Matrix-bound plasminogen activator inhibitor type 1 inhibits the invasion of human monocytes into interstitial tissue.

Invasion of tissue by monocytes in the course of cellular immune reactions is a multistep process that is thought to be based on the action of urokinase type plasminogen activator (u-PA), an ubiquitous serine protease able to convert the zymogen plasminogen into the active protease plasmin. Expression and occupation of urokinase-type plasminogen activator receptors (u-PA-R) are known to be up-regulated by IFN-gamma and TNF-alpha, and endogenously occupied u-PA-R were found to be instrumental in monocyte invasiveness. We used the amnion invasion assay to investigate whether monocyte invasiveness is affected by matrix-bound plasminogen activator inhibitors (PAI) and by fluid phase u-PA. We show in this study that preincubation of amnion membranes with 1.5 U/cm2 PAI-1 decreases invasion of IFN-gamma activated monocytes by 70% compared with controls. Anti-vitronectin antibodies, which block PAI-1 binding to the matrix, abrogate the inhibitory effect of PAI-1 on monocyte invasiveness, indicating that active PAI-1 is bound via matrix-associated vitronectin. In contrast, preincubation of the amnion membrane with PAI-2 which does not bind to the extracellular matrix has no effect on monocyte invasiveness. Finally, the inhibitory action of matrix-bound PAI-1 can be abrogated by addition of 5 IU/ml u-PA to the monocytes in the invasion chamber. These findings indicate that monocyte invasiveness might be regulated not only by expression and occupation of u-PA-R but also by matrix-bound PAI-1.     

IFN-gamma, tumor necrosis factor-alpha, and urokinase regulate the expression of urokinase receptors on human monocytes

The ability of macrophages to reach inflammatory loci is crucial in the function of cellular immunity. Invasive properties of macrophages may be due to the proteinase urokinase which binds to cell surface receptors, and thereby confers on macrophages the capacity for localized proteolysis of the interstitium. Here, we investigated the role of the macrophage-activating factors IFN-gamma, TNF-alpha, and granulocyte-macrophage-CSF and of urokinase on the expression of urokinase receptors by human cultured monocytes. IFN-gamma and TNF-alpha induced increased urokinase binding to human cultured monocytes in a time- and dose-dependent fashion. At optimal concentrations, IFN-gamma (200 U/ml) increased the number of receptors/cell from 14,000 to 64,000, TNF-alpha (50 U/ml) to 30,000, and combinations of IFN-gamma and TNF-alpha to 90,000. Granulocyte-macrophage-CSF had no effect. The enhanced urokinase binding is due to increased numbers of urokinase receptors and not an increased affinity of the receptor for urokinase. In the presence of urokinase during monocyte activation, IFN-gamma induced only 25,000 receptors/cell. However, urokinase does not inhibit increased receptor expression when the cells are activated with TNF-alpha. The effect of urokinase on induction of urokinase receptors by combinations of IFN-gamma and TNF-alpha varied with the dosage of TNF-alpha: A combination of IFN-gamma (200 U/ml) and TNF-alpha (15 U/ml) induced 38,000 receptors/cell in the presence and 90,000 receptors/cells in the absence of urokinase, whereas IFN-gamma (200 U/ml) and TNF-alpha (20 U/ml) induced 90,000 receptors/cell in the absence and presence of urokinase. These studies demonstrate that IFN-gamma, TNF-alpha, and urokinase collectively regulate the number of urokinase receptors on human monocytes. The induction of urokinase receptors may be responsible for increased invasiveness of the activated macrophage.     

Synergistic activation of human monocytes by granulocyte-macrophage colony-stimulating factor and IFN-gamma. Increased TNF-alpha but not IL-1 activity

TNF-alpha and IL-1 activities and PGE2 levels were investigated in the supernatants of highly purified human monocytes cultured for 18 h with recombinant human granulocyte-macrophage CSF (GM-CSF). GM-CSF alone did not stimulate IL-1 or TNF-alpha activities or the production of PGE2. GM-CSF with IFN-gamma, but not with LPS, consistently activated the monocytes for TNF-alpha activity. In contrast, for increased IL-1 activity, GM-CSF synergized weakly and irregularly with LPS, but not at all with IFN-gamma. For the third monocyte product investigated, GM-CSF was a weak and inconsistent inducer of PGE2 and only in the co-presence of IFN-gamma. Thus, GM-CSF can elicit different responses in human monocytes depending both on the co-stimulus as well as the monocyte product being investigated.     

