Regulation of Alveolar Macrophage Production of Chemoattractants by Leukotrine B4 and Prostaglandin E2 John W. Christman, Brian W. Christman, Virginia L. Shepherd, and Jean E. Rinaldo Department of Medicine, Vanderbilt University and the Veterans Administration Medical Center, Nashville, Tennessee
Alveolar macrophages (AM) appear to influence the recruitment of neutrophils into the lung by the elaboration of both lipid and peptide chemotactic molecules for neutrophils. Little is known about the mechanisms that regulate production or release of chemotactic molecules by AM or the interaction between these classes of chemotactic molecules. We investigated the hypothesis that the lipid mediator leukotriene B4 (LTB4) has an in vitro regulatory action on the production of chemotactic proteins by AM. In these experiments, the chemotactic activity in AM culture supernatants was measured in a modified Boyden chamber. LTB4 treatment increased AM production of chemotactic activity in excess of what might be attributed to the amount of LTB4 measured in the culture supernatant after the incubation period. This effect was magnified by in vivo administration of endotoxin prior to AM harvesting. Pretreatment with LTB4 caused a sustained 250 % increase in AM production of chemotactic activity, yet only negligible amounts of LTB4 were measured by gas chromatography/mass spectrometry in the Lf'Bz-pretreated AM culture supernatants, indicating that LTB4 alone did not account for the chemotactic activity observed in our studies. A chemotactic peptide in Ll'Bz-treated AM culture supernatant could be isolated and separated from LTB4 by molecular sieve chromatography. Purified column fractions contained 80 % of the chemotactic activity of endotoxin-stimulated AM culture supernatant and had a molecular mass of 10,000 D. In contrast to LTB4 , prostaglandin E, (PGEz) suppressed chemotactic activity production by endotoxin-stimulated AM by 70 %. Pretreatment with PGEz was not effective; PGEz had to be present in the AM culture medium during endotoxin exposure in order to exert a suppressive effect. The effect on chemotactic activity production appears to be specific since neither LTB4 or PGEz altered tumor necrosis factor production by AM. These data indicate that ambient levels of LTB4 and PGEz may regulate the synthesis, release, or bioactivity of certain chemotactic cytokines by AM.
The alveolar macrophage (AM) has the potential to influence the evolution of pulmonary inflammation by elaborating lipid (1-4) and protein (5, 6) chemotactic molecules for neutrophils (PMN). Chemotactic molecules cause the directed transendothelial migration of peripheral blood neutrophils into the lung interstitiurrrand alveolar space. The appearance of these factors appears to herald the onset of acute inflammation, and, conversely, the disappearance of these factors may be associated with the resolution of acute inflammation. Neutrophilic inflammation plays an essential role in defending the lung from microbial pathogens, yet excessive or inappropriate inflammation may result in injury to lung endothelial, interstitial, and epithelial cells and extracellular (Received in original form January 2, 1991 and in revisedform March 11, 1991) Address correspondence to: John William Christman, M.D., Center for Lung Research, Vanderbilt University, Nashville, TN 37232. Abbreviations: alveolar macrophage(s), AM; gas chromatography/mass spectrometry, GC/MS; interleukin, IL; lipopolysaccharide, LPS; leukotriene B4 , LTB4 ; pentaftuorobenzyl, PFB; prostaglandin Ez, PGEz; polymorphonuclear leukocyte(s), PMN; tumor necrosis factor, TNF. Am. J. Respir. Cell Mol. BioI. Vol. S. pp. 297-304, 1991
matrix. Thus, a better understanding of the mechanisms governing synthesis and release by AM of chemotactic molecules is essential to understand both pulmonary host defense and the pathogenesis of acute lung injury. A prominent AM-derived chemotactic lipid mediator is leukotriene B4 (LTB4 ) (1-4). Although the chemotactic proteins produced by rat AM have not been completely characterized, peripheral blood mononuclear leukocytes produce at least two protein factors that affect neutrophil movement, monocyte-derived neutrophil chemotactic protein (also referred to as interleukin [ILl-8) (7-13) and macrophage inflammatory protein-2 (MIP-2) (14-18). Several studies evaluating LTB4 production have used short incubation times of 1 to 4 h (1-4), while experiments that have suggested the presence of chemotactic proteins have more frequently employed longer incubation times of6 to 24 h (5-18). Given this time course, it is reasonable that LTB4 might feed back and interact with the AM to promote chemotactic protein release. Furthermore, if the general hypothesis that LTB4 increases the production of chemotactic proteins is correct, homeostasis would propose that such a stimulating effect would be modulated by the effect of other lipid mediators. Prostaglandin E, (PGEz) seems to be a likely candidate for
