only in the pyruvate-perfused heart and not in the lactateperfused heart. The authors concluded that the intracellular concentrations of these metabolites were very dependent upon the substrate utilized. Therefore, another possible mechanism for the changes in tissue amino acids would be the following: an increase in plasma lactate in the septic rats, causing increased myocardial lactate uptake and utilization leading to decreased alanine and aspartate formation. Evidence consistent with this hypothesis comes from two studies that found that myocardial lactate extraction and utilization are increased during sepsis/endotoxemia (7, 34). A well recognized and extensively validated effect of sepsis is impairment of intrinsic myocardial contractility, which is independent of changes in preload or afterload. This myocardial depression is a hallmark of sepsis and has been extensively documented in humans and in various animal models, including the rat. Our findings of decreased stroke volume index and cardiac index in the septic rats are compatible with such an impairment in myocardial contractility. It is also possible, however, that decreased preload due to intravascular volume deficiency was contributing to or responsible for these decreases in myocardial performance. The fact that the hematocrit of the septic rats (39.6 t 0.8) was not increased relative to the control rats (42.0 t 0.9) provides some evidence against significant intravascular volume depletion as a cause of the decreased stroke volume index in sepsis (3). It is interesting to note that the values for stroke volume index of the septic and control rats in the present study, 0.56 t 0.01 and 0.82 t 0.01, respectively, are quite comparable to the values of 0.5 t 0.1 and 0.9 t 0.1 obtained by Rackow et al. (29), who used a similar model of sepsis. Thus, although we cannot exclude decreased preload as the cause of decreased cardiac function in this model, our findings are consistent with the large number of studies that have documented the effect of sepsis to impair myocardial contractility. It should also be noted that halothane, the inhalational anesthetic agent used in the present study, decreases both systemic vascular resistance and myocardial contractility. The arterial blood pressure in the control rats (78.5 t 2.5 mmHg) and septic rats (84.8 t 3.7 mmHg) reflects these effects of halothane, and this level of blood pressure is commonly reported in rats that are appropriately anesthetized with this agent (12). The fact that cardiac index in the control rats (26.0 t 1.0 ml. min. 100 g-‘) is similar to the cardiac index for awake control rats reported in other studies indicates that halothane did not excessively depress myocardial contractility (24). Finally, as previously discussed, the group of septic and control rats used to determine cardiac hemodynamics was different from the group of septic and control rats in which the metabolite assays were performed. The hearts in which the metabolite assays were performed were obtained from septic and control rats that were

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (196.011.235.237) on August 2, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

C56

EFFECT

OF SEPSIS ON TRICARBOXYLIC

anesthetized with halothane for a period of only l-l.5 min (minimum time sufficient to permit opening of the chest and freeze clamping of the heart). The rat cecal ligation and perforation model reproduces many of the metabolic changes observed in septic patients. Both pyruvate and lactate were increased severalfold in the plasma of septic rats. This increase in both pyruvate and lactate, with maintenance of the normal lactate-to-pyruvate ratio, is typical of sepsis and in contrast to the changes during ischemia in which there is a disproportinate increase in lactate (21). 31P-NMR spectroscopy was employed both as an analytical tool and as a useful screening technique for a variety of diverse metabolites that are present in the phosphorus spectrum. An important finding of the 31PNMR method was the marked increase in phosphocholine in the septic rat hearts. Phosphocholine is a major component of the phospholipids that are located in the plasma and subcellular organelle membranes. (It should also be noted that the region of the spectrum in which phosphoethanolamine, another major membrane phospholipid, resides was also markedly increased in the septic rats, but because phosphoethanolamine overlapped with other metabolites in this region, it was not possible to adequately quantitate the increase.) The change in phosphocholine could be due to either increased uptake of phosphocholine into the membranes or secondary to increased membrane disruption with release of phosphocholine. There is evidence that endotoxin activates phospholipases in cardiac membranes (17). Liu and Kang (17) examined the effects of endotoxin on phospholipids in cardiac sarcolemma and found a 283% increase in lysophatidylcholine and a 131% increase in lysophosphatidylethanolamine. Our results are consistent with these findings and compatible with sepsis-induced membrane degradation. In summary, there is no evidence that sepsis results in impairment of the myocardial TCA cycle or in the generation of high-energy phosphates. Additionally, there is no evidence that the disturbance in microcirculation or in coronary blood flow autoregulation, which is typical of sepsis, results in myocardial ischemia. Sepsis appears to have an important effect on membrane phospholipid composition. This study was supported by grants from the Edward Mallinckrodt, Jr., Foundation, the American Medical Association Education and Research Fund, the American Society of Anesthesiology, National Institutes of Health Biomedical Research Support Shared Instruments Grant 1 SlO RR-02004, American Cancer Society Grant BC-42, National Institutes of Health Grants NS-08862 and GM-30331, and a gift from the Monsanto Company. This work was presented in preliminary form at the regional meeting of the American Chemical Society, St. Louis, Missouri, 1989. Address for reprint requests: R. S. Hotchkiss, Dept. of Anesthesiology, Washington University School of Medicine, Box 8054, 660 S. Euclid, St. Louis, MO 63110. Received 14 June 1990; accepted in final form 16 August 1990. REFERENCES L. T. Myocardial dysfunction in endotoxin- and E. co& induced shock: pathophysiological mechanisms. Circ. Shock 15:

1. ARCHER,

261-280,1985. 2. BURNS. A.

