Lethal Hypoglycemia and Hypothermia Induced by Administration of Low Doses of Tumor Necrosis Factor to Adrenalectomized Rats Tova Chajek-Shaul,

Varda Barash, Joseph Weidenfeld,

Gideon Friedman,

Ehud Ziv, Esther Shohami,

and Eitan Shiloni

An increased sensitivity of adrenalectomized (Adex) rats to intravenous (IV) injection of recombinant human tumor necrosis factor (rHuTNF) was manifested by a marked increase in the rate of mortality. The rats that died exhibited severe hypoglycemia and hypothermia. Administration of 2.5 or 10 fig/l00 g body weight (3% or 12%) of the lethal dose in sham-operated rats (90 pg/lW g body weight) rHuTNF caused a mortality rate of 50% or 100%. respectively, within 4 hours of its injection. Pre-administration of dexamethasone or intermittent glucose infusion protected the animals from the lethal effect of rHuTNF. lndomethacin did not change the mortality rate in rHuTNF-treated Adex rats, but prevented it in sham-operated rats. The rats that died exhibited a marked decrease in body temperature, but only Adex rats developed hypoglycemia after low doses of TNF. Pretreatment with dexamethasone prevented the hypothermia in both Adex and sham-operated rats, while indomethacin was effective only in sham-operated rats and did not prevent the hypothermia or the hypoglycemia in Adex rats. In the surviving rHuTNF-treated Adex rats, a rapid increase in body temperature occurred, blood glucose decreased to 30 mg/dL, serum insulin concentration decreased to 6 #J/mL, liver glycogen content was reduced by 96%. and a significant reduction in liver phosphoeonolpyruvate carboxykinase (PEPCK) and liver microsomal glucose-6-phosphatase activities was observed. Repeated administration of glucose IV to rHuTNF-treated Adex rats caused an increase in blood glucose and insulin concentrations, and some repletion in liver glycogen content. Injection of rHuTNF, 2.6 to 10 pg/lW g body weight, to sham-operated rats caused a significant but slower increase in body temperature. It also induced a 50% and 66% reduction in liver glycogen content, respectively, while reduction in liver PEPCK and microsomal glucose-6-phosphatase activities was observed only after injection of IO pg/lOO g body weight rHuTNF. Serum glucose and insulin levels remained unchanged even after injection of 40 Fg/lOO g body weight, but severe hypoglycemia developed following administration of 90 pg/lW g body weight rHuTNF. In both Adex and sham-operated rats, rHuTNF injection increased serum triacylglycerol and decreased adipose tissue lipoprotein lipase. It is concluded that adrenalectomy sensitizes rats to the lethal effect of rHuTNF. The increased mortality was associated with severe hypoglycemia resulting from the combined effect of rHuTNF administration and the lack of glucocorticoids on glucose homeostasis, and mav be reversed bv treatment with either dexamethasone or glucose. 0 1990 by W.B. Saunders Company.

T

UMOR NECROSIS FACTOR (TNF) is released by mononuclear phagocytes in response to infection or endotoxin. It has been implied that TNF mediates the lethal effect of endotoxic shock.’ It also causes profound physiological and metabolic alterations, characterized by fever, pituitary and stress hormone elaboration, impaired glucose homeostasis, lipid metabolism, and increase in protein turnover and peripheral aminoacid mobilization. The increased secretion of stress hormones (glucocorticoids, catecholamines, and glucagon) may modify the cytotoxic and metabolic effects of TNF.*13 Therefore, it seems of interest to study the in vivo effect of TNF in adrenalectomized rats (Adex) where glucocorticoid secretion is completely abolished. In the present study, we studied the lethal manifestations and the metabolic effects of recombinant human TNF (rHuTNF) injected to Adex rats to elucidate the role of endogenously secreted corticosteroids in the metabolic effects of TNF. We found that adrenalectomy increased the rats’ susceptibility to the lethal effect of TNF. These findings

From the Departments of Internal Medicine B and Lipid Research Laboratory, Clinical Biochemistry, Neurology and Surgery B. Hadassah University Hospital, and the Department of Pharmacology, Hebrew University Medical School, Jerusalem, Israel. Supported in part by a Hebrew University Medical Grant to T.C-S. Address reprint requests to Tova Chajek-Shaul, MD, Lipid Research Laboratory, Department of Medicine B, Hadassah University Hospital, POB 12000, Jerusalem 91120. Israel. @ 1990 by W.B. Saunders Company. 0026-0495/90/3903-OOf35$03.00/0 242

are in accordance with the known increased sensitivity of Adex rats to endotoxin.4 We demonstrated that the lethal consequences of the TNF administration were associated with the impaired glucose homeostasis and thermoregulation observed in these rats. MATERIALS AND METHODS

Male rats of the Hebrew University strain, weighing 80 to 120 g, were used. They were maintained at 22°C with 12/12 lighting schedule and ad libitum access to a pellet diet (Am Rod 931, Hadera, Israel), and water. The rats received intravenous (IV) injection of 2.5 pg, 10 rg, 40 peg, or 90 pg/lOO g body weight rHuTNF dissolved in 1 mL of 0.9% NaCl containing 0.1% bovine serum albumin. The control rats received 1 mL of vehicle IV. The rats were deprived of food during the period following TNF injection and were killed by decapitation between 2:00 and 4:00 PM, 4 hours after TNF administration. Adrenalectomy was performed under phenobarbital (3.5 mg/ 100 g body weight) anesthesia 10 days before TNF treatment. Adex rats were given 0.9% NaCl instead of drinking water. We have determined basal serum corticosterone and adrenocorticotropic hormone (ACTH) levels in sham-operated and Adex rats 10 days after surgery. In sham-operated rats, serum corticosterone levels were 2 * 0.03 pg/dL and ACTH levels 51 + 5 pg/mL. In Adex rats, serum corticosterone levels were less than 0.5 pg/dL and ACTH levels 400 * 59 pg/mL. Four hours after injection of rHuTNF (2.5 rg/mL), serum corticosterone levels increased two- to threefold in sham-operated rats, while they were less than 0.5 pg/dL in Adex rats. These data demonstrate that the entire adrenal gland was removed and no significant functional adrenal tissue was present 10 days after surgery. In some animals, dexamethasone 0.1 mg/ 100 g body weight was injected intraperitoneally 20 hours and 2 hours prior to TNF administration. Indomethacin 0.3 mg/lOO g body Metabolism, Vol 39, No 3 (March), 1990: pp 242-250

