IL-6 Regulates Neutrophil Trafficking during Acute Inflammation via STAT3¹

Transition from innate to adaptive immunity is critical for the successful resolution of any inflammatory response, and represents a fundamental requirement for competent host defense. At the cellular level, this immunological switch is characterized by an initial influx of neutrophils which is subsequently replaced by inflammatory monocytes and T cells (1). Inappropriate regulation of this leukocyte trafficking can lead to impaired neutrophil clearance and increased tissue damage from the accumulation of neutrophil-secreted proteases and reactive oxygen species at the site of inflammation (2). Indeed, this situation has been implicated in the pathogenesis of many inflammatory diseases, such as chronic peritonitis, sepsis, inflammatory bowel disease, chronic obstructive pulmonary disease, and renal injury (3, 4, 5, 6, 7). Thus, identifying key inflammatory mediators and the pathophysiological consequences of their action in modulating neutrophil recruitment and clearance is a crucial step in understanding the molecular pathogenesis of these diseases.

Recent advances have identified certain inflammatory cytokines, TLR signaling, and complement activation as being pivotal in directing the tightly coordinated shift from innate to adaptive immunity (8). In particular, the cytokine IL-6 can dictate the profile of leukocyte recruitment during the inflammatory response via selective regulation of inflammatory chemokines and apoptotic events (1, 9, 10). The classical responsiveness to IL-6 is governed by a receptor complex consisting of two membrane-bound subunits, an 80-kDa cognate α-chain (IL-6Rα), and a ubiquitously expressed 130-kDa β-chain receptor (gp130) which acts as the universal signal-transducing element for all IL-6 family cytokines (11). Alternatively, IL-6 regulation of leukocyte trafficking relies upon signaling via its soluble IL-6Rα (termed IL-6 trans-signaling) (12). However, in either case, the receptor-mediated signaling events orchestrating this inflammatory response remain undefined.

Cytokine signaling through gp130 is triggered by receptor-associated JAKs, which phosphorylate specific tyrosine residues in the gp130 cytoplasmic domain to enable activation of STAT1 and STAT3 (13), as well as Src-homology phosphatase 2 (SHP2)⁵-mediated MAPK and PI3K pathways (14). The phosphotyrosine residue at position 757 (pY₇₅₇) of murine gp130 (759 in human gp130) not only serves as a binding site for SHP2, but also plays a crucial role in the negative regulation of gp130 signaling by recruiting suppressor of cytokine signaling 3 (SOCS3) (15).

To genetically dissect the contribution of these gp130-mediated signaling events in directing neutrophil trafficking during inflammation, studies presented here have adopted two strains of gp130 knockin mice, gp130^{ΔSTAT/ΔSTAT} and gp130^Y757F/Y757F, in which gp130-mediated activation of STAT1/3 and SHP2-ERK MAPK and -PI3K pathways, respectively, have been genetically abolished (16, 17). Notably, in gp130^{ΔSTAT/ΔSTAT} mice, the absence of STAT1/3-binding motifs on gp130 results in impaired STAT3-mediated induction of SOCS3, leading to increased activation of the ERK MAPK and PI3K pathways (16, 17). Conversely, in gp130^Y757F/Y757F mice, the phenylalanine substitution at Y₇₅₇ in gp130 also prevents negative regulation of gp130-mediated STAT1/3 signaling by SOCS3, resulting in STAT1/3 hyperactivation (17). Using an established model of acute peritoneal inflammation which mimics many of the clinical features of bacterial peritonitis (9, 10), our studies conclude that the control of neutrophil clearance by IL-6 is reliant on gp130 signaling via STAT3, but not STAT1. Consequently, we propose that gp130-mediated STAT3 signaling represents a critical checkpoint regulator in the successful transition from innate to adaptive immunity, and therefore in the resolution of acute inflammation.

