Modulation of Macrophage Phenotype by Soluble Product(s) Released from Neutrophils
http://www.100md.com
免疫学杂志 2005年第4期
Abstract
The regulation of macrophage phenotype by neutrophils was studied in the s.c. polyvinyl alcohol sponge wound model in mice made neutropenic by anti-Gr-1 Ab, as well as in cell culture. Wounds in neutropenic mice contained 100-fold fewer neutrophils than those in nonneutropenic controls 1 day after sponge implantation. Wound fluids from neutropenic mice contained 68% more TNF-, 168% more IL-6, and 61% less TGF-1 than those from controls. Wound fluid IL-10 was not different between the two groups, and IL-4 was not detected. Intracellular TNF- staining was greater in cells isolated from neutropenic wounds than in those from control wounds. The hypothesis that wound neutrophil products modulate macrophage phenotype was tested in Transwell cocultures of LPS-stimulated J774A.1 macrophages and day 1 wound cells (84% neutrophils/15% macrophages). Overnight cocultures accumulated 60% less TNF- and IL-6 than cultures of J774A.1 alone. The suppression of cytokine release was mediated by a soluble factor(s), because culture supernatants from wound cells inhibited TNF- and IL-6 release from LPS-stimulated J774A.1 cells. Culture supernatants from purified wound neutrophils equally suppressed TNF- release from LPS-stimulated J774A.1 cells. Wound cell supernatants also suppressed TNF- and superoxide release from murine peritoneal macrophages. The TNF- inhibitory factor has a molecular mass <3000 Da and is neither PGE2 nor adenosine. The present findings confirm a role for neutrophils in the regulation of innate immune responses through modulation of macrophage phenotype.
Introduction
Neutrophils are the first blood-borne nucleated cells to infiltrate sites of injury or infection, where they produce multiple effector molecules, including reactive oxygen and nitrogen intermediates, proteolytic enzymes, and cytokines. The anti-infectious capacity of neutrophils provides obvious functional advantage to their recruitment into infected tissue. Neutrophils, however, also accumulate in large numbers at sites of sterile inflammation, such as uninfected wounds, myocardial and cerebral infarctions, and fractures where, in the absence of infectious challenge, their role in the inflammatory response is less readily apparent.
Experiments reported in this study were initially designed to define the contribution of neutrophils to the cytokine profile of an acute sterile inflammatory site, such as that provided by the polyvinyl alcohol (PVA)3 sponge wound model. The findings demonstrated that neutrophil depletion results in increased proinflammatory cytokine concentrations in early tissue inflammation. This finding suggested that neutrophils produce a factor(s) that inhibits the release of proinflammatory cytokines by macrophages. The results supported and extended the predictions of this hypothesis by demonstrating that wound neutrophils produce a soluble factor(s) that suppresses LPS-mediated activation of a macrophage cell line and primary peritoneal macrophages. The present results thus reveal an additional capacity of neutrophils in inflammation, namely, the ability of these cells to modulate macrophage inflammatory phenotype.
Materials and Methods
Animals
B6D2F1 male mice (Taconic Farms), 8–12 wk of age, were housed at the Central Research Facilities of Rhode Island Hospital and fed mouse chow and water ad libitum. Mice were certified free of common pathogens by the supplier and were monitored by Brown University/Rhode Island Hospital veterinary personnel. Animal protocols were approved by the animal care committee at Rhode Island Hospital.
PVA sponge wound model
Mice were anesthetized with pentobarbital (50 mg/kg i.p.; Abbott Laboratories). Five PVA sponges (PVA Unlimited), measuring 1 x 1 x 0.6 cm, were inserted into individual s.c. pockets through a midline dorsal incision under sterile conditions, and the skin was closed with clips (1). Mice were euthanized by CO2 asphyxiation, and the sponges were removed under sterile conditions. Culture of randomly selected PVA sponges at the time of removal confirmed sterility in this experimental model. Wound cells were isolated by rapid repeated compression of sponges in a Stomacher (Tekmar). When necessary, cells were subjected to hypotonic lysis of erythrocytes, followed by reconstitution in PBS (Invitrogen Life Technologies). Viable cells were identified by trypan blue exclusion and were counted using a hemocytometer; viability was >95%. Differential cell counts were performed on Hema-3 (Biochemical Sciences)-stained Cytospins (Shandon), and cell phenotype was confirmed by flow cytometry, as described below. Extracellular fluid from the PVA sponges (wound fluid) was obtained by centrifugation (400 x g for 5 min) of two or three sponges in the barrel of a syringe that was seated in a sterile tube. Wound fluid was stored at –80°C until analysis.
Induction of neutropenia
Anti-mouse Gr-1 mAb (RB6.8C5 hybridoma originally produced by R. L. Coffman, DNAX Research Institute) was produced under serum-free conditions using a bioreactor (Cell Pharm Micro Mouse; UniSyn Technologies). Mice were rendered neutropenic by a single i.p. injection of 0.5 mg of anti-Gr-1 Ab 3 days before PVA sponge insertion. Control animals were injected i.p. with an equal volume of normal saline or with 0.5 mg of rat IgG (Rockland Immunochemicals).
Circulating leukocyte counts
Heparinized blood was obtained from the lateral tail vein or the inferior vena cava. Leukocyte count was determined using a hemocytometer after dilution and erythrocyte lysis with 0.01% crystal violet in 3% acetic acid. Differential cell counts were performed on Hema-3 stained blood smears.
Purification of neutrophils from the wound cell suspension
Wound cells were isolated 1 day after PVA sponge insertion, suspended in biotin-free FcR-blocking agent (Accurate Chemical & Scientific) at 0.3 ml/106 cells, washed, and incubated with biotin-conjugated Ab directed against the macrophage cell surface Ag F4/80 (1 μg/106 cells; Caltag Laboratories). Cells were washed in PBS and incubated with MACS Anti-Biotin MicroBeads (Miltenyi Biotec) according to the manufacturer’s recommendations, and macrophages were depleted using the MiniMACS magnetic cell separator (Miltenyi Biotec). The resulting cell suspension was 96% neutrophils by light microscopy.
Cell culture
All cell cultures were performed in complete medium (CM; RPMI 1640 (Invitrogen Life Technologies) supplemented with 10 mM MOPS, 5 x 10–5 M 2-ME, 100 U/ml penicillin-streptomycin, and 1% FBS (HyClone)) unless noted otherwise. The murine macrophage cell line J774A.1 was obtained from American Type Culture Collection. Murine resident peritoneal macrophages were obtained by peritoneal lavage with HBSS and 1% FBS. Thioglycolate-elicited peritoneal macrophages were harvested 4 days after i.p. injection of 3 ml of 3% Brewer modified thioglycolate medium (BD Biosciences). J774A.1 macrophages or peritoneal macrophages were allowed to adhere to tissue culture-treated plastic for 2 h, and nonadherent cells were removed before initiation of experiments.
Three culture formats were used for coculture experiments: 1) J774A.1 cells, 5 x 105 in 1 ml of CM; 2) wound cells, 2.5 x 106 in 1 ml of CM; and 3) cocultures in Transwell plates (0.4-μm pore size; Costar) containing 5 x 105 J774A.1 cells in the lower chamber and 2.5 x 106 wound cells in the upper chamber in a total volume of 1 ml of CM. LPS (Escherichia coli serotype 055:B5, Sigma-Aldrich; 0.1 μg/ml) was added 1 h after the initiation of all cultures. Where noted, indomethacin (Sigma-Aldrich) or NS-398 (Cayman Chemicals) was added at the start of cell incubation.
J774A.1 cells or peritoneal macrophages were also cultured with supernatants conditioned by day 1 wound cells. The wound cell-conditioned supernatant was generated by isolating wound cells from PVA sponges 1 day after insertion and culturing these cells in CM at 2.5 x 106/ml for 24 h. Cells were then removed from the medium by centrifugation, and the supernatant was stored at –80°C. Supernatants conditioned by purified wound neutrophils were generated in similar manner. J774A.1 cells (1 x 105) or primary peritoneal macrophages (3 x 105) were incubated in 200 μl of CM or a 1/1 (v/v) dilution of wound cell-conditioned supernatant in CM for 1 h before addition of LPS, then cultured for an additional 24 h. Wound cell-conditioned supernatant was fractionated using Amicon Centriplus YM-3 centrifugal filter devices (Millipore). When indicated, wound cell-conditioned supernatant or an equal volume of CM was treated with adenosine deaminase (type X from bovine spleen; Sigma-Aldrich) at 0.02 U/ml for 20 min at 25°C, and adenosine deaminase was removed, or not, using a Centricon-10 concentrator (Millipore) before addition of the treated supernatant to cell cultures. Indomethacin, NS-398, AH6809 (Cayman Chemicals), L161982 (gift from Dr. R. Young, Merck Frosst Canada), or 8 phenyl-theophylline (Sigma-Aldrich) was added to cell cultures where noted.
Conditioned culture supernatants were also prepared from splenic lymphocytes. Splenocytes were obtained by mechanical dissociation of spleens and elimination of fibrous tissue using wire mesh. Erythrocytes were removed by hypotonic lysis; macrophages and dendritic cells were depleted by adherence to plastic for 2 h. Supernatants were prepared from the resulting cell suspension (90% lymphocytes, 4% monocytes, and 4% neutrophils) by culture at 2.5 x 106/ml for 24 h. Conditioned supernatants were also prepared from murine L929 fibroblasts and from fibroblasts obtained from a murine wound by culture at a density of 5 x 105/ml for 24 h. Cells were removed by centrifugation, and the supernatants were used as described above.
Macrophage viability was assessed after 24 h of incubation by trypan blue exclusion, reduction of MTT (Sigma-Aldrich) (2), and release of lactate dehydrogenase (3). Protein synthesis was determined by measuring the 24-h incorporation of L-[ring-2,6-N-3H]phenylalanine into TCA-precipitable protein. When indicated, cells were lysed in buffer containing 0.01 M Tris, 0.15 M NaCl, and 0.5% Igepal CA 630 (Sigma-Aldrich) with Complete (Roche) protease inhibitor.
Cell staining and flow cytometry
Wound cells were isolated from neutropenic or control animals 1 day after sponge insertion and pooled. Abs used for cell staining and flow cytometry were obtained from BD Biosciences unless noted otherwise.
Surface Ag staining. Erythrocytes were removed from the wound cell suspension by hypotonic lysis. Cells were incubated with FcR-blocking agent (Accurate Chemical & Scientific), washed in PBS with 1% FBS, and incubated with predetermined optimal concentrations of fluorochrome-conjugated Abs to B220 and IgM (for B cells), CD3 (for T cells), or the appropriate isotype control Abs.
Intracellular Ag staining. Erythrocytes were removed from the wound cell suspension by hypotonic lysis. Cells were incubated with FcR-blocking agent (Accurate Chemical & Scientific), washed in PBS with 1% FBS, fixed in paraformaldehyde with saponin (Cytofix/Cytoperm; BD Biosciences) according to the manufacturer’s recommendations, washed in Permwash (BD Biosciences), and incubated with a predetermined optimal concentration of fluorochrome-conjugated Ab directed against the intracellular macrophage Ag macrosialin (anti-CD68; Serotec) or the appropriate isotype control Ab.
Intracellular cytokine staining. Cells were incubated (106 cells/ml) in CM for 6 h with brefeldin A (Golgiplug; BD Biosciences) at 1 μl/106 cells and stained for intracellular Ag as described above, using anti-CD68, anti-TNF-, and/or the appropriate isotype control Abs.
