Effects of systemic infusions of endotoxin, tumor necrosis factor, and interleukin-1 on glucose metabolism in the rat: Relationship to endogenous glucose production and peripheral tissue glucose uptake

Effects of Systemic Infusions of Endotoxin, Tumor Necrosis Factor, and Interleukin-1 on Glucose Metabolism in the Rat: Relationship to Endogenous Glucose Production and Peripheral Tissue Glucose Uptake Pei Ra Ling, Bruce R. Bistrian, Beatrice Mendez, and Nawfal W. lstfan This study was performed to characterize and compare the actions of insulin on hepatic glucose production and peripheral glucose utilization during infusions of endotoxin, tumor necrosis factor (TNF), interleukin-1 (IL-l), and a combination of IL-1 and TNF in the rat. The euglycemic hyperinsulinemic clamp technique was combined with a primed-constant tracer infusion of high-performance liquid chromatography (HPLC)-purified 3H-3-glucose for estimation of whole-body glucose appearance and utilization rates; 14C-deoxyglucose (14C-DG) uptake was also measured in specific tissues following intravenous bolus administration. As expected, acute endotoxemia resulted in a significant reduction of glucose infusion during the clamp procedure (insulin concentration, 100 pU/mL), suggesting decreased insulin action. Similarly, infusion of TNF decreased the rate of glucose infusion necessary to maintain euglycemia. However, differences between endotoxin- and cytokine-treated rats were noted in whole-body glucose appearance (or disappearance) rates. Whereas endotoxin infusion predominantly decreased whole-body glucose uptake, suggesting diminished utilization in skeletal muscles, cytokine infusions were associated with a measurable hepatic glucose output despite hyperinsulinemia. In contrast, both cytokine and endotoxin administration decreased the rate of 14C-DG uptake by muscle tissue. These results demonstrate that TNF, IL-l, and endotoxin can induce a state of insulin resistance when infused continuously; the results also emphasize the complexity of regulation of glucose homeostasis during infection and sepsis. Copyright C.I1994 by W.B. Saunders Company

T

HE OCCURRENCE OF INSULIN resistance during experimental and natural infection is evidenced by diminished glucose tolerance and hyperinsulinemia.‘-5 Factors that have been implicated in the development of insulin resistance during infection include the stress response of fever, catabolism, and elevated levels of hormonal antagonists to insulin action such as growth hormone, cortisol, and glucagon.h,7 Despite the wide acceptance of these concepts, there is a relative paucity of information about the site of alteration of insulin action during infection, since hyperglycemia can result from either impaired suppression of glucose production by insulin and/or a decrease in insulin-stimulated glucose disposal. Tumor necrosis factor (TNF) and inter&kin-l (IL-l) have attracted considerable attention for their role as possible endogenous mediators and coordinators of the metabolic, as well as the immunologic, host response to infection and sepsisK-I” The metabolic and hormonal alterations in glucose metabolism observed in animals injected with recombinant or biologically derived IL-1 and TNF in combination are similar to those observed during infection.x-“’ However, current evidence suggests that substrate metabolism is affected differently when animals are exposed to the cytokines individually.iiHowever, these conclusions are based on the changes in concentrations of glucose and insulin and in glucose kinetics in the postabsorptive state. Currently, the effects of TNF and IL-l infusion on insulin action arc not well understood. We conducted this study to characterize the glucoregulatory action of insulin on hepatic glucose production and peripheral glucose utilization during infusion of endotoxin, tumor necrosis factor alpha (TNF). interleukin-1 alpha (IL-l), a.nd a combination of IL-1 and TNF in the rat. In this study. the euglycemic hyperinsulinemic clamp technique was combined with primed-constant tracer infusion using high-performance liquid chromatography (HPLC)purified “H-3-glucosei and bolus injection of i4C-deoxyglucost (Y-DC).’ i This model allows simultaneous estimaMetabohsm,

Vol 43, No 3 (March), 1994: pp 279-284

tion of glucose metabolism on the whole-body level and in peripheral tissues. Using the clamp technique, the effects of endotoxin and cytokine administration on hepatic glucose production, whole-body glucose disposal, and tissue glucose utilization were compared under similar hyperinsulinemic conditions. MATERIALS