EFFECTS OF TNF IN ADRENALECTOMIZED

RATS

weight was injected intraperitoneally 2 hours prior to TNF administration. In some experiments Adex rats received 1 mL of 10% glucose solution IV before rHuTNF administration and every 20 minutes thereafter. Thirty-microliter specimens of blood were taken at hourly intervals through an incision in the tail vein. Rectal temperature was measured at hourly intervals with a digital thermometer (ELLA A/S, Type DM-3, Copenhagen, Denmark) accurate to +O.l“C. After decapitation, blood was collected into ice-cooled tubes. Liver homogenates, 20% (wt/vol) were prepared at YC with 0.25 mol/L sucrose, 10 mmol/L HCl, and 0.5 mmol/L Na-EDTA, pH 7.4, using a Teflon/glass homogenizer. The homogenates were centrifuged at 12,000 xg for 20 minutes, and the resulting supernatant was recentrifuged for 45 minutes at 105,000 xg. The 105,000 “g” supematant was used for the determination of phcsphoeonolpyruvate carboxykinase (PEPCK) activity. The microsomes were resuspended in 0.25 mol/L sucrose, 10 mmol/L Tris HCl buffer, pH 7.4, to give 1 mL of suspension per gram wet liver and stored at - 70°C for determination of glucose-6-phosphatase activity. Glycogen Determination Pieces of liver, 100 to 200 mg, were removed and immediately digested with 2 mL of 33% potassium hydroxide. The glycogen was isolated by ethanol precipitation and subsequently hydrolyzed to glucose by amyloglucosidase; glucose was determined by glucose oxidase.’

243

Ehle and Schotz.8 Enzyme activity was calculated according to the formula of Nilsson-Ehle and Schotz* and was expressed per milligram of tissue. One millunit of enzyme activity represents the release of 1 nmol fatty acid per minute. Analytical

Materials rHuTNF 10’ U/mg was from Cetus, Emeryville, CA. It contains 0.08 ng lipopolysaccharide (LPS) per milligram of protein. Bovine serum albumin, indomethacin, dexamethasone, and triolein were obtained from Sigma Chemical, St Louis, MO. Glycerol tri[ 19,,10(n)‘] oleate, specific activity 1 Ci/mmol, and sodium [Ylbicarbonate, specific activity 50 to 60 Ci/mmol, were obtained from Amersham International, Little Chalfont, Bucks, UK. Statistical

Enzyme Assays Glucose-6-phosphatase activity was measured in isolated microsomes. Microsomes were diluted to 2 mg of protein per milliliter and incubated with 0.2% Na-deoxycholate for 20 minutes at O°C in order to obtain fully disrupted microsomes. Glucose-6-phosphatase activity was assayed for 10 minutes at 30°C in a final volume of 0.1 mL reaction mixture containing 10 mmol/L glucose-6-phosphate, 10 mg/mL albumin, and 50 mmol/L Tris-cacodylate, pH 6.5. Inorganic phosphate (Pi) produced was measured calorimetrically according to Bruchell et al6 PEPCK activity was determined according to the method of Chang and Lane,’ measuring exchange between KH14C0, and unlabeled oxoloacetate. Determination

of Lipoprotein

Lipase Activity

The heart was weighed, and 80-mg aliquots were homogenized in 4 mL 0.025 mol/L NH,-HCl buffer (pH 8.1), containing heparin 5 U/mL, using the Polytron (Kinematica GmbH, Lucerne, Switzerland) homogenizer with a PT 10 probe at maximum speed for two minutes at OOC.Adipose epididymal fat tissue was treated as follows: 100 to 500 mg of fresh tissue was homogenized in 50 mL ice-cold acetone using the Polytron homogenizer; the homogenates were centrifuged at 10,000 rpm for 20 minutes at 4°C and the supernatant discarded; residue was re-extracted three times with 50 mL ice-cold acetone and twice with diethyl ether; defatted preparation was dried at O°C under nitrogen. This preparation was designated acetone powder. Lipoprotein lipase activity was determined on homogenates of fresh tissue or acetone powder prepared in 0.025 mol/L NH,-HCl buffer (pH 8.1) containing heparin, 5 U/mL. The system consisted of 0.1 mL homogenate or 0.1 mL serum and 0.1 mL of substrate containing labeled triolein, prepared according to the method of Nilsson-Ehle and Schotz.* The specific radioactivity of the triolein moiety was 300 to 350 dpm/nmol triacylglycerol. Incubations were carried out at 37“C for 60 minutes, The reaction was stopped by addition of methanol/chloroform/heptane (1.4:1.25:1, vol/vol) and the extraction of fatty acids was performed according to Nilsson-

Procedures

Triacylglycerol and total cholesterol were determined by enzymatic methods according to the Lipid Research Clinics (LRC) method9 using a Vitatron autoanalyzer (Vital Scientific, Dieren, The Netherlands). Serum glucose concentration was determined by the glucose oxidase method, the reagent was obtained from Boe.hringer (Mannheim, FRG). Serum insulin was determined by insulin MAIA RIA kit (using human insulin standards) obtained from Serono Diagnostic (Coinsins, Switzerland). Serum lactate was determined by an enzymatic method.“’

Analysis

Significance of differences between experimental evaluated by Student’s t test.