Cells were analyzed and sorted using a FACSort and CellQuest software (BD Biosciences). Cells were stained with propidium iodide after sorting, and nuclear morphology was determined by light microscopy.
Assays
TNF- (BioSource), IL-6 (BD Biosciences), TGF-1 (R&D Systems), IL-4, and IL-10 (4) were assayed by ELISA. PGE2 was quantified by enzyme immunoassay (Cayman Chemical). ATP, ADP, AMP, and adenosine were assayed by HPLC (5) with a limit of detection of 1 μM. TNF- bioactivity was determined by lysis of actinomycin D (Sigma-Aldrich)-treated L929 murine fibroblasts (NCTC clone 929; American Type Culture Collection) (6), using a recombinant murine TNF- standard (R&D Systems). Superoxide release was measured from the reduction of ferricytochrome c (Sigma-Aldrich) after cell stimulation with 200 nM PMA (LC Services) over 90 min (7).
Data presentation and analysis
Data reported are the mean ± 1 SD unless otherwise noted. Results from cell culture experiments are from triplicate wells in a representative experiment. Statistical analysis was performed using Student’s unpaired t test, Mann-Whitney U test, or ANOVA with Newman-Keuls test, as appropriate.
Results
Characterization of the cellular infiltrate in the PVA sponge wound model
The leukocytic infiltration of the wound increased over the first 7 days after PVA sponge insertion (Fig. 1A). Day 1 wound cells were 84 ± 4% neutrophils, 15 ± 4% macrophages, <1% lymphocytes, and <1% eosinophils in six experiments with at least five animals per experiment (mean ± SEM). Neutrophils remained the most abundant cell type through at least 3 days after wounding (Fig. 1B). Macrophages constituted 52% of the cells isolated from the wound 7 days after wounding. All subsequent experiments using wound cells or wound fluids were performed using cells or fluids isolated 1 day after PVA sponge insertion.
FIGURE 1. Characterization of the cellular infiltrate in the PVA sponge wound. PVA sponges were inserted as described in Materials and Methods, and six animals each were euthanized at 1, 3, and 7 days after sponge insertion. Wound cells were isolated, identified, and counted as described in Materials and Methods. A, Wound cell count per animal. Counts differed at all time points (p < 0.05, by ANOVA and Newman-Keuls test). B, Differential cell count. , neutrophils; , macrophages.
Effect of anti-Gr-1 Ab on circulating and wound leukocyte counts
Intraperitoneal injection of 0.5 mg of mAb directed against the neutrophil surface Ag Gr-1 reduced neutrophils from 13.3 ± 1.5 to 1.4 ± 1.0% of the total circulating leukocytes. Neutropenia persisted throughout the time course of these experiments (4 days). Differential blood lymphocyte and monocyte counts were not altered by Ab treatment (data not shown).
The effect of anti-Gr-1 Ab on leukocyte infiltration of the wounds was quite marked: wounds from neutropenic animals contained 67-fold fewer total cells than controls, with 100-fold less neutrophils and 10-fold less macrophages (Fig. 2).
FIGURE 2. Anti-Gr-1 Ab depletes leukocyte populations in the PVA sponge wound. Mice were injected i.p. with 0.5 mg of anti-Gr-1 Ab or an equal volume of saline, and PVA sponges were inserted into control (n = 6) and neutropenic (n = 7) mice 3 days later. Wound cells were isolated 1 day after sponge insertion. The graph shows cell counts per animal () and group means (–). Neutropenic wounds had lower total cell, neutrophil, and macrophage counts than control wounds (p < 0.05, by Mann-Whitney U test).
Neutrophil depletion increases proinflammatory cytokines in wound fluids
Table I shows the cytokine content of wound fluids from control and neutropenic animals. Despite the decrease in the inflammatory cell infiltrate, immunoreactive TNF- and IL-6 concentrations in wound fluid from neutropenic animals were 68 and 168% higher, respectively, than those from controls, whereas TGF-1 concentrations were 61% lower. Wound fluid IL-10 concentrations were not different between the two groups. IL-4 was below the limit of detection of the assay (10 pg/ml) in wound fluids from either group.
Table I. Cytokine concentration in wound fluids: neutrophil depletion selectively increases proinflammatory cytokinesa
The increases in proinflammatory cytokines in wound fluids from neutropenic animals were independent of plasma cytokines. Plasma TNF- concentrations at the time of sponge removal were below the limit of detection of the assay (30 pg/ml) in both groups. There was no difference in the plasma concentrations of IL-6 (control, 59 ± 21 pg/ml; neutropenic, 48 ± 14 pg/ml; p > 0.05, by Mann-Whitney U test). Neutropenic animals did not exhibit obvious differences in the appearance of the surgical site or in postoperative behavior compared with wounded control animals.
Cells from neutropenic wounds contain more TNF- than those from control wounds
To determine the source of increased TNF- in the wound fluids of neutropenic animals, cells were isolated from control and neutropenic wounds, cultured with brefeldin to inhibit intracellular cytokine transport, then stained for intracellular TNF- and the intracellular macrophage Ag CD68. The validity of CD68 staining for the identification of wound macrophages was confirmed by microscopic examination of the cells after FACS. In control animals, CD68high cells were 98.5% macrophages/mononuclear cells and 1.5% polymorphonuclear leukocytes; CD68–/low cells were 95.0% polymorphonuclear leukocytes (including bands) and 4.0% macrophages/mononuclear cells. As will be discussed below, in neutropenic animals the CD68–/low cells included a greater proportion of mononuclear cells with morphologic characteristics of monocytes, macrophages, and lymphocytes than those in controls.
The intensity of staining for TNF- was greater in CD68high cells than in CD68–/low cells in both neutropenic and control animals (Fig. 3). Macrophages from neutropenic wounds stained more intensely for TNF- than those from controls. The blood leukocytes of control and neutropenic animals did not differ in intracellular TNF- staining (data not shown).
FIGURE 3. Intracellular TNF- staining in cells from control and neutropenic wounds. Mice were injected i.p. with 0.5 mg of anti-Gr-1 Ab or saline, and PVA sponges were inserted 3 days later. Wound cells isolated 1 day after sponge insertion were cultured with brefeldin A and stained for intracellular TNF- and macrophage intracellular Ag CD68 as described in Materials and Methods. MCF, mean channel fluorescence of the quadrant.
The CD68–/low cells from neutropenic wounds contained more TNF- than those from control wounds. These CD68–/lowTNF-+ cells comprised 30% of the total cells from neutropenic animals. Morphologic examination of these cells after FACS showed 59% polymorphonuclear leukocytes and 41% mononuclear cells. In control animals, the CD68–/lowTNF-+ population of cells constituted only 6.7% of the wound cell suspension and was comprised almost entirely of polymorphonuclear leukocytes (96%), with the remaining 4% being mononuclear cells. The absolute number of these CD68–/lowTNF-+ mononuclear cells, as calculated from the respective total cellularity of control vs neutropenic wounds, was actually higher in controls (4800/animal) than in neutropenic wounds (2100/animal). These cells probably represent extravasated monocytes, found in the laboratory to be CD68–/low (data not shown). In this regard it has been shown that CD68 Ag expression is enhanced during macrophage recruitment (8).
Wound cells inhibit TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages
The hypothesis that wound cells, specifically wound neutrophils, suppress cytokine release by macrophages in vivo was modeled and tested in vitro. The murine macrophage cell line J774A.1 used in these experiments was confirmed in the laboratory to exhibit minimal release of TNF- in unstimulated culture and submaximal cytokine release when stimulated with 0.1 μg/ml LPS. J774A.1 macrophages were cultured in Transwells with or without day 1 wound cells (86% neutrophils) from nonneutropenic animals. Fig. 4A shows that supernatants from J774A.1/wound cell cocultures contained 60% less immunoreactive TNF- than those from cultures of J774A.1 cells alone. Results obtained by ELISA for TNF- were confirmed in a bioassay (Fig. 4B). Similarly, supernatants from J774A.1/wound cell cocultures contained 60% less IL-6 than cultures of J774A.1 cells (Fig. 4C). Comparison of the kinetics of TNF- release in cultures of LPS-stimulated J774A.1 cells vs those in J774A.1/wound cell cocultures showed that suppression of TNF- release in cocultures was not evident until after 8 h of incubation (Fig. 5).
FIGURE 4. Wound cells inhibit TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages. Wound cells were isolated from nonneutropenic mice 1 day after sponge insertion. Wound cells (2.5 x 106) or J774A.1 cells (J774; 5 x 105) were cultured in 1 ml of CM. Transwell cocultures (J774 and wound cells) contained 5 x 105 J774A.1 cells in the lower chamber and 2.5 x 106 wound cells in the upper chamber in a total volume of 1 ml. Cells were cultured with LPS (0.1 μg/ml) for 24 h. Culture supernatants were assayed for TNF- by ELISA (A), for TNF- bioactivity by lysis of actinomycin-treated L929 cells (B), or for IL-6 by ELISA (C). TNF- and IL-6 concentrations were lower in cocultures than in J774A.1 cells cultured alone (p < 0.05, by Student’s unpaired t test).
FIGURE 5. Kinetics of TNF- release in J774A.1/wound cell cocultures stimulated with LPS. Transwell cocultures of J774A.1 cells and day 1 wound cells were performed as described in Fig. 4. TNF- concentrations were lower in cocultures than in J774A.1 cells cultured alone at 20 and 24 h (p < 0.001, by two-factor ANOVA).
Culture supernatants conditioned by day 1 wound cells inhibit TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages
LPS-stimulated J774A.1 macrophages were also cultured in supernatants previously conditioned by day 1 wound cells. As shown in Fig. 6, J774A.1 cultures containing wound cell-conditioned supernatants accumulated 56% less TNF- and 60% less IL-6 at 24 h than those in CM. The kinetics of TNF- release differed from those in J774A.1/wound cell cocultures, because the inhibition of TNF- release by wound cell supernatants was evident as early as 1 h after stimulation with LPS (Fig. 7). The suppression of TNF- release was not due to cellular accumulation of the cytokine during culture, because lysates from J774A.1 cells incubated with wound cell-conditioned supernatants contained less TNF- than those in CM (J774A.1 in supernatants, 133 ± 6 pg/106 cells; J774A.1 in CM, 147 ± 6 pg/106 cells; p < 0.05, by Student’s unpaired t test.). IL-6 was not assayed in these samples because there is no cell-associated form of the cytokine. The suppression of macrophage TNF- release by wound cell supernatants was also evident in the absence of LPS (J774A.1 cells in CM, 2.1 ± 0.2 ng/ml; J774A.1 in wound cell supernatants, 1.6 ± 0.1 ng/ml; p < 0.05, by Student’s unpaired t test.).
FIGURE 6. Culture supernatants from wound cells inhibit TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages. A, Wound cells were isolated 1 day after PVA sponge insertion, and macrophages were depleted from the wound cell suspension as described in Materials and Methods. The unfractionated wound cells (86% neutrophils) or the purified wound neutrophils (96% neutrophils) were cultured for 24 h at 2.5 x 106/ml. Supernatants from these cultures were diluted 1/1 with CM and added to J774A.1 cells (5 x 105/ml) along with LPS (0.1 μg/ml). The TNF- concentration of the culture supernatant was assayed by ELISA after 24 h. J774A.1 cultures with either wound cell supernatant or wound neutrophil supernatant contained less TNF- than cultures in CM (p < 0.05, by ANOVA and Newman-Keuls test). B, J774A.1 cells were cultured with LPS (0.1 μg/ml), with or without wound cell-conditioned supernatants. IL-6 was assayed by ELISA after 24-h culture. J774A.1 cultures with wound cell-conditioned supernatants contained less IL-6 than cultures in CM (p < 0.01, by Student’s unpaired t test).