Interferon-gamma Stimulates Monocyte Chemotactic Protein-1 Expression by Monocytes

Monocyte chemotactic protein (MCP-1) is a specific monocyte chemoattractant and activating factor produced by both immune cells (mononuclear phagocytes and lymphocytes) and non-immune cells (parenchymal and stromal cells). In order to define the conditions under which human monocytes express MCP-1, monocytes were exposed to IFN-gamma, IL- lbeta, TNF-alpha, IL-4 or PHA under serum free conditions. There was no significant MCP-1 production by monocytes following exposure to IL-lbeta, TNF-alpha or IL-4. In contrast, stimulation with IFN-gamma resulted in a dose dependent increase in MCP-1 protein and mRNA expression. Simultaneous stimulation with IFN-gamma and IL-1beta or TNF-alpha resulted in no further increase in MCP-1 production. It is concluded that IFN-gamma, primarily a product of T(H)1 T lymphocytes, stimulates the expression of MCP-1 by monocytes.     

IFN-gamma inhibits inflammatory cell recruitment and the evolution of bacterial cell wall-induced arthritis

Localization of streptococcal cell wall Ag (SCW) in the synovial tissue of treated rats induces an influx of leukocytes and a cell-mediated immune response leading to arthritis and joint destruction. Systemic administration of the T cell product, IFN-gamma (10(6) U/kg/day), suppresses the recruitment of leukocytes into the synovium and effectively inhibits the inflammation and pathology characteristic of SCW-induced arthritis (articular index 10.4 +/- 0.6 for SCW vs 2.0 +/- 0.7 for SCW with IFN-gamma, p less than 0.005). Monocyte-macrophages from animals treated with IFN-gamma exhibited defective chemotactic responses when tested in vitro and furthermore, monocytes cultured with IFN-gamma (25 to 500 U/ml) in vitro had significantly suppressed chemotactic responses to the complement fragment C5a (p less than 0.005). The decreased ability to migrate to C5a was associated with decreased binding of fluorochrome-conjugated C5a indicative of reduced expression of C5a receptors. Based on these data, IFN-gamma that induces monocyte maturation as reflected by increased Ia expression conversely inhibits C5a receptor expression. Although locally elevated IFN-gamma levels may serve to inhibit recruitment away from an inflammatory site, systemic exposure to IFN-gamma appears to inhibit leukocyte recruitment to the inflammatory site by its ability to induce premature maturation and concomitant inability to respond to certain chemotactic ligands. Inasmuch as monocyte recruitment to the synovium is pivotal in the development of SCW-induced polyarthritis, the ability of IFN-gamma to inhibit this event effectively inhibits the synovial pathology.     

Murine monocytes express transferrin receptors: evidence for similarity to inflammatory macrophages

Nonionic detergent extracts of murine monocytes contain a specific, high-affinity binding site for the serum glycoprotein transferrin. The binding activity was saturable and specific for 125I-labeled transferrin. The transferrin-receptor content of monocytes was compared with that of resident peritoneal macrophages and thioglycollate-elicited macrophages. Whereas resident cells showed no detectable activity, inflammatory macrophages and monocytes both bound transferrin to a similar degree. The calculated dissociation constant (2.3 X 10(-10)) and the number of sites per monocyte (11,400) compared favorably with those reported for transferrin receptors on inflammatory macrophages. Thus, based on the expression of the transferrin receptor, murine monocytes resemble murine inflammatory peritoneal macrophages rather than resident tissue macrophages.     

Recombinant interferon-gamma induces interleukin 2 receptors on human peripheral blood monocytes

The present report describes the inducibility of IL 2 receptors on human peripheral blood monocytes. Although freshly isolated monocytes are IL 2 receptor negative, approximately one-third of the cells react with the anti-Tac antibody within 18 hr of culture. IFN-gamma is found to double both the number of positive cells and the number of binding sites, whereas IL 2 has no influence on the IL 2 receptor expression on monocytes. Anti-Tac precipitates from monocyte lysates several protein bands of similar m.w. to those previously found with activated T and B cells. Finally, IFN-gamma-induced, but not resting, monocytes are found to bind recombinant IL 2. We conclude that IFN-gamma induces peripheral blood monocytes to express IL 2 receptors similar in structure to those found on activated T and B lymphocytes.     

