Abstract
Data from several inflammation/autoimmunity models indicate that GM-CSF can be a key inflammatory mediator. Convenient models in readily accessible tissues are needed to enable the GM-CSF-dependent cellular responses to be elaborated. In this study, we show that, in contrast to the response to the commonly used i.p. irritant, thioglycolate medium, an Ag-specific methylated BSA-induced peritonitis in GM-CSF−/− mice was severely compromised. The reduced response in the latter peritonitis model was characterized by fewer neutrophils and macrophages, as well as by deficiencies in the properties of the remaining macrophages, namely size and granularity, phagocytosis, allogeneic T cell triggering, and proinflammatory cytokine production. B1 lymphocytes were more evident in the GM-CSF−/− Ag-specific exudates, indicating perhaps that GM-CSF can act on a common macrophage-B1 lymphocyte precursor in the inflamed peritoneum. We propose that these findings contribute to our understanding of how GM-CSF acts as a proinflammatory cytokine in many chronic inflammatory/autoimmune diseases. Of general significance, the findings also indicate that the nature of the stimulus is quite critical in determining whether a particular inflammatory mediator, such as GM-CSF, plays a role in an ensuing inflammatory reaction.
Initially discovered for its role in the differentiation of hemopoietic cells into granulocytes and macrophages (1), GM-CSF can also affect mature cell function and may be considered proinflammatory (2, 3, 4, 5, 6, 7, 8, 9, 10, 11). It has been shown to have proinflammatory effects on granulocytes (12), monocyte activation (12), and T cell function (13), and is required for normal function of at least some populations of dendritic cells (DC)3 (14, 15). GM-CSF gene-deficient (GM-CSF−/−) mice do not show myelopoietic deficiency, the major steady-state defect being in alveolar macrophage function (16). However, we and others have shown in mice that GM-CSF is critical for the development of a number of inflammatory and autoimmune diseases (reviewed in Ref.17), including arthritis (18, 19, 20), multiple sclerosis (experimental autoimmune encephalomyelitis) (21), glomerulonephritis (22), and LPS-induced lung inflammation (23). However, in these models, analysis of the mode of action of GM-CSF as a proinflammatory cytokine is hampered by the difficulty in isolating cell populations in sufficient quantity and purity from the site(s) of inflammation.
The peritoneal cavity in mice is an ideal site in which to study the cellular composition of a developing inflammatory reaction, because it is sterile and large numbers of individual cells can be isolated easily. Thioglycolate medium (TM) has for many years been the most common stimulus to elicit polymorphonuclear neutrophil (PMN) and macrophage infiltration into the peritoneal cavity; however, it is a nonspecific reaction that may not be that appropriate as a representation of the types of stimuli likely to be responsible for inflammatory/autoimmune reactions of clinical interest. Also, the peritonitis resulting from its injection appears to be independent of GM-CSF (24) and therefore unsuitable to explore how GM-CSF might be acting as a proinflammatory cytokine.
Recently, we described a T cell-dependent, Ag-specific peritonitis model using methylated BSA (mBSA) as the stimulus (25), involving prior sensitization and a booster injection intradermally of mBSA in adjuvant, followed by an i.p. challenge with mBSA. The inflammatory macrophages present following mBSA-induced peritonitis differed in their morphology, phenotype, and functional capabilities compared with TM-induced inflammatory macrophages (25). The Ag-driven model was proposed to be more relevant for the study of the inflammatory reactions, including the role(s) of different specific mediators such as GM-CSF, observed in many diseases. We show that, in contrast to TM-induced peritonitis, the corresponding reaction in the mBSA model was compromised in GM-CSF−/− mice. In these gene-deficient mice, the PMN and macrophage responses could not be sustained, and the remaining macrophages displayed altered properties consistent with a reduced activation state. Thus, in murine peritonitis, and perhaps other inflammatory reactions, the nature of the stimulus determines whether GM-CSF is a key regulator or not.
Materials and Methods
Mice
GM-CSF gene-deficient (GM-CSF−/−) mice, backcrossed onto the C57BL/6 background for 11 generations, were originally provided by the Ludwig Institute for Cancer Research (Parkville, Victoria, Australia). As wild-type controls, C57BL/6 mice, originally obtained from Central Animal Services, Monash University (Clayton, Victoria, Australia), were used. Both strains were bred in our on-site animal facility. BALB/c mice were obtained from the Walter and Eliza Hall Institute Animal Supplies (Kew, Victoria, Australia). Mice, 8–12 wk of age, were used in all experiments.
Reagents
mAbs against the following Ags were used: CD11b (Mac-1 α-chain; M1/70-allophycocyanin and -biotin; BD Pharmingen, San Diego, CA), F4/80 (CI:A3-1-biotinylated; Caltag Laboratories, Burlingame, CA), Ly-6G (1A8-PE; BD Pharmingen), B220 (CD45R; RA3-6B2-PerCP and -FITC; BD Pharmingen), and CD3e (145-2C11-PE; BD Pharmingen). mAbs derived from the following hybridomas were obtained from the American Type Culture Collection (Manassas, VA): CD11a (LFA-1 α-chain), MHC class II (M5/114.15.2), and ICAM-1 (YN1/1.7.4). mAbs against the following Ags were gifts: Ly-6C (ER-MP20) and ER-MP58 (26, 27) (Dr. P. Leenen, Erasmus University, Rotterdam, The Netherlands); c-Fms (M-CSFR; ASF-98) (28) (Dr. S.-I. Nishikawa, Kyoto University, Kyoto, Japan); and CD19 (1D3-PE labeled) and IgK (187.1-biotin labeled) (Dr. S. Nutt, Walter and Eliza Hall Institute). PE-conjugated donkey anti-rat IgG (H+L, F(ab′)2; Jackson ImmunoResearch Laboratories, West Grove, PA), FITC-conjugated anti-rat IgG (Chemicon International, Temecula, CA), or streptavidin-PE (BD Pharmingen) were used as secondary Abs.
