The role played by resident macrophages (Mφ) in the initiation of peritoneal inflammation is currently unclear. We have used a conditional Mφ ablation strategy to determine the role of resident peritoneal Mφ in the regulation of neutrophil (PMN) recruitment in experimental peritonitis. We developed a novel conditional Mφ ablation transgenic mouse (designated CD11bDTR) based upon CD11b promoter-mediated expression of the human diphtheria toxin (DT) receptor. The murine DT receptor binds DT poorly such that expression of the human receptor confers toxin sensitivity. Intraperitoneal injection of minute (nanogram) doses of DT results in rapid and marked ablation of F4/80-positive Mφ populations in the peritoneum as well as the kidney, and ovary. In experimental peritonitis, resident Mφ ablation resulted in a dramatic attenuation of PMN infiltration that was rescued by the adoptive transfer of resident nontransgenic Mφ. Attenuation of PMN infiltration was associated with diminished CXC chemokine production at 1 h. These studies indicate a key role for resident peritoneal Mφ in sensing perturbation to the peritoneal microenvironment and regulating PMN infiltration.
Macrophages (Mφ)7 are dispersed throughout the tissues and have an important role in innate immunity (1), apoptotic cell clearance (2), development (3), and morphogenesis (4, 5). Resident tissue Mφ and dendritic cells are regarded as sentinels of the innate immune system. Various strategies, such as Mφ depletion induced by administration of liposomal clodronate (6), have been used to examine Mφ function in vivo. However, previous work using this method to study the role of resident peritoneal Mφ in experimental peritonitis has produced conflicting results because they have suggested a key role (7), an inhibitory role (8), or no role at all (8). In the current study, we have used a conditional ablation strategy to clarify the role of resident Mφ in the sensing of peritoneal injury and the regulation of neutrophil (PMN) infiltration.
A conditional Mφ ablation strategy has advantages over the available naturally occurring and induced Mφ-deficient mutant mice because the timing of Mφ elimination can be chosen. Despite their limitations, nonconditional Mφ-deficient mice have proven valuable for analysis. For example, the Csfmop/Csfmop (osteopetrosis) mouse is a naturally occurring mutant of the CSF-1 gene and exhibits a Mφ deficiency at a level that permits viability in homozygotes (9). These mice have provided insight into Mφ function during development (3). In addition, mice targeted at the PU.1 locus exhibit multiple defects in development of hemopoietic lineages including a complete absence of tissue Mφ (10, 11). Although PU.1 mutation results in perinatal lethality, these mice have been used to demonstrate that mesenchymal cells are able to clear apoptotic cells during embryonic regression of interdigital tissues (12).
Previous strategies used to eliminate specific cell types in a living organism include the generation of transgenic lines that express diphtheria toxin A-chain (13, 14) or the ricin polypeptide (15). However, even low levels of unanticipated transgene expression can give unpredictable consequences (13). The alternative ablation strategy of killing thymidine kinase-expressing cells with gancyclovir (16, 17) only permits the elimination of proliferating cells. More recently, conditional Mφ ablation has been achieved using transgenic expression of Fas under the control of the c-fms promoter coupled with drug-inducible Fas dimerization to induce cell death (18). Identification of the human receptor for diphtheria toxin (DT) (also known as heparin-binding epidermal growth factor (hbEGF) (19)) created an opportunity for a unique ablation strategy. The murine form of hbEGF binds DT poorly, but mouse cells can be rendered sensitive through transgenic expression of human hbEGF. In transgenic mice expressing human hbEGF lineage specifically, cell ablation results following toxin injection. In addition, DT is a protein synthesis inhibitor and kills both mitotic and terminally differentiated cells. This strategy has recently been used to generate transgenic mice in which hepatocytes (20) or dendritic cells (21) may be conditionally ablated. In the current report, we describe how we have used this strategy to generate a conditional Mφ ablation mouse to assess the role of resident peritoneal Mφ in the initiation of acute inflammation in experimental sterile peritonitis induced by Brewer’s thioglycolate (BTG). We demonstrate that resident peritoneal Mφ are essential for PMN recruitment through Mφ-dependent CXC chemokine production.
