Tumor necrosis factor-a stimulates superoxide anion generation by perfused rat liver and Kupffer cells ABRAHAM P. BAUTISTA, AGNES SCHULER, ZOLTAN SPOLARICS, AND JOHN J. SPITZER Department of Physiology, Louisiana State University Medical Center, New Orleans, Louisiana 70112

BAUTISTA,ABRAHAM P., AGNESSCHULER,ZOLTANSPOLARICS, AND JOHN J. SPITZER. Tumor necrosis factor-a stimulates superoxide anion generation by perfused rat liver and Kupffer cell. Am. J. Physiol. 261 (Gastrointest. Liver Physiol. 24): G891-G895, 1991.-Tumor necrosis factor (TNF) has been implicated as one of the mediators of the immunologic and metabolic changes in endotoxemia. Under adverse conditions, TNF can also be cytotoxic, and its effects can ultimately contribute to organ failure. This study shows that a 30-min infusion of a nonlethal dose of TNF induced the release of superoxide anion (0.9 nmol min. g-l) by the in situ perfused rat liver. TNF also primed the liver to generate more superoxide anion (2.0 nmol min-l *g-l) in response to an in vitro challenge with phorbol 12-myristate 13-acetate (PMA). Kupffer cells are most likely responsible for the superoxide anion production under these conditions, because the isolated Kupffer cells from TNF-infused rats produced increased quantities of superoxide anion (4-8 nmol/106 cells) when subsequently treated in vitro with either PMA or opsonized zymosan (control 90%, while for the hepatic PMN it was 9599%, as assessed by morphology (Giemsa- Wrights stain) and peroxidase stain. Isolation of blood granulocytes. Whole blood (5-7 ml) was anticoagulated with 100 U/ml heparin and diluted 1:l with Hanks’ balanced salt solution containing 0.8% bovine serum albumin (Sigma), 5 mmol glucose, 10 mmol N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid (HEPES), and 11.9 mmol NaHC03. One milliliter of 6% dextran (500,000 mol wt, Sigma) was added to 10 ml of diluted whole blood. The red blood cells were allowed to sediment and the leukocyte-rich supernatant was collected. Blood granulocytes were separated from the mononuclear cell fraction by Ficoll-hypaque density gradient centrifugation using two gradients: 1.077 and 1.119. Granulocytes were collected at the interface between these two gradients. Using this method, purity of PMN was >95%, as assessedby morphology (Giemsa-Wright stain) and peroxidase stain. Peroxidase staining. Peroxidase-positive cells were identified using a modified peroxidase staining technique described previously (10). Twenty microliters of cell suspension (1 x lo6 cells/ml) was added to 180 ~1 of Gey’s balanced salt solution containing 1 mg/ml of diaminobenzidine (Sigma) and 1% hydrogen peroxide (Fisher Scientific, Fair Lawn, NJ). The cells were incubated for 10 min at 37OC.An aliquot (5 ~1) was taken and examined under a microscope. Brown-stained cells were considered peroxidase-positive cells. Measurement of superoxide anion release by in situ perfused liuer. The portal vein was cannulated with a 20gauge Teflon catheter (Delmed, Canton, MA). The liver was perfused with oxygen-saturated calcium-free Hanks’ balanced salt solution. The superior vena cava cannula served as the outflow tract, after ligation of the inferior vena cava. The blood was removed from the liver by continuous perfusion of 200 ml of the buffer at 2-3 ml. mine1 l g wet liver-l at 37°C. Once the liver was cleared of blood, Hanks’ balanced salt solution containing 32 pmol ferricytochrome c (Sigma), 5 mmol glucose, and 0.8% bovine serum albumin was perfused through this organ. Phorbol 12-myristate 13-acetate (PMA; 10V7 M) was added at 4-5 min, while superoxide dismutase (4,000-7,500 U/l iver) was introduced into the perfusion medium 8-10 min after the substrate had entered the liver. Aliquots (2.5 ml) of the perfusates were immediately placed in an ice bath, and when all the tubes were collected, they were centrifuged to remove contaminating cells and debris. The change in absorbance was measured at 550 nm. The slope of the delta absorbance of the reduced ferricytochrome c (minus with superoxide dismutase) was taken and converted into nanomoles using the molecular extinction coefficient of 21.1 M-l cm-’ (12). The value is then expressed as nanomoles per minute (of perfusion) per gram wet weight. This procedure has been described previously (4).