The generation of gp130^{ΔSTAT/ΔSTAT}, gp130^Y757F/Y757F, and compound gp130^Y757F/Y757F mice on either a Stat3 heterozygous (gp130^Y757F/Y757F:Stat3^+/−) or IL-6 homozygous (gp130^Y757F/Y757F:IL-6^−/−) background has been previously described (16, 17, 18, 19). Stat1^−/− mice (20) were crossed with gp130^Y757F/Y757F mice to generate gp130^Y757F/Y757F:Stat1^−/− mice (21). All experiments were performed following approval by the Monash Medical Centre “A” Animal Ethics Committee and the U.K. Home Office, and included wild-type (gp130^+/+) littermate controls. All mice were maintained under specific pathogen-free conditions and were age matched for each experiment.

Peritoneal inflammation was induced in vivo by i.p. injection of SES (9, 10). At the indicated times, animals were sacrificed, and their peritoneal cavities were lavaged with ice-cold PBS. Total cell counts were assessed using an automated hematology analyzer (Sysmex), and the composition of the leukocyte infiltrate was assessed by differential cell counting of Wright-Giemsa-stained slides. Lavage fluids were rendered cell-free by centrifugation for analysis of inflammatory mediators.

Murine KC, MCP-1 (R&D Systems), and IL-6 (BD Pharmingen) were quantified using commercial ELISA kits according to the manufacturers’ instructions.

Nuclear extracts were prepared from peritoneal biopsies using a modified version of a rapid technique for the extraction of nuclear proteins (22, 23). Protein concentrations were determined using the Bradford method.

EMSAs were performed using 10 μg of nuclear extract as previously described (23, 24). Oligonucleotides containing the serum-inducible element STAT-binding element were annealed and labeled with [α-³²P]dTTP (Amersham Biosciences) using the Klenow fragment of DNA polymerase I.

ERK1/2, STAT1, and STAT3 Abs were purchased from Santa Cruz Biotechnology. Phospho-STAT1 (Tyr⁷⁰¹), phospho-STAT3 (Tyr⁷⁰⁵), and phospho-ERK1/2 Abs were obtained from Cell Signaling Technology.

Protein extracts from frozen peritoneal biopsies were prepared using ice-cold lysis buffer, following which they were precleared of cellular debris before separation by SDS-PAGE and immunoblotting with specific Abs (18, 25, 26). Immunolabeled proteins were detected using either the ECL detection system (Amersham Biosciences) or the Odyssey Infrared Imaging System (LI-COR) with the appropriate secondary Abs as per the manufacturer’s instructions.

For assessment of apoptosis in vivo, at the indicated time points postadministration of SES, mice were culled and the peritoneal cavity was lavaged. Total leukocytes were recovered, and neutrophils were stained using a FITC-labeled Gr1 Ab (BD Biosciences). The extent of apoptosis in Gr1-positive neutrophils was determined by double staining with Annexin V^PE and 7-aminoactinomycin D (7-AAD; BD Biosciences), and subsequent analysis using a FACSCantoII flow cytometer (BD Biosciences). Quadrants were assigned according to autofluorescence controls, and 5000 events acquired for each sample. Cell surface expression levels of CXCR2 on murine peripheral blood neutrophils were determined on a FACSCalibur flow cytometer (BD Biosciences) by staining leukocytes from mouse whole blood that were subjected to red cell lysis with the appropriate fluorescence-labeled Ly6G (BD Biosciences) and CXCR2 (R&D Systems) Abs.

The resolution index was calculated as previously reported (27). Briefly, the maximal number of neutrophils recruited (neutrophil ψ_max) following SES administration was quantified and the optimal time (T_max) of infiltration was recorded in each genetic strain. These values were then used to calculate the time point, which reflected 50% of the peak neutrophil infiltration (T₅₀), and the theoretical neutrophil numbers present at that time point (R₅₀). The resolution index (R_i) therefore reflects the difference between T₅₀ and T_max.

Data are expressed as means ± SEM, and statistical analysis was performed using an independent samples t test with SPSS version 11 statistical software. A p value of <0.05 was considered statistically significant.