AND METHODS

Animal Preparation Male Sprague-Dawley rats (350 to 380 g) were housed in a temperature-controlled room, exposed to a l?-hour light/dark cycle, and maintained on standard rat chow with access to tap water ad libitum for 4 days. One day before the experiment, the animals underwent a surgical procedure under ether inhalation for catheter placement. One catheter (Polyethylene Tubing, PE 50, ID 0.58 mm, OD 0.965 mm. Becton Dickinson, Parsippany. NJ) was inserted in the left carotid artery for sampling blood; two other silastic catheters were placed in both jugular veins for administration of the infusions described below. The animals were then allowed to recover in individual cages.

Experimental Design After overnight fasting. the animals were randomly divided into five groups. These five groups separately received infusions of saline, 20 kg/kg recombinant human TNF containing less than 200 pg endotoxining protein (Genentech, San Francisco. CA), 20 kg/kg recombinant human IL-1 containing less than 0.74 endotoxin unit (EU) endotoxining protein (Hoffmann-L.a Roche. Nut-

From the Laboratory of Nut&ion llnfection, New England Deaconess Hospital. Harvard Medical School, Boston MA. Submitted July 9. 1992; accepted April 12, 1993. Supported in part by Grants No. CA 45768+ DK 30492. DK 31933. and DK 41128fiom the National Institutes of Health. Address reprint requests to Nawfal W. lstfan, MD, PhD, Cancer Research Institute. New England Deaconess Hospital, 194 Pilgtim Rd, Boston, MA 02215. Copyright 0 I994 by W.B. Saunders Company 00260495/94lJ303-0003$03.00l0 279

280

ley. NJ), a 1: 1 combination of TNF and IL-l (10 kg/kg TNF and 10 pgikg IL-l). or 200 pg/kg endotoxin derived from Escherickia coli (Difco, Detroit. MI) for 3 hours through the right jugular vein. The saline and cytokine solutions were freshly prepared on the day of the experiment and contained 0.1% human albumin. One half of the dose (10 pg/kg) of TNF. IL-l. and the combination of TNF + IL-1 was administered by intravenous holus, and the remaining (10 pg/kg) was constantly infused over 3 hours. Endotoxin was constantly infused into animals at 66.6 pg/kg h for 3 hours, During the infusion of cytokines and endotoxin, no animal died. Basal arterial blood samples were drawn for determination of blood glucose and insulin levels before the infusions. By the end of the first hour of infusion of saline. cytokine, or endotoxin, a modification of the hyperinsulinemic euglycemic clamp technique was used (Fig I). For all animals, an intravenous bolus of 5 FCi purified aH-3-glucose (20 to 30 Ciimmol. Du Pont, Wilmington DE) was administered, followed by a constant infusion of 5 l&i/h through the same jugular vein for 2 hours. Insulin was continuously infused at 5 mU/min kg beginning 1 hour after the initial infusions, followed within I minute by 20% dextrose (Astra Pharmaceutical Products, Westborough. MA) infused at a variable rate through another jugular vein by use of a variable syringe infusion pump (Harvard Apparatus, South Natick, MA). Arterial blood was sampled every 10 minutes for measurement of glucose concentration. The rate of glucose infusion was adjusted empirically after each arterial plasma glucose determination to maintain euglycemic basal conditions. At 138 minutes of the initial infusion, a bolus of 5 uCi i4C-DG was injected intravenously: then 400-&L arterial blood samples were drawn at 140. 150. 160. 170, and 180 minutes for measurement of iJCmDG levels in plasma. One hundredmicroliter blood samples were also collected at 140. 160. and 180 minutes for measurements of plasma glucose specific activity. The data confirmed that the isotopic steady state was achieved during this period of time. ie. between 60 to 120 minutes during the clamp period. At the end of infusions (180 minutes of total infusion time or 120

LING ET AL

minutes during hyperinsulinemic and ruglycrmic clamp), animals were decapitated. Blood was collected for measurement

the of

plasma insulin level. Pieces of liver. rectus abdominus muscle, and abdominal mesenteric adipose tissues were removed and weighed. The whole heart. including ventricles and atria. was removed and weighed. All of the tissues were stored at -20°C for later determination

of lJC-DG counts.