FIGURE 7. Kinetics of TNF- release in LPS-stimulated J774A.1 macrophages incubated with wound cell-conditioned supernatants. J774A.1 cells were cultured with LPS (0.1 μg/ml), with or without wound cell-conditioned supernatants, as described in Fig. 6. J774A.1 cultures with wound cell-conditioned supernatants contained less TNF- than cultures in CM at all time points (p < 0.05, by two-factor ANOVA).
To specifically assess the contribution of neutrophils to the TNF--suppressive capacity of wound cells, LPS-stimulated J774A.1 macrophages were cultured in supernatants conditioned by neutrophils purified from day 1 wound cells (96% neutrophils). The neutrophil supernatant and the wound cell supernatant equally suppressed TNF- release by J774A.1 macrophages (Fig. 6A).
Suppression of TNF- release from LPS-stimulated J774A.1 macrophages was not observed when supernatants were conditioned by splenic lymphocytes or L929 murine fibroblasts. (J774A.1 in CM, 94.2 ± 14.1 ng/ml; J774A.1 in splenic lymphocyte-conditioned supernatant, 83.5 ± 15.6 ng/ml; J774A.1 in L929-conditioned supernatant, 114.4 ± 13.8 ng/ml; p > 0.05, by ANOVA and Newman-Keuls test). Supernatants conditioned by fibroblasts isolated from a murine wound suppressed TNF- release from LPS-stimulated J774A.1 cells (J774A.1 in mouse wound-derived fibroblast-conditioned supernatant, 18.0 ± 0.8 ng/ml; p < 0.05 vs J774A.1 in CM (above), by ANOVA and Newman-Keuls test).
The suppression of proinflammatory cytokine release by wound cell-conditioned supernatants did not result from cytotoxicity or from nonspecific inhibition of protein synthesis. There was no difference between cells cultured in CM and those cultured in wound cell supernatant with respect to cell viability at the end of culture, as assessed by trypan blue exclusion, lactate dehydrogenase release, or reduction of MTT (Table II). Protein synthesis, as measured by incorporation of radiolabeled phenylalanine into protein, was identical in both culture conditions.
Table II. Wound cell conditioned supernatants are not toxic to J774A.1 macrophagesa
Wound cell-conditioned supernatants inhibit release of TNF- and production of superoxide by murine peritoneal macrophages
Wound cell-conditioned supernatants also inhibited LPS-stimulated TNF- release from thioglycolate-elicited murine peritoneal macrophages. As shown in Fig. 8A, macrophages incubated in wound cell supernatants released 71% less TNF- than those cultured in CM. Resident peritoneal macrophages cultured in wound cell supernatants also produced less superoxide after PMA stimulation than those cultured in CM at all tested LPS concentrations (Fig. 8B).
FIGURE 8. Wound cell-conditioned supernatants inhibit TNF- release and superoxide production by LPS-stimulated primary murine peritoneal macrophages. A, Thioglycolate-elicited peritoneal macrophages (3 x 105 in 200 μl) were cultured with LPS (0.1 μg/ml) with or without wound cell-conditioned supernatants, as described in Fig. 6. TNF- was assayed by ELISA after 24-h culture. Macrophages cultured with wound cell supernatants released less TNF- than those cultured in CM (p < 0.01, by Student’s unpaired t test). B, Resident peritoneal macrophages (3 x 105 in 200 μl) purified by overnight adherence were cultured with or without wound cell-conditioned supernatants at varying doses of LPS. Macrophage superoxide production was assayed after 24-h culture. Macrophages cultured with wound cell-conditioned supernatants () produced less superoxide than those in CM (; p < 0.05, by two-factor ANOVA).
Partial characterization of the neutrophil-derived factor(s) that inhibits proinflammatory cytokine release from J774A.1 macrophages
The neutrophil-derived factor(s) that inhibits proinflammatory cytokine release by LPS-stimulated J774A.1 macrophages is resistant to freezing and thawing and has a molecular mass <3000 Da. Fig. 9 shows that the <3000-Da filtrates of wound cell-conditioned supernatants inhibited TNF- release by LPS-stimulated J774A.1 macrophages similar to unfractionated supernatants. IL-6 release was also inhibited by the <3000-Da fraction of wound cell-conditioned supernatants (J774A.1 in CM, 19.5 ± 1.4 ng/ml; J774A.1 in supernatants, 11.4 ± 0.7 ng/ml; p < 0.01, by Student’s unpaired t test).
FIGURE 9. The <3000-Da fraction of wound cell supernatant inhibits TNF- release by LPS-stimulated J774A.1 macrophages. J774A.1 cells were cultured in CM () or wound cell supernatants () as described in Fig. 6. The <3000-Da fraction of supernatant or CM was also diluted 1/1 with CM and added to J774.1 cultures. All cultures were treated with LPS (0.1 μg/ml). The TNF- concentration of the J774A.1 culture supernatant was assayed by ELISA after 24-h culture. Both the unfractionated and the <3000-Da fraction of wound cell supernatants inhibited TNF- release compared with CM (p < 0.01, by two-factor ANOVA).
PGE2 (9) and adenosine (10, 11, 12, 13) have been shown to suppress macrophage TNF- production and meet the criteria of stability to freezing/thawing and molecular mass <3000 Da. The results presented in Fig. 10 show that cyclooxygenase inhibitors did not alter TNF- release in cocultures of day 1 wound cells and J774A.1 macrophages. Neither the selective cyclooxygenase-2 inhibitor NS398 (10 μM), nor the nonselective cyclooxygenase inhibitor indomethacin (10 μM) reversed the inhibition of TNF- release in the cocultures (Fig. 10A). The effectiveness of cyclooxygenase inhibitors was confirmed by measurement of PGE2 in the culture medium (Fig. 10B).
FIGURE 10. Effect of cyclooxygenase inhibitors on wound neutrophil inhibition of TNF- release by LPS-stimulated J774A.1 macrophages. J774A.1 cells were cultured alone or with day 1 wound cells in Transwells as described in Fig. 4. NS-398 (10 μM) or indomethacin (10 μM) was added to cultures 1 h before stimulation with LPS (0.1 μg/ml). TNF- (A) and PGE2 (B) concentrations in the J774A.1 culture medium were assayed after 24-h culture as described in Materials and Methods. J774A.1/wound cell cocultures accumulated less TNF- than J774A.1 cultures (p < 0.05, by two-factor ANOVA). Neither cyclooxygenase inhibitor had an effect on TNF- release.
The use of PG receptor antagonists confirmed that PGE2 is not the inhibitory factor under study. Neither the EP4 PG receptor antagonist L161982 (0.5–10 μM) nor the EP1/EP2/DP PG receptor antagonist AH6809 (0.1–10 μM) reversed the inhibition of TNF- release by LPS-stimulated J774A.1 macrophages cultured with wound cell-conditioned supernatants (data not shown).
Stimulation of adenosine receptors has been reported to inhibit the production of TNF- by LPS-stimulated macrophages. No ATP, ADP, AMP, or adenosine was detected in wound cell-conditioned supernatant by HPLC. To rule out adenosine receptor activation by concentrations of adenosine lower than the limit of detection of the assay, wound cell-conditioned supernatants were treated with adenosine deaminase before addition to J774A.1 cells. Adenosine deaminase did not reverse the inhibition of TNF- release by LPS-stimulated J774A.1 macrophages. Furthermore, the adenosine receptor antagonist, 8-phenyl theophylline (1–20 μM), had no effect on TNF- release by LPS-stimulated J774A.1 macrophages (data not shown).
Discussion
The results reported in this study provide in vivo and in vitro evidence that neutrophils modulate macrophage inflammatory phenotype through the release of a soluble mediator(s). In vivo findings in a model of acute sterile inflammation demonstrate that neutrophil depletion results in selective alterations in local cytokine concentrations that include increases in TNF- and IL-6 and reduction in TGF-1. Intracellular staining evidenced increased TNF- staining in both neutrophils and macrophages from the wounds of neutropenic animals compared with those from control animals. The observed increases in TNF- and IL-6 in the wound fluid of neutropenic animals reflect a local, rather than a systemic, effect, because plasma TNF- and IL-6 were not different in neutropenic and control animals. Moreover, intracellular staining of TNF- in circulating leukocytes revealed no differences between the groups.
The increases in TNF- and IL-6 concentrations of wound fluids from neutropenic animals were unexpected in view of the 67-fold reduction in the number of total wound leukocytes. The reciprocal changes in cell number vs proinflammatory cytokine content in the wounds of control and neutropenic animals suggested that cell density-dependent mechanisms regulate proinflammatory cytokine production at sites of inflammation, with a higher cell number correlating to decreased, rather than increased, cytokine levels. One well-described mechanism that could account for the present findings in vivo is the phagocytosis of apoptotic neutrophils and the subsequent reduction in proinflammatory cytokine release by macrophages at the inflammatory site (14, 15). Alternatively, the present observations could be explained by a soluble factor(s) released from wound cells that down-regulates local proinflammatory cytokine release through an autocrine and/or paracrine mechanism. To test for both mechanisms simultaneously, a coculture system that mimicked the cellularity of the early wound and prevented direct cell-to-cell contact was used.
The data in Fig. 4 show results from one such coculture experiment and demonstrate that day 1 wound cells from naive animals, which are predominantly neutrophils, inhibited TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages by 60%. The results depicted in Fig. 5 suggest synthesis and release of the inhibitory factor(s) into the medium during coculture, because TNF- accumulation in the medium was not suppressed until after 8 h in coculture. Inhibition of TNF- release was detectable by 1 h after the addition of conditioned supernatant to J774A.1 cells (Fig. 7), suggesting that those supernatants contain a preformed inhibitory factor(s). Fig. 6 provides evidence that neutrophils are, among wound cells, the most likely source for the soluble mediator(s) that inhibited macrophage cytokine release in cocultures, because cell culture supernatants from purified wound neutrophils suppressed TNF- release by LPS-stimulated J774A.1 macrophages to the same extent as supernatants from unfractionated wound cells. Moreover, the inhibition of proinflammatory cytokine release is not due to a nonselective inhibition of protein synthesis, because total protein synthesis was not altered by wound cell-conditioned supernatants.
The suppressive effects of wound cell culture supernatants were not limited to the J774A.1 macrophage cell line. Fig. 8A demonstrates that the supernatants also inhibited TNF- release from primary peritoneal macrophages. Moreover, the suppression of LPS-mediated macrophage activation was not restricted to the production of proinflammatory cytokines. The results shown in Fig. 8B illustrate the reduction in macrophage superoxide release that resulted from culture in wound cell-conditioned supernatant.
