JOURNAL
OF SURGICAL
RESEARCH
49,256-261 (1990)
Tumor Necrosis Factor Production by Kupffer Cells Requires Protein Kinase C Activation1*2 PAUL BANKEY, M.D., ANN CARLSON,B.A.,MARIAESTELA ORTIZ,B.S., RAVINDER SINGH,M.S.,AND FRANK CERRA M.D. Department
of Surgery, University Submitted
of Minnesota,
for publication
Tumor necrosis factor (TNF) has been proposed as a primary inflammatory mediator of septic shock. In vitro and in vivo studies indicate that endotoxinor lipopolysaccharide (LPS)-activated macrophages are a principle source of TNF; however, membrane signal transduction and intracellular pathways by which LPS triggers TNF production in macrophages are unclear. Recent evidence indicates that specific protein phosphorylation via activation of protein kinase C (PKC) is an early, critical step in the signaling of macrophage TNF production by phorbol esters. We hypothesize that PKC activation is also required in LPS-signaled Kupffer cell (KC) TNF production. Murine KCs were obtained by liver perfusion and digestion and then stimulated with LPS (Escherichia coEi 011 l:B4) or LPS in the presence of H-7, a selective PKC inhibitor. Conditioned media was collected at 3 hr for assay of TNF utilizing the L929 cytolysis bioassay standardized to murine-rTNF-cY. We found that H-7 inhibited significantly LPS signaled TNF release at a concentration of 10 FM, while H-8 (a cyclic nucleotide specific inhibitor) had no effect. The effect of H-7 was dose dependent and present at varying concentrations of LPS. Down regulation of PKC activity by preincubation of KCs with phorbol myristate acetate (PMA, a direct activator of PKC) also resulted in significantly reduced TNF release after LPS stimulation. The inhibitor H-7 (10 KM) also significantly inhibited LPS signaled prostaglandin Ez release in Kupffer cells. Total and specific intracellular protein phosphorylation was determined by trichloroacetic acid precipititation and SDS-polyacrylamide gel electrophoresis after labeling stimulated Kupffer cells with 32Pi. Total protein phosphorylation was not significantly altered by LPS stimulation; however, autoradiograms from PMA- and LPS-stimulated KCs demonstrate enhanced phosphorylation of a 40-kDa protein (2.7 it_0.9fold) and a 33-kDa protein (3.1 + l.O-fold) which were
1 Research supported by the VFW Cancer Research Center, Minneapolis. * Presented at the Annual Meeting of the Association for Academic Surgery, Louisville, KY, November 15-18, 1989. 0022-4804/90
$1.50
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
Minneapolis,
November
Minnesota
55455
20. 1989
inhibited by H-7. We conclude that activation of PKC and protein phosphorylation are required steps in the signal transduction pathway of LPS-stimulated TNF production in Kupffer cells. o 1990 Academic press, I~C.
INTRODUCTION Tumor necrosis factor-a (TNF) is a macrophage/monoq&e-derived cytokine that has been proposed as an early proximal mediator of many of the metabolic and physiologic responses during septic shock. The recent studies in which antibodies to TNF-(U protected baboons from lethal gramnegative infection and the induction of hypotension, fever, acute phase protein synthesis, and cortisol release in human volunteers administered recombinant TNF-cu support the role of TNF in the inflammatory response [ 1, 21. Further in uiuo and in vitro studies have demonstrated that endotoxin is a potent signal for macrophage TNF production [2, 31. Hepatic macrophages or Kupffer cells are readily exposed to endotoxin through the phagocytosis of circulating gramnegative bacteria or from the direct leakage of endotoxin from the injured GI tract during septic shock [4]. The endotoxin-activated Kupffer cell (KC) has been hypothesized to alter hepatocyte function during sepsis by a number of mechanisms, including the release of TNF [5]. Molecular mechanisms by which Kupffer cells are activated by endotoxin are poorly characterized [6]. Protein phosphorylation or dephosphorylation via activated kinases or phosphatases are generalized mechanisms by which extracellular signals are transduced into appropriate cellular responses. Recent evidence indicates that specific protein phosphorylation via activated protein kinase C (PKC) may have a significant role in the activation of inflammatory cells including macrophages [ 7-91. In vitro stimulation of macrophages with phorbol esters that specifically activate PKC has been demonstrated to signal superoxide production, prostaglandin release, and TNF production [lo-131. Since endotoxin results in similar macrophage responses, we hypothesize that PKC acti256
BANKEY
ET AL.: TNF
PRODUCTION
vation and protein phosphorylation are required during endotoxin-signaled TNF production. To test this hypothesis we have utilized the recently developed isoquinoline-sulfonamide derivatives, H-7 and H-8, which function as specific protein kinase inhibitors. The PKC inhibitory activity of H-7 has been demonstrated to be several times greater than that of H-8 [14]. Kupffer cell TNF production and protein phosphorylation following endotoxin stimulation were then determined with or without the presence of the inhibitors. Our results indicate that PKC activation is a required step in endotoxin-activated TNF-cu production by Kupffer cells. MATERIALS
AND
METHODS
Materials. Balb/c mice were obtained from the Research Animal Resources (RAR) at the University of Minnesota. Escherichia coli Olll:B4 lipopolysaccharide (LPS), phorbol myristate acetate (PMA; Sigma, St. Louis), murine-rTNF-cY (Mu-rTNF-a; Genzyme, Boston), H-7 and H-8 inhibitors (Calbiochem, La Jolla, CA) 32Pi (New England Nuclear, Boston), electrophoresis reagents, and MW standards (Bio-Rad Laboratories, Richmond, CA) were also purchased. Kupffer cells were obtained from Balb/ Cell harvests. c mice by modified pronase/DNase liver perfusion and digestion as described by Knook and Sleyster [ 151. Liver nonparenchymal cell preps were enriched for KCs by adherence. Resulting suspensions were >90% nonspecific esterase (Spec. Heme. Dept., University of Minnesota) positive and >95% viable by exclusion of trypan blue. Isolated KCs were suspended in RPM1 1640 tissue culture medium supplemented with pen/strep and 5% low endotoxin calf serum prior to incubation for determining TNF and prostaglandin Ez (PGE2) release. Measurement of tumor necrosis factor and prostaglandin E2 release. To determine bioactive TNF release, KCs were plated in RPM1 + 5% low endotoxin calf serum on 6-well tissue culture approved plates (Co-Star, Inc.) and allowed to adhere for 40 hr. After washing to remove nonadherent cells, the cultures were triggered with specific agents in the presence or absence of protein kinase inhibitors and conditioned supernatants were collected at 3 or 24 hr. The supernatants were assayed for tumor necrosis factor activity using the L929 cytolysis bioassay [16] and quantitated by comparing lytic activity of the supernatant to a standard curve prepared from murine recombinant TNF-a. Briefly, L929 cells are grown in MEM supplemented with 5% fetal calf serum, NEAA (1 ml/100 ml MEM), and pen/strep and passed with 0.5% trypsin. Freshly passed cells are counted using a hemocytometer and 5 X lo4 are plated in 100 ~1 media in 96well plates. After 3-4 hr, 50 ~1 of the supernatant to be tested is added to each well with 50 ~1 of actinomycin D with a final concentration of 0.25 pg/ml. After 18 hr cell lysis is quantified by measuring the absorbance at 590 of
