Tumor Necrosis Factor: A Cytokine Involved in Toxic Effects of Endotoxin

(1)

REVIEWS OF INFECTIOUS DISEASES • VOL. 9, SUPPLEMENT 5 • SEPTEMBER-OCTOBER 1987

Tumor Necrosis Factor: A Cytokine Involved in Toxic Effects of Endotoxin

D. N. Mannel, H . Northoff, F. Bauss, and W. Falk From the Institut fiir Immunologic und Genetik, Deutsches Krebsforschungszentrum, Heidelberg; and Deutsches Rotes Kreuz Blutspendezentrale, Ulm, Federal Republic of Germany

Endotoxin-induced tumor necrosis has been shown to be mediated by a factor termed tumor necrosis factor (TNF). The biochemical nature, source, and mode of induction of T N F have been clarified. T N F is a mediator of activated macrophages that is released into the supernatant by these cells after their stimulation with endotoxin. A number of biologic functions of T N F other than its tumoricidal activity have been demonstrated.

In vivo, T N F induces reactions similar to those induced by endotoxin. Hypothermia, elevated hematocrit and plasma lactate levels, and reduced plasma glucose levels have been measured in mice injected with TNF. These reactions typically occur soon after endotoxin injection and are induced with purified recombinant T N F in mice that exhibit a low response to endotoxin. No TNF-induced production of interleukin 1 (which can induce similar effects) was detected in macrophage/monocyte cultures. Therefore, T N F appears to mediate endotoxin effects directly.

Bacterial endotoxin has been shown to exert detri- mental as well as beneficial effects i n various ex- perimental models i n w h i c h different means were used for c o m p r o m i s i n g the host. Toxic effects, fever, t u m o r regression, a n d enhancement o f n o n - specific resistance to infection and lethal irradiation are some examples o f endotoxin reactions. In recent years it has become apparent that some o f these effects are accomplished by the interaction o f endotoxin with its m a i n target cell — the macrophage. Sen- sitivity to endotoxin depends o n the state o f acti- v a t i o n o f the host's l y m p h o r e t i c u l a r system.

E n d o t o x i n is extremely toxic for activated macrophages i n B C G - i n f e c t e d animals [1] o r for macrophages that have been activated i n vitro [2]. A l s o , release o f mediators is qualitatively a n d quantita- tively related to the degree o f macrophage activa- t i o n [3]. T h u s , the status o f the reactivity o f macrophages is o f great importance to the lethal effects o f endotoxin. M a n y investigations have addressed the role o f macrophages and their release o f humoral mediators i n endotoxic effects.

T u m o r necrosis factor ( T N F ) is a protein released by activated macrophages u p o n their stimulation

We thank Dr. Peter Krammer for his critical reading of the manuscript.

Please address requests for reprints to Dr. Daniela N . Mannel, Institut fur Immunologic und Genetik, Deutsches Krebsforschung- szentrum, Im Neuenheimer Feld 280, 6900 Heidelberg, Federal Republic of Germany.

with bacterial lipopolysaccharide ( L P S ) [2]. The mo- lecular c l o n i n g o f the c D N A for T N F a n d the expression o f this c D N A i n Escherichia coli [4-6] have made available large amounts o f the recombinant protein for b i o l o g i c research. Before the nature o f the T N F molecule was k n o w n , T N F was defined by its t u m o r necrotizing activity i n vivo a n d its t u m o r cell cytotoxic activity i n vitro. Investigations o f the biologic activities o f p u r i f i e d recombinant T N F revealed that T N F not o n l y serves as a cytotoxic m o l - ecule for t u m o r cells but also has a n array o f effects o n different target cells. C o m p a r i s o n o f the a m i n o acid sequences o f T N F a n d cachectin, another m o n o k i n e that is secreted i n response to endotoxin or other bacterial a n d p r o t o z o a l products, revealed the identity o f the two proteins [7], T h e activity o f cachectin leads to a decrease i n lipoprotein lipase activity and to the metabolic derangements that occur d u r i n g infection and cachexia [8]. A n t i b o d i e s to the purified c a c h e c t i n / T N F molecule inhibited the lethal activity o f endotoxin, a finding indicating a role for this mediator i n endotoxic effects [9, 10].

Direct analysis o f the b i o l o g i c effects o f T N F i n different test systems further demonstrated the i m - portance o f T N F i n i n f l a m m a t o r y events. P r o d u c - t i o n o f prostaglandin E² a n d collagenase by h u m a n synovial cells and dermal fibroblasts [11]; i n d u c t i o n o f procoagulant activity a n d tissue factor by the endothelium; a n d propagation o f the coagulation pathway leading to the deposition o f f i b r i n [12], bone resorption, a n d i n h i b i t i o n o f bone f o r m a t i o n

S602

(2)

[13] are effects o f T N F t y p i c a l for inflammatory responses.

