Cyclophosphamide stimulates lung fibroblasts to release neutrophil and monocyte chemoattractants

Cyclophosphamide is an alkylating antineoplastic agent used in several conditions. However, little is known about the mechanism of its pulmonary toxicity. In the present study, we determined that human lung fibroblasts release activity for neutrophils and monocytes in response to cyclophosphamide in a dose- and time-dependent manner. Checkerboard analysis revealed that both neutrophil and monocyte activities were chemotactic. The release of chemotactic activity was inhibited by lipoxygenase inhibitors and cycloheximide. Molecular-sieve column chromatography revealed that both neutrophil (NCA) and monocyte (MCA) chemotactic activities had multiple peaks. NCA was inhibited by a leukotriene B(4) receptor antagonist and anti-interleukin-8 and anti-granulocyte colony-stimulating factor antibodies. MCA was attenuated by a leukotriene B(4) receptor antagonist and anti-monocyte chemoattractant protein-1 and anti-granulocyte-macrophage colony-stimulating factor antibodies. The concentrations of interleukin-8, granulocyte colony-stimulating factor, monocyte chemoattractant protein-1, and granulocyte-macrophage colony-stimulating factor significantly increased in response to cyclophosphamide. These data suggest that lung fibroblasts may modulate inflammatory cell recruitment into the lung by releasing NCA and MCA in response to cyclophosphamide.     

Bronchial epithelial cells release monocyte chemotactic activity in response to smoke and endotoxin

An increase in mononuclear phagocytes occurs within the airways during airway inflammation. Bronchial epithelial cells could release monocyte chemotactic activity and contribute to this increase. To test this hypothesis, bovine bronchial epithelial cells were isolated and maintained in culture. Bronchial epithelial cell culture supernatant fluids were evaluated for monocyte chemotactic activity. Epithelial cell culture supernatant fluids attracted significantly greater numbers of monocytes compared to media alone and the number of monocytes attracted increased in a time dependent manner. Endotoxin and smoke extract induced a dose and time dependent release of monocyte chemotactic activity compared with cells cultured in media (52.5 +/- 2.6 (endotoxin), 30.5 +/- 2.3 (smoke) vs 20.5 +/- 2.2 cells/high power field (HPF) p less than 0.001). The released activity was chemotactic by checkerboard analysis. Stimulation of the epithelial cells by opsonized zymosan, calcium ionophore (A23187), and PMA also resulted in an increase in monocyte chemotactic activity (p less than 0.01). Because the release of activity was blocked by the lipoxygenase inhibitors, nordihydroguaiaretic acid and diethycarbamazine, epithelial cell monolayers were cultured with 3 microCi [3H]arachidonic acid for 24 h and then exposed to A23187, PMA, or both stimuli, for 4, 8, and 24 h. Analysis of the released 3H activity was performed with reverse-phase HPLC and revealed that the major lipoxygenase product was leukotriene B4. These data suggest that monocytes may be recruited into airways in response to chemotactic factors released by bronchial epithelial cells.     

Porcine spermadhesin PSP-I/PSP-II stimulates macrophages to release a neutrophil chemotactic substance: modulation by mast cells

The complex of porcine seminal plasma heterodimers I and II (PSP-I/PSP-II), which are heterodimers of glycosylated spermadhesins, is the major component of porcine seminal fluid. The proinflammatory and immunostimulatory activities of this spermadhesin complex suggest its participation in modulation of the uterine immune activity that may ensure reproductive success. Spermadhesin PSP-I/PSP-II induced the migration of neutrophils into the peritoneal cavity of rats via activation of resident cells. In the present study, we have investigated the involvement of macrophages and mast cells in the neutrophil chemotactic activity of PSP-I/PSP-II and the underlying mechanism. Macrophages and mast cells were isolated, cultured, and stimulated with purified PSP-I/PSP-II. Pharmacological modulation was performed using the glucocorticoid dexamethasone, indomethacin (cyclooxygenase inhibitor), MK886 (leukotriene inhibitor), and the supernatant of spermadhesin-stimulated mast cells. Macrophages stimulated with PSP-I/PSP-II released into the culture supernatant a neutrophil chemotactic substance. This activity was partly inhibited by both dexamethasone (85%) and the supernatant of spermadhesin-stimulated mast cells (74%) but not by indomethacin and MK886. An anti-tumor necrosis factor (TNF) alpha antibody neutralized (by 68%) the neutrophil chemotactic activity of PSP-I/PSP-II-stimulated macrophages. An anti-interleukin (IL)-4 antibody blocked the inhibitory activity of spermadhesin-stimulated mast cells on release of a neutrophil chemotactic substance by PSP-I/PSP-II-stimulated macrophages. As a whole, these data indicate that the neutrophil migration-inducing ability of spermadhesin PSP-I/PSP-II involves the release of the inflammatory cytokine TNFalpha by stimulated macrophages and that this activity is modulated by the lymphokine IL-4 liberated by mast cells. The balance between these two cytokines may control onset of the local inflammatory reaction, avoiding excessive neutrophil recruitment that would lead to tissue damage.     