TM-induced peritonitis
Peritoneal exudate cells were elicited by i.p. injection of 1 ml of Brewer’s TM (Difco Laboratories, Detroit, MI) as previously described (25, 27). Cells were harvested at various time points after injection by washing cells from the peritoneal cavity by lavage with 5 ml of ice-cold, sterile PBS.
mBSA-induced peritonitis
mBSA-induced peritonitis was induced as previously described (25). Briefly, mice were immunized intradermally in the base of the tail with 100 μg of mBSA (Sigma-Aldrich, St. Louis, MO) emulsified in an equal volume of CFA containing 5 mg/ml heat-killed Mycobacterium tuberculosis (H37 Ra; Difco Laboratories); 14 days later, the primary immunization protocol was repeated as a boost, and after a further 7 days, peritoneal exudate cells were elicited by i.p. injection as above. Cells were harvested at various time points as described above, with day 0 being the day of the i.p. injection eliciting the peritonitis.
Quantification of peritoneal cavity cell infiltration
Total and differential cell counts (Diff-Quik; Lab Aids, Narrabeen, New South Wales, Australia) were performed on the peritoneal exudate cells (25).
T cell proliferation assay
Four days following i.p. challenge with mBSA, spleens were removed and the cells were cultured (5 × 105 cells/well) for 72 h, at 37°C (5% CO2), with 0–100 μg/ml denatured mBSA (boiled for 10 min) in RPMI 1640 containing 50 μΜ 2-ME and 5% (v/v) FCS (200 μl/well). Sixteen hours before harvesting, cells were pulsed with 1 μCi [3H]TdR (Amersham Biosciences, Little Chalfont, U.K.). Cells were harvested using an Inotech (Lansing, MI) cell harvester, and DNA synthesis was measured by [3H]TdR incorporation using a Beckman beta scintillation counter (Beckman Instruments, Irvine, CA). Results are expressed in counts per minute.
Flow cytometry analysis
Cell staining for flow cytometry analysis, including isotype controls and FcR blocking, was conducted as described previously (25). Briefly, cells were incubated with the primary Ab (unlabeled or biotin labeled), followed by an appropriate secondary Ab and directly conjugated Abs (for more than one-color staining). A typical forward- and side-scatter gate was set to exclude dead cells and aggregates; a total of 104 events in the gate were collected and analyzed using a FACSort (BD Biosciences, San Jose, CA).
Phagocytosis assay
Cells were incubated with fluorescent latex Fluospheres (L-5281 carboxylate-modified, 1.0-μm diameter, 2% solids; Molecular Probes, Eugene, OR) precoated with 1% BSA, for 60 min at 37°C, and analyzed by flow cytometry (25). Control cultures with no addition of latex beads were included.
Cell sorting
Cell populations from the peritoneal cavity and T lymphocytes from the spleen were sorted based on their forward-side scatter and cell surface expression using a FACSVantage SE cell sorter (BD Biosciences) (25). For peritoneal exudates, macrophages found in region 1 (R1; see Results for definition) from C57BL/6 mice and R1a (see Results) from GM-CSF−/− mice were sorted based on c-Fms+F4/80+Mac-1high expression. T lymphocytes from BALB/c mice were sorted by negative selection (25).
Mixed leukocyte reaction
The MLR was performed as previously described (25). Purified T cells (2 × 105 cells/well) and macrophages (2 × 105 cells/well) were cultured at 37°C and 5% CO2 for 96 h. [3H]TdR labeling, harvesting, and counting were as described for T cell proliferation.
In vitro LPS stimulation
Peritoneal cells or sorted macrophages (see above) (5 × 105/ml) were cultured in 1 μg/ml LPS (Escherichia coli serotype 0127:B8; Sigma-Aldrich) for 0, 4, and 24 h. Following culture, supernatants were collected and stored at −20°C for the detection of cytokines.
Cytokine ELISAs
IL-1β, IL-6, IL-10, TNF-α, and MCP-1 levels were measured in peritoneal exudate fluids and LPS-stimulated culture supernatants by ELISA (25). For IL-1β and IL-6, the coating and capture Abs were as follows: IL-1β, polyclonal anti-IL-1β Ab and a biotinylated anti-IL-1β mAb (Endogen, Woburn, MA); IL-6, anti-IL-6 mAb (20F3.11; American Type Culture Collection) and a biotinylated anti-IL-6 mAb (MP5-32C11; BD Pharmingen). To detect the cytokines, a streptavidin-HRP conjugate (BD Pharmingen), followed by tetramethylbenzidine-peroxidase substrate (Kirkegaard and Perry Laboratories, Gaithersburg, MD), was used. A standard curve was constructed using serial dilutions of purified IL-1β or IL-6, starting at a concentration of 2 ng/ml. For MCP-1, TNF-α, and IL-10, the appropriate mouse OptEIA Set (BD Pharmingen) was used according to the manufacturer’s instructions. Each ELISA was sensitive down to 7 pg/ml.