BY KUPFFER
CELLS
257
the remaining SDS solubilized L929 cells after washing and staining with crystal violet. Supernatant lysis is then compared to lysis of a known amount of Mu-rTNF-a. Fifty percent lysis routinely corresponded to 0.7 units of Mu-rTNF in our assay. Supernatant prostaglandin Ez content was measured with a RIA kit (Seragen, Inc, Boston). Measurement of Kupffercellproteinphosphoryt’ation and gel electrophoresis. Nonparenchymal liver cells (4 X lo6 cells/well) are plated in RPM1 1640 with 5% low endotoxin calf serum for 40 hr in 6-well plates. After washing, the cultures are triggered with LPS or PMA for 3 hr and then washed several times before labeling for 45 min with low phosphate Kreb’s-Ringers buffer supplemented with 500 PCi of 32Pi (1.0 mCi/ml). On completion of labeling the cultures are washed with cold buffer and the monolayers are solubilized in 500 ~1 of homogenization buffer containing phosphatase inhibitors NaF (50 mM) and EGTA (10 mM) as described by Weiel et al. [17]. The samples are stored frozen (-70°C) until electrophoresis. Label incorporation into cellular protein is determined by precipitation of 50-~1 aliquots of the labeled homogenate with cold 20% trichloroacetic acid and, after washing, pellet resolubilizing it with 0.5 M NaOH. Samples are then counted using a liquid scintillation counter (Beckman, Inc.). The protein concentration of the homogenates is determined by the method of Bradford using reagents purchased from Bio-Rad Laboratories. Individually labeled proteins were detected on autoradiograms of the samples after separation by SDS-PAGE (13.5% polyacrylamide-sodium dodecyl sulfate gel electrophoresis) according to Laemelli [18]. Equal amounts of protein were added to each lane of the gel from the cell homogenate samples. The gel samples were cell homogenate aliquots separate from those used for protein determination and measurement of label incorporation. After electrophoresis the gels were stained with Coomassie blue to detect molecular weight standards, destained, and dried before exposure to Kodak XAR-5 X-ray film for 510 days at -70°C using intensifying screens before development. Autoradiograms were scanned using MacVision and Scanning software to quantitate the density of individual protein bands. Statistical analysis. Data were analyzed by ANOVA using MacAnova, a statistics package developed at the University of Minnesota for use on Macintosh computers. Treatment means were compared using Tukey’s test, with a significance level of 0.05. In other experiments Student’s paired t test was used to determine P values. RESULTS
Effect of protein kinase C inhibitor H-7 on LPS-stimulated TNF production. We initially assessed the ability of LPS to signal TNF release in cultured liver nonparenchymal cells enriched for Kupffer cells by adherence.
258
JOURNAL
OF SURGICAL
RESEARCH:
VOL. 49, NO. 3, SEPTEMBER
1990
150
5
---0--
“-7
-
H-8
T
I
n
MEDIA ALONE
IB
~7UD)JV
q H-8UqW
’ pco.05 YS.H-a or mediaalone
CONTROL
0.4
2
10
CONTROL
50
1
100
10000
LPS CONCENTRATION(NGIML)
INHIBITORCONCENTRATION(FM)
FIG. 1. Dose-dependent inhibition of LPS-stimulated TNF production by protein kinase C inhibitor (H-7) mean + SEM). KC cultures were stimulated with or without H-7 or H-8 at increasing concentrations. Conditioned supernatants were collected at 4 hr and assayed by L929 cytolysis.
FIG. 2. Specific effect of H-7 inhibitor (10 &f) on TNF production after stimulation with increasing LPS concentrations (mean f SEM). KC cultures were stimulated with increasing LPS doses with or without H-7 or H-8 at 10 ait4 concentration. Conditioned supernatants were collected at 4 hr and assayed by L929 cytolysis.
We found a dose-dependent release of TNF bioactivity in cultures stimulated with LPS and that the protein kinase C inhibitor (H-7) significantly reduced this response (Figs. 1 and 2). The inhibition of LPS-signaled TNF production was dependent on the concentration of H-7 with 10 j&f providing the greatest specificity of inhibition. Higher concentrations also inhibited TNF production but this effect was also seen by the non-PKC kinase inhibitor H-8, suggesting mechanisms other than PKC inhibition (Fig. 1). Furthermore, H-8 at the 10 &f concentration did not significantly inhibit LPS-signaled TNF release at any LPS dose tested, while H-7 significantly reduced TNF production following stimulation with 100 rig/ml and 10 pg/ml LPS doses (Fig. 2). Effect of protein kinase C inhibitor (H-7) on KC prostaglandin E2 release. To assess the role of PKC in another synthetic pathway of the activated Kupffer cell we determined the effect of PKC inhibition on prostaglandin Ez release. We found that LPS-activated KCs released significant amounts of PGEz over controls and that this activity was significantly inhibited (approximately 70%) by the presence of H-7 at 10 &f (Fig. 3). Effectof down regulation ofprotein kinase C by treatment with PMA, a phorbol ester, on LPS-stimulated TNF pl;oduction. The specificity of protein kinase inhibitors is difficult to demonstrate conclusively. To further demonstrate the role of PKC in LPS-signaled TNF production we attempted to down regulate PKC activity by pretreating the cells for 48 hr with the phorbol ester phorbol myristate acetate (PMA). Phorbol esters are thought te produce their cellular effects exclusively by activating PKC
and prolonged incubation is thought to down regulate the PKC activity. We found that preincubation of KCs with PMA significantly reduced subsequent LPS-stimulated TNF production (Fig. 4). However, the degree of inhibition was not as great as observed by treatment with H7. Interestingly, in separate experiments with no PMA preincubation the direct activation of KCs with PMA (5
150
MEDIA+H-7
LPS ALONE
0.1
10
INHIBITORCONCENTRATION (uM) E&t of H-7 inhibitor on KC PGEz release after LPS stimulation (mean 2 SEMI. KC cultures were LPS (10 @g/ml)-stimulated with or without H-7 at increasing concentrations. Conditioned supernatants were collected at 24 hr and assayed by RIA.