Results from this study demonstrate that neutrophils mediate the inhibition of proinflammatory cytokine and superoxide release from macrophages via a soluble mediator(s) of <3000 Da. Characterization of the inhibitory molecule(s) is not yet complete. Known inhibitors of macrophage TNF- release that are <3000 Da include PGE2 and adenosine. The early accumulation of PGs at inflammatory sites (16) and the suppressive effects of PGE2 on macrophage TNF- production (9) suggested that cyclooxygenase products might mediate wound neutrophil suppression of macrophage TNF- release in response to LPS. However, the data in Fig. 10 show that inhibition of PGE2 synthesis by NS 398 or indomethacin did not alter the neutrophil-mediated inhibition of TNF- release by LPS-stimulated macrophages. Also, neither the EP4 receptor antagonist L161982 nor the EP2 receptor antagonist AH6809 affected the capacity of wound cell-conditioned supernatants to inhibit TNF- release by J774A.1 macrophages. In this regard the receptor antagonists were selected because J774A.1 cells express only PG receptors of the EP4 and EP2 subtypes (17). PGE2, therefore, is not the mediator of wound cell suppression of TNF- release by LPS-stimulated macrophages. Adenosine and adenosine receptor agonists also inhibit macrophage inflammatory responses, including the release of TNF- (10, 11, 12, 13). Neither adenosine nor its metabolic precursors (ATP, ADP, and AMP) were found in wound cell-conditioned supernatants. Also, the lack of effect of adenosine deaminase or an adenosine receptor antagonist argues against involvement of the adenosine receptor in the wound cell-mediated inhibition of TNF- release by LPS-stimulated macrophages.
Supernatants from wound-derived fibroblasts, but not those from splenic lymphocytes or from L929 immortalized murine fibroblasts, also suppressed LPS-stimulated TNF- release from J774A.1 cells. Although it is not known whether the suppressive factor produced by wound fibroblasts is identical with that from wound neutrophils, it is of interest that other cell types that come into contact with macrophages within inflammatory sites are also capable of inhibiting TNF- release.
These results confirm a role for neutrophils in the regulation of proinflammatory responses by macrophages. This conclusion finds support and extension in published studies of neutrophil depletion. Rats made neutropenic with vinblastine and subjected to hemorrhagic shock manifested decreased neutrophil sequestration and increased expression of IL-6 mRNA and CD14 mRNA in the lung compared with nonneutropenic controls (18). Mice lacking the transcriptional repressor Gfi1 are severely neutropenic and have increased lethality and higher plasma levels of TNF-, IL-10, and IL-1 after LPS administration (19). Macrophages isolated from neutropenic Gfi1-deficient mice exhibit increased LPS-stimulated TNF- release. The findings of the present study suggest that neutropenia in Gfi1-deficient animals resulted in up-regulation of macrophage cytokine production. The latter conclusion is also supported by Steinshamn’s investigation (20) of the TNF- response to endotoxin in mice made neutropenic with cyclophosphamide, showing increased lethality and increased serum TNF- concentration compared with control animals after LPS stimulation. A proinflammatory bias in neutropenia has thus been shown to be independent of species, mechanism of neutropenia, and stressor imposed upon the animals.
The work reported in this study resulted in an unexpected and paradoxical finding: that neutrophils, which are considered prototypical inflammatory cells, could act as anti-inflammatory cells through suppression of the macrophage inflammatory phenotype. The present observations together with reports indicating that neutrophils modulate T cell subset selection in candidiasis (21) and dendritic cell activation in Toxoplasma gondii infection (22), shed new light on the capacity of neutrophils to regulate multiple cellular immune responses.
As stated in the introduction, the role of neutrophils in sterile inflammation has not been completely clarified. With specific regard to the role of neutrophils in the healing of wounds, the current paradigm, which is based on the work of Simpson and Ross (23), states that neutrophils are dispensable to the repair process, and that their function in injured tissue is only to contain or eliminate infection. Only recently has this paradigm been challenged. Dovi et al. (24) demonstrated accelerated re-epithelialization of skin wounds in transiently neutropenic mice, without effects on collagen content or tensile strength of the wound. These findings do not appear to apply to all models of wound healing, because re-epithelialization of oral wounds in guinea pigs was neither accelerated nor retarded in neutropenic animals (25). Thus, the role of neutrophils in repair remains undefined.
The present observations are unlikely to be limited to wounds. The early inflammatory response in the PVA sponge wound model is identical with that observed in other models of sterile tissue injury, with the early development of a neutrophil-rich cellular infiltrate that is later replaced by a macrophage-predominant infiltrate. The suppression of the macrophage proinflammatory phenotype by neutrophil products is probably common to all acute sterile inflammatory processes. Characterization of a neutrophil-derived soluble mediator(s) that suppresses macrophage proinflammatory phenotype as well as their mechanism of action may provide opportunities for therapeutic intervention to suppress macrophage inflammatory phenotype, especially during spontaneous or therapy-induced neutropenia.
Acknowledgments
We thank Jill Rose for assistance in preparing the manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was funded by National Institutes of Health Grants GM42859 (to J.E.A.) and GM66194 (to J.S.R.) and by funds allocated to the Department of Surgery by Rhode Island Hospital, a Lifespan partner. J.M.D. was supported by a National Institutes of Health Supplement to Promote Reentry into Biomedical and Behavioral Research Careers. E.J.M. was supported by the Carter Family Charitable Trust (Armand D. Versaci Research Scholar in Surgical Sciences Award).
2 Address correspondence and reprint requests to Dr. J. M. Daley, Division of Surgical Research, Rhode Island Hospital, NAB 214, 593 Eddy Street, Providence, RI 02903. E-mail address: jdaley@lifespan.org
3 Abbreviations used in this paper: PVA, polyvinyl alcohol; CM, complete medium.
Received for publication April 27, 2004. Accepted for publication December 1, 2004.
References
Meszaros, A. J., J. S. Reichner, J. E. Albina. 1999. Macrophage phagocytosis of wound neutrophils. J. Leukocyte Biol. 65:35.
Mosmann, T.. 1983. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J. Immunol. Methods 65:55.
Korzeniewski, C., D. M. Callewaert. 1983. An enzyme-release assay for natural cytotoxicity. J. Immunol. Methods 64:313.
Ayala, A., Z. K. Deol, D. L. Lehman, C. D. Herdon, I. H. Chaudry. 1994. Polymicrobial sepsis but not low-dose endotoxin infusion causes decreased splenocyte IL-2/IFN- release while increasing IL-4/IL-10 production. J. Surg. Res. 56:579.
Micheli, V., H. A. Simmonds, M. Bari, G. Pompucci. 1993. HPLC determination of oxidized and reduced pyridine coenzymes in human erythrocytes. Clin. Chim. Acta 220:1.
Cui, S., J. S. Reichner, R. B. Mateo, J. E. Albina. 1994. Activated murine macrophages induce apoptosis in tumor cells through nitric oxide-dependent or -independent mechanisms. Cancer Res. 54:2462.
Nessel, C. C., W. L. Henry, Jr, B. Mastrofrancesco, J. S. Reichner, J. E. Albina. 1999. Vestigial respiratory burst activity in wound macrophages. Am. J. Physiol. 276:R1587.
Rabinowitz, S. S., S. Gordon. 1991. Macrosialin, a macrophage-restricted membrane sialoprotein differentially glycosylated in response to inflammatory stimuli. J. Exp. Med. 174:827.
Kunkel, S. L., M. Spengler, M. A. May, R. Spengler, J. Larrick, D. Remick. 1988. Prostaglandin E2 regulates macrophage-derived tumor necrosis factor gene expression. J. Biol. Chem. 263:5380.
Hasko, G., C. Szabo, Z. H. Nemeth, V. Kvetan, S. M. Pastores, E. S. Vizi. 1996. Adenosine receptor agonists differentially regulate IL-10, TNF-, and nitric oxide production in RAW 264.7 macrophages and in endotoxemic mice. J. Immunol. 157:4634
Hasko, G., D. G. Kuhel, J. F. Chen, M. A. Schwarzschild, E. A. Deitch, J. G. Mabley, A. Marton, C. Szabo. 2000. Adenosine inhibits IL-12 and TNF- production via adenosine A2a receptor-dependent and independent mechanisms. FASEB J. 14:2065.
McWhinney, C. D., M. W. Dudley, T. L. Bowlin, N. P. Peet, L. Schook, M. Bradshaw, M. De, D. R. Borcherding, C. K. Edwards, III. 1996. Activation of adenosine A3 receptors on macrophages inhibits tumor necrosis factor-. Eur. J. Pharmacol. 310:209
Sajjadi, F. G., K. Takabayashi, A. C. Foster, R. C. Domingo, G. S. Firestein. 1996. Inhibition of TNF- expression by adenosine: role of A3 adenosine receptors. J. Immunol. 156:3435.
Fadok, V. A., D. L. Bratton, A. Konowal, P. W. Freed, J. Y. Westcott, P. M. Henson. 1998. Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-, PGE2, and PAF. J. Clin. Invest. 101:890.
Murai, N., K. Nagai, H. Fujisawa, K. Hatanaka, M. Kawamura, Y. Harada. 2003. Concurrent evolution and resolution in an acute inflammatory model of rat carrageenin-induced pleurisy. J. Leukocyte Biol. 73:456.
Gilroy, D. W., P. R. Colville-Nash, D. Willis, J. Chivers, M. J. Paul-Clark, D. A. Willoughby. 1999. Inducible cyclooxygenase may have anti-inflammatory properties. Nat. Med. 5:698
Katsuyama, M., R. Ikegami, H. Karahashi, F. Amano, Y. Sugimoto, A. Ichikawa. 1998. Characterization of the LPS-stimulated expression of EP2 and EP4 prostaglandin E receptors in mouse macrophage-like cell line, J774.1. Biochem. Biophys. Res. Commun. 251:727.
Omert, L., K. Tsukada, C. Hierholzer, V. A. Lyons, T. M. Carlos, A. B. Peitzman, T. R. Billiar. 1998. A role of neutrophils in the down-regulation of IL-6 and CD14 following hemorrhagic shock. Shock 9:391.
Karsunky, H., H. Zeng, T. Schmidt, B. Zevnik, R. Kluge, K. W. Schmid, U. Dührsen, T. M?r?y. 2002. Inflammatory reactions and severe neutropenia in mice lacking the transcriptional repressor Gfi1. Nat. Genet. 30:295.
Steinshamn, S., M. H. Bemelmans, W. A. Buurman, A. Waage. 1995. Granulocytopenia reduces release of soluble TNF receptor p75 in endotoxin-stimulated mice: a possible mechanism of enhanced TNF activity. Cytokine 7:50.
Romani, L., A. Mencacci, E. Cenci, G. Del Sero, F. Bistoni, P. Puccetti. 1997. An immunoregulatory role for neutrophils in CD4+ T helper subset selection in mice with candidiasis. J. Immunol. 158:2356.
Bennouna, S., S. K. Bliss, T. J. Curiel, E. Y. Denkers. 2003. Cross-talk in the innate immune system: neutrophils instruct recruitment and activation of dendritic cells during microbial infection. J. Immunol. 171:6052.
Simpson, D. M., R. Ross. 1972. The neutrophilic leukocyte in wound repair. J. Clin. Invest. 51:2009
Dovi, J. V., L.-K. He, L. A. DiPietro. 2003. Accelerated wound closure in neutrophil-depleted mice. J. Leukocyte Biol. 73:448.