Intracellular cytokine staining
Intracellular IL-6 levels were measured in LPS-stimulated mBSA-induced peritoneal cells. Brefeldin A (10 μg/ml; Sigma-Aldrich) was added to the cultures for the final 2 h of LPS stimulation to limit cytokine secretion (29). Following FcR blocking, cells were stained for CD11b (Mac-1 α-chain; M1/70-allophycocyanin) as above, fixed in 4% paraformaldehyde, and permeabilized using 0.5% saponin and 2% FCS in PBS. Cells were subsequently stained with biotinylated anti-IL-6 mAb (MP5-32C11; BD Pharmingen), and detected using a streptavidin-PE (BD Pharmingen) conjugate. Appropriate isotype controls were used. Samples were analyzed by flow cytometry as described above. Results are presented as the geometric mean fluorescence intensity (GMFI), a measure of the amount of cytokine being produced per cell (30).
Statistics
Data are presented as mean ± SEM for indicated experiments. Differences between GM-CSF−/− mice and C57BL/6 mice were performed using the Student’s two-tailed t test. A value of p ≤ 0.05 was considered statistically significant.
Results
Induction of an Ag-specific peritonitis in GM-CSF−/− mice
We confirmed and extended previous findings showing that the inflammatory reaction in the peritoneal cavity of GM-CSF−/− mice following TM injection was no different from that seen in C57BL/6 mice in terms of numbers and types of cells present (Fig. 1⇓). Both the initial, predominantly PMN, influx and the subsequent macrophage one were similar despite the lack of GM-CSF, although the macrophages were, on average, smaller in GM-CSF−/− mice compared with those in C57BL/6 mice (Fig. 1⇓B).
The development of TM-induced peritonitis in GM-CSF−/− vs C57BL/6 mice. A, Total peritoneal cell numbers over time. B, Diff-Quik stains of 4-day TM-induced exudate cells. For A, results are expressed as the mean ± SEM for four independent experiments.
We next assessed the mBSA-dependent peritonitis in GM-CSF−/− mice. On day 1 post-mBSA challenge, the number of peritoneal exudate cells increased from 4.5 ± 0.3 × 106 on day 0 (the day of mBSA i.p. challenge) to 13.6 ± 1.2 × 106, which was similar to that seen for C57BL/6 mice (4.6 ± 0.6 to 11.6 ± 0.4 (×106)) (Fig. 2⇓A). However, although the number of cells in the peritoneal cavity of C57BL/6 mice continued to increase, peaking 3–4 days post-mBSA challenge, the number of cells in the peritoneal cavity of GM-CSF−/− mice remained relatively constant on days 1–4 post-mBSA challenge (Fig. 2⇓A). The number of peritoneal cells gradually returned to resident levels by day 20 post-mBSA challenge in both C57BL/6 and GM-CSF−/− mice; however, there were still significantly more inflammatory exudate cells in C57BL/6 mice compared with GM-CSF−/− mice on days 7 and 14 post-mBSA challenge (Fig. 2⇓A). Saline challenge failed to induce a cellular infiltrate in either GM-CSF−/− or C57BL/6 mice.
The development of mBSA-induced peritonitis in GM-CSF−/− vs C57BL/6 mice. A, Total peritoneal cell numbers over time. B and C, Total number of PMN (B) and total number of macrophages (C), counted using Diff-Quik stains. D, Diff-Quik stains of 4-day mBSA-induced exudate cells. For A–C, results are expressed as the mean ± SEM for four independent experiments.
Morphologic characterization of cells during mBSA-induced peritonitis in GM-CSF−/− mice
On day 1 post-mBSA challenge, there was no difference in the number of PMN present in the peritoneal cavity of GM-CSF−/− compared with C57BL/6 mice (5.0 ± 0.3 vs 5.0 ± 0.8 (×106), respectively) (Fig. 2⇑B). However, by day 2 post-mBSA challenge, although there was an increase in the number of PMN in C57BL/6 mice, there was a significant decrease in PMN numbers in GM-CSF−/− mice (Fig. 2⇑B).
By days 3–4 post-mBSA challenge, the predominant cell type in C57BL/6 mice is the macrophage, with a rapid increase in their numbers from days 2 to 4 (Fig. 2⇑C) (25). However, as assessed by morphology, this influx of macrophages on days 2 to 4 was dramatically reduced in GM-CSF−/− mice, with less than half the number of macrophages present in the peritoneal cavity of GM-CSF−/− mice on day 4 postchallenge compared with C57BL/6 mice (8.7 ± 0.7 vs 18.4 ± 4 (×106), respectively) (Fig. 2⇑C). Furthermore, the macrophages present in 4-day mBSA exudates from GM-CSF−/− mice were smaller, less vacuolated, and less foam-like compared with C57BL/6 mice (Fig. 2⇑D), noting again that the macrophages from C57BL/6 mice were not particularly large and foam-like when compared with TM-elicited macrophages (compare Figs. 1⇑B and 2⇑D) (25).
The number of lymphocytes did not significantly differ between GM-CSF−/− and C57BL/6 mice at any time point during the mBSA peritonitis (data not shown). We have previously reported that eosinophils enter the peritoneal cavity following mBSA-induced peritonitis (25). Fewer eosinophils entered the peritoneal cavity following mBSA challenge in GM-CSF−/− mice compared with C57BL/6 mice; the kinetics of the eosinophil response also differed with a more gradual increase over the first 4 days post-mBSA challenge, followed by a more rapid decline in the numbers of eosinophils in GM-CSF−/− mice compared with C57BL/6 mice (data not shown).