BANKEY
ET AL.: TNF
PRODUCTION
BY KUPFFER
259
CELLS
TABLE
1
Kupffer Cell Total Protein Phosphorylation Response to LPS and H-7 Inhibitor (10 p&f) in Vitro Control Total phosphorylated proteins (3zPi counts x 1000) Total protein @g/culture)
-I
CONTROL
PMA
PMA/H-7
LPS
PMAiLPS
TREATMENT FIG. 4. Down regulation of LPS-stimulated TNF production by preincubation with PMA, a specific PKC agonist. (Mean + SEM). KC cultures were incubated for 48 hr with PMA (5 pg/ml) and then stimulated with LPS. Conditioned supernatants were collected at 4 hr and assayed by L929 cytolysis.
pg/ml) for 4 hr did not produce significant TNF release versus controls. (PMA, 13 -t 5 units, vs control, 3 + 4 units, Fig. 4). Effect of LPS on total and specific protein phosphoryEation in Kupfier cells. Treatment of the KC cultures with LPS or H-7 for 4 hr did not alter total cellular protein (Table 1). Total cellular protein phosphorylation measured by incorporation of labeled t3’Pi] also was not significantly changed by treatment with LPS, PMA, or H7 over 4 hr. (Table 1). However, LPS and PMA did result in the intensification in the radiolabeling of specific proteins at 40 kDa (2.7 +- 0.9-fold increase by densitometry) and 33 kDa (3.1 & l.O-fold by densitometry) compared to control (Fig. 5). This enhanced phosphorylation was inhibited by H-7 (Fig. 5).
H-7
LPS
LPS + H-7
35.3 +_5.1
32.1 t- 7.7
41.0 + 2.8
36.6 + 4.1
15.5 f 2.0
16.1 f 1.6
16.4 k 1.9
15.2 f 2.1
patients. Recent experimental evidence indicates that TNF is an early mediator released primarily by monocytes/macrophages that is capable of inducing many of the metabolic and physiologic responses characteristic of septic shock [2, 191. In vitro studies indicate that endotoxin or bacterial LPS is a potent signal for macrophage TNF release. Molecular mechanisms as to how the endotoxin signal is transduced into the appropriate intracellular responses resulting in macrophage TNF release remain elusive. Defining the mechanisms of TNF production provides a potential interventional approach for the future treatment of septic shock. Protein phosphorylation is a generalized mechanism by which extracellular signals are transduced into appropriate cellular responses. Specific extracellular signals initiate the production of membrane-derived middle messengers (cyclic nucleotides, inositol phosphates, or diacylglycerol) that activate specific protein kinases, providing appropriate intracellular protein phosphorylation. The
Effect
68
kD
45
kD
36
kD
29
kD
21
kD
20
kD
14
kD
of
PKC
Inl
DISCUSSION
Our results indicate that protein kinase C activity is necessary for LPS-activated Kupffer cells to produce tumor necrosis factor-a. Supporting this hypothesis are the findings that the PKC-specific inhibitor H-7 significantly reduces LPS-stimulated TNF production over the inhibition observed for the CAMP-dependent kinase inhibitor H-8. In addition, the down regulation of PKC by phorbol ester stimulation also reduces LPS-stimulated TNF production, Gel electrophoresis indicates that LPS stimulation increases the phosphorylation of specific KC intracellular proteins and that this increased phosphorylation is at least partially sensitive to H-7 inhibition. Septic shock due to gram-negative bacteremia or endotoxemia continues to be a major problem in surgical
LPS(H.l’PMAtH.7 “1
FIG. 5. Specific KC phosphorylated proteins following LPS and PMA stimulation. Representative autoradiogram from three replicates of 32Pi-labeled KCs as described under Materials and Methods.
260
JOURNAL
OF SURGICAL
RESEARCH:
protein phosphorylation then contributes to the desired cellular response. Protein kinase C is a specific kinase activated by signal-induced hydrolysis of membrane inositol phospholipids producing diacylglycerol and inositol phosphates [20]. Our results indicate that protein phosphorylation via protein kinase C is required for TNF release by LPS-stimulated Kupffer cells. The role of PKC in TNF production has been implicated from studies in other macrophage populations. The human monocyte-like cell line HL-60 was shown to produce TNF-(Y when stimulated with the phorbol ester PMA. [13] Thus, direct activation of PKC by phorbol esters was a sufficient signal in these transformed cells. Our results indicate that PMA was an insufficient direct signal for KC TNF release. This difference is most likely due to differences between Kupffer cells and monocytes and primary cultures versus cell lines. Two reports have studied the in vitro activity of PKC in response to LPS and lipid A (the active portion of LPS) in murine macrophages with differing results [21,22]. Lipid A enhanced PKC activity, while LPS had no effect. Recently, production of diacylglycerol and inositol phosphate has been reported in response to LPS stimulation in murine peritoneal macrophages ]23]. Our data support the role of PKC in LPS-stimulated macrophage TNF production and are the first investigating Kupffer cells or liver nonparenchymal cells. We have utilized the isoquinoline-sulfonamide-derived inhibitors H-7 and H-8 to demonstrate the role of PKC. These inhibitors compete for ATP with the ATP binding site on the kinase. H-7 has been demonstrated to be a more potent inhibitor of PKC with a Ki of 6 pM, while H-8 is more potent for cyclic nucleotide-dependent kinases (Ki for PKC is 15 pM) [14]. However, as our results indicate, the specificity of the inhibition is dependent on the concentration used and at sufficiently high levels H8 will also inhibit LPS-stimulated TNF production significantly. The inhibition of TNF production by H-8 at concentrations above the Ki for PKC suggests that this may be due to PKC inhibition; however, inhibition of other intracellular signals necessary for Kupffer cell TNF production cannot be ruled out. These inhibitors have previously been demonstrated to inhibit LPS-signaled B lymphocyte activation and thrombin-activated protein phosphorylation in platelets, both responses attributed to PKC activation [8, 91. Our results that the PKC inhibitor reduces PGEz release by LPS-stimulated Kupffer cells is consistent with reports demonstrating phorbol ester-stimulated PGEz production in macrophage populations [ 121. Since nonspecific inhibitory effects cannot totally be excluded in these experiments, we attempted to down regulate PKC by directly stimulating it with PMA for 48 hr and then determining LPS-stimulated TNF release. The finding of reduced TNF release is consistent with PKC down regulation and a necessary role for PKC in TNF release. The similiar protein levels excludes a cytotoxic effect from the inhibitors.