ANION

PRODUCTION

Measurement of superoxide anion by isolated hepatic cells. Approximately 0.75-l x lo6 cells/well were plated on a six-well culture plate (Costar). After 1 h, the supernatant was aspirated and replaced by Hanks’ balanced salt solution containing 50 PM ferricytochrome c. PMA or opsonized zymosan (Sigma) was added at a final concentration of 1 PM and 1.98 mg/well, respectively. Negative controls contained 500 U/well of superoxide dismutase (Sigma). Tumor necrosis factor assay. TNF activity in serum was measured using the L929 cell line cytotoxicity assay as described previou .sly (25) . Statistics. Data presented in this paper represent means t SE of 5-7 independent experiments. Statistical significance was assessedby Student’s t test and nonparametric statistical method. RESULTS

Tumor necrosis factor actiuity. Within 1 min after the infusion of TNF, a small amount of TNF activity (260 t 75 U/ml) could be detected in the serum. At 15 min, the blood TNF content reached 54,000 t 3,500 and did not change significantly at 30 min (58,000 t 8,250). The animals were killed at the end of the 30-min TNF infusion period, since our previous observations have shown that at this tim .e macrophage -rich organs and the hepatic nonparenchym .a1cells are in a hypermetabolic state (18, 23). In addition, TNF-induced migration of neutrophils into the liver is not fully manifested at this time interval. Superoxide anion production in the in situ perfused liuer. Thirty minutes after the infusion of TNF in vivo, a considerable reduction of ferricytochrome c was detected in the perfused liver compared with the parallel control (Fig. 1). This reduction was inhibited in the presence of superoxide dismutase. The amount of superoxide dismutase-inhibitable reduction of the ferricytochrome c by the perfused liver was estimated to be 0.9 nmol min-l l g superoxide anion? The effect of TNF was manifested in this preparation by the increased superoxide anion release and also by the priming of the cells to the subsequent effect of PMA (Fig. 2). Thus l

3.0

n 0E

1 2.5

c t .-0

2.0

2

1.5

.-s g b

1.0

F 0.5 WI

0

2

PERFUSION

TIME (MINUTES)

1. Reduction of ferricytochrome c by in situ perfused liver after infusion of tumor necrosis factor (TNF). 0, Control; l and V, TNFtreated. Arrow indicates addition of superoxide dismutase (0). Error bars denote GE; n = 5-7. FIG.

Downloaded from www.physiology.org/journal/ajpgi at Macquarie Univ (137.111.162.020) on February 13, 2019.

TNF 3.0

EFFECT

ON

SUPEROXIDE

1

2. Phorbol 12-myristate 13-acetate or opsonized zymosan-induced superoxide anion production by blood and hepatic polymorphonuclear neutrophils after infusion of TNF

0

E 2.5 c w c 2.0 .-0 c a 1.5-

Treatment Group

.-G!i ox l.Ob

Control TNF

0.5

4

0.0

1 2

4

PERFUSION

6

8

10

TIME (MINUTES)

2. Effect of phorbol 12-myristate 13-acetate (PMA) on the reduction of ferricytochrome c by in situ perfused liver. Arrow indicates addition of PMA to both control (0) and TNF-treated (0) groups. * P < 0.05 by paired nonparametric statistical analysis (n = 7). FIG.

1. Total sequestered hepatic neutrophil and Kupffer cell yield after TNF infusion TABLE

Control TNF

PMA Blood

PMN

6kO.5 7t1.8

Zymosan Hepatic

PMN

NT 8t2

Blood

PMN

4.2k0.3

6.5k1.5

Hepatic

PMN

NT 4.1kO.8

Values are means t SE expressed as lo6 cells; n = 5-7. NT, superoxide anion release by hepatic PMN from control rats was not tested due to insufficient number of cells in the control livers.

0

Treatment

G893

PRODUCTION

TABLE

n

;

ANION

Group

PMN 2.5&l llt2*

KC1

KC2

KC3

38k4.5

32t3.5

3t0.8

34k6

3Ok6.2

6k1.4

Values are means t SE expressed as lo6 cells/liver; n = 5-7. The liver was perfused with collagenase, followed by pronase digestion. Nonparenchymal cells were separated by centrifugal elutriation and Ficoll- hypaque density gradient centrifugation as described in MATERIALSAND METHODS. * P < 0.001, from control.