Materials and Methods
The CD11b promoter from coordinates −1704 to + 83 (22) was used to drive expression of the human hbEGF cDNA (19). Splicing and polyadenylation signals were provided by a region of the human growth hormone gene that had previously worked effectively with the CD11b promoter (22). The fusion protein between hbEGF and GFP was generated by continuing the open reading frame from the final residue of hbEGF with the first residue of GFP.
The CD11b-DTR construct was used to generate transgenic mice on the FVB/N background (23) using conventional techniques; transgene expression was detected using an RT-PCR assay. The primer sequences used were 5′-AAGATCCGCCACAACATCG for the forward primer and 5′-GCAGCTCTAGGTTGGATTTCTG for the reverse primer. Because the reverse primer sequence was taken from the base pairs flanking intron III of hGH, no PCR product could be amplified from genomic DNA.
Flow cytometry analysis
Bone marrow-derived Mφ were prepared as previously described (4). Resident peritoneal cells were isolated by peritoneal lavage. Elicited peritoneal Mφ were lavaged from the peritoneal cavity 4 days after i.p. injection of 1 ml of 3% BTG (Difco). For flow cytometric analysis, ∼2 × 106 cells were incubated for 10 min in FACS buffer containing rat anti-mouse FcR blocker CD16/CD32 (FcγRIII/II). FITC- or PE-conjugated Abs at concentrations suggested by the supplier were then added, and the mixture was incubated at 4°C for 45–60 min. The cells were then washed and fixed in 2% formaldehyde before undergoing flow cytometry analysis. A fixed volume of whole blood was obtained by tail vein bleed into 3.9% sodium citrate. Blood was aliquoted into flow cytometry tubes, blocked, and stained as for peritoneal lavage cells. One milliliter of FACSLysis buffer (BD Biosciences) was then added to lyse erythrocytes, and samples were spun and processed by flow cytometry. The Abs used were anti-CD11b FITC conjugate (eBiosciences), anti-GR1 PE (eBiosciences), anti-B220 (mouse CD45R) FITC and PE conjugates (BD Pharmingen), mouse anti-CD3 PE conjugates (BD Pharmingen), and F4/80 allophycocyanin, FITC, and PE conjugates (Caltag and Serotec). Annexin VFITC conjugate (US Biological) and propidium iodide (PI; Sigma-Aldrich) were used to determine the levels of apoptosis and necrosis. A known amount of fluorescent Flow-Check fluorospheres (Beckman Coulter) was added to peritoneal lavage samples before analysis, and the ratio of cells to beads was used to calculate the absolute number of any cell type in peritoneal lavage fluid. Analyses were performed using a FACScan or FACSCalibur instrument.
Immunolabeling of tissues
Organs were fixed in 4% formaldehyde in PBS at 4°C overnight, and paraffin-embedded tissue sections were prepared according to conventional methods. Rehydrated sections were immersed in acetone for 10 min and rinsed in PBS, and a limited trypsin (Sigma-Aldrich) digestion was performed for 20 min. Sections were then washed in PBS, and endogenous peroxidase activity was quenched with a 30-min incubation in 0.3% H2O2 in methanol. Sections were then incubated with F4/80 mAb (Caltag Laboratories) at a 1/100 dilution. Further steps were performed according to the recommendations of the manufacturer of the VECTASTAIN Elite ABC (avidin/biotin complex) system used in the labeling (Vector Laboratories). Sections were then washed in water, lightly counterstained with Mayer’s hematoxylin, dehydrated, and mounted.
Peritonitis was induced by i.p. injection of 1 ml of 3% BTG. Resident Mφ ablation was induced in transgenic mice by i.p. injection of DT (25 ng/g body weight) 24 h before the administration of BTG, with DT-treated FVB/N wild-type (WT) mice serving as control. Mice underwent peritoneal lavage at various time points following BTG injection. In some experiments, peritonitis was induced by the injection of 0.2 mg of zymosan (Sigma-Aldrich) with peritoneal lavage being performed at 8 h. All experiments were performed in accordance with institutional and U.K. Home Office guidelines.