Asstlys Glucose specitic activities in plasma were determined as previously described.” Brirtly, plasma samples were first deproteinized with equal volumes of Ba(OH)l and ZnSO4 and immediately centrifuged. The supernatant was then passed through a 1.5 x 0.5.cm column

of analytical

grade

anion

(200-400

mesh, Bio-Rad

Laboratories, Richmond. CA) resin and a 1.5 x 05cm column of cation (200-400 mesh, Bio-Rad) resin. This procedure resulted in adherence of labeled organic acids to the resin. Each column was then washed with 3 mL distilled water to elute the labeled glucose. The eluate was collected and evaporated to dryness in an oven at 60°C. and the residue was dissolved in 1 mL distilled water. Aliquots were counted for ‘H-glucose content hy tise of Beckman Ready Gel scintillation fluid (Beckman Instruments. Fullerton. CA) and external standards for efficiency determination. Glucose concentrations were determined by the glucose oxidase method using a Beckman glucose analyzer (Glucose Analyzer 2). Plasma “C-DG level and the accumulation of ‘V-DG in the tissues were determined using the method described r1sewhere.i~~” In brief, blood samples were deproteinized with Ba(OH)?-ZnSOa and immediately centrifuged. An aliquot (100 JLL) of the supernatant was used for counting rJC-DG radioactivity (Beckman Instruments). Tissue samples were placed in 0.5 or 1 mL I mol/L NaOH (according to the size), digested at 60°C for I hour. and then neutralized with 1 mol/L HCL. Separate 2Ot)-FL aliquots of digested tissues were treated with either 1 mL HCLO, (4% wtivol) or 1 mL Ba(OH):-ZnSOj and centrifuged. The supernatants were counted for iJC-DG plus ‘“C-DG-h-phosphate (HCLO.,-treated fractions) and ‘IC-DG-&phosphate [Ba(OH):-ZnSOJ-treated fractions]. Plasma insulin level, basal and 120-minute. was determined by a radioimmunoassay kit using porcine insulin standards (Binax. South Portland, ME).

Cdcdrrtiotzs

Fig 1. Study protocol for hyperinsulinemic glucose clamp. Primedconstant infusion of purified 3H-3-glucose was administered to label the glucose pool for 2 hours after 1 hour of initial infusions of saline, TNF (20 kg/kg), IL-1 (20 pg/kg), TNF + IL-1 (total 20 pg/kg), and endotoxin (66.6 kg/kg h). Insulin (5 mu/kg min) was infused to elicit a higher insulin level similar to that of the fed state; glucose was infused at a variable rate to maintain a fasting euglycemic level for 2 hours. In addition, 5 &i ‘%XtG was injected at 138 minutes; blood samples were drawn at 140, 150. 160, 170, and 180 minutes for determination of W-DG in plasma. Blood samples were also obtained at 140, 160, and 180 minutes to determine plasma glucose specific activity.

The rate of glucose appearance (R,i) was calculated using the infusion rate of ‘H-labeled tracer (I) and the steady-state plasma 3H-glucose specific activity, as follows: R,, = I/plasma 3H-glucose specific activity. The rate of endogenous glucose production was calculated from the difference between the determined total rate of glucose appearance (isotope dilution) and the exogenous glucose infusion rate during the clamp. The apparent rate of glucose uptake in tissues was determined according the following equation’? Apparent

glucose

uptake

lJC-DG,,,,,,,

= l:;,

‘JC-DG,,,t,,,fi( t ) dt

“““““J

where G is the plasma glucose concentration at steady state, i is sampling time. t is time. and dt indicates the change in DG plasma in relation to time. The quantity l”,, “‘C-DG pi,l\m:,(t) dt was determined by numerical integration (trapezoid rule) of iJC-DG blood measurements with a computer program.