H.. M. E. GIAIMO.

AND

W. R. SUMMER.

Dichloroacetic

ACID CYCLE OF HEART

acid improves in vitro myocardial function following in vivo endotoxin administration. J. Crit. Cure 1: 11-17, 1986. 3. CHIEN, S., R. J. DELLENBACK, S. USAMI, AND M. I. GREGERSEN. Hematocrit changes in endotoxin shock. Proc. Sot. Exp. Biol. Med. 118: 1182-1187,1965. 4. COALSON, J. J., H. K. WOODRUFF, L. J. GREENFIELD, C. A. GUENTER, AND L. B. HINSHAW. Effects of digoxin on myocardial ultrastructure in endotoxin shock. Surg. Gynecol. Obstet. 135: 908912,1972. 5. CUNNION, R. E., G. L. SCHAER, M. M. PARKER, C. NATANSON, AND J. E. PARRILLO. The coronary circulation in human septic shock. Circulation 73: 637-644, 1986. 6. DAWSON, K. L., E. R. GELLER, AND J. R. KIRKPATRICK. Enhancement of mitochondrial function in sepsis. Arch. Surg. 123: 241244,1988. 7. DHAINAUT, J.-F., M.-F. HUYGHEBAERT, J.-F. MONSALLIER, G. LEFEVRE, J. DALL’AVA-SANTUCCI, F. BRUNET, D. VILLEMANT, A. CARLI, AND D. RAICHVARG. Coronary hemodynamics and myocar-

dial metabolism of lactate, free fatty acids, glucose, and ketones in patients with septic shock. Circulation 75: 533-541, 1987. 8. DRAKE, A. J., J. R. HAINES, AND M. I. M. NOBLE. Preferential uptake of lactate by the normal myocardium in dogs. Cardiouusc. Res. 14: 65-72, 1980. 9. DRAPIER, J. C., AND

J. B. HIBBS. Differentiation of murine macrophages to express nonspecific cytotoxicity of tumor cells results in L-arginine-dependent inhibition of mitochondrial iron-sulfur enzymes in the macrophage effector cells. J. Immunol. 140: 2829-

2838,1988. 10. FERNANDES, C. AND E. KNOBEL.

tients-analysis

J., M. IERVOLINO, R. A. NEVES, E. M. SAMPAIO, Myocardial lesions in postmortem septic paof 71 cases (Abstract). Crit. Cure Med. 17: S54,

1989. 11. GERTZ, SEARLE,

E. W., J. A. WISNESKI, R. NEESE, J. D. BRISTOW, G. L. AND J. T. HANLON. Myocardial lactate metabolism: evidence of lactate release during net chemical extraction in man.

Circulation 12. GIBBS, N.

63: 1273-1279,

1981.

M., D. R. LARACH, T. M. SKEEHAN, AND H. G. SCHULER. Halothane induces depressor responses to noxious stimuli in the rat. Anesthesiology 70: 503-510, 1989. 13. GOLDBERG, N. D., J. V. PASSONNEAU, AND 0. H. LOWRY. Effects of changes in brain metabolism on the levels of citric acid cycle intermediates. J. Biol. Chem. 241: 3997-4003, 1966. 14. HEYMANN, M. A., B. D. PAYNE, J. I. E. HOFFMAN, AND A. M. RUDOLPH. Blood flow measurements with radionuclide-labeled particles. Prog. Cardiovasc. Dis. 20: 55-79, 1977. 15. HIBBS, J. B., 2. VAVRIN, AND R. R. TAINTOR. L-Arginine is required for expression of the activated macrophage effector mechanism causing selective metabolic inhibition in target cells. J. Immunol. 138: 550-565,1987. 16. HOTCHKISS, R. S., AND I. E. KARL.

S.-K. SONG, C. S. LING, J. J. H. ACKERMAN, Sepsis does not alter red blood cell glucose metabolism or Na’ concentration: a 2H, 23Na NMR study. Am. J. Physiol. 258 (Regulative Integrative Comp. Physiol. 27): R21-R31, 1990. 17. LIU, M.-S. Mechanisms of myocardial membrane alterations in endotoxin shock: roles of phospholipase and phosphorylation. Circ. Shock 30: 43-49, 1990. 18. LOWRY, 0. H., AND J. V. PASSONEAU. matic Analysis. New York: Academic, 19. MCDONOUGH, K. H., J. J. HENRY, C.