Monokine-induced neutrophil chemotactic factor gene expression in human fibroblasts

Although fibroblasts are important in providing a structural framework for most tissues, they also appear to be active participants in the inflammatory process via the production of specific mediators. The production of inflammatory mediators by fibroblasts is especially important in relation to their strategic location within connective tissue as they may act as a cellular communication bridge between the interstitium and vasculature. In this paper, we demonstrate that fibroblasts may participate in these inflammatory reactions by the production of a neutrophil chemotactic factor (NCF) with characteristics similar to a recently isolated and cloned monocyte-derived NCF. Either tumor necrosis factor-alpha-, interleukin-1 alpha-, or interleukin-1 beta-stimulated fibroblasts showed both a time- and dose-dependent increase in steady-state levels of NCF mRNA and secretion of chemotactic activity. In contrast, lipopolysaccharide and interleukin-6 failed to induce fibroblast-derived NCF. The expression of fibroblast-derived NCF mRNA was first detectable by 30 min poststimulation, whereas chemotactic activity was significantly observed 3-4 h postchallenge. Heat-inactivated monokine (100 degrees C) failed to induce NCF mRNA expression, suggesting that only the active proteins are capable of inducing NCF. Gel filtration analysis using high pressure liquid chromatography indicated peak chemotactic activity with an approximate molecular mass of 8000 daltons. This peak of NCF activity was found to be relatively stable to both heat and trypsin inactivation. Specificity of the fibroblast-derived neutrophil chemotactic activity was demonstrated with inhibition of chemotaxis by the addition of neutralizing antibody directed against recombinant human neutrophil chemotactic factor. These data provide evidence that monokine-treated fibroblasts can synthesize a potent chemotactic agent with molecular and physicochemical characteristics similar to monocyte-derived NCF and that this factor may contribute to neutrophil-mediated disease processes.     

Cyclic strain stimulates monocyte chemotactic protein-1 mRNA expression in smooth muscle cells

Hemodynamic forces are important determinants for the formation of atherosclerotic plaques. The recruitment of circulating monocytes into the arterial wall is an important step during atherogenesis. Monocyte chemotactic protein-1 (MCP-1) has been shown to be a key factor for monocyte transmigration. This study examined the effects of cyclic strain on MCP-1 mRNA expression levels of cultured rat aortic smooth muscle cells. The MCP-1 mRNA levels of aortic smooth muscle cells first increased as the duration of cyclic strain increased, reaching the maximum at 6-12 h, maintained at high levels throughout the 48-h strain period. To explore signaling pathways mediating cyclic strain-stimulated MCP-1 mRNA expression, we examined the involvement of tyrosine kinase and protein kinase C (PKC). Tyrosine kinase inhibitors, genistein and tyrphostin 51, at 50 microM blocked cyclic strain-stimulated MCP-1 mRNA expression. Preincubation with a PKC activator, phorbol 12-myristate 13-acetate (PMA), 2 microM, for 24 h to downregulate PKC did not decrease cyclic strain-induced MCP-1 mRNA expression. A 6-h incubation with 0. 1 microM PMA to activate PKC, which stimulated MCP-1 expression when applied alone, abolished the stimulatory effects of cyclic strain. A specific PKC inhibitor, calphostin C (0.1 microM), diminished cyclic strain-stimulated MCP-1 mRNA expression. Angiotensin II at 10 or 1,000 nM induced a moderate upregulation of MCP-1 mRNA, and no synergistic effects were observed between angiotensin II and cyclic strain. These results indicate that cyclic strain stimulates MCP-1 mRNA expression in smooth muscle cells through signaling pathway(s) mediated by tyrosine kinase activation.     