TNF-alpha plus IFN-gamma induce connexin43 expression and formation of gap junctions between human monocytes/macrophages that enhance physiological responses

In this work, the effects of bacterial LPS, TNF-alpha, and IFN-gamma on gap junctional communication (dye coupling) and on the expression of connexin43 (immunofluorescence, immunoblotting, and RT-PCR) in monocytes/macrophages were studied. Freshly isolated human monocytes plated at high density and treated either with LPS plus IFN-gamma or TNF-alpha plus IFN-gamma became transiently dye coupled (Lucifer yellow) within 24 h. Cells treated with LPS, TNF-alpha, or IFN-gamma alone remained dye uncoupled. In dye-coupled cells, the spread of Lucifer yellow to neighboring cells was reversibly blocked with 18 alpha-glycyrrhetinic acid, a gap junction blocker, but it was unaffected by oxidized ATP or probenecid, which block ionotropic ATP-activated channels and organic anion transporters, respectively. Abs against TNF-alpha significantly reduced the LPS plus IFN-gamma-induced increase in dye coupling. In dye-coupled monocytes/macrophages, but not in control cells, both connexin43 protein and mRNA were detected, and their levels were higher in cells with an elevated incidence of dye coupling. In dye-coupled cells, the localization of connexin43 immunoreactivity was diffuse at perinuclear regions and thin cell processes. The addition of 18-alpha-glycyrrhetinic acid induced a profound reduction of monocyte/macrophage transmigration across a blood brain barrier model. It also induced a significant reduction in the secretion of metalloproteinase-2 in cells treated with TNF-alpha plus IFN-gamma. We propose that some monocyte/macrophage responses are coordinated by connexin-formed membrane channels expressed transiently at inflammatory sites in which these cells form aggregates.     

Monocyte chemotaxis and activating factor production by keratinocytes in response to IFN-gamma

Monocytes accumulate in the epidermis and along the dermo-epidermal junction in several different inflammatory skin diseases. To determine whether human epidermal keratinocytes elaborate a specific chemotaxin responsible for the accumulation of monocytes at these anatomic sites, monocyte chemotactic activity in conditioned 16-h cultured keratinocyte supernatants were assayed using human peripheral blood monocytes as the target cell. Dilutional analysis revealed directed monocyte migration in IFN-gamma-treated (100 U/ml) keratinocyte supernatants (80% maximal FMLP response) which was 10-fold more than IFN-gamma itself or untreated keratinocyte activity alone. Gel filtration chromatography revealed that this activity eluted just ahead of a 12.5-kDa molecular mass marker. Blocking studies demonstrated that a rabbit polyclonal antibody to monocyte chemotaxis and activating factor (MCAF) inhibited all monocyte chemotaxis by greater than 80%. Keratinocytes were metabolically labeled with 35S-cysteine/methionine, and after 16 h incubation the supernatants immunoprecipitated with the same anti-MCAF antibody. MCAF was detected as a protein doublet of 12 and 9 kDa only in IFN-gamma-treated (100 U/ml) keratinocyte supernatants. Incubation with IFN-gamma and TNF-alpha (250 U/ml) in combination resulted in increased production of MCAF protein. By Northern blot analysis, MCAF mRNA was constitutively expressed in keratinocytes and upregulated only in the presence of IFN-gamma. TNF-alpha, IL-1 beta, transforming growth factor-beta and phorbol esters had no positive or negative influence on MCAF mRNA. These studies demonstrate that biologically active MCAF is elaborated by human epidermal keratinocytes upon activation by IFN-gamma, a cytokine also required for the induction of adherence between monocytes and keratinocytes. Keratinocyte-derived MCAF is likely to be important in the regulation of cutaneous monocyte trafficking and may also be responsible for the recruitment of Langerhans cells and dermal dendrocytes, which share many phenotypic features with monocytes/macrophages, to their anatomic locations in skin.     