Cellular response to mBSA in GM-CSF−/− mice
Because GM-CSF is implicated in the development of APC (6, 14) and mBSA is a T cell-dependent Ag (31), the in vitro T cell response to mBSA in GM-CSF−/− mice was examined. Spleen cells were isolated from mice 4 days post i.p. challenge with mBSA, i.e., 25 days post the initial primary immunization. There was no difference in the proliferative response of T cells to mBSA from GM-CSF−/− mice compared with C57BL/6 mice (Fig. 3⇓). These data suggest that GM-CSF is not involved in the Ag-priming phase of mBSA-induced peritonitis.
Effect of GM-CSF deficiency on mBSA-specific T cell response. Spleen cells from mice 4 days post i.p. challenge with mBSA were cultured in the presence of increasing concentrations (0–100 μg/ml) of denatured mBSA for 72 h. Cultures were pulsed with [3H]TdR for the last 16 h (see Materials and Methods). Results are expressed as the amount of [3H]TdR incorporation (mean ± SEM) from eight independent samples from two experiments.
Properties of macrophage-lineage cells elicited during mBSA-induced peritonitis in GM-CSF−/− mice
From the morphologic analysis above, there was a dramatic reduction in the number of macrophages post-mBSA challenge in GM-CSF−/− mice compared with C57BL/6 mice. The effect of a lack of GM-CSF expression on several parameters of the macrophage lineage infiltrate in 4-day mBSA-induced peritoneal exudates was next studied.
Forward-side scatter profile and lineage markers
As previously reported by ourselves (25, 27), c-Fms (CSF-1R, M-CSFR) is a useful surface marker for defining macrophage-lineage cells in vivo, including during inflammatory reactions. Using predominantly size and granularity (forward-side scatter by flow cytometry), in combination with c-Fms, F4/80, and Mac-1 expression to detect macrophage-lineage cells, we previously defined four regions to help analyze the different cell populations in resident and stimulated murine peritoneal cavities (25). The four regions for the forward-side scatter profile of a 4-day post-mBSA exudate from C57BL/6 mice are reproduced in Fig. 4⇓A (see also Ref.25). As before (25), macrophages are found in R1, with smaller macrophages being in R2 with B1 lymphocytes; B and T lymphocytes are predominantly in R3 and eosinophils are in R4.
Flow cytometric analysis of 4-day mBSA-induced peritonitis exudate cells in C57BL/6 mice (A) and GM-CSF−/− mice (B). On the basis of c-Fms, F4/80 and Mac-1 expression, as well as size and granularity (forward-side scatter profile), four regions (R1–R4) were defined (25 ). Cells in R1 are predominately c-Fms+ macrophages; cells in R2 are c-Fms−/low and consist of macrophages and B1 lymphocytes; cells in R3 are c-Fms− and consist of lymphocytes; cells in R4 are predominately c-Fms− eosinophils (25 ). Cells in R1 from GM-CSF−/− mice are further subdivided into R1a and R1b, based on Mac-1high and Mac-1low expression, respectively (see Fig. 5⇓A). R1a cells are predominately c-Fms+ macrophages, and R1b cells are c-Fms− B1 lymphocytes (see Fig. 5⇓B). The percentages shown for each region are averages from four independent experiments.
For GM-CSF−/− mice, using these same parameters, cells from the resident peritoneal cavity could be allocated to the same regions and with a similar distribution to that found for the equivalent cells from C57BL/6 mice (data not shown) (25). However, at 4 days post i.p. challenge with mBSA, there were differences in the exudate populations between GM-CSF−/− and C57BL/6 mice, in line with what was observed with the morphologic analysis. These differences are described below for the myeloid cell-containing regions defined above, namely R1, -2, and -4.
R1 (macrophage rich).
On day 4 post-mBSA challenge, only 25 ± 2% of the exudate cells from GM-CSF−/− mice were found in R1 compared with 39 ± 2% of exudate cells from C57BL/6 mice (p < 0.001) (Fig. 4⇑). The reduced percentage of cells in R1 from GM-CSF−/− mice represented a significant reduction in the number of cells in this region (2.2 ± 0.2 vs 9.4 ± 0.6 (×106); p < 0.001; GM-CSF−/− vs C57BL/6 mice, respectively). Also, the majority of the R1 cells from the GM-CSF−/− exudates were smaller and less granular. As discussed above, we have previously shown (25) that the cells in R1 from C57BL/6 mice are predominantly macrophages (Mac-1highc-Fms+F4/80+) (Table I⇓). In GM-CSF−/− mice, the cells in R1 were also Mac-1+ (Table I⇓); however, a lower percentage of cells were positive for c-Fms and F4/80 (Table I⇓). In C57BL/6 mice, R1 cells had high levels of Mac-1 expression (Mac-1high), whereas in GM-CSF−/− mice, only 66 ± 4% of cells were Mac-1high, the rest (30 ± 4%) having lower levels of Mac-1 expression (Mac-1low) and being similar to those seen on cells in R2 (see below) (25). Also, by forward-side scatter, the Mac-1low cells were less granular than the Mac-1high cells. Therefore, to aid further analysis of the R1 population in GM-CSF−/− mice, the Mac-1high and Mac-1low populations were arbitrarily subdivided into R1a and R1b regions, respectively, for these mice (Fig. 4⇑B).