Adoptive transfer of peritoneal cells
Peritoneal lavage samples from groups of FVB/N WT mice were pooled, spun, and either resuspended in 1 ml of sterile PBS (Mφ-rich peritoneal cells) or plated on tissue culture plastic for 2 h to deplete Mφ by adhesion (Mφ-depleted peritoneal cells). A total of 97 ± 2.8% of Mφ was removed from the cell suspension. Four hours before induction of BTG peritonitis, Mφ-rich or Mφ-depleted peritoneal cells suspended in 1 ml of PBS were injected i.p. Experimental groups consisted of 1) CD11b-DTR transgenic mice depleted of resident peritoneal Mφ by prior administration of DT, 2) DT-treated FVB/N WT mice, 3) Mφ-depleted mice reconstituted with Mφ-rich peritoneal cells, and 4) Mφ-depleted mice reconstituted with Mφ-depleted peritoneal cells. Groups 1 and 2 were injected with 1 ml of PBS 4 h before BTG administration as an injection control for the adoptive cell transfer procedure. Animals were sacrificed 8 h following the initiation of peritonitis.
Mice underwent peritoneal lavage at 1 and 3 h following the i.p. injection of BTG. Lavage fluid was centrifuged, aliquoted, and stored at −80°C until analyzed by ELISA for MIP-2 and keratinocyte-derived chemokine (KC) (R&D Systems). Chemokine production by peritoneal cell populations that had been depleted of either Mφ or mast cells (MC) was also determined in vitro. Peritoneal cells were incubated with PE-conjugated anti-F4/80 Ab or an Ab to the MC marker c-kit (CD117) to stain Mφ or MC, respectively. Peritoneal cells were then incubated with anti-PE-conjugated MACS magnetic beads, and Mφ or MC were removed by passing the cells over a magnetic MACS column (Miltenyi Biotec). As a control, total peritoneal cells were incubated with an isotype control Ab followed by magnetic beads and subsequently passed over the magnetic MACS column. This method resulted in >97% depletion of Mφ or MC. Control peritoneal cells and Mφ- or MC-depleted peritoneal cells were then plated in 48-well plates (5 × 105 cells per well) and exposed to 1% BTG for 3 h. Peritoneal cell-conditioned supernatants were harvested, spun, and stored at −80°C until analyzed by specific ELISA for MIP-2 and KC (R&D Systems).
The Student t test with a tailed distribution or ANOVA was used to analyze data. A value of p < 0.05 was deemed statistically significant. Data are presented as mean ± SE.
Generation of transgenic mice
Conditional ablation transgenic mice were generated using an established strategy (21) and a construct (designated CD11b-DTR) that used the CD11b promoter (22) to provide Mφ expression specificity. CD11b-DTR expresses the DT receptor (alternatively named hbEGF (19)) (Fig. 1 A) as a GFP fusion protein. The hbEGF-GFP construct conferred sensitivity to DT in transiently transfected murine cells indicating that it was functional (data not shown).
Six transgenic lines were produced with the CD11b-DTR construct. Although the fluorescence signal from the hbEGF-GFP fusion protein was insufficient to permit detection of transgene expression by FACS analysis, a RT-PCR assay indicated that four lines exhibited detectable transgene expression in peritoneal cells and spleen with line 34 exhibiting high expression (Fig. 1, B and C). WT and CD11b-DTR-34 bone marrow-derived Mφ expressed the Mφ-specific gene F4/80 as expected (Fig. 1,D), indicating that expression of the DT receptor was unlikely to have disrupted normal Mφ differentiation. Transgene expression was also observed in day 4 and 8 CD11b-DTR-34 bone marrow-derived Mφ (Fig. 1 E). Because line CD11b-DTR line 34 showed the highest levels of transgene expression, further analysis was restricted to this line.