Selective neutrophil desensitization to chemotactic factors

In the presence of extracellular calcium and magnesium, a series of chemotactic oligopeptides and C5a caused aggregation of human polymorphonuclear neutrophils (PMNs). This cellular response developed rapidly and began to reverse 2 min after exposure to the chemotactin. In the absence of the bivalent cations, none of the chemotactins stimulated the aggregation response. If cells were first exposed to a chemotactin and then treated with calcium and magnesium, aggregation was detected only after addition of the cations, and the magnitude of the response fell sharply as the interval between the addition of chemotactin and addition of cations was lengthened: when this interval exceeded 2 min, aggregation was barely detectable. This loss of reactivity persisted even when cells were re-exposed to fresh chemotactic factor and washed between the first and second exposures. In all instances, however, loss of cellular reactivity was highly selective: cells preincubated with any chemotactic oligopeptide were hyporesponsive to subsequent stimulation with an oligopeptide but remained fully responsive to C5a; cells preincubated with C5A were hyporesponsive to C5a but retained their responsitivity to the oligopeptides. Because this selectivity parallels the known specificities of these chemotactic factors for their receptors in or on the neutrophil, desensitization may reflect functional loss of receptors after stimulation. Alternatively, this selectivity may indicate that morphologically identical neutrophils contain subpopulations of cells with varying reactivities to receptor-bound chemotactic factors. In either event, desensitization may be useful in functionally defining chemotactic factors and their respective receptors. The rapidity of development of desensitization suggests that it may operate to limit or moderate various in vitro and in vivo neutrophil responses to chemotactic factors.     

Selectins and monocyte chemotactic peptide as the markers of atherosclerosis activity

The role of adhesive selectin molecules in the process of atherogenesis is an open question. These molecules are known as markers of atherosclerosis activity, however, only some biological mechanisms are known up to now. In this study we examined the levels of soluble forms of E-, P-selectin and monocyte chemotactic protein (MCP-1) in the process of extracorporeal cholesterol elimination by LDL-apheresis. We measured the levels of sE-, sP-selectin and MCP-1 in the plasma before and after LDL-apheresis and in the washout solution from immunoabsorption columns Lipopak. Eighty measurements were performed repeatedly in 6 patients with severe familial hypercholesterolemia (FH) on long-term LDL-apheresis treatment. Before the procedure P-selectin levels were 204+/-179 ng/ml, E-selectin 32.1+/-33.7 ng/ml, MCP-1 323.8+/-121 pg/l, whereas after the procedure we found P-selectin levels 131.6+/-34 ng/ml, E-selectin 33.1+/-51 ng/ml, and MCP-1 200.4+/-15 pg/l. Levels of P-selectin were increased in the blood of patients with FH in spite of long-term intensive extracorporeal LDL-elimination, documenting thus the activity of atherosclerosis. The levels of P-selectin and MCP-1 decreased significantly after the hypolidemic procedure and could be used as another marker showing the effectivity of the extracorporeal LDL-cholesterol elimination (immediately after the procedure), and, after further verification, may serve as a marker for controlling the therapy efficacy.     

Induction of macrophage antiprotozoal activity by monocyte chemotactic and activating factor

Abstract To determine if monocyte chemotactic and activating factor (MCAF) induces intracellular antimicrobial activity, human monocyte-derived macrophages were treated with MCAF and challenged with Toxoplasma gondii and Leishmania donovani . Pretreatment with MCAF induced macrophages to inhibit protozoal replication by approximately 50%. These findings suggest a potential host defense role for MCAF in the inflammatory response to intracellular pathogens.     