direct addition of PMA to the perfused liver further increased superoxide anion production to 2.0 nmol (Fig. 2). The likely sources of the superoxide anion released by the liver are the Kupffer cells and inmigrating neutrophils, as the results of subsequent experiments indicate. Neutropenia and neutrophil migration into the liver. The total neutrophil count was reduced in the blood from 6,700 t 1,100 to 3,780 t 89O/pl 30 min after the administration of TNF. In the liver, the total number of neutrophils obtained after collagenase-pronase digestion and centrifugal elutriation was 11 t 2 X 106/liver compared with 2.5 t 1 x lo6 in the control. The total number of Kupffer cells (the sum of the 3 subgroups, arbitrarily classified as KCl, 2, and 3) did not change significantly after TNF treatment (Table 1). Superoxide anion production by blood and hepatic neutrophils. Table 2 shows that in vivo administration of TNF did not significantly enhance the production of superoxide anion by blood neutrophils in the presence of PMA or zymosan. The amount of superoxide anion released by hepatic neutrophils in vitro after in vivo TNF treatment did not change significantly from that released by control blood neutrophils (Table 2). Because few hepatic neutrophils were obtained from saline controls, direct comparisons with hepatic neutrophils could not be made. Superoxide anion production by Kupffer cells. Although the nonlethal dose of TNF in vivo increased the spontaneous release of superoxide anion both in the in situ

perfused liver (Fig. 1) and isolated Kupffer cells (Fig. 3), its main effect was to prime these cells to a subsequent challenge with either PMA or zymosan. Figure 3 shows that TNF infusion in vivo sensitized all three fractions of Kupffer cells to release large amounts of superoxide in response to either PMA or zymosan. Superoxide anion released by isolated hepatic endothelial cells was Cl.0 nmol/106 cells and the presence of PMA or zymosan did not result in a significant increase of superoxide anion. Isolated hepatocytes did not generate detectable amounts of superoxide anion. DISCUSSION

To our knowledge, this is the first demonstration that in vivo infusion of a nonlethal dose of TNF stimulates the release of superoxide anion in the isolated perfused rat liver. In addition, the generation of superoxide anion was further enhanced when the perfused liver was treated with PMA, a known activator of protein kinase C and superoxide anion release (12). The TNF-induced priming of enhanced superoxide anion release by the liver is most likely mediated by the increased activation of the NADPH oxidase system and hexose monophosphate shunt activity in the hepatic phagocytes. Isolated Kupffer cells also produced large amounts of superoxide anion in vitro when challenged with either PMA or opsonized zymosan, suggesting that the priming effect of TNF was maintained throughout the cell isolation procedures. TNF pretreatment failed to show a priming effect on blood PMN to either PMA or zymosan. Sequestered hepatic neutrophils also did not appear to be responsive to the same secondary stimuli. The lack of effect was in contrast to our previous findings after the administration of a sublethal dose of endotoxin (2), which showed a marked priming of these cells. The most likely sources of the superoxide anion in the perfused rat liver after TNF treatment are the Kupffer cells. Although a number of PMN were detected in the liver after TNF infusion, their contribution under these experimental conditions is not likely to be significant, because isolated PMN from the liver or blood did not release superoxide anion spontaneously. The contribution of endothelial cells or hepatocytes is also insignificant, since these cells failed to release superoxide anion in vitro either spontaneously or in the presence of PMA or zymosan. A number of in vitro studies have shown that TNF enhanced the production of superoxide anion by human

Downloaded from www.physiology.org/journal/ajpgi at Macquarie Univ (137.111.162.020) on February 13, 2019.

G894

TNF

KUPFFER

CELL 1

EFFECT

ON

SUPEROXIDE

KUPFFER CELL 2

ANION

PRODUCTION

KUPFFER CELL 3

8

FIG. 3. PMAperoxide anion from control (hatched bars)

or opsonized zymosan-induced suproduction by isolated Kupffer cells (open bars) and TNF-treated rats; n = 5-7.

0

neutrophils (13), bone marrow- erived macrophages, and resident macrophages from other tissues when these cells were challenged with other stimulants (20,26). However, in vitro treatment of macrophages, including rat Kupffer cells, with TNF alone does not induce the release of superoxide anion (19), which supports the idea that increased superoxide anion production in pathological conditions is likely to be mediated by the sensitizing effect of TNF on these cells and not by a direct effect. However, TNF can stimulate arachidonic acid metabolism, leading to the production of prostaglandins by Kupffer cells (19). The presence of neutrophils in the liver after TNF infusion can be attributed to the known effects of this cytokine as an inflammatory agent (6, 24). TNF may have stimulated the Kupffer cells to generate chemotactic factors that enhance the infiltration of neutrophils into the liver. As a consequence, hepatic macrophages, and subsequently the neutrophils as well, are sensitized to release toxic oxygen-derived radicals to kill invading pathogens. However, such release may also induce tissue injury that can contribute at least partially to the development of organ failure. Oxygen-derived radicals have been recently implicated as the mediators of TNF-induced cytotoxicity (15). TNF acts directly on its target, e.g., L929 cell line, by binding to the surface receptors that initiate a cascade of events, including G protein-coupled activation of phospholipases, generation of free radicals from xanthine oxidase and mitochondrial sources, and damage to nuclear DNA by endonucleases (15). Indirectly, it may act via the macrophages or neutrophils by sensitizing these cells to release toxic free radicals, which may target the endothelial cells or parenchymal cells in organs like the liver. The present work sh.ows that in vivo TNF infusion the Kupffer cells to generate a significant amount of superoxide anion after an appropriate second stimulus. A similar sequence of events was also observed after a relatively mild dose (LD& of endotoxin in vivo treatment (2). A nonlethal dose of TNF was used in this study to avoid dealing with preterminal events and also to limit the contribution of inmigrating neutrophils in the liver to the superoxide anion release by the intact liver. The short-term nonlethal TNF treatment (30 min) of this study did not appear to prime the blood and hepatic