groups was

RESULTS

Table 1 shows the mortality rate of animals injected with rHuTNF under various experimental conditions. All shamoperated animals survived the 4 hours following the injection of 2.5 or 10 fig/100 g body weight, but died following injection of 90 rg/lOO g body weight rHuTNF. Three quarters of the sham-operated rats died following injection of 40 rg/lOO g body weight. In contrast, in Adex rats, 100% of the rats injected with 10 rg/ 100 g body weight died within 4 hours following rHuTNF injection. Approximately 50% died following the administration of the smaller rHuTNF dose (2.5 rg/lOO g body weight). When Adex or sham-operated rats were injected with dexamethasone and subsequently given rHuTNF (2.5, 10, or 40 pg/lOO g body weight, respectively), the lethal effect was completely abolished. Pretreatment with indomethacin had no effect on the mortality rate of Adex rats injected with rHuTNF, but abolished the lethal effect of rHuTNF injection (40 rg/lOO g body weight) in sham-operated rats. The pharmacological effectiveness of indomethacin was verified by measuring brain prostaglandin E, (PGE,) content, whose content decreased by 40% with indomethacin treatment (data not shown). On the other hand, an intermittent infusion of glucose (every 20 minutes), completely prevented the lethal rHuTNF effect (2.5 pg/ 100 g body weight) and improved the survival rate of Adex rats injected with rHuTNF (10 Mg/ 100 g body weight) (Table 1). Figure 1 illustrates the change in body temperature under various experimental conditions where sublethal doses of rHuTNF were injected. In sham-operated (control) animals (Fig 1B), rHuTNF caused a significant increment in body temperature; at 1, 2, and 4 hours after injection of 2.5

CHAJEK-SHAUL ET AL

244

Table 1. Effect of Dexamethasone. Indomethacin. and Glucose on the Lethal Effect of rHuTNF in Adex Rats Adrenslsctomy

rHuTNF ~~g/lOO~bodywt~

Treatment

SurvivalRate I4 h)

Mean Survival Time (hl

O/4

4.8 + 0.1

l/4

4.2 + 0.2

414

-

so

_

40

-

-

40

Dexamethasone

_

40

-

10

lndomethacin -

414 15115

-

616 016 414 314

4

S/l6

2.9 + 0.4 3.3 * 0.2

_

2.5

+

10

+

10

Dexamethasone

+

10

Glucose -

+

2.5

+

2.5

+

2.5

lndomethacin

616 316

+

2.5

Glucose

616

Not measured

NOTE. Rats were injected IV with rHuTNF at doses of 2.5, 10, 40, or 90 pg/lOO g body weight. Dexamethasone was injected intraperitoneally, 0.1 mg/lOO g body weight at 20 hours end 2 hours before rHuTNF administration. lndomethacin (0.3 mg/lOO g body weight) wes injected intraperitoneally 2 hours prior to rHuTNF administration. One milliliter of 10% glucose solution wss injected IV at 0 time and thereafter every 20 minutes. Results are cumulative data from two to four different experiments.

and 10 rg/ 100 g body weight, the temperature increased by 0.3 + 0.25OC, 0.8 + 0.26W and 1.8 + 0.2OC and 1 + O.l°C, 1.6 * 0.2V and 2 + 0.4V, respectively. In Adex rats, a rapid rise in body temperature occurred as early as 1 hour following rHuTNF (2.5 pg/lOO g body weight) injection (1.7 + O.O4‘C, Fig 1A). The elevation in temperature was maintained at this level for up to 4 hours following rHuTNF injection. Pretreating of both sham-operated and Adex rats with dexamethasone and indomethacin prior to rHuTNF injection significantly attenuated the increase in body temperature (Fig 1A and B). The increase in body temperature following 2.5 pg/lOO g body weight rHuTNF injection in Adex rats was unaffected by glucose injection, being 1.75 f 0.18% 4 hours after injection. By contrast, the seven Adex rats that died following rHuTNF treatment (2.5 pg/lOO g body weight) (Table 1) exhibited transient increase in body temperature, followed by profound hypothermia (Fig 2A). Hypothermia also developed in Adex rats following injection of rHuTNF, 10 pg/lOO g body weight. The body temperature of the Adex rats that received indomethacin and rHuTNF and died, also declined. Intermittent glucose infusion to Adex rats treated with rHuTNF, 10 pug/100 g body weight, did not prevent a decline in body temperature (from 36 f 0.1% before to 33.8 + 0.1% 4 hours after rHuTNF injection). Pretreatment with dexa-

methasone to Adex rats followed by injection of rHuTNF, 2.5 or 10 c(g/lOO g body weight, was associated with an increase in body temperature, already observed 1 hour after injection of the monokine. Hypothermia also developed in sham-operated rats that received rHuTNF, 40 fig/100 g body weight (Fig 2B). Pretreatment with either dexamethasone or indomethacin abolished the decrease in body temperature. Following injection of 90 pgLg/ 100 g body weight, body temperature also declined, being 36 + 0.1% before and 35.5 2 0.2, 32.9 + 0.7, 32.5 + 0.8 at 2, 3, and 4 hours after rHuTNF administration, respectively. Blood glucose levels were not affected in the shamoperated rats receiving 2.5, 10, or 40 rg/lOO g body weight dose of rHuTNF (Table 2, Fig 3) but declined following injection of 90 pg/lOO g body weight, being 96 k 7.5 mg/dL before, and 73 f 9 mg/dL, 31 f 6 mg/dL, 10 + 5 mg/dL, and 5 * 1 mg/dL at 1, 2, 3, and 4 hours after rHuTNF treatment. Injection of rHuTNF to Adex rats elicited profound hypoglycemia even in the surviving rats (Fig 3, Table 2). Indomethacin did not affect the change in blood glucose levels induced by rHuTNF treatment in Adex rats. In contrast, dexamethasone or glucose prevented the hypoglycemia (Fig 3, Table 2). The Adex rats that died following rHuTNF injection developed severe hypoglycemia. Blood glucose decreased from a mean baseline value of 75 mg/dL

HOURS AFTER rHuTNF INJECTION

Fig 1. Effect of rHuTNF injection on body temperature. Rectal temperature was determined with a digital thermometer. Resuits are mean + SE of three to nine rats. (A) Adsx rats were injected IV with rHuTNF, 2.5 pg/lOO g body weight. 10) rHuTNFtreated Adex rats that survived; (0) rHuTNF-treated Adex rats pretreated with dexamothasone; (A) surviving rHuTNFtreated Adex rats pretreated with indomethacin. (Bj Sham-operated rats treated with rHuTNF: (0) 2.6 or jgj 10 fig/100 g body weight. Pretreatment with indomethatin and rHuTNF: (WI 2.6 or j-j 10 w/100 g body weight; pretreatment with dexamethacone and rHufNF: (A) 10 fig/l00 g body weight.