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 5 1991
such a regulatory effect because PGE 2 has been shown in several studies to downregulate the induction of other monocyte/macrophage-derived cytokines such as tumor necrosis factor (TNF) (19, 20) and IL-l (21, 22). However, PGE 2 has not been shown to inhibit the release of chemotactic proteins by AM. Thus, in these experiments, we addressed two hypotheses: (1) LTB4 upregulates or stimulates the production of chemotactic proteins by AM; (2) this stimulatory effect is opposed by the action of PGE2 •
Material and Methods Source of AM Male Sprague-Dawley specific pathogen-free rats weighing 200 to 280 g were used exclusively for these experiments. After light pentobarbital anesthesia, the chest and abdomen were opened and the rats were exsanguinated by right ventricular puncture. The trachea was exposed and cannulated with a 16 gauge teflon catheter that was tied into place with 2-0 silk suture. The lungs were then excised and lavaged extracorporally with 6-rnl aliquots of pyrogen-free normal saline until a total return of 45 ml of lavage fluid was collected. The lungs were gently massaged with each infusate in order to maximize the cell yield. The total cell yield ranged from 106 to 107 cells/rat. Cell Preparation and Production of AM Culture Supernatant The lavage fluid was centrifuged at 400 X g at room temperature for 10 min, and the supernatant was decanted and discarded. The cell pellets from four to eight rats were pooled and suspended in RPMI-1640 culture medium (M.A. Bioproducts, Grand Island, NY) with 10% fetal bovine serum, 20 mM Hepes buffer, L-glutamine (2.9 mglrnl), penicillin (200 JLg/rnl) , and streptomycin (100 JLg/ml). Total cell count was determined on a grid hemocytometer. Cell viability was monitored by counting the percentage of cells that excluded 0.4 % trypan blue dye (OrnCO, Chagrin Falls, OR). A differential cell count was determined on a cytocentrifuge slide preparation (Cytospin, Shandon Southern Products, Runcom, Cheshire, UK) stained with a modified Wright stain (Diff-Quik'"; Dade Diagnostic, Aquada, PR). Four hundred consecutive cells were judged as being AM, lymphocytes, PMN, or eosinophils under oil immersion (l,OOOX). A cell suspension containing 1.2 X 106 viable AM was plated in 35-mm culture, six-cluster wells (Falcon, Oxnard, CA) and incubated at 37° C in 5 % CO 2 for 60 min. The wells were then vigorously washed 3 times with serum-free RPMI to remove the nonadherent cells. The adherent cells were incubated in 1.2 rnl of serum-free fresh RPMI at 37° C in 5 % CO 2 for 15 h. After incubation, the culture supernatant was decanted, centrifuged at 400 X g, and stored at - 20° C until assayed for chemotactic activity. Determination of the Chemoattractant Activity Rat peritoneal PMN were used as the responding cells in the chemotactic assay. PMN were obtained by peritoneal lavage 4 h after the intraperitoneal instillation of 10 rnl of thioglycolate broth culture medium without indicators (Difco Laboratories, Detroit, MI). The responding cells were suspended in Gey's balanced salt solution (GIBCO) with 1% bovine se-
rum albumin (Sigma Chemical Co., S1. Louis, MO) in a concentration of 0.8 X 106 viable PMN/rnl. Chemotactic assays were performed utilizing a multi-well microchamber as previously described (23). Responding cells were placed in the upper chamber, and the potential chemoattractant was placed in the lower chamber separated by polycarbonate filters with 3-JLm pore diameters. Bovine serum albumin was added to all test media to give a final protein concentration of 1%. Cell migration took place during a 30-min incubation period in a humidified incubator at 37° C. After incubation, the filters were fixed and stained with Diff-Quik. The cells that had migrated through the filter were counted by oil immersion (l,OOOX) light microscopy. The chemotactic activity was determined by counting the PMN that migrated toward the tested chemoattractant in 10 oil immersion light microscopic fields. Each individual value represents the average of triplicate measurements. The chemotactic activity is reported as the mean ± SEM of three to five independent experiments. Raw data are expressed as the number of rat peritoneal PMN that migrated toward the tested chemoattractant as enumerated by counting 10 oil immersion light microscopic fields. In order to control for interassay variability, we have normalized some of the data as the measured value divided by the positive control. The background chemotactic response (negative control) of fresh RPMI medium was subtracted from both the measured and positive control values. For the LTB4 experiments, the positive control was unstimulated AM culture supernatant. Thus, a stimulatory effect is represented by values greater than 100%. For the PG~ experiments, the positive control was