Cyclic strain stimulates monocyte chemotactic protein‐1 mRNA expression in smooth muscle cells

Hemodynamic forces are important determinants for the formation of atherosclerotic plaques. The recruitment of circulating monocytes into the arterial wall is an important step during atherogenesis. Monocyte chemotactic protein‐1 (MCP‐1) has been shown to be a key factor for monocyte transmigration. This study examined the effects of cyclic strain on MCP‐1 mRNA expression levels of cultured rat aortic smooth muscle cells. The MCP‐1 mRNA levels of aortic smooth muscle cells first increased as the duration of cyclic strain increased, reaching the maximum at 6–12 h, maintained at high levels throughout the 48‐h strain period. To explore signaling pathways mediating cyclic strain‐stimulated MCP‐1 mRNA expression, we examined the involvement of tyrosine kinase and protein kinase C (PKC). Tyrosine kinase inhibitors, genistein and tyrphostin 51, at 50 μM blocked cyclic strain‐stimulated MCP‐1 mRNA expression. Preincubation with a PKC activator, phorbol 12‐myristate 13‐acetate (PMA), 2 μM, for 24 h to downregulate PKC did not decrease cyclic strain‐induced MCP‐1 mRNA expression. A 6‐h incubation with 0.1 μM PMA to activate PKC, which stimulated MCP‐1 expression when applied alone, abolished the stimulatory effects of cyclic strain. A specific PKC inhibitor, calphostin C (0.1 μM), diminished cyclic strain‐stimulated MCP‐1 mRNA expression. Angiotensin II at 10 or 1,000 nM induced a moderate upregulation of MCP‐1 mRNA, and no synergistic effects were observed between angiotensin II and cyclic strain. These results indicate that cyclic strain stimulates MCP‐1 mRNA expression in smooth muscle cells through signaling pathway(s) mediated by tyrosine kinase activation. J. Cell. Biochem. 76:303–310, 1999. © 1999 Wiley‐Liss, Inc.     

Acetylcholine and substance P stimulate bronchial epithelial cells to release eosinophil chemotactic activity

Weinvestigated a role of neuroregulation in the release of eosinophilchemotactic activity (ECA) from bovine bronchial epithelial cells(BBEC). BBEC were stimulated with acetylcholine (ACh) and substance P(SP), and the supernatant fluids were tested for ECA by a blind-wellchemotactic chamber technique. BBEC released ECA in response to ACh andSP in a dose- and time-dependent manner. Checkerboard analysis showedthat ECA in regard to ACh and SP was chemotactic rather thanchemokinetic. Partial characterization revealed that ECA involved bothlipids and peptides. The release of ECA in response to ACh and SP wasinhibited by nonspecific and 5-specific lipoxygenase inhibitors and bycycloheximide (P < 0.01). Molecular-sieve columnchromatography revealed that these mediators induced three molecularmass peaks (near 25 kDa, 9 kDa, and 400 Da, respectively). The lowestpeak, which represented the predominant activity, was blocked byleukotriene B₄-receptor antagonist (P < 0.01) but not by platelet-activating factor-receptor antagonist. The releaseof leukotriene B₄ in the supernatant fluids was increased in response to ACh and SP stimulation (P < 0.01).Platelet-activating factor was not detected. These results raise thepossibility of a role of neuroregulation for the elaboration of ECA inthe airway.

     

Endotoxin stimulates bronchial epithelial cells to release chemotactic factors for neutrophils. A potential mechanism for neutrophil recruitment, cytotoxicity, and inhibition of proliferation in bronchial inflammation.

To test the effect of endotoxin on bronchial epithelial cells (BEC), BEC were isolated from bovine lungs and cultured in the presence of bacterial endotoxin. The BEC culture supernatant fluids were harvested, and neutrophil chemotactic activity (NCA) was determined with a blindwell chamber technique; cytotoxicity determined by lactate dehydrogenase release and BEC proliferation determined by Coulter counting. Endotoxin caused a dose- and time-dependent release of NCA from BEC cultures compared with media alone (82.3 +/- 8.1 vs 12.0 +/- 3.1 cells/high power field, p less than 0.001). To further characterize this activity, reverse phase HPLC analysis of release eicosanoid metabolites after [3H]arachidonic acid incorporation was performed. Endotoxin stimulated the release of the neutrophil chemoattractants, leukotriene B4 and 12-hydroxyeicosatetraenoic acids. Endotoxin also resulted in a dose and time dependent release of lactate dehydrogenase (42.9 +/- 4.2 vs 20.2 +/- 2.2 U/liter, p less than 0.001) although higher doses were required to cause cytotoxicity than to stimulate chemotaxis. Finally, endotoxin resulted in a dose dependent inhibition of BEC proliferation (176 x 10(3) +/- 16 x 10(3) vs 1,080 x 10(3) +/- 38 x 10(3) cells/ml measured at day 14, p less than 0.001). These data suggest that bacterial release of endotoxin may contribute to the pathophysiologic changes observed in bronchial inflammation by stimulating BEC to release NCA, denuding airway epithelium by causing cytotoxicity of BEC, and inhibiting epithelial repair by inhibiting BEC proliferation.     