CYTOKlNEi INFUSION AND INSULIN RESISTANCE

281

Statistical Anulysis

Table 2. Apparent Rate of Glucose Uptake in Tissues

Data are presented

as the mean ? standard error of the mean. Group means were compared by two-way ANOVA using the BMDP statistical software package (UCLA. Los Angeles. CA). with significance defined asPless than .Wi. Comparison of multiple groups was performed according to the Bonferroni procedure when the ANOVA was found to be significant at the 95”; conlidence level.

RESULTS

Plasma Glucose

and Insulin Concentrations

During Infusion

of Cytokines or Endotoxin Arterial Blood Glucose Basal Group (nl Saline

(7)

Clamp

ImgidL) 119 2

6

(60-120 min. mg/dL)

100 g) Under Hyperinsulinemic Condition with Infusion of Cytokines or Endotoxin

Group (n) Control

(7)

Heart

5.4 t- 0.4

Muscle

2.3 % 0.3

LlVtY

1.4 2 0.1

Fat

0.3 2 0.1

TNF (8)

3.2 + 0.4*

1.0 ? O.lf

0.8 2 O.O*

0.2 t 0.0

IL-I (7)

2.9 2 0.4*

1.4 * 0.2*

0.7 2 0.1*

0.4 2 0.111

TNF + IL-I (8)

1.7 2 0.2*t

0.9 -t 0.2*

0.7 r 0.1x

0.2 2 0.0

Endotoxin (7)

3.3 ? 0.1*

0.6 + O.l*§

0.9 & O.l*

0.1 t O.O*

NOTE. Values are means ? SE; n = number of rats. Significance was

Blood glucose and insulin concentrations are summarized in Table I. The basal arterial blood glucose level ranged from 107 to 129 mg/dL, and the basal insulin concentration ranged from 27 to 31 @/mL, both of which were not significantly different among groups. With insulin infusion of 5 mu/kg. min, plasma insulin concentration was significantly increased to approximately 100 FUlmL, ranging from 94 to 112 PUlmL (Table 2). During hyperinsulinemic condition, arterial blood glucose concentration was clamped to the basal level between 114 and 125 mg/dL with a mean coefficient of variation of less than 6.5% (3.2% to 6.5%). In all five groups, the exogenous glucose infusion needed to maintain euglyccmia appeared to reach plateau level within 60 minutes of the clamp procedure. Body weights were not significantly different among the five treatment groups. Figure 2 shows the exogenous glucose infusion rate achieved in the last hour of the clamp, as well as the rates of glucose appearance and endogenous production. As can bc seen in this figure, the glucose infusion rate maintaining similar glucose levels at similar hyperinsulinemic conditions varied significantly, indicating differences in the response to insulin fiallowing the various treatments. Animals receiving saline infusion required 4.2 ? 2.5 mmol glucose/kg. h to maintain normoglycemia, which was the highest level among the five groups. Animals that received IL-I infusion required slightly but not significantly less glucose, 3.4 & 0.3 mmol glucose/kg. h, as compared with saline infusion. With TNF infusion, the requirement for maintenance of normoglycemia was 2.3 ? 0.5 mmol glucose/kg h (P < .05). Animals receiving endotoxin infusion required only 1.9 ? 0.5 mmol glucose/kg. h. The lowest amount of glucose, I.0 + 0.3 mmol glucose/kg. h, was required by animals treated with the combination of TNF and IL-l. These results demonstrated that the exogenous glucose utilization

Table 1.

(mg/min

Plasma Insulin Basal (pU/mL)

120 Minutes WJ/mL)

114+2

30 + 3

94 ? 6

TNF (8)

107 * 5

115 r 2

31 +3

109 f 5

IL-1 (7)

126 -t 10

112 + 3

27 + 2

100 + IO

TNF + IL-I (8)

128 -t 4

125 + 2

27 + 3

112 2 8

Endotoxln (7)

129?6

122 ? 2

28 + 3

97 + 7

NOTE. Values are means -t SE; n = number of rats. Basal data were obtained before the rats received infusions of cytokines or endotoxin.

determined by two-way ANOVA with Bonferroni correction. *P < .OOl v control. tP < .05 vTNF, IL-l, and endotoxin. SP < .05 Y control. §P < .05” IL-l. IP < .05 vTNF and endotoxin.