Transgenic Mφ are killed by DT in vitro and in vivo
Treatment of BTG-elicited peritoneal Mφ from CD11b-DTR-34 mice with concentrations of DT between 1 ng/ml to 1 mg/ml over a period of 48 h induced cell death at concentrations as low as 25 ng/ml. In contrast, Mφ from WT mice or other transgenic lines were resistant (data not shown). We then asked whether i.p. injection of DT ablated resident peritoneal Mφ in vivo. DT was injected at 25 ng/g mouse weight, and FACS analysis of peritoneal cells was performed 24 h later. Normal numbers of resident peritoneal Mφ (F4/80 positive, CD11b positive, and Ly6C/G negative) were evident in WT mice receiving DT (Fig. 2,A) as well as CD11b-DTR-34 transgenic mice injected with either PBS (Fig. 2,B) or the inactive form of the toxin DTmut (Fig. 2,C). However, CD11b-DTR-34 transgenic mice showed an almost complete absence of F4/80-positive peritoneal Mφ after a single dose of DT (Fig. 2 D). Administration of DT doses of 6.25 and 12.5 ng/g mouse weight resulted in Mφ ablation of ∼72 and 82%, respectively, whereas lower doses resulted in <50% ablation. We therefore chose to use a dose of 25 ng/g body weight for the in vivo studies.
Time course of Mφ elimination in vivo
We then examined the kinetics of Mφ ablation in the peritoneal cavity following administration of a single dose of DT (25 ng/g body weight). The appearance of apoptotic and necrotic cells was monitored using Annexin VFITC and PI staining, respectively. Peritoneal lavages and flow cytometric analyses were conducted on a series of mice 4, 6, 8, and 12 h after DT injection (Fig. 3). After 6 h, 65% of the peritoneal population was annexin V positive, indicating a dramatic increase in early-stage apoptotic cells (Fig. 3,A). The maximal numbers of PI-positive cells occurred ∼2 h later at 8 h after DT injection and represented ∼20% of the total peritoneal cells (Fig. 3,B). The number of F4/80-positive cells was nearly zero at 12 h (Fig. 3 C), and this corresponded to very low levels of PI-positive and annexin V-positive cells (A and B). These data suggest that DT induces sensitive cells to undergo apoptosis and that some of these dying cells then undergo secondary necrosis. All F4/80-positive cells were cleared by 12 h. Restoration of the peritoneal Mφ population occurs by day 4 following a single i.p. dose of DT (3.94 × 105 ± 1.7 × 105 Mφ vs 5.8 × 105 ± 1.3 × 105; day 4 following DT treatment vs day 1 following PBS treatment; n = 5 per group, p > 0.05).
Specificity of Mφ elimination in vivo
To test the specificity of Mφ elimination, we examined CD3+ T cells in spleen and peritoneal cavity and B220+ B cells in the spleen. In this case, we injected two doses of DT at 25 ng/g at 48-h intervals and assessed ablation 24 h later. The F4/80-positive peritoneal population was unaffected in WT mice (Fig. 2,E, lower-right quadrant), but was eliminated in CD11b-DTR-34 animals (G, lower-right quadrant). Despite complete peritoneal Mφ elimination in CD11b-DTR-34 animals, the peritoneal CD3+ T cells were present in both WT and transgenic animals injected with DT (Fig. 2, E and F, upper-left quadrants). The relative increase in CD3-positive and double-negative cells in the DT-treated CD11b-DTR-34 animals is due to the plotting of equal numbers of detection events in the FACS analyses. B220+ and CD3+ populations in spleen were unaffected in either WT or CD11b-DTR-34 mice receiving two doses of DT (data not shown).
Because CD11b is expressed on both granulocytes and Mφ, we asked whether both of these cell types were sensitive to DT. CD11b-DTR-34 mice were injected with BTG. DT (25 ng/g body weight) or PBS was injected 8 h after initiation of peritonitis with peritoneal lavage being performed 12 h later. Despite 90% Mφ ablation (0.5 × 106 ± 0.15 × 106 vs 4.8 × 106 ± 0.23 × 106; DT injection vs PBS; p < 0.005), there was no difference in PMN number (2.3 × 106 ± 0.22 × 106 vs 1.6 × 106 ± 0.39 × 106; DT injection vs PBS; p > 0.05). In addition, flow cytometric analysis of whole blood performed 24 h following DT administration indicated that circulating PMN numbers were unaffected by DT administration (1.02 × 106 ± 0.18 × 106 PMNs/ml whole blood vs 0.97 × 106 ± 0.22 × 106; DT injection vs PBS; p > 0.05). In contrast, DT administration induced significant depletion of circulating monocytes (0.117 × 106 ± 0.059 × 106 monocytes/ml whole blood vs 0.52 × 106 ± 0.073 × 106; DT injection vs PBS; p < 0.05).