Andersen, L., R. Attstrom, O. Fejerskov. 1978. Effect of experimental neutropenia on oral wound healing in guinea pigs. Scand. J. Dent. Res. 86:237.(Jean M. Daley, Jonathan S)
The regulation of macrophage phenotype by neutrophils was studied in the s.c. polyvinyl alcohol sponge wound model in mice made neutropenic by anti-Gr-1 Ab, as well as in cell culture. Wounds in neutropenic mice contained 100-fold fewer neutrophils than those in nonneutropenic controls 1 day after sponge implantation. Wound fluids from neutropenic mice contained 68% more TNF-, 168% more IL-6, and 61% less TGF-1 than those from controls. Wound fluid IL-10 was not different between the two groups, and IL-4 was not detected. Intracellular TNF- staining was greater in cells isolated from neutropenic wounds than in those from control wounds. The hypothesis that wound neutrophil products modulate macrophage phenotype was tested in Transwell cocultures of LPS-stimulated J774A.1 macrophages and day 1 wound cells (84% neutrophils/15% macrophages). Overnight cocultures accumulated 60% less TNF- and IL-6 than cultures of J774A.1 alone. The suppression of cytokine release was mediated by a soluble factor(s), because culture supernatants from wound cells inhibited TNF- and IL-6 release from LPS-stimulated J774A.1 cells. Culture supernatants from purified wound neutrophils equally suppressed TNF- release from LPS-stimulated J774A.1 cells. Wound cell supernatants also suppressed TNF- and superoxide release from murine peritoneal macrophages. The TNF- inhibitory factor has a molecular mass <3000 Da and is neither PGE2 nor adenosine. The present findings confirm a role for neutrophils in the regulation of innate immune responses through modulation of macrophage phenotype.
Introduction
Neutrophils are the first blood-borne nucleated cells to infiltrate sites of injury or infection, where they produce multiple effector molecules, including reactive oxygen and nitrogen intermediates, proteolytic enzymes, and cytokines. The anti-infectious capacity of neutrophils provides obvious functional advantage to their recruitment into infected tissue. Neutrophils, however, also accumulate in large numbers at sites of sterile inflammation, such as uninfected wounds, myocardial and cerebral infarctions, and fractures where, in the absence of infectious challenge, their role in the inflammatory response is less readily apparent.
Experiments reported in this study were initially designed to define the contribution of neutrophils to the cytokine profile of an acute sterile inflammatory site, such as that provided by the polyvinyl alcohol (PVA)3 sponge wound model. The findings demonstrated that neutrophil depletion results in increased proinflammatory cytokine concentrations in early tissue inflammation. This finding suggested that neutrophils produce a factor(s) that inhibits the release of proinflammatory cytokines by macrophages. The results supported and extended the predictions of this hypothesis by demonstrating that wound neutrophils produce a soluble factor(s) that suppresses LPS-mediated activation of a macrophage cell line and primary peritoneal macrophages. The present results thus reveal an additional capacity of neutrophils in inflammation, namely, the ability of these cells to modulate macrophage inflammatory phenotype.
Materials and Methods
Animals
B6D2F1 male mice (Taconic Farms), 8–12 wk of age, were housed at the Central Research Facilities of Rhode Island Hospital and fed mouse chow and water ad libitum. Mice were certified free of common pathogens by the supplier and were monitored by Brown University/Rhode Island Hospital veterinary personnel. Animal protocols were approved by the animal care committee at Rhode Island Hospital.
PVA sponge wound model
Mice were anesthetized with pentobarbital (50 mg/kg i.p.; Abbott Laboratories). Five PVA sponges (PVA Unlimited), measuring 1 x 1 x 0.6 cm, were inserted into individual s.c. pockets through a midline dorsal incision under sterile conditions, and the skin was closed with clips (1). Mice were euthanized by CO2 asphyxiation, and the sponges were removed under sterile conditions. Culture of randomly selected PVA sponges at the time of removal confirmed sterility in this experimental model. Wound cells were isolated by rapid repeated compression of sponges in a Stomacher (Tekmar). When necessary, cells were subjected to hypotonic lysis of erythrocytes, followed by reconstitution in PBS (Invitrogen Life Technologies). Viable cells were identified by trypan blue exclusion and were counted using a hemocytometer; viability was >95%. Differential cell counts were performed on Hema-3 (Biochemical Sciences)-stained Cytospins (Shandon), and cell phenotype was confirmed by flow cytometry, as described below. Extracellular fluid from the PVA sponges (wound fluid) was obtained by centrifugation (400 x g for 5 min) of two or three sponges in the barrel of a syringe that was seated in a sterile tube. Wound fluid was stored at –80°C until analysis.
Induction of neutropenia
Anti-mouse Gr-1 mAb (RB6.8C5 hybridoma originally produced by R. L. Coffman, DNAX Research Institute) was produced under serum-free conditions using a bioreactor (Cell Pharm Micro Mouse; UniSyn Technologies). Mice were rendered neutropenic by a single i.p. injection of 0.5 mg of anti-Gr-1 Ab 3 days before PVA sponge insertion. Control animals were injected i.p. with an equal volume of normal saline or with 0.5 mg of rat IgG (Rockland Immunochemicals).
Circulating leukocyte counts
Heparinized blood was obtained from the lateral tail vein or the inferior vena cava. Leukocyte count was determined using a hemocytometer after dilution and erythrocyte lysis with 0.01% crystal violet in 3% acetic acid. Differential cell counts were performed on Hema-3 stained blood smears.
Purification of neutrophils from the wound cell suspension
Wound cells were isolated 1 day after PVA sponge insertion, suspended in biotin-free FcR-blocking agent (Accurate Chemical & Scientific) at 0.3 ml/106 cells, washed, and incubated with biotin-conjugated Ab directed against the macrophage cell surface Ag F4/80 (1 μg/106 cells; Caltag Laboratories). Cells were washed in PBS and incubated with MACS Anti-Biotin MicroBeads (Miltenyi Biotec) according to the manufacturer’s recommendations, and macrophages were depleted using the MiniMACS magnetic cell separator (Miltenyi Biotec). The resulting cell suspension was 96% neutrophils by light microscopy.
Cell culture
All cell cultures were performed in complete medium (CM; RPMI 1640 (Invitrogen Life Technologies) supplemented with 10 mM MOPS, 5 x 10–5 M 2-ME, 100 U/ml penicillin-streptomycin, and 1% FBS (HyClone)) unless noted otherwise. The murine macrophage cell line J774A.1 was obtained from American Type Culture Collection. Murine resident peritoneal macrophages were obtained by peritoneal lavage with HBSS and 1% FBS. Thioglycolate-elicited peritoneal macrophages were harvested 4 days after i.p. injection of 3 ml of 3% Brewer modified thioglycolate medium (BD Biosciences). J774A.1 macrophages or peritoneal macrophages were allowed to adhere to tissue culture-treated plastic for 2 h, and nonadherent cells were removed before initiation of experiments.
Three culture formats were used for coculture experiments: 1) J774A.1 cells, 5 x 105 in 1 ml of CM; 2) wound cells, 2.5 x 106 in 1 ml of CM; and 3) cocultures in Transwell plates (0.4-μm pore size; Costar) containing 5 x 105 J774A.1 cells in the lower chamber and 2.5 x 106 wound cells in the upper chamber in a total volume of 1 ml of CM. LPS (Escherichia coli serotype 055:B5, Sigma-Aldrich; 0.1 μg/ml) was added 1 h after the initiation of all cultures. Where noted, indomethacin (Sigma-Aldrich) or NS-398 (Cayman Chemicals) was added at the start of cell incubation.
J774A.1 cells or peritoneal macrophages were also cultured with supernatants conditioned by day 1 wound cells. The wound cell-conditioned supernatant was generated by isolating wound cells from PVA sponges 1 day after insertion and culturing these cells in CM at 2.5 x 106/ml for 24 h. Cells were then removed from the medium by centrifugation, and the supernatant was stored at –80°C. Supernatants conditioned by purified wound neutrophils were generated in similar manner. J774A.1 cells (1 x 105) or primary peritoneal macrophages (3 x 105) were incubated in 200 μl of CM or a 1/1 (v/v) dilution of wound cell-conditioned supernatant in CM for 1 h before addition of LPS, then cultured for an additional 24 h. Wound cell-conditioned supernatant was fractionated using Amicon Centriplus YM-3 centrifugal filter devices (Millipore). When indicated, wound cell-conditioned supernatant or an equal volume of CM was treated with adenosine deaminase (type X from bovine spleen; Sigma-Aldrich) at 0.02 U/ml for 20 min at 25°C, and adenosine deaminase was removed, or not, using a Centricon-10 concentrator (Millipore) before addition of the treated supernatant to cell cultures. Indomethacin, NS-398, AH6809 (Cayman Chemicals), L161982 (gift from Dr. R. Young, Merck Frosst Canada), or 8 phenyl-theophylline (Sigma-Aldrich) was added to cell cultures where noted.
Conditioned culture supernatants were also prepared from splenic lymphocytes. Splenocytes were obtained by mechanical dissociation of spleens and elimination of fibrous tissue using wire mesh. Erythrocytes were removed by hypotonic lysis; macrophages and dendritic cells were depleted by adherence to plastic for 2 h. Supernatants were prepared from the resulting cell suspension (90% lymphocytes, 4% monocytes, and 4% neutrophils) by culture at 2.5 x 106/ml for 24 h. Conditioned supernatants were also prepared from murine L929 fibroblasts and from fibroblasts obtained from a murine wound by culture at a density of 5 x 105/ml for 24 h. Cells were removed by centrifugation, and the supernatants were used as described above.
Macrophage viability was assessed after 24 h of incubation by trypan blue exclusion, reduction of MTT (Sigma-Aldrich) (2), and release of lactate dehydrogenase (3). Protein synthesis was determined by measuring the 24-h incorporation of L-[ring-2,6-N-3H]phenylalanine into TCA-precipitable protein. When indicated, cells were lysed in buffer containing 0.01 M Tris, 0.15 M NaCl, and 0.5% Igepal CA 630 (Sigma-Aldrich) with Complete (Roche) protease inhibitor.
Cell staining and flow cytometry
Wound cells were isolated from neutropenic or control animals 1 day after sponge insertion and pooled. Abs used for cell staining and flow cytometry were obtained from BD Biosciences unless noted otherwise.
Surface Ag staining. Erythrocytes were removed from the wound cell suspension by hypotonic lysis. Cells were incubated with FcR-blocking agent (Accurate Chemical & Scientific), washed in PBS with 1% FBS, and incubated with predetermined optimal concentrations of fluorochrome-conjugated Abs to B220 and IgM (for B cells), CD3 (for T cells), or the appropriate isotype control Abs.
Intracellular Ag staining. Erythrocytes were removed from the wound cell suspension by hypotonic lysis. Cells were incubated with FcR-blocking agent (Accurate Chemical & Scientific), washed in PBS with 1% FBS, fixed in paraformaldehyde with saponin (Cytofix/Cytoperm; BD Biosciences) according to the manufacturer’s recommendations, washed in Permwash (BD Biosciences), and incubated with a predetermined optimal concentration of fluorochrome-conjugated Ab directed against the intracellular macrophage Ag macrosialin (anti-CD68; Serotec) or the appropriate isotype control Ab.
Intracellular cytokine staining. Cells were incubated (106 cells/ml) in CM for 6 h with brefeldin A (Golgiplug; BD Biosciences) at 1 μl/106 cells and stained for intracellular Ag as described above, using anti-CD68, anti-TNF-, and/or the appropriate isotype control Abs.