VOL. 49, NO. 3, SEPTEMBER
1990
The induction of increased specific protein phosphorylation in LPS-stimulated macrophages has been reported previously [17]. Our results as well as those of Weiel et al. [17] suggest that a 37-kDa protein is a specific phosphorylation target of PKC in macrophages. The function of this phosphoprotein is unknown. In summary, activation of protein kinase C would appear to be a necessary step in the activation of Kupffer cells for the release of TNF in response to endotoxin. Whether PKC activation is signaled via LPS-induced inositol phospholipid hydrolysis and diacylglycerol production is unclear. Furthermore, the role of specific proteins phosphorylated by activated protein kinase C is unknown; however, since LPS stimulation is associated with TNF gene transcription and increased mRNA levels, it is tempting to speculate a role for these phosphoproteins in the regulation of these responses [3]. Defining these mechanisms may have important applications in altering macrophage function not only during the septic response but also in other disease processes in which the macrophage participates, such as atherosclerosis, tumor killing and cachexia, and wound healing. REFERENCES 1.
Beutler, B., and Cerami, A. Cachetin and tumor necrosis factor as two sides of the same biologic coin. Nature (London) 320: 584, 1986.
2.
Michie, H. R., Spriggs, D. R., Manogue, K. R., et al. Tumor necrosis factor and endotoxin induce similar metabolic responses in human beings. Surgery 104: 280, 1988.
3.
Beutler, B., Krochin, N., Milsark, I., Luedke, C., and Cerami, A. Control of cachetin (tumor necrosis factor) synthesis: Mechanisms of endotoxin resistance. Science 232: 977, 1986. Dietch, E., Winterton, J., et al. The gut as a portal of entry for bacteremia: Role of protein malnutrition. Ann. Surg. 206: 681, 1987. West, M. A., Billiar, T. R., Mazuski, J. E., et al. Endotoxin modulation of hepatocyte secretory and cellular protein synthesis is mediated by kupffer cells. Arch. Surg. 123: 1400, 1988. Morrison, D. C., and Rudbach, J. Endotoxin-cell-membrane interactions leading to transmembrane signaling. Contemp. Top. MOE. Zmmunol. 81: 187, 1981. Magnuson, D., Maier, R., and Pohlman, T. Protein kinase C: A potential pathway of endothelial cell activation by endotoxin, tumor necrosis factor, and interleukin-1. Surgery 106: 216,1989.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Rush, J., and Waechter, C. Inhibitors of protein kinase C block activation of B lymphocytes by bacterial lipopolysaccharide. Biochem. Biophys. Res. Commun. 145: 1315,1987. Kawamoto, S., and Hidaka, H. I-(5-Isoquinolinesulfonylj-2-methylpiperazine (H-7) is a selective inhibitor of protein kinase C in rabbit platelets. Biochem. Biophys. Res. Commun. 126: 258, 1984. Kiyotaki, C., and Bloom, B. Activation of murine macrophage cell lines. Possible involvement of protein kinases in stimulation of superoxide production. J. Zmmunol. 133: 923, 1984. Castagna, M., Takai, Y., Kaibuchi, K., Sano, K., Kikkawa, U., and Nishizuka, Y. Direct activation of calcium-activated, phospholipiddependent protein kinase by tumor-promoting phorbol esters. J. Biol. Chem. 267: 784, 1982. Hoffman, T., Lizzie, E., Suissa, J., Rotrosen, D., Sullivan, J., Mandell, G., and Bonvini, E. Dual stimulation of phospholipase activity
BANKEY
13.
14.
15.
16.
17.
ET AL.: TNF
PRODUCTION
in human monocytes. Role of calcium-dependent and calcium-independent pathways in arachidonic acid release and eicosanoid formation. J. Zmmunol. 140: 3912, 1988. Pennica, D., Nedwin, G., et al. Human tumor necrosis factor: Precursor structure, expression, and homology to lymphotoxin. Nature (London) 312: 724,1984. Hidaka, H., Inagaki, M., et al. Isoquinolinesulfonamides, novel and potent inhibitors of cyclic nucleotide dependent protein kinase and protein kinase C. Biochemistry 23: 5036, 1984. Knook, D. L., and Sleyster, E. Isolated parenchymal, Kupffer and endothelial rat liver cells characterized by their lysosomal enzyme content. Biochem. Biophys. Res. Commun. 96: 250, 1980. Aggarwal, B., Kohr, W., Hass, P., Moffat, B., Spencer, S., Henzel, W., Bringman, T., Nedwin, G., Goeddel, D., and Harkins, R. Human tumor necrosis factor: Production, purification, and characterization. J. Biol. Chem. 260: 2345, 1985. Weiel, J. E., Hamilton, T., and Adams, D. LPS induces altered phosphate labeling of proteins in murine peritoneal macrophages. J. Zmmunol. 136: 3012, 1986.
BY KUPFFER
CELLS
261
18.
Laemmli, U. K. Protein electrophoresis gels. Nature (London) 227: 680, 1970.
in SDS polyacrylamide
19.
Starnes, H. F., Warren, R. S., Jeevanandam, M., et al. Tumor necrosis factor and the acute metabolic response to tissue injury in man. J. Clin. Znuest. 82: 1321, 1988.
20.
Nishizuka, Y. The family of protein kinase C for signal transduction. JAMA 262: 1826,1989.
21.
Hamilton, T., Becton, D., et al. Interferon-y modulates protein kinase C activity in murine peritoneal macrophages. J. Biol. Chem. 260: 1378, 1985.
22.
Wightman, P., and Raetz, C. The activation of protein kinase C by biologically active lipid moieties of lipopolysaccharide. J. Biol. Chem. 259: 10,048,1984.
23.
Prpic, V., Weiel, J., Sommers, S., DiGuiseppi, J., Gonias, S., P&o, S., Hamilton, T., Herman, B., and Adams, D. Effects of bacterial lipopolysaccharide on the hydrolysis of phosphatidyl-4,5-bisphosphate in murine peritoneal macrophages. J. Zmmunol. 139: 526, 1987.