T N F also exerts biologic effects on the neutrophil, another cell that plays a central role i n i n f l a m m a - tory processes. Enhancement o f phagocytic and antibody-dependent cytotoxic activity [14] and o f adherence o f neutrophils to endothelial cells [15] i n response to T N F has been reported. Furthermore, T N F is an important i m m u n o l o g i c mediator. It enhances eosinophil toxicity [16] and increases surface expression o f class I major histocompatibility c o m - plex antigens o n endothelial cells and dermal fibroblasts [17].

In this c o m m u n i c a t i o n we compare some effects o f recombinant T N F i n vivo w i t h typical effects o f endotoxin. T N F and L P S induced similar thermoregulatory effects and changes i n b l o o d parameters when administered iv to mice. T N F induced the observed effects rather directly and not v i a the induction o f interleukin 1 (IL-1) secretion, since T N F d i d not induce IL-1 production i n our macrophage culture systems. These observations, therefore, further support the n o t i o n that T N F is involved i n en- d o t o x i n reactions and may act as an endogenous mediator o f endotoxin.

Materials, Methods, and Results

In an effort to show the connection between L P S and T N F , we measured some parameters that change significantly after iv administration o f bacterial L P S . These changes were compared with the effects i n - duced by the iv administration o f purified recombinant human T N F . The T N F preparation was k i n d - ly provided by B A S F (Ludwigshafen, F R G ) and

Table 1. Blood parameters in mice after the injection of bacterial lipopolysaccharide (LPS) or tumor necrosis factor (TNF).

Hematocrit Glucose Lactate

(%) (mg/100 ml) (mg/100 ml)

Stimulus ± SD ± SD ± SD

Control 42.3 ± 1.7 82.3 ± 23.2 62.9 ± 12.1 L P S 56.6 ± 7.3 32.9 ± 11.5 58.6 ± 8.9 T N F 63.0 ± 1.0 14.8 ± 0.4 85.3 ± 7.9 N O T E . Each group comprised three age-, sex-, and weight- matched C 3 H / H e mice. Animals were injected iv with L P S (Salmonella montevideo strain SH94, 200 ng), T N F (60 ng), or PBS (control, 100 Hematocrit, plasma lactate, and plasma glucose levels were determined from blood samples taken 6 h after injection.

contained <0.03 ng o f e n d o t o x i n / m g o f protein, as determined i n the limulus lysate assay. W h e n T N F was injected iv into mice at a concentration o f 40

\xg per a n i m a l (2 m g / k g ) , the first symptoms resem- bled the toxic effects seen after endotoxin applica- t i o n ; w i t h i n hours ruffled fur, diarrhea, hypothermia, and loss o f b o d y weight were noted. T h e extent of the observed effects was dependent o n the mouse strain used. The sensitivity o f a given mouse strain ( C 3 H / H e or C 3 H / H e J ) differed considerably from experiment to experiment. Therefore, only values obtained i n the same experiment were compared.

Changes in Blood Parameters

B l o o d parameters o f LPS-sensitive animals ( C 3 H / He) that had received 100 u.g o f L P S {Salmonella montevideo strain S H 9 4 prepared according to the modified phenol-water extraction method [18]) iv changed significantly w i t h i n 6 h . A n i m a l s were bled from the retroorbital plexus at intervals, and b l o o d or plasma parameters were enzymatically determined with use o f commercial kits (GlucoQuant and M o n o - test-Lactate kit, b o t h from Boehringer, M a n n h e i m , F R G ) . Hematocrit levels increased within 30 m i n (table 1). Plasma glucose levels dropped during the first 6 h after administration o f L P S . P l a s m a lactate levels were not changed in the animals injected with L P S despite the fact that very large and almost lethal amounts o f L P S are reported to enhance the lactate levels [19]. Similar results were obtained i n mice i n - jected with T N F . A f t e r T N F injection the plasma glucose levels were even lower than those after L P S injection, and the plasma lactate levels were en- hanced. The contaminating amount o f endotoxin i n the T N F preparation (<2 pg) was not able to induce these effects (data not shown). A l s o , mice w i t h a low response to L P S ( C 3 H / H e J strain) reacted to T N F with the same changes as C 3 H / H e mice (data not shown), a finding that indicates these effects were due to T N F itself.

Thermoregulatory Activity of TNF

M i c e react with hypothermia when injected w i t h L P S , whereas most other species develop a fever [20].

