Recombinant mouse interleukin 10 (IL-10) was exceedingly potent at suppressing the ability of mouse peritoneal macrophages (m phi) to release tumor necrosis factor alpha (TNF-alpha). The IC50 of IL-10 for the suppression of TNF-alpha release induced by 0.5 microgram/ml lipopolysaccharide was 0.04 +/- 0.03 U/ml, with as little as 1 U/ml suppressing TNF-alpha production by a factor of 21.4 +/- 2.5. At 10 U/ml, IL-10 markedly suppressed m phi release of reactive oxygen intermediates (ROI) (IC50 3.7 +/- 1.8 U/ml), but only weakly inhibited m phi release of reactive nitrogen intermediates (RNI). Since TNF-alpha is a T cell growth and differentiation factor, whereas ROI and RNI are known to inhibit lymphocyte function, it is possible that m phi exposed to low concentrations of IL-10 suppress lymphocytes. m phi deactivated by higher concentrations of IL-10 might be permissive for the growth of microbial pathogens and tumor cells, as TNF-alpha, ROI, and RNI are major antimicrobial and tumoricidal products of m phi. IL-10's effects on m phi overlap with but are distinct from the effects of the two previously described cytokines that suppress the function of mouse m phi, transforming growth factor beta and macrophage deactivation factor. Based on results with neutralizing antibodies, all three m phi suppressor factors appear to act independently.

We report the identification and purification of a new inflammatory monokine synthesized by the macrophage tumor cell line RAW 264.7 in response to endotoxin. This monokine, which we term "macrophage inflammatory protein" (MIP), is a doublet with an apparent molecular mass of approximately 8,000 daltons on SDS-PAGE but forms aggregates of greater than 2 x 10(6) daltons as assessed by gel filtration. Partial NH2-terminal amino acid sequence data reveal no significant homology with any previously described protein. Although the monokine is anionic under physiological conditions, it is one of two major macrophage-secreted proteins that bind to heparin at high salt concentrations. At 100 ng/ml or greater, MIP is chemokinetic for human polymorphonuclear cells and triggers hydrogen peroxide production. Subcutaneous injection of 10 ng or greater of MIP into footpads of C3H/HeJ mice elicits an inflammatory response, characterized by neutrophil infiltration. These findings suggest that MIP is an endogenous mediator that may play a role in the host responses that occur during endotoxemia and other inflammatory events.

SummaryThe capacity of human neutrophils (PMN) to bind tumor necrosis factor (TNF) was rapidly lost when the cells were incubated in suspension with agents that can stimulate their migratory and secretory responses . Both physiological (poly)peptides (FMLP, C5a, CSF-GM) and pharmacologic agonists (PMN, calcium ionophore A23187) induced the loss ofTNF receptors (TNF-R) from the cell surface. Half-maximal loss in TNF-R ensued after only ti2 min with 10 -7 M FMLP at 37°C, and required only 10-9 M FMLP during a 30-min exposure . However, there were no such changes even with prolonged exposure of PMN to FMLP at 4°or 16°C. Scatchard analysis revealed loss of TNF-binding sites without change in their affinity (Kd -0.4 nM) as measured at incompletely modulating concentrations of FMLP, C5a, PMA, or A23187. The binding of antiTNF-R mAbs to PMN decreased in parallel, providing independent evidence for the loss of TNF-R from the cell surface. At the same time, soluble TNF-R appeared in the medium of stimulated PMN . This inference was based on the PMN-and FMLP-dependent generation of a nonsedimentable activity that could inhibit the binding of TNF to fresh human PMN or to mouse macrophages, and the ability of mAbs specific for human TNF-R to abolish inhibition by PMN-conditioned medium of binding of TNF to mouse macrophages. Soluble TNF-R activity was associated with a protein of Mr -28,000 by ligand blot analysis of cell-free supernatants of FMLP-treated PMN . Thus, some portion of the FMLP-induced loss of TNF-R from human PMN is due to shedding of TNF-R . Shedding was unaffectedby inhibitors of serine and thiol proteases and could not be induced with phosphatidylinositol-specific phospholipase C. Loss of TNF-R from PMN first stimulated by other agents may decrease their responsiveness to TNF. TNF-R shed by PMN may be one source of the TNF-binding proteins found in body fluids, and may blunt the actions of the cytokine on other cells.