Peritoneal cell surface marker distribution within R1a
Mac-1high (R1a) cells from GM-CSF−/− mice were predominantly c-Fms+F4/80+ and B220−CD19−IgK− (Table II⇓), consistent with a macrophage phenotype and similar to R1 cells in C57BL/6 mice (see Table I⇑). In contrast, Mac-1low (R1b) cells from GM-CSF−/− exudates were predominantly c-Fms−F4/80− but B220+CD19+IgK+ (Table II⇓), and hence similar to B1 lymphocytes found in R2 (see below) (25). The presence of this latter population accounted for the lower overall percentage of R1 cells positive for c-Fms and F4/80 and higher percentage of R1 cells positive for B220, CD19, and IgK in GM-CSF−/− mice (Table I⇑).
Peritoneal cell surface marker distribution within R1a and R1b of GM-CSF−/− micea
Ly-6C (ER-MP20) and ER-MP58, which are expressed on immature cells of the macrophage lineage (26, 27), were present on a very low proportion of R1 cells for GM-CSF−/− mice, as we reported previously for C57BL/6 mice (data not shown) (25). All R1 cells were negative for the PMN marker, Ly-6G, and the DC marker, CD11c (data not shown) (25).
In conclusion, the fewer macrophages present in the 4-day post-mBSA exudate from GM-CSF−/− mice are smaller and less granular; also, a population of B1 lymphocytes appears to be more prominent (see Discussion).
R2 (macrophage and B1 lymphocyte rich) and R4 (eosinophil rich).
We have previously reported (25) for C57BL/6 mice, on day 4 post-mBSA challenge, that, in R2, both macrophages (c-Fms+F4/80+Mac-1low) and B1 lymphocytes (Mac-1lowB220+CD19+IgK+) are present. We now find, at this time, there appeared to be an increase in the percentage of cells in this region for GM-CSF−/− cells compared with C57BL/6 mice (26 ± 1 vs 20 ± 2%, respectively; p = 0.05) (Fig. 4⇑). In GM-CSF−/− mice, there was a lower percentage of c-Fms+ cells in R2 from 4-day mBSA-induced exudates compared with C57BL/6 mice (17 ± 2 vs 34 ± 4%, respectively; p < 0.05) equating to a significant reduction in the number of macrophages in this region (0.4 ± 0.1 vs 1.7 ± 0.2 (×106), GM-CSF−/− vs C57BL/6 mice, respectively; p < 0.001) and accounting for the reduced cell number (2.6 ± 0.1 vs 4.3 ± 0.9 (×106), respectively; p < 0.001). In contrast, in GM-CSF−/− mice, there was a higher percentage of B220+ cells in R2 from 4-day mBSA-induced exudates compared with C57BL/6 mice (48 ± 4 vs 22 ± 4%; p < 0.001), which equates to similar numbers of B1 lymphocytes in this region for GM-CSF−/− and C57BL/6 mice (1.25 ± 0.2 vs 0.9 ± 0.2 (×106), respectively).
By morphology, cells in R4 were previously found to be predominately eosinophils in C57BL/6 mice (25), and this is also the case for GM-CSF−/− mice. They represent 7 ± 1 and 9 ± 1% of the exudate populations from C57BL/6 (Fig. 4⇑A) and GM-CSF−/− mice (Fig. 4⇑B), respectively. However, even though somewhat variable, the eosinophil numbers post-mBSA challenge in GM-CSF−/− mice were overall less than those observed in C57BL/6 mice (data not shown), leading to a higher proportion of macrophages in R4 on day 4 post-mBSA challenge as judged by c-Fms expression (37 ± 4 vs 25 ± 5%; p = 0.05; GM-CSF−/− vs C57BL/6 mice, respectively).
MHC class II and integrin expression
There was MHC class II expression on both macrophages and B lymphocytes. There was no significant difference in the percentage of R1 cells expressing class II following mBSA-induced peritonitis in GM-CSF−/− mice compared with C57BL/6 mice (75 ± 9 vs 80 ± 6%, GM-CSF−/− vs C57BL/6, respectively) (Table I⇑). There was a similar percentage of macrophages (R1a) and the B1 lymphocytes (R1b) from GM-CSF−/− mice positive for MHC class II (Table II⇑), and they expressed similar levels.
Given that the absence of GM-CSF results in fewer macrophages being present in mBSA-induced exudates, we examined integrin expression because of their likely involvement in trafficking/adhesion. In R1, there was a trend for fewer ICAM+ cells in GM-CSF−/− mice (45 ± 10 vs 54 ± 15%, GM-CSF−/− vs C57BL/6, respectively) (Table I⇑). In GM-CSF−/− mice, the ICAM+ cells were present in R1a; cells in R1b were essentially ICAM-1− (Table II⇑). Similar observations were found in R2 (data not shown). For LFA-1, there was a significant increase in the percentage of R1-positive cells from GM-CSF−/− mice (48 ± 5 vs 31 ± 9%; p < 0.05; GM-CSF−/− vs C57BL/6 mice, respectively) (Table I⇑). This was due to a higher percentage of B1 lymphocytes (R1b cells) expressing LFA-1 compared with macrophages (R1a cells) from GM-CSF−/− mice (Table II⇑). When positive, the level of expression of LFA-1 on R1 macrophages from C57BL/6 mice was higher per cell than for GM-CSF−/− mice (data not shown).
In summary, for these markers, the major difference in R1 seems to be in LFA-1 expression, reflecting the increased proportion of B1 lymphocytes.
Phagocytosis
R1a macrophages from mBSA-induced GM-CSF−/− exudates were just as capable as the corresponding R1 macrophages from C57BL/6 exudates at phagocytosing fluorescent latex beads when monitored by flow cytometry (Fig. 5⇓A) and fluorescence microscopy (data not shown). However, they were less efficient at ingesting multiple beads (Fig. 5⇓A), an observation in line with their smaller size. R1b cells (B1 lymphocytes) from GM-CSF−/− mice were not phagocytic.