endotoxin-stimulated AM culture supernatant and the negative control was fresh RPMI that contained 30 JLg/ml of endotoxin. Thus, for the PG~ experiments, values less than 100% represent a suppression of endotoxin-stimulated AM production of chemotactic activity. Measurement of TNF TNF bioactivity was quantitated in the cell-free supernatant. Bioactivity of TNF was measured using a semiautomated L-929 fibroblast lytic assay. Briefly, L-929 cells (5 X 104/100 JLI) were cultured in 96-well, flat-bottom microtiter plates overnight at 37° C, 5% CO 2/95 % humidified air in the presence of 15 I-tg/ml of actinomycin D with a 1:2 dilution of the test sample. After incubation, plates were washed with normal saline, and the remaining cells were stained with 0.1% crystal violet (in 100% methanol). The quantity of cell lysis was determined using a microElisa plate reader (read at optical density of 540 nm). Units of TNF activity were defined by interpolating points on a curve constructed with different concentrations of human recombinant TNF (Cetus Corp., Emeryville, CA). The standard curve was linear over a range of 100 to 10,000 Vim!. The reported data were calculated as the mean ± SEM (n = 3) of the percentage of control for each group of experiments. The positive control for the LTB4 experiments is unstimulated AM culture supernatant. A stimulatory effect is represented by values greater than 100 %. The positive control for the PGE 2 experiments is endotoxin-stimulated AM culture supernatant. For these experiments, values less than 100 % represents a suppression on endotoxin-stimulated AM production of TNF activity.
Christman, Christman, Shepherd et al.: Regulation of Chemoattractants by Eicosanoids
Source and Preparation of LTB4 and PGEz LTB4 and PGE2 were a generous gift of the Upjohn Co. (Kalamazoo, MI). Lyophilized PGEz was reconstituted in 95 % ethanol and diluted to a stock solution concentration of 10-3 M with 0.9% NaCl, which was divided into 50-,u1 aliquots and frozen at -70 0 C under nitrogen. In a similar fashion, a stock solution concentration of 10-4 M LTB4 was prepared, placed in 50-,ulaliquots, and frozen under nitrogen at -70 0 C. Stimulation with LTB4 During the 15- to l8-h incubation, the unstimulated AM were exposed to 10-1l to 10-6 M concentrations of LTB4 • In selected samples, LTB4 concentrations were measured by gas chromatography/mass spectrometry (GC/MS). In order to test the effect of prior treatment, the LTB4 was added for 2 and 4 h prior to the incubation period. After this "pretreatment," the adherent cells were vigorously washed with three l-ml aliquots of RPMI medium. We have previously reported that AM from endotoxin-treated rats are primed to produce chemotactic proteins when stimulated in vitro with endotoxin (23). This observation was exploited in these experiments by comparing the effect of LTB4 on the production of chemotactic activity by AM from control versus endotoxintreated rats. In these experiments, 5.0 mg/kg of Escherichia coli endotoxin was administered by the intraperitoneal route 15 h prior to whole lung lavage. At this time point, there are less than 5 % PMN in the lavage fluid which are 80 % removed by selective adherence properties. Suppression with PGE 2 AM can be stimulated to produce near maximal production of chemotactic activity by incubation with 30 ,ug/ml of E. coli endotoxin (23). Endotoxin has no intrinsic chemotactic activity and does not interfere with our biologic assay. We therefore used endotoxin as an in vitro pharmacologic stimulus for chemotactic activity production. For the purposes of this research, we attempted to suppress endotoxin-stimulated chemotactic activity production by exposing the endotoxin-stimulated macrophages to PG~ in concentrations ranging from 10-1D to 10-5 M. We also tested the effect of 2 and 4 h of "pretreatment" with PGE z in identical manners as described above. In these experiments, the PGE ztreated AM culture supernatant was tested for chemotactic activity. These experiments were controlled by testing the chemotactic activity of RPMI culture medium to which lipopolysaccharide (LPS) and PGE z were added as a negative control and LPS-stimulated AM culture medium to which PGE z was added after the incubation period as a positive control. This method allows us to assess the intrinsic positive and negative effects of PGE z and LPS on PMN migration in this in vitro bioassay. Measurement of LTB4 LTB4 concentrations were quantified by a stable isotope dilution method in conjunction with GC/MS as previously described (24). Five nanograms of tetradeuterated LTB4 was added to each sample as an internal standard. Samples were stored at - 20? C until analyzed. Protein was precipitated by the addition of cold acetone. After increasing the aqueous volume and acidification, samples were extracted with a