OF SURGICAL
RESEARCH
49,256-261 (1990)
Tumor Necrosis Factor Production by Kupffer Cells Requires Protein Kinase C Activation1*2 PAUL BANKEY, M.D., ANN CARLSON,B.A.,MARIAESTELA ORTIZ,B.S., RAVINDER SINGH,M.S.,AND FRANK CERRA M.D. Department
of Surgery, University Submitted
of Minnesota,
for publication
Tumor necrosis factor (TNF) has been proposed as a primary inflammatory mediator of septic shock. In vitro and in vivo studies indicate that endotoxinor lipopolysaccharide (LPS)-activated macrophages are a principle source of TNF; however, membrane signal transduction and intracellular pathways by which LPS triggers TNF production in macrophages are unclear. Recent evidence indicates that specific protein phosphorylation via activation of protein kinase C (PKC) is an early, critical step in the signaling of macrophage TNF production by phorbol esters. We hypothesize that PKC activation is also required in LPS-signaled Kupffer cell (KC) TNF production. Murine KCs were obtained by liver perfusion and digestion and then stimulated with LPS (Escherichia coEi 011 l:B4) or LPS in the presence of H-7, a selective PKC inhibitor. Conditioned media was collected at 3 hr for assay of TNF utilizing the L929 cytolysis bioassay standardized to murine-rTNF-cY. We found that H-7 inhibited significantly LPS signaled TNF release at a concentration of 10 FM, while H-8 (a cyclic nucleotide specific inhibitor) had no effect. The effect of H-7 was dose dependent and present at varying concentrations of LPS. Down regulation of PKC activity by preincubation of KCs with phorbol myristate acetate (PMA, a direct activator of PKC) also resulted in significantly reduced TNF release after LPS stimulation. The inhibitor H-7 (10 KM) also significantly inhibited LPS signaled prostaglandin Ez release in Kupffer cells. Total and specific intracellular protein phosphorylation was determined by trichloroacetic acid precipititation and SDS-polyacrylamide gel electrophoresis after labeling stimulated Kupffer cells with 32Pi. Total protein phosphorylation was not significantly altered by LPS stimulation; however, autoradiograms from PMA- and LPS-stimulated KCs demonstrate enhanced phosphorylation of a 40-kDa protein (2.7 it_0.9fold) and a 33-kDa protein (3.1 + l.O-fold) which were
1 Research supported by the VFW Cancer Research Center, Minneapolis. * Presented at the Annual Meeting of the Association for Academic Surgery, Louisville, KY, November 15-18, 1989. 0022-4804/90
$1.50
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
Minneapolis,
November
Minnesota
55455
20. 1989
inhibited by H-7. We conclude that activation of PKC and protein phosphorylation are required steps in the signal transduction pathway of LPS-stimulated TNF production in Kupffer cells. o 1990 Academic press, I~C.
INTRODUCTION Tumor necrosis factor-a (TNF) is a macrophage/monoq&e-derived cytokine that has been proposed as an early proximal mediator of many of the metabolic and physiologic responses during septic shock. The recent studies in which antibodies to TNF-(U protected baboons from lethal gramnegative infection and the induction of hypotension, fever, acute phase protein synthesis, and cortisol release in human volunteers administered recombinant TNF-cu support the role of TNF in the inflammatory response [ 1, 21. Further in uiuo and in vitro studies have demonstrated that endotoxin is a potent signal for macrophage TNF production [2, 31. Hepatic macrophages or Kupffer cells are readily exposed to endotoxin through the phagocytosis of circulating gramnegative bacteria or from the direct leakage of endotoxin from the injured GI tract during septic shock [4]. The endotoxin-activated Kupffer cell (KC) has been hypothesized to alter hepatocyte function during sepsis by a number of mechanisms, including the release of TNF [5]. Molecular mechanisms by which Kupffer cells are activated by endotoxin are poorly characterized [6]. Protein phosphorylation or dephosphorylation via activated kinases or phosphatases are generalized mechanisms by which extracellular signals are transduced into appropriate cellular responses. Recent evidence indicates that specific protein phosphorylation via activated protein kinase C (PKC) may have a significant role in the activation of inflammatory cells including macrophages [ 7-91. In vitro stimulation of macrophages with phorbol esters that specifically activate PKC has been demonstrated to signal superoxide production, prostaglandin release, and TNF production [lo-131. Since endotoxin results in similar macrophage responses, we hypothesize that PKC acti256
BANKEY
ET AL.: TNF
PRODUCTION
vation and protein phosphorylation are required during endotoxin-signaled TNF production. To test this hypothesis we have utilized the recently developed isoquinoline-sulfonamide derivatives, H-7 and H-8, which function as specific protein kinase inhibitors. The PKC inhibitory activity of H-7 has been demonstrated to be several times greater than that of H-8 [14]. Kupffer cell TNF production and protein phosphorylation following endotoxin stimulation were then determined with or without the presence of the inhibitors. Our results indicate that PKC activation is a required step in endotoxin-activated TNF-cu production by Kupffer cells. MATERIALS
AND
METHODS
Materials. Balb/c mice were obtained from the Research Animal Resources (RAR) at the University of Minnesota. Escherichia coli Olll:B4 lipopolysaccharide (LPS), phorbol myristate acetate (PMA; Sigma, St. Louis), murine-rTNF-cY (Mu-rTNF-a; Genzyme, Boston), H-7 and H-8 inhibitors (Calbiochem, La Jolla, CA) 32Pi (New England Nuclear, Boston), electrophoresis reagents, and MW standards (Bio-Rad Laboratories, Richmond, CA) were also purchased. Kupffer cells were obtained from Balb/ Cell harvests. c mice by modified pronase/DNase liver perfusion and digestion as described by Knook and Sleyster [ 151. Liver nonparenchymal cell preps were enriched for KCs by adherence. Resulting suspensions were >90% nonspecific esterase (Spec. Heme. Dept., University of Minnesota) positive and >95% viable by exclusion of trypan blue. Isolated KCs were suspended in RPM1 1640 tissue culture medium supplemented with pen/strep and 5% low endotoxin calf serum prior to incubation for determining TNF and prostaglandin Ez (PGE2) release. Measurement of tumor necrosis factor and prostaglandin E2 release. To determine bioactive TNF release, KCs were plated in RPM1 + 5% low endotoxin calf serum on 6-well tissue culture approved plates (Co-Star, Inc.) and allowed to adhere for 40 hr. After washing to remove nonadherent cells, the cultures were triggered with specific agents in the presence or absence of protein kinase inhibitors and conditioned supernatants were collected at 3 or 24 hr. The supernatants were assayed for tumor necrosis factor activity using the L929 cytolysis bioassay [16] and quantitated by comparing lytic activity of the supernatant to a standard curve prepared from murine recombinant TNF-a. Briefly, L929 cells are grown in MEM supplemented with 5% fetal calf serum, NEAA (1 ml/100 ml MEM), and pen/strep and passed with 0.5% trypsin. Freshly passed cells are counted using a hemocytometer and 5 X lo4 are plated in 100 ~1 media in 96well plates. After 3-4 hr, 50 ~1 of the supernatant to be tested is added to each well with 50 ~1 of actinomycin D with a final concentration of 0.25 pg/ml. After 18 hr cell lysis is quantified by measuring the absorbance at 590 of