EFFECTS OF TNF IN ADRENALECTOMIZED

RATS

245

Fig 2. Decrease in body temperature following injection of lethal doses of rHuTNF. Results are mwns * SE of three to six rats. (A) Adex rats treated with rHuTNF: (0) 10 ug/loO g body weight: IV) 2.5 wg: (A) indomethecinand 2.5pg/loOg body weight; dexemethesone and rHuTNF. (‘1 2.6 cg or (0) 10 pg/lW g body weight. (El Shemoperated rats treated with rHuTNF. 40 pg/lDtl g body weight (+I; and indomethetin (A,); or dexamethasone (m).

+2 HOURS AFTER

to 6.2 + 1.6 mg/dL 1 to 3 hours following rHuTNF administration. The Adex rats that died following indomethatin and rHuTNF injection also developed severe hypoglycemia, with serum glucose levels of 7.2 + 1 mg/dL 2 hours after rHuTNF treatment. The rHuTNF-treated Adex and sham-operated rats first had piloerection and this was followed by lethargy, somnolescence, and gasping; diarrhea developed only in sham-operated rats that received rHuTNF, 90 rg/ 100 g body weight. These data demonstrate that Adex rats developed severe hypoglycemia following low doses of rHuTNF administraTable 2. Alteration Treatment Sham-operated

Sham-operated + daxamethasone

Sham-operated + indomethacin

Adex

tion. Since the hypoglycemia appeared to be an important factor contributing to the mortality of the Adex rats, we further studied the effect of rHuTNF on other relevant biochemical parameters in Adex rats, including serum insulin and lactate concentrations, liver glycogen content, and the key regulatory gluconeogenetic enzymes, PEPCK and glucase-Cphosphatase. Serum insulin concentration in sham-operated rats did not change after rHuTNF administration (Table 2), whereas a significant decrease in serum insulin levels was observed in Adex rats treated with rHuTNF with or without indometha-

SsrumInsulin(pU/mL)

With rHuTNF

SerumGlucose(mg/d_)

SerumLsctsts (mmd/L)

29 * 2.0

108 zt 5.3

1.5 f 0.1

2.5

31 f 2.7

118 f 6.7

3.1 * 0.2.

10.0

2.9 f 0.2.

34 + 2.3

119 f 2.0

-

82 f 5.1

140 * 6.0

3.6 f 0.2

2.5

75 f 2.0

163 f

17.0

4.3 * 0.1.

160 i

10.0

6.6 * 0.6.

10.0

80 i 8.5

40.0

108 f 5.9.

151 f 9.5

-

37 * 3.1

114 * 5.0

1.7 zt 0.3

2.5

52 zt 8.0

105 i 3.2

4.2 + 0.1’

40.0

34 * 5.5

97 + 8.7

6.1 + 0.6.

-

24 f 4.6

90 + 12

1.7 * 0.1

32 f 2.4”

3.3 * 0.2*

6 * 1.0.

6.6 * 0.2.

-

107 + 7.6

135 * 9.1

2.5

105 f 4.3

167 + 17

7.0 * 0.3*

81 + 8.0

145 f 5.8

6.6 f 0.6. 1.6 f 0.2 2.7 f O.l*

-

30 f 3.5

115 zt 6.4

2.5

35 * 4.9

120 * 15

10 Adex + indomethacin

INJECTION

-

10 Adax + glucose

rHuTNF

in Serum Insulin, Glucose, snd Lactate in Adex Rats Treated rHuTNF(jq/lOO g bodywt)

2.5 Adex + dexamethasone

+4

43 f 7.4

96k

11

5.0 * 0.4

6.5 f 0.3.

-

35 f

1.0

102 * 5.1

1.1 + 0.2

2.5

10 f

1.6.

41 * 5’

2.2 * 0.2.

NOTE. Aa in Table 1. Rats were injected IV with rHuTNF, as indicated. Four hours later, the rats were killed and blood was collected. Results are mean f SE of four to 16 animals. lP < .05 when compared with its untreated control.

246

CHAJEK-SHAUL ET AL

18C

weight, was injected (data not shown). In the surviving Adex rats (2.5 pg/lOO g body weight), rHuTNF caused a 98% depletion of liver glycogen content compared with the untreated Adex rats (Fig 4). This glycogen depletion was not affected by indomethacin administration. IV glucose infusion produced a sevenfold increase in liver glycogen content compared with liver glycogen content of rHuTNF-treated Adex rats (2.5 pg/ 100 g body weight). Pretreatment of Adex rats with dexamethasone resulted in a threefold increase in liver glycogen content compared with liver glycogen content of sham-operated or Adex rats, which was diminished by only 30% 4 hours after rHuTNF (2.5 pg/ 100 g body weight) administration (Fig 4). In Adex rats, rHuTNF treatment caused a significant reduction in liver PEPCK and liver microsomal glucose-6 phosphatase activities (Table 3). No effect on the activity of these enzymes was observed in sham-operated rats receiving the same dose of rHuTNF (2.5 rg/lOO g body weight). A 40% to 50% reduction was observed following administration of 10 pg/lOO g body weight (V. Barash, T. Chajek-Shaul, unpublished results). rHuTNF administration (2.5 to 10 kg/100 g body weight) to sham-operated and Adex rats caused a twofold increase in serum triacylglycerol concentrations without a change in serum cholesterol (Table 4). Lipoprotein lipase activity in