To define the gp130-mediated signaling events regulating the trafficking of neutrophils during acute peritoneal inflammation, gp130^+/+, gp130^{ΔSTAT/ΔSTAT}, and gp130^Y757F/Y757F mice were administered with a defined dose of SES known to elicit a robust and reproducible inflammatory response in the peritoneal cavity (9, 10). In gp130^Y757F/Y757F mice, the absolute numbers of infiltrating neutrophils peaked at 3 h (3.1 ± 0.9 × 10⁶) compared with 12 h (3.7 ± 0.4 × 10⁶) in gp130^+/+ mice, and neutrophils were more rapidly cleared and significantly reduced in numbers at 6 and 12 h (p < 0.01) in gp130^Y757F/Y757F mice relative to gp130^+/+ mice (Fig. 1,A). By contrast, the profile of neutrophil trafficking in SES-challenged gp130^{ΔSTAT/ΔSTAT} mice was comparable to that observed in gp130^+/+ mice (Fig. 1,A), although at the 3-, 6-, and 12-h time points absolute neutrophil numbers appeared consistently elevated, albeit not significantly, in gp130^{ΔSTAT/ΔSTAT} mice (e.g., gp130^+/+, 2.48 ± 0.39 × 10⁶ vs gp130^{ΔSTAT/ΔSTAT}, 3.06 ± 0.78 × 10⁶ at 6 h). Further analysis of the proportion of neutrophils within the inflammatory infiltrates revealed that the trend of augmented and impaired clearance in gp130^Y757F/Y757F and gp130^{ΔSTAT/ΔSTAT} mice, respectively, compared with gp130^+/+ mice was even more pronounced (Fig. 1 B). We note that the differences between absolute numbers of neutrophils and their relative proportion in the SES-induced inflammatory infiltrates within each genotype are a likely consequence of the impact that altered gp130 signaling also has on the production of other cell types within the infiltrates, namely mononuclear cells and lymphocytes (18, 25).

Compared with gp130^+/+ mice, the rapid termination of the neutrophil infiltrate in gp130^Y757F/Y757F mice following SES administration correlated with an enhanced and prolonged nuclear STAT DNA-binding activity in nuclear extracts from the peritoneal membrane (Fig. 1,C). Consistent with the normal neutrophil trafficking profile of gp130^{ΔSTAT/ΔSTAT} mice following SES challenge, the kinetics of SES-induced STAT DNA-binding activity in the peritoneum of gp130^+/+ and gp130^{ΔSTAT/ΔSTAT} mice was similar, although maximal STAT binding in gp130^{ΔSTAT/ΔSTAT} mice (3 h) appeared slightly delayed compared with gp130^+/+ mice (1 h). Western blot analyses of peritoneal extracts confirmed that SES-induced STAT1 and STAT3 tyrosine phosphorylation was greatest in gp130^Y757F/Y757F mice, while gp130^{ΔSTAT/ΔSTAT} mice showed the least extent of STAT1 and STAT3 activation (Fig. 2).

Because we have previously shown that gp130-mediated activation of ERK1/2 via SHP2 is diminished in gp130^Y757F/Y757F mice and elevated in gp130^{ΔSTAT/ΔSTAT} mice (17), we next examined the phosphorylation status of ERK1/2 in response to SES (Fig. 2). In gp130^+/+ mice, SES induced maximal ERK1/2 phosphorylation between 1 and 3 h before subsiding at 6 h. Although the magnitude of ERK1/2 phosphorylation in gp130^Y757F/Y757F mice resembled gp130^+/+ mice, ERK1/2 phosphorylation was sustained and remained maximal at 6 h post-SES challenge. By contrast, in gp130^{ΔSTAT/ΔSTAT} mice, the high basal ERK1/2 activity (0 h) was comparable to that induced by SES in gp130^+/+ and gp130^Y757F/Y757F mice at 1 and 3 h, and was even further induced by SES, peaking at 1 h. Collectively, these data suggest that deregulated SES-induced STAT (increased magnitude and duration) and/or ERK1/2 (increased duration) signaling in gp130^Y757F/Y757F mice has a major influence on the rate of neutrophil trafficking during acute peritoneal inflammation.