was significantly decreased by TNF, TNF + IL. and endotoxin administration, but not by IL-l infusion. Total glucose appearance measured by isotope dilution was significantly decreased only in the endotoxin group (1.9 ? 0.3 mmolikg h, P < .05 1’ control). Because of steady-state conditions during the clamp, total glucose utilization was equivalent to total glucose appearance. Thus, only endotoxin infusion was consistent with a net decrease in glucose utilization (50% reduction). Rates of endogenous glucose production were calculated by subtracting the rate of infusion of exogenous glucose from the tracer-determined total rate of glucose appearance (R,). In the present study, although the rate of hepatic glucose production showed a negative value (-0.3 it 1.4 mmolikg h) in the control group, this was not significantly different from zero. Thus, hepatic glucose production was completely suppressed by the exogenous insulin in this group. A similar situation, a small but positive number not significantly different from zero, was also found in the endotoxin-treated group, indicating that hepatic glucose production was also completely suppressed. However, in TNF, IL-l, and TNF + IL-l groups, the rate of hepatic glucose production was significantly greater than zero, suggesting that the ability of insulin to suppress hepatic glucose output was significantly decreased in these groups. Table 2 lists the apparent rates of insulin-stimulated glucose uptake by individual tissues. Endotoxin infusion significantly decreased insulin-stimulated glucose utilization in adipose tissue. liver, muscle, and heart, as compared with saline infusion (P < .05). As shown in Table 2, the responses of these tissues to infusion of TNF and IL-I. either alone or in combination, were similar to those following endotoxin administration. However, the degree of the decrease in glucose utilization in these tissues varied by treatment (ANOVA, P < .0.5). At similar glucose and insulin concentration, TNF appeared to exert a greater effect than IL-l. Administration of a combination of half the individual doses of each cytokine appeared to decrease DG uptake to a level similar to that induced by the full dose of TNF. Furthermore, DG uptake in the heart was significantly lower in the group receiving combined cytokines in comparison to the groups receiving those cytokines individu-

LING ET AL

Glucose Infusion (60- 120’)

Rate

Glucose

Hepatic

Appearance

Glucose

Production I

II

i b

Fig 2. (I) Glucose infusion rate in rats with administration of saline, TNF, IL-l, TNF + IL-l, and endotoxin throughout 60 to 120 minutes during hyperinsulinemic glucose clamp procedure; (II) Plasma glucose appearance rate after 2 hours of primed-constant infusion of purified 3H-3-glucose during hyperinsulinemic glucose clamp procedure; (Ill) Rate of hepatic glucose production during hyperinsulinemic glucose clamp procedure. (0) Control; (m) TNF; (0) IL-l; (9) TNF/IL-1; ( ) endotoxin. 9 < .05vTNF, TNF/IL-1, endotoxin; bP c .05vTNF/IL-1; “P c .05 vcontrol, TNF, IL-l, and TNF/IL-1; dP < .06 v control, IL-l, and endotoxin; *P < ,005 v control, TNF, IL-l, and endotoxin.

ally (P < .05). Only TNF and endotoxin significantly decreased DG uptake in adipose tissue, whereas cndotoxin and cytokine treatments (TNF, IL-l, and combination) resulted in similar decreases in liver tissue. DISCUSSION

The current study demonstrates that infusion of the cytokines IL-l and TNF, either alone or in combination, can affect glucose homeostasis in the rat. This is manifested as a decreased glucose infusion necessary to maintain euglycemia at a high physiologic insulin level of approximately 100 kU/mL in circulation. Since this level stimulates the usual plasma insulin concentration in the fed rat, the results of the current study suggest that TNF and possibly IL-1 can rapidly induce a state of impaired insulin action during feeding. We have also compared the effects of these cytokines with those observed during a nonlethal endotoxin infusion. As expected, acute endotoxemia resulted in a significant reduction of glucose infusion during the hyperinsulinemic and euglycemic clamp procedure, also suggesting decreased insulin-stimulated glucose utilization. However, despite the similarities in diminished insulin action on glucose metabolism between the cytokines and endotoxin, site-specific differences in tissue glucose metabolism were also noted. Impairment in the glucoregulatory action of insulin can theoretically result from a decrease in glucose utilization in peripheral tissues, an increase in endogenous glucose production, or a combination of both factors. In rats similar to those of the current study, estimates of basal endogenous glucose production rates are in the range of 3 mmol/