Phosphorylase a activity as an indicator of neutrophil activation by chemotactic peptide

The activity of glycogen phosphorylase, an enzyme that is activated by both cAMP and calcium, was used as an indicator of the state of the cytoplasm after chemotactic stimulation of polymorphonuclear leukocytes (neutrophils). The activity of the enzyme showed a clear dependence on cytoplasmic calcium. Addition of the calcium ionophore A23187 caused a 4-5-fold increase in activity of phosphorylase a. In the absence of external Ca2+, A23187 caused only brief transient activation of phosphorylase; probably reflecting release of sequestered intracellular Ca2+. Addition of the chemotactic peptide N-formylnorleucylleucylphenylalanine (FNLLP) caused a transient 2-3-fold activation of the enzyme. The dose-dependence of activation by FNLLP showed a peak at 10(-8) M, near the Kd of the receptor for FNLLP. The phosphorylase activity peaks by 90 s and then declines, returning to basal levels by 20 min after stimulation with 10(-8) M peptide and by 60 min with 10(-7) M peptide. This finding suggests that the cells do not need to maintain elevated cytoplasmic calcium levels to exhibit stimulated locomotion. Thus, if calcium continues to modulate the motility, there either must be highly localized changes that are not detected in measures of the total cytoplasm, or the sensitivity to calcium must be variable such that basal levels are sufficient to maintain locomotion. Cells loaded with the fluorescence calcium probe quin2 (0.6 mM) in the presence or absence of external Ca2+ had elevated phosphorylase levels before addition of FNLLP. Thus, the presence of quin2 may alter the cytoplasmic Ca2+ level, and it clearly alters some aspects of the neutrophil physiology. Phosphorylase a appears to be a sensitive, nonperturbing indicator of the cytoplasmic calcium levels.     

Interleukin 1 stimulates endothelial cells to release multilineage human colony-stimulating activity

Cultured human umbilical vein endothelial cells, when exposed to soluble products of peripheral blood monocytes, elaborate granulocyte-macrophage colony-stimulating activity (GM-CSA) and erythroid burst-promoting activity (BPA). We have performed studies to determine if the monokine IL 1 can stimulate endothelial cells to release hematopoietic growth factors and whether such factors can also support human megakaryocyte (Meg) and mixed-cell colony growth. Various concentrations of recombinant human IL 1 beta (rIL 1) and media conditioned by monocytes (MCM), endothelial cells (ECM), and endothelial cells cultured for 3 days in 50% MCM (ECMM) or rIL 1 (ECMrIL 1) were added to marrow mononuclear cells cultured in methylcellulose. ECMM and ECMrIL 1 stimulated, in a dose-dependent fashion, the growth of Meg, mixed-cell, and GM colonies and erythroid bursts. In contrast, ECM, MCM, and rIL 1 displayed little or no activity in the colony-forming assays. Preincubation with specific antisera to native human IL 1 or rIL 1 reduced by 75 to 100% the activity of MCM in stimulating endothelial cell release of BPA, GM-CSA, Meg-CSA, and mixed-cell CSA. Meg-CSA, although readily detectable at ECMM and ECMrIL 1 concentrations in culture of 1 to 5%, was partially masked by lineage-specific inhibitors of Meg colony growth. When ECMM was analyzed by gel filtration chromatography, the megakaryocytopoietic inhibitory activity eluted in the high Mr fractions (greater than 75 kD). Meg-CSA co-eluted with GM-CSA and BPA in a single peak of 30 kD. We conclude that endothelial cells, in response to IL 1, produce one or more growth factors that probably act on multiple classes of progenitor cells.     