BY KUPFFER
CELLS
257
the remaining SDS solubilized L929 cells after washing and staining with crystal violet. Supernatant lysis is then compared to lysis of a known amount of Mu-rTNF-a. Fifty percent lysis routinely corresponded to 0.7 units of Mu-rTNF in our assay. Supernatant prostaglandin Ez content was measured with a RIA kit (Seragen, Inc, Boston). Measurement of Kupffercellproteinphosphoryt’ation and gel electrophoresis. Nonparenchymal liver cells (4 X lo6 cells/well) are plated in RPM1 1640 with 5% low endotoxin calf serum for 40 hr in 6-well plates. After washing, the cultures are triggered with LPS or PMA for 3 hr and then washed several times before labeling for 45 min with low phosphate Kreb’s-Ringers buffer supplemented with 500 PCi of 32Pi (1.0 mCi/ml). On completion of labeling the cultures are washed with cold buffer and the monolayers are solubilized in 500 ~1 of homogenization buffer containing phosphatase inhibitors NaF (50 mM) and EGTA (10 mM) as described by Weiel et al. [17]. The samples are stored frozen (-70°C) until electrophoresis. Label incorporation into cellular protein is determined by precipitation of 50-~1 aliquots of the labeled homogenate with cold 20% trichloroacetic acid and, after washing, pellet resolubilizing it with 0.5 M NaOH. Samples are then counted using a liquid scintillation counter (Beckman, Inc.). The protein concentration of the homogenates is determined by the method of Bradford using reagents purchased from Bio-Rad Laboratories. Individually labeled proteins were detected on autoradiograms of the samples after separation by SDS-PAGE (13.5% polyacrylamide-sodium dodecyl sulfate gel electrophoresis) according to Laemelli [18]. Equal amounts of protein were added to each lane of the gel from the cell homogenate samples. The gel samples were cell homogenate aliquots separate from those used for protein determination and measurement of label incorporation. After electrophoresis the gels were stained with Coomassie blue to detect molecular weight standards, destained, and dried before exposure to Kodak XAR-5 X-ray film for 510 days at -70°C using intensifying screens before development. Autoradiograms were scanned using MacVision and Scanning software to quantitate the density of individual protein bands. Statistical analysis. Data were analyzed by ANOVA using MacAnova, a statistics package developed at the University of Minnesota for use on Macintosh computers. Treatment means were compared using Tukey’s test, with a significance level of 0.05. In other experiments Student’s paired t test was used to determine P values. RESULTS
Effect of protein kinase C inhibitor H-7 on LPS-stimulated TNF production. We initially assessed the ability of LPS to signal TNF release in cultured liver nonparenchymal cells enriched for Kupffer cells by adherence.
258
JOURNAL
OF SURGICAL
RESEARCH:
VOL. 49, NO. 3, SEPTEMBER
1990
150
5
---0--
“-7
-
H-8
T
I
n
MEDIA ALONE
IB
~7UD)JV
q H-8UqW
’ pco.05 YS.H-a or mediaalone
CONTROL
0.4
2
10
CONTROL
50
1
100
10000
LPS CONCENTRATION(NGIML)
INHIBITORCONCENTRATION(FM)
FIG. 1. Dose-dependent inhibition of LPS-stimulated TNF production by protein kinase C inhibitor (H-7) mean + SEM). KC cultures were stimulated with or without H-7 or H-8 at increasing concentrations. Conditioned supernatants were collected at 4 hr and assayed by L929 cytolysis.
FIG. 2. Specific effect of H-7 inhibitor (10 &f) on TNF production after stimulation with increasing LPS concentrations (mean f SEM). KC cultures were stimulated with increasing LPS doses with or without H-7 or H-8 at 10 ait4 concentration. Conditioned supernatants were collected at 4 hr and assayed by L929 cytolysis.
We found a dose-dependent release of TNF bioactivity in cultures stimulated with LPS and that the protein kinase C inhibitor (H-7) significantly reduced this response (Figs. 1 and 2). The inhibition of LPS-signaled TNF production was dependent on the concentration of H-7 with 10 j&f providing the greatest specificity of inhibition. Higher concentrations also inhibited TNF production but this effect was also seen by the non-PKC kinase inhibitor H-8, suggesting mechanisms other than PKC inhibition (Fig. 1). Furthermore, H-8 at the 10 &f concentration did not significantly inhibit LPS-signaled TNF release at any LPS dose tested, while H-7 significantly reduced TNF production following stimulation with 100 rig/ml and 10 pg/ml LPS doses (Fig. 2). Effect of protein kinase C inhibitor (H-7) on KC prostaglandin E2 release. To assess the role of PKC in another synthetic pathway of the activated Kupffer cell we determined the effect of PKC inhibition on prostaglandin Ez release. We found that LPS-activated KCs released significant amounts of PGEz over controls and that this activity was significantly inhibited (approximately 70%) by the presence of H-7 at 10 &f (Fig. 3). Effectof down regulation ofprotein kinase C by treatment with PMA, a phorbol ester, on LPS-stimulated TNF pl;oduction. The specificity of protein kinase inhibitors is difficult to demonstrate conclusively. To further demonstrate the role of PKC in LPS-signaled TNF production we attempted to down regulate PKC activity by pretreating the cells for 48 hr with the phorbol ester phorbol myristate acetate (PMA). Phorbol esters are thought te produce their cellular effects exclusively by activating PKC
and prolonged incubation is thought to down regulate the PKC activity. We found that preincubation of KCs with PMA significantly reduced subsequent LPS-stimulated TNF production (Fig. 4). However, the degree of inhibition was not as great as observed by treatment with H7. Interestingly, in separate experiments with no PMA preincubation the direct activation of KCs with PMA (5
150
MEDIA+H-7
LPS ALONE
0.1
10
INHIBITORCONCENTRATION (uM) E&t of H-7 inhibitor on KC PGEz release after LPS stimulation (mean 2 SEMI. KC cultures were LPS (10 @g/ml)-stimulated with or without H-7 at increasing concentrations. Conditioned supernatants were collected at 24 hr and assayed by RIA.