neutrophils to release this radical after PMA or zymosan stimulation, suggesting that under these conditions Kupffer cells are perhaps one of the major targets of TNF action. This also implies that Kupffer cells may be the primary cell type responsible for tissue injury in the liver during the early phase of endotoxemia, to be joined later by the sequestered neutrophils. These studies demonstrate that TNF is an important immunomodulator that can contribute significantly to the defense mechanisms of the host against an infection. It has also been shown that TNF enhances the release of nitric oxides that inhibit hepatocyte protein synthesis (8) and can be protozoocidal for Leishmania major (16). However, at some point the combination of the two free radical species may be overwhelming for the protective mechanisms surrounding the macrophages, and thus may eventually contribute to tissue damage and ultimately organ failure. The authors are grateful to Jean Carnal for excellent technical assistance and Howard Blakesly for the illustrations. We also thank Dr. Gregory J. Bagby for the TNF assays. This study was supported by National Institute of General Medical Sciences Grant GM-32654. Address reprint requests to A. P. Bautista. Received

3 May

1991; accepted

in final

form

24 July

1991.

REFERENCES 1. ARTHUR, M. J. P., I. S. BENTLEY, A. R. TANNER, P. KOWALSKI SAUNDERS, G. H. MILLWARD-SADLER, AND R. WRIGHT. Oxygenderived free radicals promote hepatic injury in the rat. Gastroenterology 89: 1114-1122, 1985. 2. BAUTISTA, A. P., K. MESZAROS, J. BOJTA, AND J. J. SPITZER. Superoxide anion generation during the early stage of endotoxemia in rats. J. Leukocyte Biol. 48: 123-128, 1990. 3. BAUTISTA, A. P., N. B. D’SOUZA, C. H. LANG, J. BAGWELL, AND J. J. SPITZER. Alcohol-induced downregulation of superoxide anion release by hepatic phagocytes in endotoxemic rats. Am. J. Physiol. 260 (Regulatory Integrative Comp. Physiol. 29): R969-R976, 1991. 4. BAUTISTA, A. P., AND J. J. SPITZER. Superoxide anion generation by in situ perfused liver: effect of in vivo endotoxin. Am. J. Physiol. 259 (Gczstrointest. Liver Physiol. 22): G907-G912, 1990. 5. BEUTLER, B., AND A. CERAMI. Cachectin and tumor necrosis factor as two sides of the same biologically active coin. Nature Lord 320: 584-588,1986. 6. CARLOS, T. M., AND J. M. HARLAN. Membrane proteins involvement in phagocyte adherence to endothelium. Immunol. Reu. 114: 5-28,199O. 7. CARSWELL, E. A., L. J. OLD, R. L. KASSEL, S. GREEN, N. FIORE,

Downloaded from www.physiology.org/journal/ajpgi at Macquarie Univ (137.111.162.020) on February 13, 2019.

TNF

8.

9.

10.

11.

12.

13.

14.

15. 16.

17.