We have previously used mice deficient in IL-6 to demonstrate the significance of IL-6 signaling in coordinating neutrophil clearance during acute peritoneal inflammation (9). As shown in Fig. 3,A, genetic ablation of IL-6 in gp130^Y757F/Y757F mice restored the profile of neutrophil infiltration and clearance in the resultant gp130^Y757F/Y757F:IL-6^−/− mice back to a similar pattern seen in gp130^+/+ mice upon SES administration, thus confirming that IL-6 is a primary gp130-acting cytokine responsible for the altered neutrophil trafficking in gp130^Y757F/Y757F mice. Furthermore, SES-induced STAT1 and STAT3 activation in gp130^Y757F/Y757F:IL-6^−/− mice were reduced back to wild-type levels (Fig. 3 B). Because IL-6-induced ERK1/2 activation via gp130 is abolished in gp130^Y757F/Y757F mice (17), these data suggest a causal role for exaggerated IL-6-induced STAT signaling in altered neutrophil trafficking.

Although the data presented above infer that the threshold of IL-6-mediated STAT signaling is critical for governing normal neutrophil trafficking during acute peritoneal inflammation, the relative involvement of STAT1 and STAT3 in this process is unclear. We therefore genetically assessed the contribution of exaggerated STAT3 activation to the altered SES-induced neutrophil trafficking in gp130^Y757F/Y757F mice by using gp130^Y757F/Y757F:Stat3^+/− mice in which the level of gp130-dependent STAT3 activation has been normalized (18, 26). The profile of neutrophil trafficking in gp130^Y757F/Y757F:Stat3^+/− mice appeared intermediate between gp130^+/+ and gp130^Y757F/Y757F mice, with the absolute numbers of infiltrating neutrophils peaking within 3 h and remaining at these levels up to 12 h. Interestingly, the magnitude of this peak neutrophil influx in gp130^Y757F/Y757F:Stat3^+/− mice at 3 h was significantly reduced (1.8 ± 0.2 × 10⁶, p > 0.05) compared with the corresponding peak neutrophil numbers at 3 and 12 h in gp130^Y757F/Y757F and gp130^+/+ mice, respectively (Fig. 4,A). Despite these observations, it is worth noting the similar proportion of neutrophils within the initial inflammatory infiltrate and early clearance phase of trafficking in gp130^+/+ and gp130^Y757F/Y757F:Stat3^+/− mice compared with gp130^Y757F:Y757F mice (Fig. 4,B), the latter of which is significantly (p > 0.01) reduced at 6 h (gp130^+/+, 43 ± 3% vs gp130^Y757F/Y757F, 18 ± 4% vs gp130^Y757F/Y757F:Stat3^+/−, 38 ± 2%). A similar trend in absolute numbers (gp130^+/+, 0.78 ± 0.24 × 10⁶, and gp130^Y757F/Y757F:Stat3^+/−, 0.62 ± 0.14 × 10⁶ vs gp130^Y757F/Y757F, 0.36 ± 0.09 × 10⁶; p > 0.01) and proportions (gp130^+/+, 18.2 ± 5.2%, and gp130^Y757F/Y757F:Stat3^+/− 12.5 ± 2.25% vs gp130^Y757F/Y757F, 5.8 ± 2.7%; p > 0.05) of neutrophils in gp130^+/+ and gp130^Y757F/Y757F:Stat3^+/− mice compared with gp130^Y757F:Y757F mice was also apparent at the later 24-h time point. This rescue in the pattern of neutrophil recruitment in gp130^Y757F/Y757F:Stat3^+/− mice was characterized by a reduction in the level of SES-induced STAT3 tyrosine phosphorylation which was similar to that seen in gp130^+/+ mice (Fig. 4 C).

We have previously shown that STAT1 hyperactivity remained unchanged in gp130^Y757F/Y757F:Stat3^+/− mice (25), which is consistent with the notion that the specific reduction of STAT3 in these mice was responsible for alleviating the neutrophil trafficking phenotype. However, to assess whether exaggerated STAT1 activation also contributed to the aberrant neutrophil trafficking phenotype observed in gp130^Y757F/Y757F mice, we next analyzed SES-induced neutrophil recruitment and clearance in gp130^Y757F/Y757F mice crossed onto a Stat1-deficient background (gp130^Y757F/Y757F:Stat1^−/−). As expected, STAT3 activity remained elevated in gp130^Y757F/Y757F:Stat1^−/− mice (Fig. 4,C), and the neutrophil trafficking profile in these mice resembled that observed in gp130^Y757F/Y757F mice (Fig. 4, A and B).