Cells in culture are exposed to marked oxidative stress, H2O2 being one of the predominant agents. Pyruvate and other alpha-ketoacids reacted rapidly, stoichiometrically, and nonenzymatically with H2O2, and they protected cells from its cytolytic effects. All five human and murine cell types studied, both malignant and nonmalignant, released pyruvate at an initial rate of 35-60 microM/h/2.5 X 10(6) cells when placed in 1 ml pyruvate-free medium. After 6-12 h a plateau of 60-150 microM pyruvate was attained, corresponding to concentrations reported for normal human serum and plasma. The rate of pyruvate accumulation was almost doubled in the presence of exogenous catalase, suggesting that released pyruvate functions as an antioxidant. The rate of pyruvate accumulation was dependent on cell number. Succinate, fumarate, citrate, oxaloacetate, alpha-ketoglutarate, and malate were not secreted in significant amounts from P815 cells; export was specific for pyruvate and lactate among the metabolites tested. Extracellular pyruvate was in equilibrium with intracellular stores. Thus, cells conditioned the extracellular medium with pyruvate at the expense of intracellular pyruvate, until homeostatic levels were attained in both compartments. We propose that cells plated at low density in the absence of exogenous pyruvate fail to thrive for two reasons: prolonged depletion of intracellular pyruvate and prolonged vulnerability to oxidant stress.

The basis of resistance to oxidative injury was studied in six murine tumor cell lines that differed 54-fold in their resistance to enzymatically generated H(2)0(2). The tumors varied 56.7-fold in their specific activity of catalase, 5.3-fold in glutathione peroxidase (GPO), 3.3-fold in glutathione reductase (GR), and 2.7-fold in glutathione. There was no correlation among the levels of the three enzymes, and tumor cell resistance to lysis by H(2)0(2). However, the logarithm of the flux of H(2)0(2) necessary to cause 50 percent lysis of the tumor cells correlated with their content of glutathione (r = 0.91). The protective role of glutathione was analyzed by blocking GR and GPO, the catalysts of the glutathione redox cycle. This was facilitated by the demonstration that the anti-neoplastic agent 1,3-bis-(2- chloroethyl)-l-nitrosourea (BCNU) was a potent inhibitor of GR in intact tumor cells. BCNU inactivated tumor cell GR with a 50 percent inhibitory dose of 11 μM and a t(l/2) of inhibition of 30 s. Complete inhibition of GR was attained with no effect on GPO or catalase. Tumor cells whose GR was inactivated by BCNU could be lysed by fluxes of H(2)0(2) to which they were otherwise completely resistant. They could be killed by phorbol myristate acetate (PMA)-stimulated, bacilli Calmette-Guerin-activated macrophages in numbers which were otherwise insufficient, and by nonactivated macrophages, which otherwise were ineffective. BCNU-treated target cells were also much more sensitive to antibody-dependent, macrophage-mediated cytolysis. However, such tumor cells were no more sensitive than controls to lysis by alloreactive T cells or by antibody plus complement. Next, we deprived tumor cells of selenium by passage in selenium-deficient mice. GPO was inhibited 85 percent in such cells, with no effect on GR or catalase. Tumor cells with reduced GPO activity were markedly sensitized to lysis by small fluxes of H(2)0(2) or by PMA-stimulated macrophages or granulocytes. In contrast, inhibition of catalase with aminotriazole had no effect on the sensitivity of three tumors to peroxide-mediated lysis, and had modest effects with two others. Thus, the oxidation-reduction cycle of glutathione serves as one of the major defense mechanisms of tumor cells against three related forms of oxidant injury: lysis by fluxes of H(2)0(2), by PMA-triggered macrophages, and by macrophages in the presence of anti-tumor antibody.