Differential deletion of Mφ populations
We also asked whether DT injection could eliminate Mφ in distant organs. Two doses of DT (25 ng/g) were administered IP at 48-h intervals, and the presence of F4/80-positive Mφ in kidney, liver, and lung was analyzed 24 h later and in the ovary 16 h later (Fig. 4). The ovary was examined at 16 h, because there was evidence of some patchy ovarian necrosis present at 20 h. WT mice injected with DT and CD11b-DTR-34 homozygote mice injected with DTmut were unaffected (Fig. 4, top two rows). Both kidney and ovary of CD11b-DTR-34 homozygote mice injected with DT exhibited an absence of F4/80+ cells. In the kidney, mesangial and interstitial Mφ were ablated in the absence of overt renal injury. However, hepatic sinusoidal Mφ and alveolar Mφ were unaffected, indicating that not all populations of tissue Mφ were susceptible. However, the rapid elimination of Mφ populations in the peritoneal cavity and kidney while leaving other cell populations intact establishes the basic validity of this approach to conditional cell ablation.
Resident Mφ ablation reduces PMN influx and CXC chemokine responses during experimental peritonitis
We used the conditional ablation strategy to investigate the role of resident tissue peritoneal Mφ in sensing perturbation of the microenvironment and subsequent initiation of acute peritoneal inflammation and PMN recruitment in experimental peritonitis. Resident Mφ ablation markedly attenuated PMN infiltration following the administration of 3% BTG (Fig. 5,A). We also performed Mφ repletion studies with either Mφ-rich or Mφ-depleted peritoneal cells derived from WT mice. Reconstitution of DT-treated CD11b-DTR-34 homozygote mice with Mφ-rich peritoneal cells 4 h before BTG treatment resulted in complete restoration of peak PMN infiltration at 8 h. In contrast, the administration of Mφ-depleted peritoneal cells was ineffective (Fig. 5 B). Previous work suggested that the nature of the inflammatory stimulus may determine the involvement of Mφ in experimental peritonitis (8), and we therefore performed Mφ depletion studies in zymosan peritonitis. We also found that depletion of resident peritoneal Mφ resulted in a significant reduction in PMN infiltration 8 h following the induction of zymosan peritonitis (2.6 × 106 ± 8.8 × 105 PMNs vs 5.4 × 106 ± 3.7 × 105; DTR plus DT vs FVB/N WT controls plus DT; n = 6 per group; p < 0.05).
In this model, we found peak levels of the PMN CXC chemokines MIP-2 and KC at the 1-h time point. Resident peritoneal Mφ ablation before the initiation of BTG peritonitis markedly reduced the elevation in MIP-2 levels (148.5 ± 34.8 vs 1762.1 ± 153.5 pg/ml; Mφ-depleted mice vs nondepleted mice: p < 0.00001). There was a slight, albeit statistically significant, difference in the much lower levels of MIP-2 between DT-treated and control mice at 3 h (204 ± 54 vs 74 ± 8 pg/ml; Mφ-depleted mice vs nondepleted mice; p < 0.05). This suggests that the production of MIP-2 in vivo is predominantly Mφ dependent. In addition, a 50% reduction in the level of KC was evident in Mφ-depleted mice at the 1-h time point (1408.2 ± 322.5 vs 2467.5 ± 264.9 pg/ml; Mφ-depleted mice vs nondepleted mice; p < 0.05). Interestingly, the levels of KC levels at 3 h are higher in Mφ-depleted mice compared with control mice (1477 ± 400 vs 74 ± 8 pg/ml; Mφ-depleted mice vs nondepleted mice; p < 0.01), thereby suggesting a source of KC other than resident Mφ.