A s shown i n figure 1, T N F caused a hypothermic reaction i n C 3 H / H e as well as C 3 H / H e J mice. T h e rectal temperature o f the animals was determined with an electronic temperature probe ( H a a k e DT-

10, Karlsruhe, F R G ) at different times after iv injec-

(3)

0 3 6 9 12 15 18 21 24 27 HOURS AFTER INJECTION

Figure 1. Hypothermia in mice after injection with bac- terial lipopolysaccharide (LPS) or tumor necrosis factor (TNF). Groups of three C 3 H / H e (——) or C 3 H / H e J ( ) mice were injected with either L P S {Salmonella montevideo strain SH94, 100 \ig\ circles) or T N F (60 ng;

squares) iv and their temperatures were determined at intervals after injection. A l l three C 3 H / H e mice injected with T N F died within 24 h (a).

t i o n o f either L P S or T N F . The T N F - i n d u c e d hypo- t h e r m i a was dose-dependent (data not shown), a finding that again indicates that T N F itself has a thermoregulatory activity similar to that o f bacterial endotoxin.

Lack of IL-1 Induction by TNF

We tested whether T N F was able to induce IL-1 activity i n murine or h u m a n monocyte/macrophage cultures, since the thermoregulatory activity o f en- d o t o x i n has been shown to be mediated by endogenous p y r o g e n / I L - 1 . M u r i n e peritoneal exudate cells ( P E C ) were induced by the injection o f 1 m l o f P B S ip 16 h before the collection o f the exudate. T h e adherent fraction was cultured i n the presence o f either L P S or T N F or without any stimulating agent.

Cell-free supernatants were tested after 24 h for their costimulator activity i n a standard IL-1 assay.

C 3 H / H e J thymocytes (3 x 10⁵) were cultured i n the presence o f s u b o p t i m a l concentrations o f phyto- hemagglutinin (50 ng/ml; P H A - M , Sigma, Deisen- hofen, F R G ) and w i t h serial dilutions o f the supernatants. I L - l - d e p e n d e n t p r o l i f e r a t i o n o f the thymocytes was determined by counting the incor- porated radioactivity at 72 h after a 16-h pulse w i t h

1 p,Ci o f [³H]thymidine ( A m e r s h a m , Braunschweig, F R G ; specific activity, 50 C i / m M ) . O n l y weak IL-1

Table 2. Induction of interleukin 1 (IL-1) in murine peritoneal exudate cell ( P E C ) cultures.

Mouse strain, stimulus IL-1

(Hg/ml)* ( U / m l ) t C 3 H / H e

L P S , 10 24

T N F

10 6

1 <4

0.1 <4

None <4

C 3 H / H e J T N F

10 <4

1 <4

0.1 <4

None <4

* Adherent P E C (2 x 10⁶) from either C 3 H / H e or C 3 H / H e J mice were cultured in the presence of either lipopolysaccharide (LPS; Salmonella montevideo strain SH94, 10 ng/ml), tumor necrosis factor (TNF; 0.1-10 ng/ml), or no stimulus in a volume of 1 ml for 24 h.

t IL-1 units were determined as the reciprocal dilution of the cell-free culture supernatant that induced [3H]thymidine incor- poration 2.5-fold higher than the background.

activity was determined i n supernatants o f T N F - stimulated C 3 H / H e - P E C (table 2). L P S , however, induced significant IL-1 activity under these c o n d i - tions. N o IL-1 activity was detected when P E C o f C 3 H / H e J mice (low response to endotoxin) were stimulated with T N F .

H u m a n peripheral b l o o d mononuclear l y m p h o -

Table 3. Induction of interleukin 1 (IL-1) by bacterial lipopolysaccharide (LPS) or tumor necrosis factor (TNF) in human monocyte cultures.

IL-1

Stimulus* ( U / m l ) t

L P S , 10 ng/ml 64

T N F

10 ng/ml <4

1 ^ig/ml <4

100 ng/ml <4

10 ng/ml <4

1 ng/ml <4

None <4

* Adherent human mononuclear cells (3 x 10⁵) were cultured in the presence of L P S (Salmonella typhimurium strain 0901, Difco, Detroit), of T N F at the indicated concentrations, or with no stimulus in a volume of 1 ml for 24 h.

t IL-1 units were determined as indicated in table 2.