60 30

4 2 HOURS AFTER rHuTNF INJECTION Fig 3. Effect of rHuTNF injection on serum glucose concentration. Results are mean f SE of three to nine rats. (0) Untreated Adex rats; (A) rHuTNF. 2.6 c(g/lOO g body weight, treated Adex rats, pretreated with indomethacin that survived. (0) rHuTNF. 2.6 wg/lOO g body weight. treated Adex rsts that survived; (Vj Sham-operated rats; (Aj sham-operated rats treated with rHuTNF. 2.6 rg/lOO g body weight: (0110 ccg/lOO g body weight, end 101 4Oxg/lOOg bodyweight: (*j Adexrats pretreated with dexamethssone only or with rHuTNF, (0) 2.6 w/lo0 g body weight, or (Wj 10 pg/lOO g body weight: (0) sham-operated rats treated with dexamethasone only or with rHuTNF, (4) 40 pg/lOO g body weight.

tin. Pre-administration of dexamethasone resulted in an increase in serum insulin concentrations in both rHuTNFtreated and untreated Adex rats. Therefore, it appears that insulin secretion responded appropriately to the changes in serum glucose levels. Serum lactate increased by about twofold following rHuTNF administration in both control and Adex rats (Table 2). Indomethacin and glucose treatment did not abolish the increased serum lactate induced by rHuTNF. In dexamethasone-treated Adex and shamoperated rats, the serum lactate concentration was significantly higher than in the respective Adex control group (P < .005); a further increase occurred 4 hours after rHuTNF treatment. Liver glycogen of sham-operated rats decreased by 50% 4 hours after injection of rHuTNF, 2.5 gg/ 100 g body weight (Fig 4) and by 85% when rHuTNF, 10 rg/lOO g body

TNF-

+

control

-

-+++

+

adrenalectomized

Fig 4. Effect of rHuTNF injection on rat liver glycogen content. Hatched bars, Adex rats treated with rHuTNF and repeated injection with glucose. Black bars, Adex rats treated with dexamethasone. Arrow, rHuTNF-treated Adex rats that recehred indomethacin. Results ere mean of four to six rets and the vertical lines represent r SE. lP ~06 when compared with its rHuTNF untreated control: l*P ~026 when compared with rHuTNFtreeted Adax rats; l**P ~005 when compared with nontreated sham-operated rats, or Adex rats.

EFFECTS OF TNF IN ADRENALECTOMIZED

RATS

247

Table 3. Effect of rHuTNF on Liver PEPCK and Glucose-6-Phosphatase

rHuTNF (Wa/l@J g body wt)

Treatment

DISCUSSION

Activity

The present results demonstrated an increased sensitivity of Adex rats to IV injection of rHuTNF as manifested by a marked increase in mortality. Administration of 3% or 12% of the rHuTNF dose that was lethal in intact rats (90 kg/100 g body weight), caused a mortality rate of 50% or 100% in the injected Adex rats, respectively (Table 1). A similar increased sensitivity has been described when endotoxin was injected into Adex rats4 Bertini et al” recently reported that adrenalectomy enhanced the sensitivity of mice to the lethal effect of both interleukin-1 (IL-l) and TNF. Pre-administration of dexamethasone protected the Adex and shamoperated rats from the lethal effect of TNF.” However, it seems that the response of Adex or sham-operated rats treated with lethal doses of TNF after indomethacin administration differed. Indomethacin, an arachidonic acid cyclooxygenase inhibitor known to decrease the inflammatory action of TNF and protect the rat against its lethal effect,‘* did not change the mortality rate in rHuTNF-treated Adex rats (Table l), confirming previous findings,” but did attenuate the hyperthermia observed in the surviving rats (Fig 1). Fever is a most prominent in vivo response to infection and is mediated via secretion of endogenous pyrogens from activated macrophages, eg, IL-l, TNF. Indeed, injection of sublethal doses of TNF to rabbits,‘3v’4rats,15 and humans”‘s” elicited a febrile response. Increase in body temperature is thought to occur as a result of the action of PGEs on the pre-optic anterior hypothalamus.‘**‘9 Indomethacin would reduce PGE production and, indeed, it reduces the febrile response to TNF in both Adex and sham-operated rats. Dexamethasone has an antipyretic effect towards LPS, IL-1,19 and TNF (Fig 1). This effect may be attributed to its

GIUCOSSBPEPCK Phosphstsse Activity Activity fnmd/min/mg pcotsin)

Sham-operated

-

258 f 44

411

253 f 31

364 f 20

AdeX

2.5 -

294 f 31

370 f 35

2.5

181 zt 15.

215 f 38.

NOTE. Rats were injected IV with 2.5 pg/lOO

f 31

g body weight. Four

hours later, the rats were killed and the liver was immediately removed and homogenized as described in Materials and Methods. PEPCK activity was measured in 105,000

x g supematant fraction and glucose-B-

phosphatase

measured

activity

was

crosomes. Results are mean f

on deoxycholate-treated

mi-

SE of six to eight animals.

+P < .05 when compared with untreated Adex rats.

heart muscle was not affected by rHuTNF administration. Pre-administration of dexamethasone in Adex and shamoperated rats induced a significant increase in heart lipoprotein lipase. Four hours after rHuTNF injection to dexamethasone-treated Adex or sham-operated rats, heart lipoprotein lipase was reduced by 30%. A 50% to 60% decrease in adipose tissue lipoprotein lipase activity was observed in both control and Adex rats treated with rHuTNF. Although glucose infusion caused an increase in adipose tissue lipoprotein lipase activity in Adex rats, rHuTNF was still able to induce a significant decrease in enzymic activity even in the glucose-treated animals. Pre-administration of dexamethasone and indomethacin to sham-operated and Adex rats did not affect the response of adipose tissue lipoprotein lipase to rHuTNF.