CXC chemokine responses are Mφ dependent in vitro
Previous studies of peritoneal and dermal inflammation have implicated the MC as playing an important role in the initiation of PMN infiltration (8, 24). We therefore performed in vitro studies to determine the production of PMN chemokines by BTG-stimulated peritoneal cell populations that had been depleted of Mφ or MC. Control peritoneal cells produced significant levels of MIP-2 and KC, which was not affected by MC depletion (Fig. 6). However, chemokine levels were dramatically reduced following the depletion of Mφ, thereby indicating that chemokine production was completely Mφ dependent with no involvement of MC (Fig. 6).
Previous analyses (20, 21) and the experiments we describe here show that expression of human hbEGF (19) in mouse cells can confer sensitivity to DT in vivo, and that, as a consequence, injection of DT will kill cells that express hbEGF. Our data indicate that Mφ populations in the peritoneal cavity and kidney can be rapidly killed or eliminated while leaving other cell populations intact, and this establishes the basic validity of this approach to conditional cell ablation. We noted that hepatic and alveolar Mφ populations were unaffected, and it may be the case that a higher dose of DT may have ablated these cells. However, we found that mice could become unwell with doses of DT >25 ng/g body weight, and we therefore did not use doses >25 ng/g body weight in this study. It is pertinent that, despite PMN expression of CD11b, the administration of DT did not induce the death of recruited or circulating PMNs, indicating that PMNs are insensitive to DT, potentially as a result of their lower level of protein synthesis.
We used the conditional ablation strategy to investigate the role of resident peritoneal Mφ in the initiation of acute peritoneal inflammation following the administration of BTG. Previous work has indicated that leukotrienes derived from resident peritoneal Mφ are involved in the development of early vascular permeability in sterile peritonitis (25). Although early work in rat models of peritonitis implicated the resident peritoneal Mφ in the orchestration of PMN recruitment (26, 27, 28), more recent studies have produced conflicting results (7, 8). Indeed, studies by Ajuebor et al. (8) suggest that resident Mφ depletion inhibits PMN influx in LPS-induced inflammation, has no effect in BTG peritonitis, and augments PMN influx in zymosan peritonitis. In the latter model, it is proposed that Mφ-derived IL-10 inhibits PMN recruitment. Conversely, work by Knudsen et al. (7) using clodronate-induced depletion of peritoneal Mφ in a rat model of sterile peritonitis demonstrated that PMN infiltration was Mφ dependent.
In this study, administration of DT resulted in a dramatic 98% Mφ ablation that markedly blunted PMN infiltration, thereby indicating a key role for the resident Mφ in the orchestration of acute peritoneal inflammation in this experimental model. It is important to note that the reduced PMN infiltration in DT-treated mice was not attributable to a systemic neutropenia, because PMNs were not sensitive to DT and the number of circulating PMNs in DT-treated mice was comparable with that of PBS-treated mice at 24 h. The importance of the resident Mφ was reinforced by experiments involving the adoptive transfer of nontransgenic peritoneal Mφ following DT-mediated Mφ ablation and before the initiation of peritonitis. The presence or absence of Mφ in the transferred peritoneal cell population directly correlated with the restoration of the PMN influx, thereby suggesting that the Mφ exerts a critical role in this process. In addition, we found that depletion of resident Mφ also significantly reduced PMN infiltration in zymosan peritonitis, thereby suggesting that the sensing function of the resident Mφ may be stimulus independent.
The magnitude of the Mφ depletion may explain the apparent discrepancy between these results and the study by Ajuebor et al. (8). Administration of a single dose of DT induced 98% Mφ ablation, whereas three doses of liposomal clodronate resulted in >85% Mφ depletion in the study by Ajuebor et al. (8). Mφ are a potent source of chemokines and cytokines, and it may be the case that, in certain circumstances, a relatively small population of residual Mφ may exert significant biological effects. Although administration of liposomal clodronate may exert marked biological effects despite depletion of only ∼80% of Mφ (29), it may be necessary to deplete almost all peritoneal Mφ to delineate their roles as sentinel cells.