Cells were analyzed and sorted using a FACSort and CellQuest software (BD Biosciences). Cells were stained with propidium iodide after sorting, and nuclear morphology was determined by light microscopy.
Assays
TNF- (BioSource), IL-6 (BD Biosciences), TGF-1 (R&D Systems), IL-4, and IL-10 (4) were assayed by ELISA. PGE2 was quantified by enzyme immunoassay (Cayman Chemical). ATP, ADP, AMP, and adenosine were assayed by HPLC (5) with a limit of detection of 1 μM. TNF- bioactivity was determined by lysis of actinomycin D (Sigma-Aldrich)-treated L929 murine fibroblasts (NCTC clone 929; American Type Culture Collection) (6), using a recombinant murine TNF- standard (R&D Systems). Superoxide release was measured from the reduction of ferricytochrome c (Sigma-Aldrich) after cell stimulation with 200 nM PMA (LC Services) over 90 min (7).
Data presentation and analysis
Data reported are the mean ± 1 SD unless otherwise noted. Results from cell culture experiments are from triplicate wells in a representative experiment. Statistical analysis was performed using Student’s unpaired t test, Mann-Whitney U test, or ANOVA with Newman-Keuls test, as appropriate.
Results
Characterization of the cellular infiltrate in the PVA sponge wound model
The leukocytic infiltration of the wound increased over the first 7 days after PVA sponge insertion (Fig. 1A). Day 1 wound cells were 84 ± 4% neutrophils, 15 ± 4% macrophages, <1% lymphocytes, and <1% eosinophils in six experiments with at least five animals per experiment (mean ± SEM). Neutrophils remained the most abundant cell type through at least 3 days after wounding (Fig. 1B). Macrophages constituted 52% of the cells isolated from the wound 7 days after wounding. All subsequent experiments using wound cells or wound fluids were performed using cells or fluids isolated 1 day after PVA sponge insertion.
FIGURE 1. Characterization of the cellular infiltrate in the PVA sponge wound. PVA sponges were inserted as described in Materials and Methods, and six animals each were euthanized at 1, 3, and 7 days after sponge insertion. Wound cells were isolated, identified, and counted as described in Materials and Methods. A, Wound cell count per animal. Counts differed at all time points (p < 0.05, by ANOVA and Newman-Keuls test). B, Differential cell count. , neutrophils; , macrophages.
Effect of anti-Gr-1 Ab on circulating and wound leukocyte counts
Intraperitoneal injection of 0.5 mg of mAb directed against the neutrophil surface Ag Gr-1 reduced neutrophils from 13.3 ± 1.5 to 1.4 ± 1.0% of the total circulating leukocytes. Neutropenia persisted throughout the time course of these experiments (4 days). Differential blood lymphocyte and monocyte counts were not altered by Ab treatment (data not shown).
The effect of anti-Gr-1 Ab on leukocyte infiltration of the wounds was quite marked: wounds from neutropenic animals contained 67-fold fewer total cells than controls, with 100-fold less neutrophils and 10-fold less macrophages (Fig. 2).
FIGURE 2. Anti-Gr-1 Ab depletes leukocyte populations in the PVA sponge wound. Mice were injected i.p. with 0.5 mg of anti-Gr-1 Ab or an equal volume of saline, and PVA sponges were inserted into control (n = 6) and neutropenic (n = 7) mice 3 days later. Wound cells were isolated 1 day after sponge insertion. The graph shows cell counts per animal () and group means (–). Neutropenic wounds had lower total cell, neutrophil, and macrophage counts than control wounds (p < 0.05, by Mann-Whitney U test).
Neutrophil depletion increases proinflammatory cytokines in wound fluids
Table I shows the cytokine content of wound fluids from control and neutropenic animals. Despite the decrease in the inflammatory cell infiltrate, immunoreactive TNF- and IL-6 concentrations in wound fluid from neutropenic animals were 68 and 168% higher, respectively, than those from controls, whereas TGF-1 concentrations were 61% lower. Wound fluid IL-10 concentrations were not different between the two groups. IL-4 was below the limit of detection of the assay (10 pg/ml) in wound fluids from either group.
Table I. Cytokine concentration in wound fluids: neutrophil depletion selectively increases proinflammatory cytokinesa
The increases in proinflammatory cytokines in wound fluids from neutropenic animals were independent of plasma cytokines. Plasma TNF- concentrations at the time of sponge removal were below the limit of detection of the assay (30 pg/ml) in both groups. There was no difference in the plasma concentrations of IL-6 (control, 59 ± 21 pg/ml; neutropenic, 48 ± 14 pg/ml; p > 0.05, by Mann-Whitney U test). Neutropenic animals did not exhibit obvious differences in the appearance of the surgical site or in postoperative behavior compared with wounded control animals.
Cells from neutropenic wounds contain more TNF- than those from control wounds
To determine the source of increased TNF- in the wound fluids of neutropenic animals, cells were isolated from control and neutropenic wounds, cultured with brefeldin to inhibit intracellular cytokine transport, then stained for intracellular TNF- and the intracellular macrophage Ag CD68. The validity of CD68 staining for the identification of wound macrophages was confirmed by microscopic examination of the cells after FACS. In control animals, CD68high cells were 98.5% macrophages/mononuclear cells and 1.5% polymorphonuclear leukocytes; CD68–/low cells were 95.0% polymorphonuclear leukocytes (including bands) and 4.0% macrophages/mononuclear cells. As will be discussed below, in neutropenic animals the CD68–/low cells included a greater proportion of mononuclear cells with morphologic characteristics of monocytes, macrophages, and lymphocytes than those in controls.
The intensity of staining for TNF- was greater in CD68high cells than in CD68–/low cells in both neutropenic and control animals (Fig. 3). Macrophages from neutropenic wounds stained more intensely for TNF- than those from controls. The blood leukocytes of control and neutropenic animals did not differ in intracellular TNF- staining (data not shown).
FIGURE 3. Intracellular TNF- staining in cells from control and neutropenic wounds. Mice were injected i.p. with 0.5 mg of anti-Gr-1 Ab or saline, and PVA sponges were inserted 3 days later. Wound cells isolated 1 day after sponge insertion were cultured with brefeldin A and stained for intracellular TNF- and macrophage intracellular Ag CD68 as described in Materials and Methods. MCF, mean channel fluorescence of the quadrant.
The CD68–/low cells from neutropenic wounds contained more TNF- than those from control wounds. These CD68–/lowTNF-+ cells comprised 30% of the total cells from neutropenic animals. Morphologic examination of these cells after FACS showed 59% polymorphonuclear leukocytes and 41% mononuclear cells. In control animals, the CD68–/lowTNF-+ population of cells constituted only 6.7% of the wound cell suspension and was comprised almost entirely of polymorphonuclear leukocytes (96%), with the remaining 4% being mononuclear cells. The absolute number of these CD68–/lowTNF-+ mononuclear cells, as calculated from the respective total cellularity of control vs neutropenic wounds, was actually higher in controls (4800/animal) than in neutropenic wounds (2100/animal). These cells probably represent extravasated monocytes, found in the laboratory to be CD68–/low (data not shown). In this regard it has been shown that CD68 Ag expression is enhanced during macrophage recruitment (8).
Wound cells inhibit TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages
The hypothesis that wound cells, specifically wound neutrophils, suppress cytokine release by macrophages in vivo was modeled and tested in vitro. The murine macrophage cell line J774A.1 used in these experiments was confirmed in the laboratory to exhibit minimal release of TNF- in unstimulated culture and submaximal cytokine release when stimulated with 0.1 μg/ml LPS. J774A.1 macrophages were cultured in Transwells with or without day 1 wound cells (86% neutrophils) from nonneutropenic animals. Fig. 4A shows that supernatants from J774A.1/wound cell cocultures contained 60% less immunoreactive TNF- than those from cultures of J774A.1 cells alone. Results obtained by ELISA for TNF- were confirmed in a bioassay (Fig. 4B). Similarly, supernatants from J774A.1/wound cell cocultures contained 60% less IL-6 than cultures of J774A.1 cells (Fig. 4C). Comparison of the kinetics of TNF- release in cultures of LPS-stimulated J774A.1 cells vs those in J774A.1/wound cell cocultures showed that suppression of TNF- release in cocultures was not evident until after 8 h of incubation (Fig. 5).
FIGURE 4. Wound cells inhibit TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages. Wound cells were isolated from nonneutropenic mice 1 day after sponge insertion. Wound cells (2.5 x 106) or J774A.1 cells (J774; 5 x 105) were cultured in 1 ml of CM. Transwell cocultures (J774 and wound cells) contained 5 x 105 J774A.1 cells in the lower chamber and 2.5 x 106 wound cells in the upper chamber in a total volume of 1 ml. Cells were cultured with LPS (0.1 μg/ml) for 24 h. Culture supernatants were assayed for TNF- by ELISA (A), for TNF- bioactivity by lysis of actinomycin-treated L929 cells (B), or for IL-6 by ELISA (C). TNF- and IL-6 concentrations were lower in cocultures than in J774A.1 cells cultured alone (p < 0.05, by Student’s unpaired t test).
FIGURE 5. Kinetics of TNF- release in J774A.1/wound cell cocultures stimulated with LPS. Transwell cocultures of J774A.1 cells and day 1 wound cells were performed as described in Fig. 4. TNF- concentrations were lower in cocultures than in J774A.1 cells cultured alone at 20 and 24 h (p < 0.001, by two-factor ANOVA).
Culture supernatants conditioned by day 1 wound cells inhibit TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages
LPS-stimulated J774A.1 macrophages were also cultured in supernatants previously conditioned by day 1 wound cells. As shown in Fig. 6, J774A.1 cultures containing wound cell-conditioned supernatants accumulated 56% less TNF- and 60% less IL-6 at 24 h than those in CM. The kinetics of TNF- release differed from those in J774A.1/wound cell cocultures, because the inhibition of TNF- release by wound cell supernatants was evident as early as 1 h after stimulation with LPS (Fig. 7). The suppression of TNF- release was not due to cellular accumulation of the cytokine during culture, because lysates from J774A.1 cells incubated with wound cell-conditioned supernatants contained less TNF- than those in CM (J774A.1 in supernatants, 133 ± 6 pg/106 cells; J774A.1 in CM, 147 ± 6 pg/106 cells; p < 0.05, by Student’s unpaired t test.). IL-6 was not assayed in these samples because there is no cell-associated form of the cytokine. The suppression of macrophage TNF- release by wound cell supernatants was also evident in the absence of LPS (J774A.1 cells in CM, 2.1 ± 0.2 ng/ml; J774A.1 in wound cell supernatants, 1.6 ± 0.1 ng/ml; p < 0.05, by Student’s unpaired t test.).
FIGURE 6. Culture supernatants from wound cells inhibit TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages. A, Wound cells were isolated 1 day after PVA sponge insertion, and macrophages were depleted from the wound cell suspension as described in Materials and Methods. The unfractionated wound cells (86% neutrophils) or the purified wound neutrophils (96% neutrophils) were cultured for 24 h at 2.5 x 106/ml. Supernatants from these cultures were diluted 1/1 with CM and added to J774A.1 cells (5 x 105/ml) along with LPS (0.1 μg/ml). The TNF- concentration of the culture supernatant was assayed by ELISA after 24 h. J774A.1 cultures with either wound cell supernatant or wound neutrophil supernatant contained less TNF- than cultures in CM (p < 0.05, by ANOVA and Newman-Keuls test). B, J774A.1 cells were cultured with LPS (0.1 μg/ml), with or without wound cell-conditioned supernatants. IL-6 was assayed by ELISA after 24-h culture. J774A.1 cultures with wound cell-conditioned supernatants contained less IL-6 than cultures in CM (p < 0.01, by Student’s unpaired t test).