A Flexible

System

of Enzy-

1972, p. 282 H. LANG, AND J. J. SPITZER. Substrate utilization and high energy phosphate levels of hearts from hyperdynamic septic rats. Circ. Shock 18: 161-170, 1986. 20. MELA, L., L. V. BACALZO, AND L. D. MILLER. Defective oxidative metabolism of rat liver mitochondria in hemorrhagic and endotoxin shock. Am. J. Physiol. 202: 571-577, 1971. 21. MIZOCK, B. Septic shock-a metabolic perspective. Arch. Intern. Med. 144: 579-585,1984. 22. MUDGE, G. H., R. M. MILLS, H. TAEGTMEYER, R. GORLIN, AND M. LESCH. Alterations of myocardial amino acid metabolism in chronic ischemic heart disease. J. Clin. Invest. 58: 1185-1192,1976. 23. NEWSHOLME, E. A., P. NEWSHOLME, AND R. CURI. The role of

the citric acid cycle in cells of the immune system and its importance in sepsis, trauma and burns. Biochem. Sot. Symp. 54: l45161,1987. 24. PALSSON,

J., S.-E. RICKSTEN,

AND

S. LUNDIN.

Changes in central

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (196.011.235.237) on August 2, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

EFFECT

OF

SEPSIS

ON

TRICARBOXYLIC

hemodynamics during experimental septic shock in conscious rats. Circ. Shock 22: 65-72, 1987. 25. PARRILLO, J. E. The cardiovascular pathophysiology of sepsis. Annu. Rev. Med. 40: 469-485,1989. 26. PEUHKURINEN, K. J. Regulation of the tricarboxylic acid cycle pool size in heart muscle. J. Mol. Cell. Curdiol. 16: 487-495, 1984. 27. PEUHKURINEN, E. HASSINEN.

28.

29. 30.

31. 32.

K. T., T. E. S. TAKALA,

E. M. NUUTINEN,

ACID

OF

HEART

c57

effect on cardiac and skeletal muscle mitochondria. Surg. Gynecol. 133: 433-436, 1971. A. D., R. L. NUNNALLY, AND R. M. PESHOCK. Metabolic studies of pyruvate- and lactate-perfused guinea pig hearts by 13C NMR. J. Biol. Chem. 260: 9272-9279,1985. SPITZER, J. J., A. A. BECHTEL, L. T. ARCHER, M. R. BLACK, AND L. B. HINSHAW. Myocardial substrate utilization in dogs following endotoxin administration. Am. J. Physiol. 227: 132-136, 1974. TAEGTMEYER, H. Metabolic responses to cardiac hypoxia. Circ. Res. 43: 808-815, 1978. TAEGTMEYER, H. Six blind men explore an elephant: aspects of fuel metabolism and the control of tricarboxylic acid cycle activity in heart muscle. Basic Res. Cardiol. 79: 322-336, 1984. TAVAKOLI, H., AND L. MELA. Alterations of mitochondrial metabolism and protein concentrations in subacute septicemia. Infect.

Obstet. 33. SHERRY,

34.

AND I.

Tricarboxylic acid cycle metabolites during ischemia in isolated perfused rat heart. Am. J. Physiol. 244 (Heart Circ. Physiol. 13): H281-H288,1983. PINSKY, M. R. Multiple systems organ failure: malignant intravascular inflammation. Crit. Care Clin. 5: 195-198, 1989. RACKOW, E. C., M. E. ASTIZ, B. E. FIELD, D. MCKEE, AND M. H. WEIB. Myocardial dysfunction early in sepsis (Abstract). Crit. Cure Med. 15: 374, 1987. RUMSEY, W. L., L. KILPATRICK, D. F. WILSON, AND M. ERECINSKA. Myocardial metabolism and coronary flow: effects of endotoxemia. Am. J. Physiol. 255 (Heart Circ. Physiol. 24): H1295-H1304, 1988. SCHUMER, W., T. K. DAS GUPTA, G. S. Moss, AND L. M. NYHUS. Effect of endotoxemia on liver cell mitochondria in man. Ann. Surg. 171: 875-882,197O. SCHUMER, W., P. R. ERVE, AND R. P. OBERNOLTE. Endotoxemic

CYCLE

35. 36.

37.

Immunol. 38: 536-541,1982. 38. VARY, T. C., J. H. SIEGEL, T. NAKATANI, T. SATO, AND H. AOYAMA. Effect of sepsis on activity of pyruvate dehydrogenase complex in skeletal muscle and liver. Am. J. Physiol. 250 (Endocrinol. Metub. 13): E634-E640,1986. 39. WICHTERMAN, K. A., A. E. BAUE, AND I. H. CHAUDRY. Sepsis and septic shock: a review of laboratory models and a proposal. J. Surg. Res. 29: 189-201, 1980.

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (196.011.235.237) on August 2, 2018. Copyright © 1991 American Physiological Society. All rights reserved.