Table 4. Effect of rHuTNF on Serum Lipids and on Tissue Lipoprotein

Lipase Activity

Swum Lipids

Treatment Sham-operated

rHuTNF (re/tgo g bodvw)

Chdestsrol kna/dL)

He&l (mu/g)

108 f

lot

71 f 5.9

4,116

f 126

1,333

f 195

2.5

185 f

15t

65 f 1.1

2,116

f 168t

1,280

zt 115

231 zt 23t

69 i 6.6

1,916

f 133t

1,516

f 126

84 zt 2.7

64 f 3.0

4.250

f 413

1,325

f 133

162 f 3.lt

60 f 3.0

1,666

f 258t

1,468

+ 81

-

82 + 3.0

102 f 10.0

6,166

f 246

2,400

f 158

2.5

97 f 6.0t

108 i 5.1

3,433

& 166t

1,883

+ 158

132 f 17.0t

101 f 1.0

2,966

i 383t

1,633

f 183t

3,433

f 350

2.5

Sham-operated + dexamethasone

LipoproteinLipsssActivity Epididymal Fst Pad WJ/e)*

10

Sham-operated + indomethacin

Triscylglycerol tmg/dL)

10 Adax

-

105 i 8

88 i 6

2.5

208 f 13

102 f 10

1,850

i

195t

871 & 103

Adex

-

110 f 13.2

64 f 2.6

3,533

f

133

983 f 82

2.5

149 * 1o.ot

58 f 7.1

2,683

* 200t

Adex + dexamethasone

-

166 * 9.3

104 f 4.7

6,393

+ 263

2,516

f

2.5

262 f 23.7t

107 f 6.7

3,350

+ 203t

1,966

f 141t

1,443

f 66t

10 Adex + glucose

206 f 13.0t

86 f 10.0

3,019

+ 217t

-

113 f 5.5

75 * 5.3

5,733

* 200

2.5

170 * ll.OT

91 + 7.7

2,986

zt 271t

220 f 29.0t

72 f 5.5

3,186

zt 18lt

10

803 f 91

950 f 75

900 f 1,283

166

111

* 143

783 f 110

NOTE. As in Table 1. Rats were injected IV with rHuTNF as indicated. Four hours later, the rats were killed, and blood, heart, and epididymal fat tissue were removed and processed as described in Materials and Methods. Results are mean f *mu/g

acetone powder.

tP -z .005 when compared with its untreated control.

SE of four to six animals.

240

ability to induce synthesis and release of phospholipase-A, (PLA,) inhibitory proteins (lipocortins). Inhibition of PLA, led to a decrease in pro-inflammatory mediators, such as PGE,.*’ A progressive reduction in body temperature was observed after injection of a lethal dose of TNF to both sham-operated and Adex rats (Fig 2), in accordance with the observations of Kettelhut et al.‘* In both studies, this hypothermic effect was inhibited by pre-administration of indomethacin to intact rats, suggesting that an increase in PGE levels is involved in the pathogenesis of the observed hypothermia. However, indomethacin had no effect on the hypothermia observed in Adex rats. This may indicate that the presence of glucocorticoids is needed for indomethacin to express its antihypothermic effect. Pre-administration of dexamethasone to both sham-operated and Adex rats prevented the TNF-induced hypothermia and a slight increase in body temperature was even observed. The mechanism of this effect may involve its antiinflammatory action via a decrease in PGE synthesis or a decrease in the TNF receptor affinity, similar to that described for dexamethasone in cultured cells.8 Glucocorticoids are also important for maintaining body temperature and for thermoregulatory responses to low environmental temperatures.*’ This effect may operate through more than one action of the glucocorticoids, eg, pressor effect, potentiation of the effect of lipolytic hormones (catecholamines), and their effect on the carbohydrate fuel reserves. Thus, dexamethasone, in addition to being a potent inhibitor of TNF synthesis in activated macrophages, is able to attenuate or even to prevent the inflammatory and toxic response of rHuTNF after being released. Another major event observed in the Adex rats that died following rHuTNF injection, was severe hypoglycemia which already developed after a low dose of rHuTNF (2.5 pg/ 100 g body weight). In our study, only sham-operated rats injected with 90 pg/ 100 g body weight rHuTNF developed hypoglycemia. Kettelhut et al’* observed severe hypoglycemia following injection of 400 rg/lOO g body weight rHuTNF, while Tracey et al** did not observe hypoglycemia in rats injected with TNF up to 190 rg/lOO g body weight. Since it is well established that body temperature falls during hypoglycemia,23-25it appears that the hypothermia observed in Adex rats treated with rHuTNF may be aggravated by the hypoglycemia. This suggests that a major reason for the increased sensitivity of Adex rats to the lethal effect of rHuTNF is associated with the impaired glucose homeostasis. Adex or low doses of rHuTNF administration did not cause hypoglycemia. However, the combination of both treatments led to a severe reduction in blood glucose concentration. Pretreatment with either glucose infusion or dexamethasone protected the rats from the hyoglycemic and the lethal effects of rHuTNF in Adex rats. Therefore, it appears that the absence of circulatory glucocorticoids, which play a crucial role in glucose counterregulation, was responsible for the hypoglycemia and the resulting lethal rHuTNF effect in the presence of adrenalectomy. The mechanism by which rHuTNF induced hypoglycemia is at present unknown. It is not mediated via PGE synthesis,