FIGURE 7. Kinetics of TNF- release in LPS-stimulated J774A.1 macrophages incubated with wound cell-conditioned supernatants. J774A.1 cells were cultured with LPS (0.1 μg/ml), with or without wound cell-conditioned supernatants, as described in Fig. 6. J774A.1 cultures with wound cell-conditioned supernatants contained less TNF- than cultures in CM at all time points (p < 0.05, by two-factor ANOVA).
To specifically assess the contribution of neutrophils to the TNF--suppressive capacity of wound cells, LPS-stimulated J774A.1 macrophages were cultured in supernatants conditioned by neutrophils purified from day 1 wound cells (96% neutrophils). The neutrophil supernatant and the wound cell supernatant equally suppressed TNF- release by J774A.1 macrophages (Fig. 6A).
Suppression of TNF- release from LPS-stimulated J774A.1 macrophages was not observed when supernatants were conditioned by splenic lymphocytes or L929 murine fibroblasts. (J774A.1 in CM, 94.2 ± 14.1 ng/ml; J774A.1 in splenic lymphocyte-conditioned supernatant, 83.5 ± 15.6 ng/ml; J774A.1 in L929-conditioned supernatant, 114.4 ± 13.8 ng/ml; p > 0.05, by ANOVA and Newman-Keuls test). Supernatants conditioned by fibroblasts isolated from a murine wound suppressed TNF- release from LPS-stimulated J774A.1 cells (J774A.1 in mouse wound-derived fibroblast-conditioned supernatant, 18.0 ± 0.8 ng/ml; p < 0.05 vs J774A.1 in CM (above), by ANOVA and Newman-Keuls test).
The suppression of proinflammatory cytokine release by wound cell-conditioned supernatants did not result from cytotoxicity or from nonspecific inhibition of protein synthesis. There was no difference between cells cultured in CM and those cultured in wound cell supernatant with respect to cell viability at the end of culture, as assessed by trypan blue exclusion, lactate dehydrogenase release, or reduction of MTT (Table II). Protein synthesis, as measured by incorporation of radiolabeled phenylalanine into protein, was identical in both culture conditions.
Table II. Wound cell conditioned supernatants are not toxic to J774A.1 macrophagesa
Wound cell-conditioned supernatants inhibit release of TNF- and production of superoxide by murine peritoneal macrophages
Wound cell-conditioned supernatants also inhibited LPS-stimulated TNF- release from thioglycolate-elicited murine peritoneal macrophages. As shown in Fig. 8A, macrophages incubated in wound cell supernatants released 71% less TNF- than those cultured in CM. Resident peritoneal macrophages cultured in wound cell supernatants also produced less superoxide after PMA stimulation than those cultured in CM at all tested LPS concentrations (Fig. 8B).
FIGURE 8. Wound cell-conditioned supernatants inhibit TNF- release and superoxide production by LPS-stimulated primary murine peritoneal macrophages. A, Thioglycolate-elicited peritoneal macrophages (3 x 105 in 200 μl) were cultured with LPS (0.1 μg/ml) with or without wound cell-conditioned supernatants, as described in Fig. 6. TNF- was assayed by ELISA after 24-h culture. Macrophages cultured with wound cell supernatants released less TNF- than those cultured in CM (p < 0.01, by Student’s unpaired t test). B, Resident peritoneal macrophages (3 x 105 in 200 μl) purified by overnight adherence were cultured with or without wound cell-conditioned supernatants at varying doses of LPS. Macrophage superoxide production was assayed after 24-h culture. Macrophages cultured with wound cell-conditioned supernatants () produced less superoxide than those in CM (; p < 0.05, by two-factor ANOVA).
Partial characterization of the neutrophil-derived factor(s) that inhibits proinflammatory cytokine release from J774A.1 macrophages
The neutrophil-derived factor(s) that inhibits proinflammatory cytokine release by LPS-stimulated J774A.1 macrophages is resistant to freezing and thawing and has a molecular mass <3000 Da. Fig. 9 shows that the <3000-Da filtrates of wound cell-conditioned supernatants inhibited TNF- release by LPS-stimulated J774A.1 macrophages similar to unfractionated supernatants. IL-6 release was also inhibited by the <3000-Da fraction of wound cell-conditioned supernatants (J774A.1 in CM, 19.5 ± 1.4 ng/ml; J774A.1 in supernatants, 11.4 ± 0.7 ng/ml; p < 0.01, by Student’s unpaired t test).
FIGURE 9. The <3000-Da fraction of wound cell supernatant inhibits TNF- release by LPS-stimulated J774A.1 macrophages. J774A.1 cells were cultured in CM () or wound cell supernatants () as described in Fig. 6. The <3000-Da fraction of supernatant or CM was also diluted 1/1 with CM and added to J774.1 cultures. All cultures were treated with LPS (0.1 μg/ml). The TNF- concentration of the J774A.1 culture supernatant was assayed by ELISA after 24-h culture. Both the unfractionated and the <3000-Da fraction of wound cell supernatants inhibited TNF- release compared with CM (p < 0.01, by two-factor ANOVA).
PGE2 (9) and adenosine (10, 11, 12, 13) have been shown to suppress macrophage TNF- production and meet the criteria of stability to freezing/thawing and molecular mass <3000 Da. The results presented in Fig. 10 show that cyclooxygenase inhibitors did not alter TNF- release in cocultures of day 1 wound cells and J774A.1 macrophages. Neither the selective cyclooxygenase-2 inhibitor NS398 (10 μM), nor the nonselective cyclooxygenase inhibitor indomethacin (10 μM) reversed the inhibition of TNF- release in the cocultures (Fig. 10A). The effectiveness of cyclooxygenase inhibitors was confirmed by measurement of PGE2 in the culture medium (Fig. 10B).
FIGURE 10. Effect of cyclooxygenase inhibitors on wound neutrophil inhibition of TNF- release by LPS-stimulated J774A.1 macrophages. J774A.1 cells were cultured alone or with day 1 wound cells in Transwells as described in Fig. 4. NS-398 (10 μM) or indomethacin (10 μM) was added to cultures 1 h before stimulation with LPS (0.1 μg/ml). TNF- (A) and PGE2 (B) concentrations in the J774A.1 culture medium were assayed after 24-h culture as described in Materials and Methods. J774A.1/wound cell cocultures accumulated less TNF- than J774A.1 cultures (p < 0.05, by two-factor ANOVA). Neither cyclooxygenase inhibitor had an effect on TNF- release.
The use of PG receptor antagonists confirmed that PGE2 is not the inhibitory factor under study. Neither the EP4 PG receptor antagonist L161982 (0.5–10 μM) nor the EP1/EP2/DP PG receptor antagonist AH6809 (0.1–10 μM) reversed the inhibition of TNF- release by LPS-stimulated J774A.1 macrophages cultured with wound cell-conditioned supernatants (data not shown).
Stimulation of adenosine receptors has been reported to inhibit the production of TNF- by LPS-stimulated macrophages. No ATP, ADP, AMP, or adenosine was detected in wound cell-conditioned supernatant by HPLC. To rule out adenosine receptor activation by concentrations of adenosine lower than the limit of detection of the assay, wound cell-conditioned supernatants were treated with adenosine deaminase before addition to J774A.1 cells. Adenosine deaminase did not reverse the inhibition of TNF- release by LPS-stimulated J774A.1 macrophages. Furthermore, the adenosine receptor antagonist, 8-phenyl theophylline (1–20 μM), had no effect on TNF- release by LPS-stimulated J774A.1 macrophages (data not shown).
Discussion
The results reported in this study provide in vivo and in vitro evidence that neutrophils modulate macrophage inflammatory phenotype through the release of a soluble mediator(s). In vivo findings in a model of acute sterile inflammation demonstrate that neutrophil depletion results in selective alterations in local cytokine concentrations that include increases in TNF- and IL-6 and reduction in TGF-1. Intracellular staining evidenced increased TNF- staining in both neutrophils and macrophages from the wounds of neutropenic animals compared with those from control animals. The observed increases in TNF- and IL-6 in the wound fluid of neutropenic animals reflect a local, rather than a systemic, effect, because plasma TNF- and IL-6 were not different in neutropenic and control animals. Moreover, intracellular staining of TNF- in circulating leukocytes revealed no differences between the groups.
The increases in TNF- and IL-6 concentrations of wound fluids from neutropenic animals were unexpected in view of the 67-fold reduction in the number of total wound leukocytes. The reciprocal changes in cell number vs proinflammatory cytokine content in the wounds of control and neutropenic animals suggested that cell density-dependent mechanisms regulate proinflammatory cytokine production at sites of inflammation, with a higher cell number correlating to decreased, rather than increased, cytokine levels. One well-described mechanism that could account for the present findings in vivo is the phagocytosis of apoptotic neutrophils and the subsequent reduction in proinflammatory cytokine release by macrophages at the inflammatory site (14, 15). Alternatively, the present observations could be explained by a soluble factor(s) released from wound cells that down-regulates local proinflammatory cytokine release through an autocrine and/or paracrine mechanism. To test for both mechanisms simultaneously, a coculture system that mimicked the cellularity of the early wound and prevented direct cell-to-cell contact was used.
The data in Fig. 4 show results from one such coculture experiment and demonstrate that day 1 wound cells from naive animals, which are predominantly neutrophils, inhibited TNF- and IL-6 release by LPS-stimulated J774A.1 macrophages by 60%. The results depicted in Fig. 5 suggest synthesis and release of the inhibitory factor(s) into the medium during coculture, because TNF- accumulation in the medium was not suppressed until after 8 h in coculture. Inhibition of TNF- release was detectable by 1 h after the addition of conditioned supernatant to J774A.1 cells (Fig. 7), suggesting that those supernatants contain a preformed inhibitory factor(s). Fig. 6 provides evidence that neutrophils are, among wound cells, the most likely source for the soluble mediator(s) that inhibited macrophage cytokine release in cocultures, because cell culture supernatants from purified wound neutrophils suppressed TNF- release by LPS-stimulated J774A.1 macrophages to the same extent as supernatants from unfractionated wound cells. Moreover, the inhibition of proinflammatory cytokine release is not due to a nonselective inhibition of protein synthesis, because total protein synthesis was not altered by wound cell-conditioned supernatants.
The suppressive effects of wound cell culture supernatants were not limited to the J774A.1 macrophage cell line. Fig. 8A demonstrates that the supernatants also inhibited TNF- release from primary peritoneal macrophages. Moreover, the suppression of LPS-mediated macrophage activation was not restricted to the production of proinflammatory cytokines. The results shown in Fig. 8B illustrate the reduction in macrophage superoxide release that resulted from culture in wound cell-conditioned supernatant.