Functional capacity of macrophages from 4-day mBSA- and 4-day TM-induced peritonitis exudates in GM-CSF−/− vs C57BL/6 mice. A, Phagocytic capability of R1 (C57BL/6 mice) or R1a (GM-CSF−/− mice) cells. Results are expressed as the percentage of cells capable of phagocytosing the latex beads (total), as well as the percentage capable of phagocytosing ≤3 or >3 beads. Results are expressed as the mean ± SEM from three independent experiments. Note that R1b cells from GM-CSF−/− were not phagocytic. B, Lymphocyte stimulatory ability of macrophages. Sorted macrophage populations (R1 for C57BL/6 mice, or R1a for GM-CSF−/− mice) (2 × 105 cells) (see Materials and Methods) were incubated with allogeneic splenic T cells (2 × 105 cells) for 96 h at 37°C and 5% CO2. Cultures were pulsed with [3H]TdR for the last 16 h. Control cultures contained medium or T cells alone. Results are expressed as the amount of [3H]TdR incorporation (mean ± SEM) from eight independent samples from two experiments. ∗, p < 0.001, GM-CSF−/− vs C57BL/6. ∗∗, p < 0.01, GM-CSF−/− vs C57BL/6.
Mixed leukocyte reaction
We previously reported that C57BL/6 macrophages from 4-day mBSA-induced peritoneal exudates were able to stimulate allogeneic T lymphocytes, whereas macrophages from 4-day TM-induced exudates were not (25). As shown above (Fig. 3⇑), the Ag-specific T cell response to mBSA developed normally in GM-CSF−/− mice when APC from the spleen were used. We next determined whether GM-CSF−/− macrophages from 4-day mBSA-induced peritoneal exudates were also able to stimulate allogeneic T lymphocytes. When purified T cells were incubated with equal numbers of GM-CSF−/− R1a macrophages from 4-day mBSA-induced exudates, they were able to proliferate (Fig. 5⇑B); however, their proliferative capacity was reduced compared with that of comparable numbers of R1 macrophages from C57BL/6 mice (Fig. 5⇑B). As for C57BL/6 macrophages from 4-day TM-induced exudates, GM-CSF−/− R1a macrophages from 4-day TM-induced exudates did not induce T cells to proliferate (Fig. 5⇑B).
Cytokine production from 4-day mBSA-induced exudate cells
TNF-α, IL-1β, IL-6, MCP-1, and IL-10 could not be detected in peritoneal exudate lavage fluid following either 4-day mBSA- or 4-day TM-induced peritonitis (data not shown). However, low levels of IL-1β, IL-6, and MCP-1 could be detected in culture supernatants from 4-day mBSA-induced exudate cells from C57BL/6 and GM-CSF−/− mice (Fig. 6⇓A), with IL-1β levels being significantly higher in C57BL/6 vs GM-CSF−/− culture supernatants. GM-CSF, by itself, is a poor stimulator of the commonly secreted cytokines; however, it can prime, in vivo and in vitro, monocytes/macrophages to up-regulate inflammatory cytokine production in response to a stimulus such as LPS (2, 3, 7, 8). Therefore, we measured the GM-CSF dependence of IL-6, IL-1β, TNF-α, MCP-1, and IL-10 production from peritoneal exudate cells in response to LPS. Following LPS stimulation in vitro (4 and 24 h), 4-day mBSA-induced exudate cells from GM-CSF−/− mice secreted less TNF-α, IL-1β, MCP-1, and IL-6 compared with 4-day mBSA-induced exudate cells from C57BL/6 cells (Fig. 6⇓A). However, there was no significant difference in the production of the anti-inflammatory cytokine IL-10 (Fig. 6⇓A).
Cytokine expression in vitro in response to LPS by cells from GM-CSF−/− vs C57BL/6 mice. A, Levels of secreted TNF-α, IL-1β, MCP-1, IL-10, and IL-6 from 4-day mBSA-induced exudate cells (5 × 105 cells/ml), unstimulated (−) or stimulated with 1 μg/ml LPS (+) in vitro for 4 h (TNF-α) or 24 h (IL-1β, MCP-1, IL-10, IL-6), as measured by ELISA. Results are expressed as the mean ± SEM from four independent experiments. For TNF-α, IL-1β, MCP-1, and IL-10, cytokine levels are shown on the left y-axis; for IL-6, on the right y-axis. B, Surface Mac-1 and intracellular IL-6 expression by 4-day mBSA-induced exudate cells from GM-CSF−/− vs C57BL/6 mice following a 24 h in vitro LPS stimulation. Data shown are for individual mice and are representative of three independent experiments. C, Level of intracellular IL-6 staining in macrophages. Mac-1high macrophages which produced IL-6 in B were gated, and results are expressed as the GMFI following subtraction of the GMFI of the control. D, Levels of secreted TNF-α, IL-1β, MCP-1, IL-10, and IL-6 from 4-day mBSA-induced macrophages (5 × 105 cells/ml) stimulated with 1 μg/ml LPS in vitro for 4 h (TNF-α) or 24 h (IL-1β, MCP-1, IL-10, IL-6), as measured by ELISA. Results are expressed as the mean ± SEM. For TNF-α, IL-1β, IL-10, and MCP-1, cytokine levels are shown on the left y-axis; for IL-6, on the right y-axis. R1 exudate cells from C57BL/6 mice and R1a exudate cells from GM-CSF−/− mice were sorted before in vitro LPS stimulation. ∗, p < 0.01, GM-CSF−/− vs C57BL/6. ∗∗, p < 0.05, GM-CSF−/− vs C57BL/6.