CHAJEK-SHAUL ET AL

as the hypoglycemia was not affected by indomethacin (Fig 3). rHuTNF administration to control rats did not cause any change in serum insulin concentration within 4 hours (Table 2) and initially it may even be increased.26 In rHuTNFtreated Adex rats, insulin levels were decreased (Table 2) in agreement with previous work,*’ suggesting that the reduction of blood glucose is not mediated by an increase in insulin output. Similar results have been described for the effect of IL-l in Adex mice and rats.*‘.*’ Available data strongly suggest that rHuTNF administration causes an increase in glucose utilization in both in vivo and in tissue-cultured cells.29*‘0It was shown that TNF stimulates hexose transport via an increased activity of glucose transporter associated with accumulation of hexose transporter mRNA.3’ This effect may be enhanced in Adex rats owing to a lack of glucocorticoids which decrease glucose uptake to peripheral cells. Glucocorticoids and insulin affect the cellular glucose uptake by altering the distribution of glucose transporters between the plasma membrane and intracellular sites of peripheral cells in an opposite direction.32.33 rHuTNF affects liver carbohydrate metabolism. Thus, a virtually complete depletion of liver glycogen content was observed after rHuTNF administration to Adex rats (Fig 4) as compared with partial depletion in sham-operated rats or dexamethasone-treated Adex rats. A small but significant increase in liver glycogen content was observed following glucose infusion to rHuTNF-treated Adex rats. Similar repletion of liver glycogen content occurred after infusion of glucose to endotoxin-treated rats._“’ Reduced liver PEPCK and glucose-6-phosphatase activities were observed after administration of rHuTNF (2.5 pg/ 100 g body weight) to Adex rats. In preliminary experiments we have shown that rHuTNF administration to intact rats inhibited PEPCK and glucose-6-phosphatase activity in a dose-dependent manner. A similar reduction in enzymic activity was described previously following endotoxin and IL-I administration.35*36 Our data indicate that rHuTNF is more potent in Adex rats. Thus, the profound decrease in glycogen content, PEPCK, and glucose-6-phosphatase activities in rHuTNFtreated Adex rats potentially altered liver glucose production and release, in a manner that was no longer sufficient to compensate for the increase in glucose utilization by extrahepatic tissues3’ However, a more detailed study on carbohydrate turnover is needed to determine the functional importance of these observations. Serum lactate concentration was moderately elevated in both control and Adex rats that received rHuTNF. It possibly resulted from an increased rate of glycolysis in extrahepatic tissues. This is similar to the report by Lang et al,” who studied carbohydrate metabolism in the hypermetabolic septic rat. Adrenalectomy did not modify the effect of rHuTNF administration on observed changes in plasma lipids, namely an increase in serum triacylglycerol and a decrease in adipose tissue lipoprotein lipase activity (Table 4). Thus, adipose tissue lipoprotein lipase activity was reduced by 50% to 60% in rHuTNF-treated rats. It is important to note that adipose tissue lipoprotein lipase activity, although reduced after

249

rHuTNF injection, did increase after glucose administration. Therefore, it seems that adrenalectomy did not alter the response of adipose tissue lipoprotein lipase activity to rHuTNF or to glucose administration. Rat heart lipoprotein lipase activity did not change following rHuTNF injection. Dexamethasone administration to Adex or sham-operated rats caused an increase in heart lipoprotein lipase activity, which was in turn inhibited by rHuTNF injection. A similar observation was reported in isolated perfused rat hearts excised from glucocorticoid-treated rats.‘* The mechanism of increase in plasma triacylglycerols is probably related to increased triacylglycerol synthesis and secretion, rather than to the observed changes in lipoprotein lipase activity.39*40 It seems that in metabolic states where glucose homeosta-

sis is impaired, such as adrenal insufficiency, addition or in vivo production of low doses of rHuTNF induces severe or even lethal hypoglycemia and hypothermia. This may also occur in other metabolic states where glucose homeostasis is impaired and carbohydrate reserves are low, eg, after prolonged starvation and severe liver disease and hyperinsulinemia.4’ Glucose infusion restores blood glucose levels, causing some repletion of liver glycogen content and this may prevent the fatal outcome. ACKNOWLEDGMENT

The authors wish to thank 0. Halimi, S. Yosha, A. Sanelvich, and B. Morag for excellent technical assistance, and J. Hollander for expert secretarial assistance.

REFERENCES

1. Sherry B, Cerami A: Cachectin/tumor necrosis factor exerts endocrine, paracrine, and autocrine control of inflammatory responses. J Cell Biol 107:1269-1277,1988 2. Ku11FC Jr: Reduction in tumor necrosis receptor affinity and cytotoxicity by glucocorticoids. Biochem Biophys Res Comm 153:402409,1988 3. Hepburn A, Boeynaems JM, Fiers W, et al: Modulation of tumor necrosis factor-a cytotoxicity in L929 cells by bacterial toxins, hydrocortisone and inhibitors of arachidonic acid metabolism. Biothem Biophys Res Commun 149:8 15822,1987 4. Abernathy RS, Halberg F, Spink WW: Studies on the mechanism of chlorpromazine protection against Brucella endotoxins in mice. J Lab Clin Med 49:708-715,1957 5. Hujing F: A rapid enzymatic method for glycogen determination in very small tissue samples. Clin Chim Acta 30:567-572, 1970 6. Bruchell A, Hume R, Bruchell B: A new microtechnique for the analysis of the human hepatic microsomal glucose-6-phosphatase system. Clin Chim Acta 173:183-192, 1988 7. Chang ME, Lane MD: The enzymatic carboxylation of phosphoenol pyruvate. II. Purification and properties of liver mitochondrial phosphoenolpyruvate carboxykinase. J Biol Chem 241:24132429,1966 8. Nilsson-Ehle P, Schotz MC: A stable, radioactive substrate emulsion for assay of lipoprotein lipase. J Lipid Res 17:536-541, 1976 9. Lipid Research Clinics Program Manual of Laboratory Operations: Vol 1: Lipid and lipoprotein analysis. DHEW Publication No. (NIH) 76-628, Washington DC, US Government Printing Office, 1976 10. Lowry OH, Passonneau JC: A Flexible System of Enzymatic Analysis. San Diego, CA, Academic, 1972 11. Bertini R, Bianchi M, Ghezzi G: Adrenalectomy sensitizes mice to the lethal effects of interleukin 1 and tumor necrosis factor. J ExpMed 167:1708-1713,1988 12. Kettelhut IC, Fiers W, Goldberg AL: The toxic effects of tumor necrosis factor in vivo and their prevention by cyclooxygenase inhibitors. Proc Nat1 Acad Sci USA 8:4273-4277,1987 13. Dinarello CA, Cannon JG, Wolff SM, et al: Tumor necrosis factor (cache&in) is an endogenous pyrogen and induces production of interleukin 1. J Exp Med 163:1433-1450, 1986 14. Morimoto A, Sakata Y, Watanabe T, et al: Characteristics of fever and acute-phase response induced in rabbits by IL-l and TNF. Am J Physiol256:R35-R41.1989 15. Kettelhut IC, Goldberg AL: Tumor necrosis factor can induce fever in rats without activating protein breakdown in muscle or lipolysis in adipose tissue. J Clin Invest 81:1384-1389, 1988