(4)

cytes (consisting o f ~ 9 5 % monocytes) that had been purified by centrifugation over f i c o l l - H y p a q u e [21]

and adherence to plastic d i d not produce IL-1 activity when stimulated with T N F at concentrations ranging from 1 n g / m l to 10 jug/ml (table 3). In further experiments we tested whether T N F was able to modulate L P S - i n d u c e d IL-1 activity and whether i n h i b i t o r y activities were generated that c o u l d inter- fere w i t h the IL-1 test system. However, T N F nei- ther modulated the induction o f IL-1 production nor interfered with the thymocyte costimulator test system (data not shown). F r o m these results we con- cluded that T N F per se is an endogenous thermoregulatory mediator similar to IL-1 and that it acts o n the thermoregulatory system without the p r i o r i n d u c t i o n o f IL-1 production by monocytes.

Conclusions

T h e effect o f macrophages i n sepsis and i n i n f l a m - m a t o r y responses is mediated by a large array o f secreted products. T h e macrophage is the most i m - portant target cell for endotoxin, and the lethal effects o f endotoxin are dependent o n the stage o f activation o f the macrophages. T h i s indicates the involvement o f endotoxin-induced monokines i n the toxic effects o f endotoxin. There is strong evidence that T N F , like IL-1, is an important mediator o f endotoxic effects. (/) It was demonstrated by Beutler et a l . [9] that reduction o f endotoxin-caused lethal- ity i n mice can be achieved with antibodies to T N F . (2) In this paper we discuss some endotoxin-like reactions i n vivo (changes i n b l o o d parameters and i n - d u c t i o n o f hypothermia) that can be evoked by i n - jection o f T N F alone. (3) It was shown that T N F p r o d u c t i o n was not induced i n endotoxin-tolerant mice [3]. In endotoxin-tolerant mice other mediators, such as IL-1, were generated i n lower amounts than i n nontolerant controls, but no T N F activity c o u l d be detected. (4) P r e l i m i n a r y data by A d e r k a et al. [22] demonstrated that peripheral b l o o d monocytes o f patients with cancer spontaneously produced larger amounts o f T N F than d i d cells o f controls, a finding indicating that T N F may also mediate cachexia i n cancer.

T h e i n d u c t i o n o f IL-1 activity by T N F in endothelial cells has been described by several investiga- tors [23, 24]. Highest levels o f IL-1 activity were detected when endothelial cells were exposed to T N F for >20 h . However, we d i d not detect any costimulator activity generated i n murine or h u m a n m o n o -

cyte/macrophage cultures after 24 h when the i n - fluence o f endotoxin was excluded by the use o f P E C from C 3 H / H e J mice or by the selection o f culture m e d i u m and fetal calf serum l a c k i n g I L - l - i n d u c i n g activity [25]. Whether IL-1 activity i n macrophages can be induced by T N F under m o d i f i e d experimen- tal conditions or is induced i n amounts smaller than those detectable i n the costimulator assay needs to be clarified. H y b r i d i z a t i o n experiments for the detection o f T N F - i n d u c e d IL-1 m R N A are currently being done. F r o m the results obtained i n this inves- tigation, it became obvious that T N F — i n contrast to L P S —did not induce the p r o d u c t i o n o f significant titers o f IL-1 i n cultures o f murine adherent P E C s or h u m a n monocytes. Therefore, T N F seems to be directly responsible for the thermoregulatory changes measured 2-3 h after T N F application.

F r o m the literature and from the data presented i n this report, it becomes clear that the tumor- necrotizing action describes an important function o f T N F . However, the name tumor necrosis factor is historical and may be misleading since it falls short o f indicating the potential importance o f this monocyte-derived mediator. Considering all the biologic activities attributed to T N F , one can rank T N F among mediators such as IL-1 and interferon-y i n its ability to modulate the i m m u n e system and af- fect the host organism.

References

1. Peavy D L , Baughn R E , Musher D M . Effects of B C G infection on the susceptibility of mouse macrophages to endotoxin. Infect Immun 1979;24:59-64

2. Mannel D N , Moore R N , Mergenhagen SE. Macrophages as a source of tumoricidal activity (tumor necrotizing factor). Infect Immun 1980;30:523-30

3. Urbaschek R, Mannel D N , Mergenhagen SE, Urbaschek B.

The role of postendotoxin serum components from B C G infected mice in the protection of compromised hosts. In:

Szentivanyi E , Nowotny A , Friedman H , eds. Immunobi- ology and immunopharmacology of bacterial endotoxins.