kg. h.“.” This level is only slightly lower than the total glucose infusion rate under the hyperinsulinemic conditions of this study, implying a small increase in insuhnstimulated glucose uptake in the control group. Consequently, the major effect of insulin in this group has been the complete suppression of endogenous glucose production, as noted in Fig 1. This conclusion is supported by the data of Lang and Dobrescu, who measured the actual increase in glucose utilization rates at several insulin levelsix According to their study, which was conducted in similar rats, the increase in insulin-induced glucose uptake was minimal for circulating insulin levels lower than 150 kU/mL. In contrast to the saline- and endotoxin-treated rats. endogenous glucose production was measurable under similar hyperinsulinemic conditions in rats treated with TNF and IL-l. These results point to a state of insulin resistance at the level of endogcnous glucose production in the liver associated with cytokine infusions. Although similar conclusions about hepatic insulin resistance have been made in the septic dog, I4 it is unlikely that tither of these experimental models fully explains all of the changes in glucose metabolism in the septic condition. For example, in our endotoxin-treated group, as well as in other experimental sepsis models,tx insulin’s ability to suppress endogenous glucose production was maintained, thus implying absence of resistance to insulin action in the liver. However, it is also possible that suppression of cndogenous glucose output in the endotoxemic group of the current study is the result of a direct effect.ix,‘“~” Although further studies will be needed to clariy this observation, it is interesting to note

283

CYTOKINE INFUSION AND INSULIN RESISTANCE

that separate in vitro evidence showing diminished gluconeogenie pathways in hepatocytes of septic rats?’ is consistent with this interpretation. Differences between the effects of endotoxin and cytokine infusions on total insulin-stimulated glucose uptake are also supported by the decrease of DG uptake in skeletal muscle, as summarized in Table 2. In contrast, TNF or IL-1 infusions did not impair whole-body glucose uptake, although significant decreases in DG were still noted in skeletal muscle. It is important to note that peripheral insulin resistance was recently documented in the rat during prolonged infusion of TNF.?’ Thus. it is possible that the DG technique is more sensitive in detecting differences in peripheral insulin action, and that the effects of TNF on this parameter are more exaggerated during prolonged infusions. Evaluation of these findings should include an understanding of the methodologic limitations of isotope dilution (whole-body glucose kinetics) and DG uptake in tissue. In deriving estimates of actual glucose utilization in specific tissues, the assumption is frequently made that a constant relationship to DG uptake exists and that this relationship can be determined in vitro.‘” This relationship, frequently referred to as “lumped constant,” is used to correct for the discrimination against DG in the glucose metabolic pathway.” Based on measurements by several investigators,tiJh.Zi.zs actual glucose utilization in incubated tissue samples exceeds the uptake of DG by a factor of approximately 2.0, accounting for the commonly used lumped constant of 0.50. These values of the lumped constant in the rat have not been shown to be affected by tissue sensitivity to insulin,z4J5 the level of insulin concentration,Z5 or experimental conditions such as the presence of sepsis.‘h,Zd,‘5 Although tissue-specific glucose utilization rates have been derived from DG uptake in vivo and compared with whole-body glucose utilization by isotope dilution,‘” our data indicate a discrepancy in such a comparison. Assuming a lumped constant of 0.5, a total muscle mass of 40% of body weight. and the values of DG uptake in Table 2, WC estimate a total glucose utilization rate of approximately 6.13 mmolih in the skeletal muscle of rats in the current study. This estimate exceeds the rate of whole-body glucose utilization derived from jH-glucose dilution (Fig 2). implying either heterogeneity of glucose utilization in various muscle tissues. or inadequacy of the lumped constant to relate DG uptake to whole-body glucose appearance rates