Interaction between transmembrane TNF and TNFR1/2 mediates the activation of monocytes by contact with T cells

Monocytes and monocytic cells produce proinflammatory cytokines upon direct cell contact with activated T cells. In the autoimmune disease rheumatoid arthritis, the pivotal role of TNF-alpha implies that the interaction between transmembrane TNF-alpha (mTNF) and the TNF receptors (TNFR1 and TNFR2) might participate in the T cell contact-dependent activation of monocytes. Accordingly, treatment of rheumatoid arthritis by administration of a TNF-alpha-blocking Ab was found to significantly decrease TNF-alpha production by monocytes. Several lines of evidence indicated that signaling through TNFR1/2 and through mTNF (reverse signaling) is involved in TNF-alpha production by monocytes after T cell contact: 1) blocking mTNF on activated T cells leads to a significant reduction in TNF-alpha production; 2) down-regulation of TNFR1/2 on monocytes by transfection with small interfering RNA results in diminished TNF-alpha production; 3) blocking or down-regulating TNFR2 on activated T cells inhibits TNF-alpha production, indicating that mTNF on the monocyte surface mediates signaling; 4) ligation of mTNF on monocytes by surface TNFR2 transfected into resting T cells induces TNF-alpha production due to reverse signaling by mTNF; and 5) ligation of mTNF on monocytes by a soluble TNFR2:Ig receptor construct induces TNF-alpha production due to reverse signaling. In conclusion, we identified mTNF and TNFR1/2 as interaction partners contributing to TNF-alpha production in monocytes. Both pathways initiated by mTNF-TNFR interaction are likely to be inhibited by treatment with anti-TNF-alpha Abs.     

Lipopolysaccharide-induced modulation of human monocyte urokinase production and activity

Cells of the monocyte/macrophage lineage are known to produce urokinase type plasminogen activator (u-PA) and are active participants in the inflammatory response. Modulation of cellular u-PA production, for instance in response to LPS, may have an important impact on the evolution of inflammatory lesions. A definitive picture of how monocyte u-PA production and activity are regulated by LPS is lacking. We addressed this issue directly by measuring u-PA Ag and activity in mononuclear cell cultures. By using a competition ELISA to quantitate u-PA Ag, we found that LPS-stimulated mononuclear cells in culture increased u-PA production in a dose-dependent manner and that all the u-PA detected was attributable to the monocytes therein. Increasing amounts of u-PA were secreted into the medium, bound to the cell surface, and found intracellularly. Although the absolute amounts of u-PA varied from donor to donor, the increases seen with LPS stimulation were a consistent and statistically significant finding. Only the cell-surface-bound u-PA was fibrinolytically active, however, with this activity increasing upon LPS stimulation. All monocyte cell-surface-associated fibrinolytic activity was attributed to u-PA, as shown by plasminogen dependence, neutralization by antibodies to u-PA, and identification of fibrinolytically active molecules eluted from the cell surface. The surface bound u-PA was not inhibited by its physiologic inhibitors, PAI-1 or PAI-2, whereas free u-PA was. Hence LPS stimulation results in monocytes exhibiting increased cell-surface-associated u-PA Ag and fibrinolytic activity, in spite of concomitant high levels of plasminogen activator inhibitor type 2 production. This surface-bound enzymatic activity may influence the ability of monocytes to migrate in and interact with an inflammatory microenvironment.     

Emergence of anti-inflammatory monocytes in long-term surviving hosts of IL-10-transduced liver allografts

We previously reported that transduction of IL-10 to rat liver allografts facilitates survival prolongation after orthotopic liver transplantation (OLT). In the current study, we examined Lewis hosts of IL-10-transduced allografts that had survived longer than 50 days in order to characterize peripheral blood mononuclear cells (PBMC). Phenotype analysis of the PBMC demonstrated an 18-fold increase in monocytes (CD11b/c(+)) with a massive increase in the monocyte/lymphocyte ratio. The monocytes expressed downregulated MHC class II (RT1B) but upregulated Fcgamma receptors in comparison with monocytes from the control hosts. The capacity of enriched monocytes to secrete TNF-alpha in response to LPS stimulation was downregulated in the survivors, while production of IL-10 was comparable. The monocytes from long-term survivors significantly inhibited the donor antigen stimulated secretion of IFN-gamma and IL-2. Monocytosis characterized by a shift to anti-inflammatory monocytes is associated with survival prolongation in the hosts of IL-10 transduced liver allografts.     

Triggering HLA-DR molecules on human peripheral monocytes induces their death

Although engagement of MHC class II molecules on human monocytes triggers various cellular events, their possible role in monocyte death is yet unknown. We demonstrate that ligation of MHC class II on primary monocytes induces a rapid cell death that has all the characteristics of monocyte apoptosis, does not require de novo protein synthesis, is independent from both Fas and TNF-alpha systems, and is not rescued by ligation of CD40. However, cell-cell interactions that involve the beta2-integrin CD18 seem to be critical for the execution of this monocyte death. Priming monocytes with IFN-gamma enhances significantly their HLA-DR-mediated death whereas LPS treatment effectively reverses this death process. Thus, our results describe the MHC class II molecules, in particular HLA-DR, as mediators of monocyte death and suggest that this novel pathway of monocyte death might have an important role in controlling the outcome of inflammatory process and the regulation of monocyte hemostasis.     