We next determined whether the macrophages were the major source of IL-1β, IL-6, TNF-α, and MCP-1 in the LPS-stimulated exudates. First, 4-day mBSA-induced exudates, following LPS stimulation in vitro for 24 h, were stained for Mac-1 and intracellular IL-6. Fig. 6⇑B shows that, for both C57BL/6 and GM-CSF−/− mice, the Mac-1high cells from 4-day mBSA-induced exudates were the major producers of IL-6. These correspond to the macrophages found in R1 (see Fig. 4⇑). Mac-1low cells, comprised predominantly of B1 lymphocytes, eosinophils, and some smaller macrophages, were essentially negative for IL-6 production (Fig. 6⇑B). The lower cytokine levels in the GM-CSF−/− supernatants following LPS stimulation (Fig. 6⇑A) could therefore be explained by the fact that the macrophages were the major source of these cytokines, and that the percentage of macrophages present in 4-day mBSA-induced exudates was lower in GM-CSF−/− mice compared with C57BL/6 mice. However, as shown for IL-6 (Fig. 6⇑C), there was less accumulation of this cytokine in the macrophages from GM-CSF−/− mice, as measured by the GMFI, compared with the macrophages from C57BL/6 mice (GMFI, 35 ± 6 vs 16 ± 5; p = 0.03; GM-CSF−/− vs C57BL/6, respectively).
Additionally, to demonstrate that the macrophages from GM-CSF−/− exudates were in fact secreting less cytokines per cell than the macrophages from C57BL/6 exudates, R1 or R1a macrophages were sorted from 4-day mBSA-induced C57BL/6 and GM-CSF−/− exudates, respectively, and a constant number of cells (5 × 105 cells/ml) were stimulated with LPS in vitro. Cytokine levels were measured in the supernatants 4 and 24 h post-LPS stimulation as above. As shown in Fig. 6⇑D, the macrophages from 4-day mBSA-induced exudates from GM-CSF−/− mice secreted significantly less TNF-α,IL-1β, MCP-1, and IL-6 following LPS stimulation compared with the macrophages from 4-day mBSA-induced exudates from C57BL/6 mice. The non-R1 and non-R1a cells, collected during the sorting for the macrophages, were also stimulated with LPS. Very low levels of cytokines could be detected in the supernatants, most likely arising from some contaminating macrophages (data not shown). Once again, there was no difference in the amount of IL-10 secreted by the macrophages from GM-CSF−/− vs C57BL/6 mice following in vitro LPS stimulation (Fig. 6⇑D).
By way of comparison, following LPS stimulation in vitro (24 h), 4-day TM-induced exudate cells from GM-CSF−/− mice secreted less MCP-1 compared with 4-day mBSA-induced exudate cells from C57BL/6 cells (6.0 ± 0.4 vs 1.1 ± 0.03, C57BL/6 vs GM-CSF−/−, respectively; p < 0.01). However, there were no differences between the levels of IL-6, TNF-α, and IL-1β secreted from the exudate cells from these mice (data not shown).
Discussion
GM-CSF has been shown in mice to be critical for the pathogenesis of a number of Ag-driven autoimmune/inflammatory diseases (18, 19, 20, 21, 22, 23). It can also exacerbate this type of condition in humans (32). In inflammation models relying on Ag priming and challenge to initiate the effector phase, the absence of GM-CSF, or reduction in its levels, resulted in a dramatic reduction in the cellular infiltrate (18, 20, 21). GM-CSF activity could be important at either or both of these Ag-dependent stages. Data from these disease models suggest that it is unlikely that inefficient activation of T cells by DC in the periphery would account for the observed resistance to disease in GM-CSF−/− mice (18, 20, 21), and data using timed addition of a neutralizing anti-GM-CSF mAb in both the collagen-induced arthritis (20) and experimental autoimmune encephalomyelitis (21) models were consistent with the key function(s) of GM-CSF being at the later (effector) stage, i.e., following Ag challenge. Such a function(s) is subsequent to T cell activation, and most probably affects mediators of inflammation (e.g., cytokines). Inflammatory mediator levels were lowered at the sites of inflammation when GM-CSF was absent or its levels lowered (20). However, from these earlier studies, the cellular composition, as well as the relative contribution of GM-CSF to cell number and/or activation, could not be easily delineated.
Prior studies involving TM-induced peritonitis in GM-CSF−/− mice failed to find a role for GM-CSF, leading to the conclusion that it is not involved in sterile peritonitis (24); other studies examining the peritoneal cavity response in these mice were complicated by the use of infectious organisms as a stimulus for the peritoneal reaction (33, 34). The above findings indicate that our Ag-induced peritonitis model (25) is a convenient one that allows us to explore the role of GM-CSF in Ag-driven autoimmune/inflammatory diseases at a cellular and molecular level. In contrast to TM injection, the mBSA-induced peritonitis in GM-CSF−/− mice was very different to that found in wild-type mice. Although a normal mBSA-specific T cell response developed in GM-CSF−/− mice following mBSA-induced peritonitis, the inflammatory reaction in the peritoneal cavity was severely compromised, reminiscent of what has been described in a number of murine models of disease (18, 19, 20, 21, 22, 23). There was an inability to sustain PMN, macrophage, and eosinophil numbers in GM-CSF−/− mice, resulting in a large decrease in the number of inflammatory cells present in these mice. These reduced myeloid cell numbers are consistent with GM-CSF being involved in myeloid cell survival, or even local proliferation, thus maintaining cells at the site of inflammation (35, 36, 37).