Although peritoneal Mφ may produce myriad mediators capable of recruiting PMNs (30, 31), we examined the effect of Mφ ablation upon the level of CXC chemokines in this model. Our data suggest that the initiation of PMN infiltration is mediated by resident peritoneal Mφ-dependent production of chemokines previously documented to play a role in orchestrating PMN recruitment in BTG peritonitis (32, 33) and in other inflammatory situations (34, 35, 36). MC are also a rich source of proinflammatory and vasoactive mediators and have been documented to play an important role in PMN recruitment during inflammation of the peritoneum (8) as well as other sites such as the skin (24).
We found that resident Mφ ablation markedly reduced the peak level of MIP-2 and significantly blunted the level of KC at 1 h. However, partial inhibition (50%) of KC production at 1 h and the persistent elevation of KC at 3 h in Mφ-depleted mice suggest that KC may be produced by other cells within the peritoneum. The persistent elevation of KC also suggests that Mφ may play a role in the negative regulation of KC production by non-Mφ cells. In this context, it is pertinent that our in vitro data indicate that both KC and MIP-2 production by peritoneal cells obtained by peritoneal lavage is almost entirely dependent upon Mφ, because peritoneal cells depleted of Mφ produced minimal levels of chemokines. These findings suggest that peritoneal cells retrievable by peritoneal lavage are not the source of KC detected in our in vivo study. This interpretation of the data is consistent with recent work in a wound model (37) demonstrating MIP-2 expression by inflammatory cells and KC expression by resident tissue cells such as endothelial cells and fibroblasts. Peritoneal mesothelial cells undoubtedly participate in peritoneal inflammation and can produce chemokines and cytokines (38), and it may be the case that mesothelial cells contribute to the KC production evident in this study. Despite this, however, we found that PMN infiltration is still markedly blunted despite the persistent presence of KC at 3 h.
We also examined the potential interaction between Mφ and MC by performing in vitro studies of peritoneal cells that had been depleted of Mφ or MC before stimulation with BTG. Interestingly, depletion of MC had no significant effect upon the production of KC and MIP-2 following BTG stimulation, suggesting that chemokine production in this model was Mφ dependent and MC independent.
In conclusion, this work has used a novel model of conditional Mφ ablation to dissect the role of resident peritoneal Mφ in the initiation of acute peritoneal inflammation. Our data indicate a key role for the resident Mφ in sensing peritoneal irritation and orchestrating PMN infiltration in BTG and zymosan peritonitis. This proinflammatory function is predominantly mediated by production of the potent PMN CXC chemokine MIP-2 and, to a lesser extent, KC. Although previous work has implicated the involvement of other cells such as MC, our study suggests that resident Mφ are critically important producers of PMN chemokines and act to orchestrate PMN recruitment in murine BTG peritonitis. We also anticipate that CD11b-DTR transgenic mice will be valuable for studying other Mφ functions in vivo in a variety of different biological contexts. The option of being able to choose the time and, with local toxin injection, perhaps the locality of ablation offers a number of advantages over other cell ablation systems.
We thank Leon Eidels for providing the cDNA to human hbEGF, Daniel Tenen for the CD11b-based transgene construct, and Michael Clay for technical assistance.
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.
The Lang Laboratory is supported by grants from the National Institutes of Health (RO1 EY10559, EY11234, EY12370, and EY14102) and by funds from the Abrahamson Pediatric Eye Institute Endowment at Children’s Hospital Medical Center of Cincinnati. J.H. is in receipt of a Wellcome Trust Senior Research Fellowship in Clinical Science (Grant 061139). J.-F.C. is supported by the Canadian Institutes of Health Research. J.S. is supported by the Wellcome Trust (Program Grant 064487).
Abbreviations used in this paper: Mφ, macrophage; PMN, neutrophil; DT, diphtheria toxin; hbEGF, heparin-binding epidermal growth factor; BTG, Brewer’s thioglycolate; WT, wild type; PI, propidium iodide; KC, keratinocyte-derived chemokine; MC, mast cell.