as measured by isotope dilution in plasma. In fact, since the abdominal muscle sampled in the current study is predominantly composed of type I fibers, and because insulin action in these fibers is higher than in other types,?6 our estimates may not adequately reflect insulin-stimulated glucose uptake in other muscle tissues. However, regardless of the actual reason, we have only reported the apparent glucose uptake rates as these are derived from DG appearance in tissues. Therefore, the decrease in this parameter in the skeletal muscle of TNF- and IL-l-treated rats may actually reflect a decrease in the distribution of glucose to that tissue rather than an actual reduction in glucose utilization in metabolic pathways. This could result from blood flow redistribution away from skeletal muscle or from a change in the glucose transport mechanism in cytokine-treated animals. Although we have not measured the hemodynamic effects of IL-l and TNF in the current study, previous investigation in our laboratory, using similar animals and cytokine infusion protocols, has demonstrated significant blood flow redistribution at 4 hours of TNF infusion.‘” However, we were unable to detect similar changes in rats infused with IL-l.“’ Therefore, at least in the IL-l-treated rats of the current study, reduction of DG uptake in skeletal muscle is unlikely to have a major hemodynamic component. The possibility that IL-1 and TNF may influence glucose transport in specific tissues and cells has been recently demonstrated in fibroblasts and macrophages.Z7,‘x The results of this study add to the existing evidenceZ9J” that secretion of cytokines during sepsis and infection may serve to ensure adequate nutrient supply to the tissues with a large number of inflammatory cells. This is usually manifested as an increase in muscle protein breakdown,“’ an increase in hepatic glucose output in the postabsorptive state,31J? and a redistribution of infused glucose away from skeletal muscle, as in the current study. .4lthough the changes in glucose metabolism may be described by the term “insulin resistance,” the mechanism of this derangement remains poorly understood and may involve different regulatory steps. The current study documents the complexity of these mechanisms and suggests the involvement of IL-1 and TNF, among other mediators. Studies directed at understanding the interaction between various cytokines, hormones, and growth factors are needed to clarify the pathophysiology of metabolic derangements in sepsis and endotoxemia.

REFERENCES 1. Shamgraw RE. Jahoor between septic and postburn 989. 1989

F, Miyoshi H. et al: Differentiation insulin resistance. Metabolism 38:983-

2. White RH, Frayn KN, Little RA, et al: Hormonal and metabohc responses to glucose infusion in sepsis studies by the hyperglycemic glucose clamp technique. JPEN 11:345-353. 1987 3. Maclane MP, Tomasik TW, Law WR, et al: Hepatic insulin resistance during canine sepsis. Circ Shock 33:207-215. 1991 4. Nelson KM. Turinsky J: Local effect of burn on skeletal muscle insulin responsiveness. J Surg Res 32:288-297, 1981 5. Westfall MV. Sayeed MM: Basal and insulin-stimulated

skeletal muscle sugar transport in endotoxin and hacteremia rats. Am J Physiol254:R673-R679, 1988 6. Gelfand RA. Matthews DE. Bier DM, et al: Role of counterregulatory hormones in the catabolic response to stress. J Clin Invest 74:2238-2248, 1984 7. Zenser TV, DeRubertis FR, George DT, et al: Infectioninduced hyperglucagonemia and altered hepatic response to glucagon in the rat. Am J Physiol227:1299-1305. 1974 8. Tredget EE, Yu YM. Zhong S. et al: Role of interleukin tumor necrosis factor in energy metabolism in rabbits. Physiol255:E760-E768. 1988