reverse-phase silica cartridge (C-18 PrepSep; Waters Associates, Milford, MA), rinsed sequentially with water and hexane, and eluted with ethyl acetate. After evaporation, samples were purified on LK6D silica thin layer chromatography plates (Waters Associates) employing a mobile phase of ethyl acetate:acetic acid:hexane:water (50:10:50:50, vol/vol/vol/ vol). Samples were eluted with ethyl acetate, evaporated, and converted to the pentafluorobenzyl (PFB) ester by incubation with 12.5% PFB bromide in acetonitrile with a catalytic amount of diisopropylethylamine. The resultant PFB derivatives were purified by thin layer chromatography with a solvent system of hexane:ethyl acetate (1:1 vol/vol). After elution and evaporation, free hydroxyl groups were converted to the corresponding trimethylsilyl ethers by treatment with N,O-bis(trimethylsilyl)trifluoroacetamide and pyridine. Samples were dried and dissolved in decane. Gas chromatography was performed on a 7.5-m SPB-l fused silica capillary column (Supelco, Folsom, CA) in a Varian Vista 6000 instrument (Sunnyvale, CA) linked to a Nermag RI010C mass spectrometer (Fairfield, NJ). Injections were made in the splitless mode with helium as a carrier gas. The temperature program ran from 170 to 280 0 C at 20° Cimino Selected ions monitored were 479 and 483 for LTB4 and ~- LTB4 , respectively. Standard curves were constructed by combining various concentrations of LTB4 in the range of 0 to 5,000 pg with 5 ng of ~-LTB4' Each standard was analyzed concurrently with the samples and subjected to identical purification and derivatization procedures. Data reduction and transformation were accomplished with a SIDAR data system. Quantification was performed by interpolation/extrapolation from the peak height ratios m/z 479/483 for the points of the standard curve using the equation resulting from standard linear regression analysis (r> 0.98). Gel Chromatography AM culture supernatant (80 rnl) was concentrated 160-fold by dialysis against polyvinylpyrrolidone (Sigma) using a dialysis membrane with a molecular weight exclusion limit of 1,000 D (Spectrum Medical Industries, Los Angeles, CA). One-half milliliter of concentrated supernatant was applied to a 100 x l-cm Sephacryl S-200 column, and eighty l-ml fractions were collected. The chemotactic activity of each elution fraction was tested at 1:1, 1:2, and 1:10dilutions. The data presented represent the chemotactic activity of the 1:2 dilution, since the background chemotactic activity was minimal and the intensity of the chemotactic peak was equal to or greater than 1:1 and 1:10 dilutions. The column was precalibrated with molecules of known molecular weight, and the location of these markers is noted in Figures 5 and 6. The column was also calibrated with LTB4 • This was performed by adding 10 ,ug of LTB4 to the column in a volume of 0.5 rnl and testing the elution fractions for chemotactic activity. The elution volume of LTB4 is also noted in Figures 5 and 6.
Results We first tested the hypothesis that in vitro treatment of AM with LTB4 would induce the production of non-LTB4 chemotactic activity. The percent increase in chemotactic activity production by AM which results from in vitro treatment
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 5 1991
Added versus measured LTB 4 concentrations in LlBs-treated AM culture supernatant*
o u 0
Added [LTB 4 J
10- 9 10- 8 10- 7 10- 6
c Q) lJ)
u ec Q) u
10- 1 1 10- 10
[LTB 4] added to Macrophage Culture (Molar)
Figure 1. The effect of leukotriene B4 (LTB4) treatment of alveolar macrophage (AM) production of chemotactic activity. Values represent the percent increase compared to the chemotactic activity of unstimulated AM culture supernatant. The data represent mean ± SEM (n = 4). Asterisks indicate significantly different from control by Student's t test (P < 0.05).
with LTB4 is shown in Figure 1. The chemotactic activity of AM culture supernatant was significantly increased by treatment with LTB4 in concentrations that ranged from 10-8 M (3.36 ng/ml) to 10-6 M (336 ng/ml). LTB4 was present in the AM culture supernatant and could have contributed to the chemotactic activity. To exclude the possibility that LTB4 itself was the chemotactic moiety, we compared the dose-response relationship shown in Figure 1 with a dose-response relationship for the chemotactic activity of LTB4 itself (Figure 2). As seen, LTB4 is significantly chemotactic compared to RPMI culture medium in concentrations that range from 10-7 M (33.6 nglml) to 10-6 M (336 nglml). Thus, we observed only a log difference between the concentrations of LTB4 which stimulated the production of chemotactic activity by culture AM and that concentration of LTB4 which was intrinsically chemotactic. We then tested the hypothesis that the LTB4 which we initially added was taken up or metabolized during the 15- to 18-h incubation period. We did this by measuring the concentrations of LTB4 in the LI'Bi-treated AM culture supernatant
The effect of PGEz and LTB 4 on AM production of TNF*