Differential regulation of human monocyte programmed cell death (apoptosis) by chemotactic factors and pro-inflammatory cytokines

In the absence of appropriate stimuli, monocytes undergo programmed cell death (PCD) or apoptosis. IL-1 beta and TNF-alpha prevent monocyte PCD, which suggests that viability may be regulated by biologically active peptides released during inflammation. To explore this possibility, we evaluated several chemotactic factors and pro-inflammatory cytokines for their ability to regulate PCD. The recruitment factors, FMLP, C fragment C5a, monocyte chemotactic protein-1, or transforming growth factor-beta 1, were incapable of rescuing monocytes from PCD nor did they enhance PCD, whereas several inflammatory cytokines in addition to IL-1 beta and TNF-alpha, including granulocyte-monocyte-CSF and IFN-gamma, prevented monocyte PCD provided that sufficient levels of these cytokines were continuously maintained in the cultures. Cytokine-mediated inhibition of PCD could be blocked by specific antisera, ruling out potential effects caused by LPS contamination. When tested at equivalent concentrations, IL-2, IL-4, and IL-6 had no effect on PCD indicating selectivity in cytokine modulation of monocyte PCD. Because monocytes produce IL-1 beta, TNF-alpha, and granulocyte-monocyte CSF when activated, the data suggest autocrine as well as paracrine control of cell survival and accumulation. The results also suggest that monocytes recruited to a site of inflammation will undergo PCD in the absence of specific cytokines and/or other stimuli that block this process.     

Normal nonmetastatic human trophoblast cells share in vitro invasive properties of malignant cells

First-trimester normal human trophoblast cells show some phenotypic similarities to malignant cells, e.g., rapid proliferation and ability to invade neighboring tissue, including basement membrane in situ, but do not have the ability for unlimited growth or metastasis. The present study examined whether the invasive ability of normal trophoblast cells is an intrinsic property of these cells, independent of the microenvironment provided by the pregnant uterus, and if so, whether they share some of the molecular mechanisms of invasion exercized by metastatic malignant cells. The ability of in vitro grown human trophoblast lines to invade an epithelium-free human amniotic membrane was measured from the temporal kinetics of retention of radioactivity within this membrane resulting from a penetration by 125I-iododeoxyuridine-labeled trophoblast cells. The magnitude of this invasion was compared to that of the highly metastatic human JAR-choriocarcinoma cell line and murine B16F10 melanoma line. Trophoblasts were found to share some of the same molecular mechanisms of invasion with the metastatic cell lines. Inhibitors of collagenase, plasmin, plasminogen, and plasminogen activators completely prevented invasion of the amnion by the trophoblast lines as well as by the metastatic JAR and B16F10 lines. Mersalyl, a compound known to activate collagenase, stimulated invasion by all cell lines tested, including under conditions in which plasmin activity was inhibited. In addition, trophoblasts produced significant levels of type IV collagenase and laminin, both of which appear to be important products of metastatic tumor cells required for basement membrane invasion. It may be concluded from these findings that the invasive property of first trimester human trophoblasts is genetically determined; that the magnitude of amnion invasion cannot differentiate between metastatic cell lines and invasive but nonmetastatic cell lines; and that invasiveness is not a sufficient prerequisite for metastatic ability.     

Activated T lymphocytes regulate hyaluronan binding to monocyte CD44 via production of IL-2 and IFN-gamma