1 and Am J

284

9. Del Rey A. Besedovsky H: Interleukin 1 affects glucose homeostasis. Am J Physiol253:R794-R798. 1987 10. Flores EA. Bistrian BR, Pomposelli JJ, et al: Infusion of tumor necrosis factoricachectin promotes muscle catabolism in the rat: A synergistic effect with interleukin 1. J Clin Invest 83:16141622.1989 11. Flares EA. Istfan N, Pomposelli JJ, et al: Effects of intrrleukin-l and tumor necrosis factoricachectin on glucose turnover in the rat. Metabolism 29:738-743, 1990 12. DeFronzo RA. Tobin JD. Andres R: Glucose clamp technique: A method for quantifying insulin secretion and resistance. Am J Physiol237:E214-E223.1979 13. Sokoloff L. Reivich M, Kennedy C. et al: The [‘%]deoxyglucase method for the measurement of local cerebral glucose utilization: Theory, procedure and normal values in the conscious and anesthetized albino rat. J Neurochem 28:X97-916. 1977 14. Horn FG, Goodner CJ, Berrie MA: A (3H)-2-deoxyglucose method for comparing rates of glucose metabolism and insulin responses among rat tissues in viva: Validation of the model and the absence of an insulin effect on brain. Diabetes 33:141-152. 1984 15. Ferre’ P. Leturgue A. Burnol AF: A method to quantity glucose utilization in viva in skeletal muscle and white adipose tissue of the anesthetized rat. Biochem J 228:103-110. 1985 16. Lang CH, Dobrescu C. Meszaros K: Insulin-mediated glucose uptake by individual tissue during sepsis. Metabolism 39:10971107.1990 17. Ling PR, Istfan N, Colon E. et al: Effect offish oil on glucose metabolism in the interleukin-1 a-treated rat. Metabolism 4?:81-

85, 1993 1X. Lang CH. Dobrescu C: In vivo insulin resistance during nonlethal hypermetabolic sepsis. Circ Shock 28:165-178. 1989 19. McLane MP, Tomasik TW, Law WR, et al: Hepatic insulin resistance during canine sepsis. Circ Shock 33:207-215, 1991 20. Clemens MG. Chandry IH. Diagneau N, et al: Insulin resistance and depressed gluconeogenic capability during early hyperglycemic sepsis. J Trauma 24:701-708, 1984 21. de Vasconcelos PR. Kettlewell MGW, Williamson DH: Time course of changes in hepatic metabolism in response to sepsis

LING ET AL

in the rat: Impairment of gluconeogenesis and ketogenesis in vitro. Clin Sci 72:683-691. I987 22. Ardawi MS. Ashy AA. Jamal YS: Metabolic control of hepatic gluconeogenesis in response to sepsis. J Lab Clin Med 114579-586. 1989 23. Lang CH, Dobrescu C. Bagby GJ: Tumor necrosis factor impairs insulin action on peripheral glucose disposal and hepatic glucose output. Endocrinology 130:43-Z 1992 24. Meszaros K. Bagby GJ, Lang CH. et al: Increased uptake and phosphorylation of 2-deoxyglucose by skeletal muscle in endotoxin-treated rats. Am J Physiol253:E33-E39. 1987 25. Mrszaros K. Lang CH. Bagby GI. et al: Contribution of different organs to increased glucose consumption after endotoxin administration. J Biol Chem 262:10965-10970. 1987 26. James DE, Jenkins AB. Kraegen EW: Heterogeneity of insulin action in individual muscles in vivo: Euglycemic clamp studies in rats. Am J Physiol248:8567-E574, 1985 27. Pekala P, Marlow M. Heuvelman D. et al: Regulation of hexose transport in aortic endothelial cells by vascular permeability factor and tumor necrosis factor-alpha. but not by insulin. J Biol Chem 265:18051-18054, 1990 28. Fuer JS, Usher P. Moller DE. et al: Induction of the glucose transport system by tumor necrosis factoricachectin: Implication for the hypoglycemia of sepsis and tumors. Am Fed Clin Res 35:502A, 1987 (abstr) 29. Lang CH. Dobrescu C: Interleukin-1 induced increases in glucose utilization are insulin-mediated. Life Sci 45:2127-2134, 1989 30. Evans DA, Jacobs DO, Wilmore DW: Tumor necrosis factor enhances glucose uptake by peripheral tissues. Am J Physiol 257:R1182-RI 189. 1989 31. Westfall MV, Sayeed MM: Effect of bacteremia on xkeletal muscle glucose transport in fed rats. Circ Shock 26:727-737. 1988 32. Filkins JP, Mathews HL, Yelich MR: Monokine and hyperinsulinemia in endotoxicosis and septic peritonitis. in Kluger MJ. Oppenheim JJ. Powanda MC (eds): The Physiologic. Metabolic and Immunologic Actions of Interleukin-1. New York. NY. Liss, 1989. pp 23 I-240