In addition to an effect on their numbers, the macrophages present in GM-CSF−/− mice following mBSA-induced peritonitis were smaller and less granular overall, less efficient at phagocytosis and at stimulating an MLR, and secreted less proinflammatory mediators in vitro in response to LPS. They also expressed lower levels of LFA-1, although Mac-1, MHC class II, and ICAM-1 expression levels were similar to wild-type mice. These differences suggest that GM-CSF is also involved in the priming/activation of the inflammatory cells within the inflamed peritoneum. There are a number of reports implicating GM-CSF in priming/activating cells for heightened immune/inflammatory responses (2, 3, 6, 7, 8, 9, 10, 35, 38, 39). One such consequence is that they produce more inflammatory cytokines, such as IL-1, IL-6, TNF-α, and MCP-1. Therefore, a lack of GM-CSF results in less of these cytokines being produced and, consequently, reduced numbers of cells at the inflammatory site, as the current results show in the mBSA-induced peritonitis model in GM-CSF−/− mice. These effects would not appear to be due to an increased production of the anti-inflammatory cytokine IL-10 in macrophages from GM-CSF−/− mice following LPS stimulation.
GM-CSF has been implicated in APC development (6, 14), and it has been reported that the source of the APC is important in determining whether an impaired T cell-proliferative response is seen to a specific Ag in GM-CSF−/− mice (13). GM-CSF is also the most commonly used cytokine for in vitro differentiation of DC (14, 15); however, GM-CSF−/− mice have only slightly reduced numbers of DC (40), thus raising doubt as to its in vivo role in the generation of DC. In support of this, it has been pointed out in a recent review that in vitro DCs, which are often generated with protocols involving GM-CSF, can be quite different from their in vivo counterparts, suggesting that there may be other cytokines that are important in vivo, e.g., Flt3 ligand (41). Our results above show that GM-CSF is not involved in the development of the APC in the spleen responsible for presentation of mBSA to primed T cells. We also found a similar result with type II collagen as the Ag (20). However, when the mBSA-induced peritoneal macrophages from GM-CSF−/− mice were used in an MLR, they were less efficient at stimulating T cell proliferation compared with the mBSA-induced peritoneal macrophages from C57BL/6 mice. We have previously shown that the mBSA-induced macrophages from C57BL/6 mice appear to take on a DC-like role, expressing higher levels of class II and being efficient stimulators of allogeneic MLR (25); however, they do not appear to be fully differentiated DC based on morphology or phenotype, and they still retain macrophage functions. Thus, although the mBSA-induced peritoneal macrophages from C57BL/6 were capable of stimulating T cell proliferation, those from GM-CSF−/− mice were less effective, in line with the cells being less efficient in a number of other functional assays described above.
Apart from the dramatic decrease in the macrophage response, 4-day mBSA-induced exudates from GM-CSF−/− mice also contained a population of cells that were large but not very granular (R1b cells; Fig. 4⇑B). By cell surface marker analysis, these were deemed to be B1 lymphocytes. They were larger than the B1 lymphocytes found in the unstimulated peritoneal cavity where they are a major population of B cells (42). It has previously been suggested that B1 lymphocytes and certain macrophage populations may have a common progenitor (43). It is possible that, in the absence of GM-CSF, there is a block in macrophage lineage development, thereby favoring B1 lymphocyte progression at the expense of the macrophage. In this connection, s.c. administration of GM-CSF led to an increase in the numbers of CD5+ macrophages in the peritoneal cavity (43). In vitro studies support this concept (43, 44).
In summary, TM-induced peritonitis has been the standard method for providing large numbers of elicited macrophages for study and for determining the importance of a particular mediator(s) in the development of an inflammatory reaction. As we previously suggested (25), this type of response may not be relevant to many inflammatory diseases, particularly where an acquired immune response is important. We showed above that, for sterile murine peritonitis, a GM-CSF dependence was observed only for an Ag-specific response, involving priming and challenge. An important general conclusion from our studies here is that any assessment of the inflammatory mediator(s) critical for the maintenance and progression of an inflammatory reaction must take into account the type of stimulus or insult that the host is responding to. We contend that the results presented in this paper contribute to our understanding of the suppressed inflammatory response observed under conditions of GM-CSF deprivation in various inflammation/autoimmunity models (18, 19, 20, 21, 22, 23).
Acknowledgments
We thank Jennifer Davis for assistance with the maintenance and care of the mice, Alice Holloway for cell sorting, and Amanda Turner for technical assistance.
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 supported by the National Health and Medical Research Council of Australia and the Cooperative Research Centre for Chronic Inflammatory Diseases.
-
↵2 Address correspondence and reprint requests to Dr. Andrew D. Cook, Department of Medicine, Royal Melbourne Hospital, University of Melbourne, Victoria, Australia, 3010. E-mail address: adcook{at}unimelb.edu.au
-
↵3 Abbreviations used in this paper: DC, dendritic cell; TM, thioglycolate medium; PMN, polymorphonuclear neutrophil; mBSA, methylated BSA; R1–R4, regions 1–4; GMFI, geometric mean fluorescence intensity.
- Received January 30, 2004.
- Accepted July 20, 2004.
- Copyright © 2004 by The American Association of Immunologists
References
In this issue