IL-1 alpha or tumor necrosis factor-alpha stimulate release of three NAP-1/IL-8-related neutrophil chemotactic proteins in human dermal fibroblasts

Human dermal fibroblasts in culture secrete three protein-like neutrophil chemotactic factors, when stimulated either with human rIL-1 alpha or IL-1 beta; not, however, after incubation with LPS. These three fibroblast-derived neutrophil-activating proteins (FINAP) could be purified by subsequently performed reversed phase and size exclusion HPLC. By high resolution SDS-PAGE, all the proteins were shown to migrate with an Mr of 6,700 (alpha-FINAP), 3,600 (beta-FINAP), and 5,300 (gamma-FINAP). All purified cytokine preparations were found to be chemotactic for human neutrophils. In addition, all FINAP induced release of lysosomal enzymes in neutrophils. Deactivation of chemotaxin-elicitable enzyme release showed cross-desensitization of all FINAP with NAP-1/IL-8. Western blot analysis of alpha-FINAP by using mAb against neutrophil-activating protein (NAP)-1/IL-8 reveals immunologic cross-reactivity with NAP-1/IL-8. By amino-terminal amino acid sequence analysis alpha-FINAP could be identified as the 77-residue extended form of NAP-1/IL-8 containing the 79-residue form as a minor contaminant. Whereas beta-FINAP has been found to be a truncation product of alpha-FINAP, gamma-FINAP shows identity with authentic melanoma growth stimulatory activity with respect to retention time upon reversed phase HPLC, high resolution SDS-PAGE, and biologic properties, as well as amino-terminal amino acid sequence. These data show that human dermal fibroblasts may actively participate in inflammatory reactions by secretion of proinflammatory cytokines.     

Reactive nitrogen and oxygen species attenuate interleukin- 8-induced neutrophil chemotactic activity in vitro

Peroxynitrite, formed by the reaction between nitric oxide and superoxide, has been shown to induce protein nitration, which compromises protein function. We hypothesized that peroxynitrite may regulate cytokine function during inflammation. To test this hypothesis, the neutrophil chemotactic activity (NCA) of interleukin-8 (IL-8) incubated with peroxynitrite was evaluated. Peroxynitrite attenuated IL-8 NCA in a dose-dependent manner (p < 0.01) but did not significantly reduce NCA induced by leukotriene B(4) or complement-activated serum. The reducing agents, dithionite, deferoxamine, and dithiothreitol, reversed and exogenous L-tyrosine abrogated the peroxynitrite-induced NCA inhibition. Papa-NONOate [N-(3-ammoniopropyl)-N-(n-propyl)amino]diazen-1-ium-1, 2-dialase or sodium nitroprusside, NO donors, or a combination of xanthine and xanthine oxidase to generate superoxide did not show an inhibitory effect on NCA induced by IL-8. In contrast, small amounts of SIN-1, a peroxynitrite generator, caused a concentration-dependent inhibition of NCA by IL-8. Consistent with its capacity to reduce NCA, peroxynitrite treatment reduced IL-8 binding to neutrophils. Nitrotyrosine was detected in the IL-8 incubated with peroxynitrite by enzyme-linked immunosorbent assay. These findings are consistent with nitration of tyrosine by peroxynitrite with subsequent inhibition of IL-8 binding to neutrophils and a reduction in NCA and suggest that oxidants may play an important role in regulation of IL-8-induced neutrophil chemotaxis.     

Serine protease inhibitors modulate chemotactic cytokine production by human lung fibroblasts in vitro

Chemotactic chemokines can be released from lung fibroblasts in response to interleukin (IL)-1beta and tumor necrosis factor (TNF)-alpha. An imbalance between proteases and antiproteases has been observed at inflammatory sites, and, therefore, protease inhibitors might modulate fibroblast release of chemotactic cytokines. To test this hypothesis, serine protease inhibitors (FK-706, alpha(1)-antitrypsin, or N(alpha)-p-tosyl-L-lysine chloromethyl ketone) were evaluated for their capacity to attenuate the release of neutrophil chemotactic activity (NCA) or monocyte chemotactic activity (MCA) from human fetal lung fibroblasts (HFL-1). Similarly, the release of the chemoattractants IL-8, granulocyte colony-stimulating factor, monocyte chemoattractant protein-1, macrophage colony-stimulating factor, and granulocyte/macrophage colony-stimulating factor, from HFL-1, were evaluated in response to IL-1beta and TNF-alpha. NCA, MCA, and chemotactic cytokines were attenuated by FK-706. However, matrix metalloproteinase inhibitors were without effect, and cysteine protease inhibitors only slightly attenuated chemotactic or cytokine release. These data suggest that IL-1beta and TNF-alpha may stimulate lung fibroblasts to release NCA and MCA by a protease-dependent mechanism and that serine protease inhibitors may attenuate the release.