16. Starnes HF Jr, Warren RS, Jecvanandam M, et al: Tumor necrosis factor and the acute metabolic response to tissue injury in man. JClin Invest 82:1321-1325,1988 17. Michie HR, Spriggs DR, Manogue KR, et al: Tumor necrosis factor and endotoxin induce similar metabolic responses in human beings. Surgery 104:280-286, 1988 18. Dascombe MJ: The pharmacology of fever. Prog Neurobiol 251327-373, 1985 19. Abul H, Davidson J, Milton AS, et al: Dexamethasone pretreatment is antipyretic toward polyinosinicpolycytidylic acid, lipopolysaccharide and interleukin l/endogenous pyrogen. Arch Pharmacol335:305-309, 1987 20. Weidenfeld J, Abu-Amer Y, Shohami E: Inhibition of prostagiandin synthesis in brain of rat by dexamethasone: Lack of effect of dexamethasone phosphate ester and various hormonal steroids. Neuropharmacology 27:1295-1299.1988 2 1. Deavers DR, Musacchia XJ: The function of glucocorticoids in thermogenesis. Fed Proc 38:2177-2181, 1979 22. Tracey KJ, Beutler B, Lowry SF, et al: Shock and tissue injury induced by recombinant human cachectin. Science 234:470474,1986 23. Dudley S, Laidlaw PP. Trevan JW, et al: The effect of insulin on the respiratory exchange. J Physiol57:47, 1923 24. Cassidy GJ, Dworkin S, Finney WH: Insulin and the mechanism of hibernation. Am J Physiol73:417-427,1925 25. Dworkin S, Finney WH: Artificial hibernation on the wocdchuck (Arctomys monas). Am J Physiol80:75-8 1, 1927 26. Ciancio MJ, Jones SB, Yelilch MR, et al: Glucoregulatory and sympathoadrenal responses to TNF in conscious rats. Fed Proc 46:561, 1987 (abstr) 27. Del Rey A, Besedovsky H: Interleukin 1 affects glucose homeostasis. Am J Physiol253:R794-R798, 1987 28. Besedovsky H, Del Rey A: Neuroendocrine and metabolic responses induced by interleukin-1. J Neurol Sci 18:172-178, 1987 29. Lee MD, Zentella A, Pekala PH, et al: Effect of endotoxininduced monokines on glucose metabolism in the muscle cell line L6. Proc Nat1 Acad Sci USA 84:2590-2594, 1987 30. Meszaros K, Lang CH, Bagby GJ, et al: Tumor necrosis factor increases in vivo glucose utilization of macrophage rich tissues. Biochem Biophys Res Commun 149:1-6,1987 3 1. Hansen LK, Jensen TL, Peterson M, et al: Heat-shock protein synthesis in mitogen activated human lymphocytes: Kinetics of protein and mRNA induction and mitogen induced thermotolerance. J Cell Biol 167:718a, 1988 (abstr) 32. Carter-Su C, Okamoto K: Effect of insulin and glucocorti-

CHAJEK-SHAUL

250

coids on glucose transporters 252:E441-E453,1987

in rat adipocytes. Am J Physiol

33. Horner HC, Munck A, Lienhard GE: Dexamethasone causes translocation of glucose transporters from the plasma membrane to an intracellular site in human Iibroblasts. J Biol Chem 262:1769617702,1987 34. Lang CH, Bagby GJ, Buday AZ, et al: The contribution of gluconeogenesis to glycogen repletion during glucose infusion in endotoxemia. Metabolism 36:180-187, 1987 35. Berry LJ, Goodrum KJ, Resnick IG, et al: Bacterial toxins, in Liehr H, Grun M (eds): The Reticuloendothelial System and the Pathogenesis of Liver Disease. Amsterdam, Elsevier/North Holland, 1980, pp 257-266 36. Hill MR, Stith RD, McCallum RE: Interleukin 1: A regula-

ET AL

tory role in glucocorticoid-regulated hepatic metabolism. J Immunol 137:858-862, 1986 37. Lang CH, Bagby GJ, Spitzer JJ: Carbohydrate dynamics in the hypermetabolic septic rat. Metabolism 33:959-963, 1984 38. Hulsmann WC, Dubelaar M-L: Effects of tumor necrosis factor (TNF) on lipolytic activities of rat heart. Mol Cell Biochem 79:147-151, 1988 39. Feingold KR, Grunfeld C: Tumor necrosis factor-alpha stimulates hepatic lipogenesis in the rat in vivo. J Clin Invest 80:184-190, 1987 40. Chajek-Shaul T, Friedman G, Stein 0, et al: Mechanism of the hypertriglyceridemia induced by tumor necrosis factor administration to rats. Biochim Biophys Acta 1001:316-324, 1989 41. Yelich MR, Filkins JP: Insulin hypersecretion and potentiation of endotoxin shock in the rat. Circ Shock 9:589-603, 1982