Interactions of the cell surface proteoglycan CD44 with the extracellular matrix glycosaminoglycan hyaluronan (HA) are important during inflammatory immune responses. Our previous studies indicated that monocyte HA binding could be induced by TNF-alpha. Moreover, monocyte HA binding could be markedly up-regulated by culturing PBMC with anti-CD3 (TCR complex) mAbs. The present study was undertaken to identify soluble factors and/or cell surface molecules of activated T lymphocytes that might regulate HA binding to monocytes. Abs to IL-1 alpha, IL-1 beta, IL-2, IL-3, IL-10, IL-15, GM-CSF, IFN-gamma, and TNF-alpha were tested for their effects on anti-CD3 mAb-, Con A-, and PMA/ionomycin-mediated monocyte HA binding in PBMC cultures. Anti-TNF-alpha, anti-IL-2, and anti-IFN-gamma Abs, when added together to PBMC cultures, completely blocked Con A- and partially blocked anti-CD3- and PMA/ionomycin-induced monocyte HA binding. Furthermore, when added together to PBMC cultures, IL-2 and TNF-alpha induced high levels of monocyte HA binding. Likewise, IFN-gamma augmented TNF-alpha-induced monocyte HA binding. To investigate the role of T cell-monocyte direct contact in induction of monocyte HA binding, we studied PMA/ionomycin-activated, paraformaldehyde-fixed CD4(+) T cells in these assays. Fixed, PMA/ionomycin-activated CD4(+) T lymphocytes induced monocyte HA binding, but direct T cell-monocyte contact was not required. Moreover, anti-IFN-gamma and anti-TNF-alpha Abs blocked fixed PMA/ionomycin-activated CD4(+) T cell-induced monocyte HA binding. Taken together, these studies indicate roles for soluble T lymphocyte-derived factor(s), such as IL-2 and IFN-gamma, and a role for monocyte-derived TNF-alpha in Con A-, TCR complex-, and PMA/ionomycin-induced HA binding to monocyte CD44.     

The principal tumor necrosis factor receptor in monocyte cytotoxicity is on the effector cell, not on the target cell.

Several tumor target cell lines, prototypically K562 cells, are resistant to lysis by recombinant tumor necrosis factor (TNF alpha) but are killed by monocytes expressing membrane-associated TNF, suggesting that membrane TNF could account for monocyte-mediated cytotoxicity. Formaldehyde-fixed monocytes or extracted monocyte membrane fragments are cytotoxic to K562 target cells. Treatment of monocytes with interferon-gamma (IFN-gamma) increases cytotoxicity by live and fixed cells or by extracted monocyte membranes. Both TNF and TNF receptors are detectable on monocyte membranes by FACS analysis, and the levels of each are modulated by treatment with IFN-gamma. Cytotoxicity can be inhibited by either anti-TNF or anti-TNF receptor antibodies. Incubation of effector cells with exogenous soluble TNF prior to fixation or membrane preparation increases their cytotoxicity. In contrast, incubation of the target cells with exogenous TNF neither increases nor decreases killing by effector cell membrane fragments or intact effector cells. The data suggest that the TNF receptors on the effector cell, but not on the target cell, play a crucial role in TNF-mediated cytotoxicity.     

Lung-specific delivery of cytokines induces sustained pulmonary and systemic immunomodulation in rats

The recombinant cytokines IFN-gamma and TNF-alpha stimulate several macrophage-mediated functions important in host defense. However, systemic administration of cytokines may be limited by significant host toxicity. We investigated whether aerosolized cytokines can stimulate alveolar macrophage and blood monocyte function, and whether they induce an inflammatory response in the lungs of normal rats. We found that aerosolized murine rIFN-gamma or recombinant human TNF-alpha increased IL-1 production by both alveolar macrophages and blood monocytes for at least 5 days after administration. Furthermore, murine rIFN-gamma increased the expression of Ia Ag on alveolar macrophages and human rTNF-alpha increased alveolar macrophage- and blood monocyte-mediated tumor lysis. Sequential aerosolization of IFN-gamma and TNF-alpha significantly increased both IL-1 release and Ia expression compared to either cytokine administered alone. Aerosolized human rTNF-alpha achieved lung levels comparable to those produced by an i.v. TNF-alpha dose reported to cause diffuse organ injury and death in rats. However, plasma TNF-alpha levels were several thousand-fold lower after aerosol administration. Aerosolized cytokines did not induce lung edema or an inflammatory cell infiltrate within the airways or alveoli. Aerosolized human rTNF-alpha alone, or murine rIFN-gamma and human rTNF-alpha, induced margination of leukocytes in pulmonary blood vessels 1 day after aerosolization, and a few small foci of pulmonary hemorrhage 5 days later. We conclude that aerosol administration of IFN-gamma or TNF-alpha enhances both pulmonary and systemic monocyte function, and that the combination of IFN-gamma and TNF-alpha produce additive or synergistic effects. Aerosolized cytokines induce only a minimal pulmonary inflammatory response. Aerosolized TNF-alpha produces high cytokine levels in the lung but very low uptake into the circulation.