BANKEY
ET AL.: TNF
PRODUCTION
BY KUPFFER
259
CELLS
TABLE
1
Kupffer Cell Total Protein Phosphorylation Response to LPS and H-7 Inhibitor (10 p&f) in Vitro Control Total phosphorylated proteins (3zPi counts x 1000) Total protein @g/culture)
-I
CONTROL
PMA
PMA/H-7
LPS
PMAiLPS
TREATMENT FIG. 4. Down regulation of LPS-stimulated TNF production by preincubation with PMA, a specific PKC agonist. (Mean + SEM). KC cultures were incubated for 48 hr with PMA (5 pg/ml) and then stimulated with LPS. Conditioned supernatants were collected at 4 hr and assayed by L929 cytolysis.
pg/ml) for 4 hr did not produce significant TNF release versus controls. (PMA, 13 -t 5 units, vs control, 3 + 4 units, Fig. 4). Effect of LPS on total and specific protein phosphoryEation in Kupfier cells. Treatment of the KC cultures with LPS or H-7 for 4 hr did not alter total cellular protein (Table 1). Total cellular protein phosphorylation measured by incorporation of labeled t3’Pi] also was not significantly changed by treatment with LPS, PMA, or H7 over 4 hr. (Table 1). However, LPS and PMA did result in the intensification in the radiolabeling of specific proteins at 40 kDa (2.7 +- 0.9-fold increase by densitometry) and 33 kDa (3.1 & l.O-fold by densitometry) compared to control (Fig. 5). This enhanced phosphorylation was inhibited by H-7 (Fig. 5).
H-7
LPS
LPS + H-7
35.3 +_5.1
32.1 t- 7.7
41.0 + 2.8
36.6 + 4.1
15.5 f 2.0
16.1 f 1.6
16.4 k 1.9
15.2 f 2.1
patients. Recent experimental evidence indicates that TNF is an early mediator released primarily by monocytes/macrophages that is capable of inducing many of the metabolic and physiologic responses characteristic of septic shock [2, 191. In vitro studies indicate that endotoxin or bacterial LPS is a potent signal for macrophage TNF release. Molecular mechanisms as to how the endotoxin signal is transduced into the appropriate intracellular responses resulting in macrophage TNF release remain elusive. Defining the mechanisms of TNF production provides a potential interventional approach for the future treatment of septic shock. Protein phosphorylation is a generalized mechanism by which extracellular signals are transduced into appropriate cellular responses. Specific extracellular signals initiate the production of membrane-derived middle messengers (cyclic nucleotides, inositol phosphates, or diacylglycerol) that activate specific protein kinases, providing appropriate intracellular protein phosphorylation. The
Effect
68
kD
45
kD
36
kD
29
kD
21
kD
20
kD
14
kD
of
PKC
Inl
DISCUSSION
Our results indicate that protein kinase C activity is necessary for LPS-activated Kupffer cells to produce tumor necrosis factor-a. Supporting this hypothesis are the findings that the PKC-specific inhibitor H-7 significantly reduces LPS-stimulated TNF production over the inhibition observed for the CAMP-dependent kinase inhibitor H-8. In addition, the down regulation of PKC by phorbol ester stimulation also reduces LPS-stimulated TNF production, Gel electrophoresis indicates that LPS stimulation increases the phosphorylation of specific KC intracellular proteins and that this increased phosphorylation is at least partially sensitive to H-7 inhibition. Septic shock due to gram-negative bacteremia or endotoxemia continues to be a major problem in surgical
LPS(H.l’PMAtH.7 “1
FIG. 5. Specific KC phosphorylated proteins following LPS and PMA stimulation. Representative autoradiogram from three replicates of 32Pi-labeled KCs as described under Materials and Methods.
260
JOURNAL
OF SURGICAL
RESEARCH:
protein phosphorylation then contributes to the desired cellular response. Protein kinase C is a specific kinase activated by signal-induced hydrolysis of membrane inositol phospholipids producing diacylglycerol and inositol phosphates [20]. Our results indicate that protein phosphorylation via protein kinase C is required for TNF release by LPS-stimulated Kupffer cells. The role of PKC in TNF production has been implicated from studies in other macrophage populations. The human monocyte-like cell line HL-60 was shown to produce TNF-(Y when stimulated with the phorbol ester PMA. [13] Thus, direct activation of PKC by phorbol esters was a sufficient signal in these transformed cells. Our results indicate that PMA was an insufficient direct signal for KC TNF release. This difference is most likely due to differences between Kupffer cells and monocytes and primary cultures versus cell lines. Two reports have studied the in vitro activity of PKC in response to LPS and lipid A (the active portion of LPS) in murine macrophages with differing results [21,22]. Lipid A enhanced PKC activity, while LPS had no effect. Recently, production of diacylglycerol and inositol phosphate has been reported in response to LPS stimulation in murine peritoneal macrophages ]23]. Our data support the role of PKC in LPS-stimulated macrophage TNF production and are the first investigating Kupffer cells or liver nonparenchymal cells. We have utilized the isoquinoline-sulfonamide-derived inhibitors H-7 and H-8 to demonstrate the role of PKC. These inhibitors compete for ATP with the ATP binding site on the kinase. H-7 has been demonstrated to be a more potent inhibitor of PKC with a Ki of 6 pM, while H-8 is more potent for cyclic nucleotide-dependent kinases (Ki for PKC is 15 pM) [14]. However, as our results indicate, the specificity of the inhibition is dependent on the concentration used and at sufficiently high levels H8 will also inhibit LPS-stimulated TNF production significantly. The inhibition of TNF production by H-8 at concentrations above the Ki for PKC suggests that this may be due to PKC inhibition; however, inhibition of other intracellular signals necessary for Kupffer cell TNF production cannot be ruled out. These inhibitors have previously been demonstrated to inhibit LPS-signaled B lymphocyte activation and thrombin-activated protein phosphorylation in platelets, both responses attributed to PKC activation [8, 91. Our results that the PKC inhibitor reduces PGEz release by LPS-stimulated Kupffer cells is consistent with reports demonstrating phorbol ester-stimulated PGEz production in macrophage populations [ 121. Since nonspecific inhibitory effects cannot totally be excluded in these experiments, we attempted to down regulate PKC by directly stimulating it with PMA for 48 hr and then determining LPS-stimulated TNF release. The finding of reduced TNF release is consistent with PKC down regulation and a necessary role for PKC in TNF release. The similiar protein levels excludes a cytotoxic effect from the inhibitors.