EFFECT

ON

SUPEROXIDE

AND B. WILLIAMSON. An endotoxin-induced serum factor that causes necrosis of tumor. Proc. Nutl. Acad. Sci. USA 72: 36663670, 1975. CURRAN, R. D., T. R. BILLIAR, J. B. STEUHER, J. B. OCHOA, B. G. HARBRECHT, S. G. FLINT, AND R. L. SIMMONS. Multiple cytokines are required to induce hepatocyte nitric oxide production and inhibit protein synthesis. Ann. Surg. 212: 462-469, 1990. FIDGON, C. G., J. M. M. LEEMANS, W. S. BONT, AND J. E. VRIES. Theory and practice of centrifugal elutriation (CE). Factor influencing the separation of human blood cells. Cell Biophys. 5: 105112, 1983. JOHNSON, G. D., AND J. DORLING. Immunofluorescence and immuno-peroxidase techniques. In: Techniques in Clinical Immunology, edited by R. A. Thompson. London: Blackwell, 1981, p. 131. JOHNSTON, R. B., C. A. GODZIK, AND 2. A. COHN. Increased superoxide anion production by immunologically activated and chemically elicited macrophages. J. Exp. Med. 148: 115-127, 1978. JOHNSTON, R. B., AND S. KITAGAWA. Molecular basis for enhanced respiratory burst of activated macrophages. Federation Proc. 44: 2927-2932, 1985. KAPP, A., G. ZECK-KAPP, AND D. BLOHM. Human tumor necrosis factor is a potent activator of oxidative metabolism in human polymorphonuclear neutrophilic granulocytes: comparison with human lymphotoxin. J. Invest. Dermatol. 92: 348-354, 1989. KATO, K., A. NAKANE, T. MINAGAWA, N. KASAI, K. YAMAMOTO, N. SATO, AND N. TSURUOKA. Human tumor necrosis factor increases the resistance against Listeria infection in mice. Med. Microbial. Immunol. Berlin 178: 337-346, 1989. LARRIC, J. W., AND S. C. WRIGHT. Cytotoxic mechanisms of tumor necrosis factor alpha. I?ASEB J. 4: 3215-3223, 1990. LIEW, F. Y., Y. LI, AND S. MILLIOTT. Tumor necrosis factor in leishmaniasis. II. TNF-induced macrophage leishmanicidal activity is mediated by nitric acid from L-arginine. Immunology 71: 556559,199o. MESZAROS, K., J. BOJTA, A. P. BAUTISTA, C. H. LANG, AND J. J.

ANION

18.

19.

20.

21.

22.

23.

24.

25.

26.

PRODUCTION

G895

SPITZER. Glucose utilization by Kupffer cells, endothelial cells, and granulocytes in endotoxemic rat liver. Am. J. Physiol. 260 (Gastrointest. Liver Physiol. 23): G7-G12, 1991. MESZAROS, K., C. H. LANG, G. J. BAGBY, AND J. J. SPITZER. Tumor necrosis factor increases in vivo glucose utilization of macrophage-rich tissues. Biochem. Biophys. Res. Commun. 149: l6, 1987. PETERS, T., T. GAILLARD, AND K. DECKER. Tumor necrosis factor alpha stimulates prostaglandin but not superoxide anion synthesis in rat Kupffer cells. Eicosanoids 3: 115-120, 1990. PHILLIPS, W. A., AND J. A. HAMILTON. Phorbol ester-stimulated production by murine bone marrow-derived macrophages requires preexposure to cytokines. J. Immunol. 142: 2445-2449, 1989. SANDERSON, R. J., AND K. E. BIRD. Cell separation by counterflow centrifugation. In: Methods in Cell Biology, edited by D. M. Prescott. New York: Academic, 1977, vol. 15, p. l-14. SPOLARICS, Z., A. P. BAUTISTA, AND J. J. SPITZER. Metabolic response of isolated liver cells to in vivo phagocytic challenge. Hepatology 13: 277-281, 1991. SPOLARICS, Z., A. SCHULER, G. J. BAGBY, C. H. LANG, K. MESZAROS, AND J. J. SPITZER. Tumor necrosis factor increases in vivo glucose uptake in hepatic nonparenchymal cells. J. Leukocyte Biol. 49: 309-312,199l. VADAS, M. A., AND J. R. GAMBLE. Regulation of the adhesion of neutrophil to endothelium. Biochem. Pharmacol. 40: 1683-1687, 1990. WANG, A. M., A. A. CREASY, M. B. LADNER, L. S. LIN, J. STRICKLER, J. N. VAN ARDSELL, R. YAMAMOTO, AND D. F. MARK. Molecular cloning of the complementary DNA for human tumor necrosis factor. Science Wash. DC 228: 149-154, 1985. WARREN, J. S., S. L. KUNKEL, T. W. CUNNINGHAM, K. J. JOHNSON, AND P. A. WARD. Macrophage-derived cytokines amplify immune complex-triggered superoxide anion responses by rat alveolar macrophages. Am. J. Pathol. 130: 489-495, 1988.

Downloaded from www.physiology.org/journal/ajpgi at Macquarie Univ (137.111.162.020) on February 13, 2019.

Tumor necrosis factor-alpha stimulates superoxide anion generation by perfused rat liver and Kupffer cells.

Tumor necrosis factor (TNF) has been implicated as one of the mediators of the immunologic and metabolic changes in endotoxemia. Under adverse conditi...
1MB Sizes 0 Downloads 0 Views