Results from this study demonstrate that neutrophils mediate the inhibition of proinflammatory cytokine and superoxide release from macrophages via a soluble mediator(s) of <3000 Da. Characterization of the inhibitory molecule(s) is not yet complete. Known inhibitors of macrophage TNF- release that are <3000 Da include PGE2 and adenosine. The early accumulation of PGs at inflammatory sites (16) and the suppressive effects of PGE2 on macrophage TNF- production (9) suggested that cyclooxygenase products might mediate wound neutrophil suppression of macrophage TNF- release in response to LPS. However, the data in Fig. 10 show that inhibition of PGE2 synthesis by NS 398 or indomethacin did not alter the neutrophil-mediated inhibition of TNF- release by LPS-stimulated macrophages. Also, neither the EP4 receptor antagonist L161982 nor the EP2 receptor antagonist AH6809 affected the capacity of wound cell-conditioned supernatants to inhibit TNF- release by J774A.1 macrophages. In this regard the receptor antagonists were selected because J774A.1 cells express only PG receptors of the EP4 and EP2 subtypes (17). PGE2, therefore, is not the mediator of wound cell suppression of TNF- release by LPS-stimulated macrophages. Adenosine and adenosine receptor agonists also inhibit macrophage inflammatory responses, including the release of TNF- (10, 11, 12, 13). Neither adenosine nor its metabolic precursors (ATP, ADP, and AMP) were found in wound cell-conditioned supernatants. Also, the lack of effect of adenosine deaminase or an adenosine receptor antagonist argues against involvement of the adenosine receptor in the wound cell-mediated inhibition of TNF- release by LPS-stimulated macrophages.
Supernatants from wound-derived fibroblasts, but not those from splenic lymphocytes or from L929 immortalized murine fibroblasts, also suppressed LPS-stimulated TNF- release from J774A.1 cells. Although it is not known whether the suppressive factor produced by wound fibroblasts is identical with that from wound neutrophils, it is of interest that other cell types that come into contact with macrophages within inflammatory sites are also capable of inhibiting TNF- release.
These results confirm a role for neutrophils in the regulation of proinflammatory responses by macrophages. This conclusion finds support and extension in published studies of neutrophil depletion. Rats made neutropenic with vinblastine and subjected to hemorrhagic shock manifested decreased neutrophil sequestration and increased expression of IL-6 mRNA and CD14 mRNA in the lung compared with nonneutropenic controls (18). Mice lacking the transcriptional repressor Gfi1 are severely neutropenic and have increased lethality and higher plasma levels of TNF-, IL-10, and IL-1 after LPS administration (19). Macrophages isolated from neutropenic Gfi1-deficient mice exhibit increased LPS-stimulated TNF- release. The findings of the present study suggest that neutropenia in Gfi1-deficient animals resulted in up-regulation of macrophage cytokine production. The latter conclusion is also supported by Steinshamn’s investigation (20) of the TNF- response to endotoxin in mice made neutropenic with cyclophosphamide, showing increased lethality and increased serum TNF- concentration compared with control animals after LPS stimulation. A proinflammatory bias in neutropenia has thus been shown to be independent of species, mechanism of neutropenia, and stressor imposed upon the animals.
The work reported in this study resulted in an unexpected and paradoxical finding: that neutrophils, which are considered prototypical inflammatory cells, could act as anti-inflammatory cells through suppression of the macrophage inflammatory phenotype. The present observations together with reports indicating that neutrophils modulate T cell subset selection in candidiasis (21) and dendritic cell activation in Toxoplasma gondii infection (22), shed new light on the capacity of neutrophils to regulate multiple cellular immune responses.
As stated in the introduction, the role of neutrophils in sterile inflammation has not been completely clarified. With specific regard to the role of neutrophils in the healing of wounds, the current paradigm, which is based on the work of Simpson and Ross (23), states that neutrophils are dispensable to the repair process, and that their function in injured tissue is only to contain or eliminate infection. Only recently has this paradigm been challenged. Dovi et al. (24) demonstrated accelerated re-epithelialization of skin wounds in transiently neutropenic mice, without effects on collagen content or tensile strength of the wound. These findings do not appear to apply to all models of wound healing, because re-epithelialization of oral wounds in guinea pigs was neither accelerated nor retarded in neutropenic animals (25). Thus, the role of neutrophils in repair remains undefined.
The present observations are unlikely to be limited to wounds. The early inflammatory response in the PVA sponge wound model is identical with that observed in other models of sterile tissue injury, with the early development of a neutrophil-rich cellular infiltrate that is later replaced by a macrophage-predominant infiltrate. The suppression of the macrophage proinflammatory phenotype by neutrophil products is probably common to all acute sterile inflammatory processes. Characterization of a neutrophil-derived soluble mediator(s) that suppresses macrophage proinflammatory phenotype as well as their mechanism of action may provide opportunities for therapeutic intervention to suppress macrophage inflammatory phenotype, especially during spontaneous or therapy-induced neutropenia.
Acknowledgments
We thank Jill Rose for assistance in preparing the manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was funded by National Institutes of Health Grants GM42859 (to J.E.A.) and GM66194 (to J.S.R.) and by funds allocated to the Department of Surgery by Rhode Island Hospital, a Lifespan partner. J.M.D. was supported by a National Institutes of Health Supplement to Promote Reentry into Biomedical and Behavioral Research Careers. E.J.M. was supported by the Carter Family Charitable Trust (Armand D. Versaci Research Scholar in Surgical Sciences Award).
2 Address correspondence and reprint requests to Dr. J. M. Daley, Division of Surgical Research, Rhode Island Hospital, NAB 214, 593 Eddy Street, Providence, RI 02903. E-mail address: jdaley@lifespan.org
3 Abbreviations used in this paper: PVA, polyvinyl alcohol; CM, complete medium.
Received for publication April 27, 2004. Accepted for publication December 1, 2004.
References
Meszaros, A. J., J. S. Reichner, J. E. Albina. 1999. Macrophage phagocytosis of wound neutrophils. J. Leukocyte Biol. 65:35.
Mosmann, T.. 1983. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J. Immunol. Methods 65:55.
Korzeniewski, C., D. M. Callewaert. 1983. An enzyme-release assay for natural cytotoxicity. J. Immunol. Methods 64:313.
Ayala, A., Z. K. Deol, D. L. Lehman, C. D. Herdon, I. H. Chaudry. 1994. Polymicrobial sepsis but not low-dose endotoxin infusion causes decreased splenocyte IL-2/IFN- release while increasing IL-4/IL-10 production. J. Surg. Res. 56:579.
Micheli, V., H. A. Simmonds, M. Bari, G. Pompucci. 1993. HPLC determination of oxidized and reduced pyridine coenzymes in human erythrocytes. Clin. Chim. Acta 220:1.
Cui, S., J. S. Reichner, R. B. Mateo, J. E. Albina. 1994. Activated murine macrophages induce apoptosis in tumor cells through nitric oxide-dependent or -independent mechanisms. Cancer Res. 54:2462.
Nessel, C. C., W. L. Henry, Jr, B. Mastrofrancesco, J. S. Reichner, J. E. Albina. 1999. Vestigial respiratory burst activity in wound macrophages. Am. J. Physiol. 276:R1587.
Rabinowitz, S. S., S. Gordon. 1991. Macrosialin, a macrophage-restricted membrane sialoprotein differentially glycosylated in response to inflammatory stimuli. J. Exp. Med. 174:827.
Kunkel, S. L., M. Spengler, M. A. May, R. Spengler, J. Larrick, D. Remick. 1988. Prostaglandin E2 regulates macrophage-derived tumor necrosis factor gene expression. J. Biol. Chem. 263:5380.
Hasko, G., C. Szabo, Z. H. Nemeth, V. Kvetan, S. M. Pastores, E. S. Vizi. 1996. Adenosine receptor agonists differentially regulate IL-10, TNF-, and nitric oxide production in RAW 264.7 macrophages and in endotoxemic mice. J. Immunol. 157:4634
Hasko, G., D. G. Kuhel, J. F. Chen, M. A. Schwarzschild, E. A. Deitch, J. G. Mabley, A. Marton, C. Szabo. 2000. Adenosine inhibits IL-12 and TNF- production via adenosine A2a receptor-dependent and independent mechanisms. FASEB J. 14:2065.
McWhinney, C. D., M. W. Dudley, T. L. Bowlin, N. P. Peet, L. Schook, M. Bradshaw, M. De, D. R. Borcherding, C. K. Edwards, III. 1996. Activation of adenosine A3 receptors on macrophages inhibits tumor necrosis factor-. Eur. J. Pharmacol. 310:209
Sajjadi, F. G., K. Takabayashi, A. C. Foster, R. C. Domingo, G. S. Firestein. 1996. Inhibition of TNF- expression by adenosine: role of A3 adenosine receptors. J. Immunol. 156:3435.
Fadok, V. A., D. L. Bratton, A. Konowal, P. W. Freed, J. Y. Westcott, P. M. Henson. 1998. Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-, PGE2, and PAF. J. Clin. Invest. 101:890.
Murai, N., K. Nagai, H. Fujisawa, K. Hatanaka, M. Kawamura, Y. Harada. 2003. Concurrent evolution and resolution in an acute inflammatory model of rat carrageenin-induced pleurisy. J. Leukocyte Biol. 73:456.
Gilroy, D. W., P. R. Colville-Nash, D. Willis, J. Chivers, M. J. Paul-Clark, D. A. Willoughby. 1999. Inducible cyclooxygenase may have anti-inflammatory properties. Nat. Med. 5:698
Katsuyama, M., R. Ikegami, H. Karahashi, F. Amano, Y. Sugimoto, A. Ichikawa. 1998. Characterization of the LPS-stimulated expression of EP2 and EP4 prostaglandin E receptors in mouse macrophage-like cell line, J774.1. Biochem. Biophys. Res. Commun. 251:727.
Omert, L., K. Tsukada, C. Hierholzer, V. A. Lyons, T. M. Carlos, A. B. Peitzman, T. R. Billiar. 1998. A role of neutrophils in the down-regulation of IL-6 and CD14 following hemorrhagic shock. Shock 9:391.
Karsunky, H., H. Zeng, T. Schmidt, B. Zevnik, R. Kluge, K. W. Schmid, U. Dührsen, T. M?r?y. 2002. Inflammatory reactions and severe neutropenia in mice lacking the transcriptional repressor Gfi1. Nat. Genet. 30:295.
Steinshamn, S., M. H. Bemelmans, W. A. Buurman, A. Waage. 1995. Granulocytopenia reduces release of soluble TNF receptor p75 in endotoxin-stimulated mice: a possible mechanism of enhanced TNF activity. Cytokine 7:50.
Romani, L., A. Mencacci, E. Cenci, G. Del Sero, F. Bistoni, P. Puccetti. 1997. An immunoregulatory role for neutrophils in CD4+ T helper subset selection in mice with candidiasis. J. Immunol. 158:2356.
Bennouna, S., S. K. Bliss, T. J. Curiel, E. Y. Denkers. 2003. Cross-talk in the innate immune system: neutrophils instruct recruitment and activation of dendritic cells during microbial infection. J. Immunol. 171:6052.
Simpson, D. M., R. Ross. 1972. The neutrophilic leukocyte in wound repair. J. Clin. Invest. 51:2009
Dovi, J. V., L.-K. He, L. A. DiPietro. 2003. Accelerated wound closure in neutrophil-depleted mice. J. Leukocyte Biol. 73:448.
Andersen, L., R. Attstrom, O. Fejerskov. 1978. Effect of experimental neutropenia on oral wound healing in guinea pigs. Scand. J. Dent. Res. 86:237.(Jean M. Daley, Jonathan S)