VOL. 49, NO. 3, SEPTEMBER
1990
The induction of increased specific protein phosphorylation in LPS-stimulated macrophages has been reported previously [17]. Our results as well as those of Weiel et al. [17] suggest that a 37-kDa protein is a specific phosphorylation target of PKC in macrophages. The function of this phosphoprotein is unknown. In summary, activation of protein kinase C would appear to be a necessary step in the activation of Kupffer cells for the release of TNF in response to endotoxin. Whether PKC activation is signaled via LPS-induced inositol phospholipid hydrolysis and diacylglycerol production is unclear. Furthermore, the role of specific proteins phosphorylated by activated protein kinase C is unknown; however, since LPS stimulation is associated with TNF gene transcription and increased mRNA levels, it is tempting to speculate a role for these phosphoproteins in the regulation of these responses [3]. Defining these mechanisms may have important applications in altering macrophage function not only during the septic response but also in other disease processes in which the macrophage participates, such as atherosclerosis, tumor killing and cachexia, and wound healing. REFERENCES 1.
Beutler, B., and Cerami, A. Cachetin and tumor necrosis factor as two sides of the same biologic coin. Nature (London) 320: 584, 1986.
2.
Michie, H. R., Spriggs, D. R., Manogue, K. R., et al. Tumor necrosis factor and endotoxin induce similar metabolic responses in human beings. Surgery 104: 280, 1988.
3.
Beutler, B., Krochin, N., Milsark, I., Luedke, C., and Cerami, A. Control of cachetin (tumor necrosis factor) synthesis: Mechanisms of endotoxin resistance. Science 232: 977, 1986. Dietch, E., Winterton, J., et al. The gut as a portal of entry for bacteremia: Role of protein malnutrition. Ann. Surg. 206: 681, 1987. West, M. A., Billiar, T. R., Mazuski, J. E., et al. Endotoxin modulation of hepatocyte secretory and cellular protein synthesis is mediated by kupffer cells. Arch. Surg. 123: 1400, 1988. Morrison, D. C., and Rudbach, J. Endotoxin-cell-membrane interactions leading to transmembrane signaling. Contemp. Top. MOE. Zmmunol. 81: 187, 1981. Magnuson, D., Maier, R., and Pohlman, T. Protein kinase C: A potential pathway of endothelial cell activation by endotoxin, tumor necrosis factor, and interleukin-1. Surgery 106: 216,1989.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Rush, J., and Waechter, C. Inhibitors of protein kinase C block activation of B lymphocytes by bacterial lipopolysaccharide. Biochem. Biophys. Res. Commun. 145: 1315,1987. Kawamoto, S., and Hidaka, H. I-(5-Isoquinolinesulfonylj-2-methylpiperazine (H-7) is a selective inhibitor of protein kinase C in rabbit platelets. Biochem. Biophys. Res. Commun. 126: 258, 1984. Kiyotaki, C., and Bloom, B. Activation of murine macrophage cell lines. Possible involvement of protein kinases in stimulation of superoxide production. J. Zmmunol. 133: 923, 1984. Castagna, M., Takai, Y., Kaibuchi, K., Sano, K., Kikkawa, U., and Nishizuka, Y. Direct activation of calcium-activated, phospholipiddependent protein kinase by tumor-promoting phorbol esters. J. Biol. Chem. 267: 784, 1982. Hoffman, T., Lizzie, E., Suissa, J., Rotrosen, D., Sullivan, J., Mandell, G., and Bonvini, E. Dual stimulation of phospholipase activity
BANKEY
13.
14.
15.
16.
17.
ET AL.: TNF
PRODUCTION
in human monocytes. Role of calcium-dependent and calcium-independent pathways in arachidonic acid release and eicosanoid formation. J. Zmmunol. 140: 3912, 1988. Pennica, D., Nedwin, G., et al. Human tumor necrosis factor: Precursor structure, expression, and homology to lymphotoxin. Nature (London) 312: 724,1984. Hidaka, H., Inagaki, M., et al. Isoquinolinesulfonamides, novel and potent inhibitors of cyclic nucleotide dependent protein kinase and protein kinase C. Biochemistry 23: 5036, 1984. Knook, D. L., and Sleyster, E. Isolated parenchymal, Kupffer and endothelial rat liver cells characterized by their lysosomal enzyme content. Biochem. Biophys. Res. Commun. 96: 250, 1980. Aggarwal, B., Kohr, W., Hass, P., Moffat, B., Spencer, S., Henzel, W., Bringman, T., Nedwin, G., Goeddel, D., and Harkins, R. Human tumor necrosis factor: Production, purification, and characterization. J. Biol. Chem. 260: 2345, 1985. Weiel, J. E., Hamilton, T., and Adams, D. LPS induces altered phosphate labeling of proteins in murine peritoneal macrophages. J. Zmmunol. 136: 3012, 1986.
BY KUPFFER
CELLS
261
18.
Laemmli, U. K. Protein electrophoresis gels. Nature (London) 227: 680, 1970.
in SDS polyacrylamide
19.
Starnes, H. F., Warren, R. S., Jeevanandam, M., et al. Tumor necrosis factor and the acute metabolic response to tissue injury in man. J. Clin. Znuest. 82: 1321, 1988.
20.
Nishizuka, Y. The family of protein kinase C for signal transduction. JAMA 262: 1826,1989.
21.
Hamilton, T., Becton, D., et al. Interferon-y modulates protein kinase C activity in murine peritoneal macrophages. J. Biol. Chem. 260: 1378, 1985.
22.
Wightman, P., and Raetz, C. The activation of protein kinase C by biologically active lipid moieties of lipopolysaccharide. J. Biol. Chem. 259: 10,048,1984.
23.
Prpic, V., Weiel, J., Sommers, S., DiGuiseppi, J., Gonias, S., P&o, S., Hamilton, T., Herman, B., and Adams, D. Effects of bacterial lipopolysaccharide on the hydrolysis of phosphatidyl-4,5-bisphosphate in murine peritoneal macrophages. J. Zmmunol. 139: 526, 1987.