New York: Plenum Press 1986:221-31

4. Pennica D, Nedwin G E , Hayflick JS, Seeburg P H , Derynck R, Palladino M A , Kohr W J , Aggarwal BB, Goeddel DV.

Human tumor necrosis factor: precursor structure, expression and homology to lymphotoxin. Nature 1984;312:724-9 7. Shirai T, Yamaguchi H , Ito H , Todd CW, Wallace RB. Clon- ing and expression in Escherichia coli of the gene for human tumour necrosis factor. Nature 1985;313:803-6 6. Fransen L , Miiller R, Marmenout A , Tavernier J, Van der

Heyden J, Kawashima E , Chollet A , Tizard R, Van Heu- verswyn H , Van Vliet A , Ruysschart M - R , Fiers W. M o - lecular cloning of mouse tumour necrosis factor c D N A and its eukaryotic expression. Nucleic Acid Res 1985;

13:4417-29

(5)

7. Beutler B, Greenwald D, Hulmes JD, Chang M , Pan Y - C E , Mathison J, Ulevitch R, Cerami A . Identity of tumor necrosis factor and the macrophage-secreted factor cachectin.

Nature 1985;316:552-4

8. Torti F M , Dieckmann B, Beutler B, Cerami A , Ringold G M . A macrophage factor inhibits adipocyte gene expression:

an in vitro model of cachexia. Science 1985;229:867-9 9. Beutler B, Milsark IW, Cerami A C . Passive immunization against cachectin/tumor necrosis factor protects mice from lethal effects of endotoxin. Science 1985;229:869-71 10. Beutler B, Cerami A . Cachectin and tumour necrosis factor

as two sides of the same biological coin. Nature 1986;320:584-8

11. Dayer J - M , Beutler B, Cerami A . Cachectin/tumor necrosis factor stimulates collagenase and prostaglandin E² production by human synovial cells and dermal fibroblasts. J Exp Med 1985;162:2163-8

12. Nawroth PP, Stern D M . Modulation of endothelial cell hemostatic properties by tumor necrosis factor. J Exp Med 1986;163:740-5

13. Bertolini D R , Nedwin G E , Bringman TS, Smith DD, Mundy GR. Stimulation of bone resorption and inhibition of bone formation in vitro by human tumour necrosis factors. Na- ture 1986;319:516-8

14. Shalaby M R , Aggarwal BB, Rinderknecht E , Svedersky LP, Finkle BS, Palladino M A Jr. Activation of human poly- morphonuclear neutrophil functions by interferon-y and tumor necrosis factors. J Immunol 1985;135:2069-73 15. Gamble JR, Harlan J M , Klebanoff SJ, Vadas M A . Stimula-

tion of the adherence of neutrophils to umbilical vein endothelium by human recombinant tumor necrosis factor.

Proc Natl Acad Sci U S A 1985;82:8667-71

16. Silberstein DS, David JR. Tumor necrosis factor enhances eosinophil toxicity to Schistosoma mansoni larvae. Proc Natl Acad Sci U S A 1986;83:1055-9

17. Collins T, Lapierre L A , Fiers W, Strominger J L , Pober JS.

Recombinant human tumor necrosis factor increases m R N A levels and surface expression of H L A - A , B antigens in vascular endothelial cells and dermal fibroblasts in vitro. Proc Natl Acad Sci U S A 1986;83:446-50 18. Galanos C , Luderitz O, Westphal O. A new method for the

extraction of R. lipopolysaccharides. Eur J Biochem 1969;9:245-9

19. Sakaguchi O, Sakaguchi S, Tsunoda N . Changes in the activities of enzymes, especially lactate dehydrogenase, in endotoxin-poisoned mice. Microbiol Immunol 1979;

23:605-16

20. Prashker D, Wardlow A C . Temperature responses of mice to Escherichia coli endotoxin. Br J Exp Pathol 1971;

52:36-46

21. Boyum A . Isolation of mononuclear cells and granulocytes from human blood. Scand J Clin Lab Invest [Suppl]

1968;21:77-89

22. Aderka D, Fisher S, Levo Y, Holtmann H , Hahn T, Wallach D. Cachectin/tumor-necrosis-factor production by cancer patients [letter]. Lancet 1986;2:1190

23. Nawroth PP, Bank I, Handley D, Cassimeris J, Chess L , Stern D. Tumor necrosis factor/cachectin interacts with endothelial cell receptors to induce release of interleukin 1.

J Exp Med 1986;163:1363-75

24. Libby P, Ordovas J M , Auger K R , Robbins A H , Birinyi L K ,^f Dinarello C A . Endotoxin and tumor necrosis factor induce interleukin-1 gene expression in adult human vascular endothelial cells. A m J Pathol 1986;124:179-85 25. Northoff H , Kabelitz D, Galanos C . Interleukin 1 produc-

tion for detection of bacterial polysaccharide in fetal calf sera and other solutions [letter]. Immunology Today 1986;

7:126-7