To further characterize changes in the kinetics of neutrophil clearance following SES stimulation in each of the genetically modified strains, we next quantified the neutrophil “resolution index” (R_i) (27). The data from these calculations (outlined in Materials and Methods) are presented in Table I. The R_i calculated for gp130^+/+ mice (8 h) was consistent with previously published results (27). Enhanced gp130-STAT signaling in the gp130^Y757F/Y757F mice reduced the R_i (2 h), whereas the R_i values for the gp130^{ΔSTAT/ΔSTAT} and gp130^Y757F/Y757F: IL-6^−/− mice (8 and 7 h, respectively) were similar to gp130^+/+ mice. By contrast, a reduction in the activation of STAT3 in gp130^Y757F/Y757F:Stat3^+/− mice dramatically increased the R_i to 18 h, whereas the absence of STAT1 in gp130^Y757F/Y757F:Stat1^−/− had no effect on the R_i. Collectively, these data therefore support a more prominent role for the IL-6/STAT3-signaling axis in controlling SES-induced neutrophil infiltration during acute peritoneal inflammation.

Neutrophil trafficking following inflammatory activation is governed by the initial production of neutrophil-activating chemokines, and the eventual clearance of neutrophils occurs via both reduced local chemotactic activity and apoptotic mechanisms. In this respect, IL-6 has been shown to inhibit the inflammatory induction of neutrophil-activating chemokines and to facilitate their apoptotic clearance (1, 10). However, the contribution of specific IL-6-signaling pathways to these processes remains to be fully elucidated. We therefore initially assessed whether the rapid clearance of neutrophils during acute peritoneal inflammation in gp130^Y757F/Y757F mice was attributable to increased neutrophil apoptosis by performing Annexin-V/7-AAD staining on the Gr1⁺ neutrophil population following SES administration. However, as shown in Fig. 5, a comparable profile of neutrophil apoptosis was recorded in gp130^+/+ and gp130^Y757F/Y757F mice at 1, 2, and 3 h post-SES administration. This result also suggests that the low proportion of neutrophils in the peritoneal lavage of gp130^Y757F/Y757F mice at the early 1 h time point (e.g., Fig. 4 B) is unlikely to result from elevated apoptosis.

We next examined whether gp130-signaling alterations influenced chemokine-directed neutrophil recruitment by measuring the protein levels of the key neutrophil-activating chemokine CXCL1/KC in the peritoneal lavage fluid of SES-challenged mice (9, 10). No significant changes were observed after 1 h (optimal time point for CXCL1/KC induction in this model (9)) among the various genotypes (Fig. 6,A), suggesting that, consistent with our earlier data (Fig. 4,B), deregulated gp130 signaling does not affect the initial CXCL1/KC-driven induction of neutrophil recruitment. However, CXCL1/KC levels were significantly reduced in gp130^Y757F/Y757F mice compared with gp130^+/+ mice at 3 h (Fig. 6,B). Importantly, this was in contrast to the elevated levels of CXCL1/KC in gp130^{ΔSTAT/ΔSTAT} mice at 3 h (Fig. 6,B). As predicted, this response was highly dependent on IL-6 as evidenced by the increased levels of CXCL1/KC detected in lavage fluid from gp130^Y757F/Y757F:IL-6^−/− mice (Fig. 6,B). Notably, the monoallelic deletion of Stat3 in gp130^Y757F/Y757F:Stat3^+/− mice restored CXCL1/KC levels similar to those observed in gp130^+/+ mice, suggesting an exclusive role for STAT3 via gp130 in attenuating SES-induced CXCL1/KC production, and therefore in promoting the clearance of neutrophils. In support of this notion, the reduced CXCL1/KC levels in gp130^Y757F/Y757F:Stat1^−/− mice (68 ± 22 pg/ml) were similar to those in gp130^Y757F:Y757F mice (45 ± 27 pg/ml) at 3 h postactivation. Furthermore, the specificity of the IL-6/STAT3-mediated effect on neutrophil-activating chemokines was supported by the normal expression level of the monocyte chemoattractant CCL2/MCP1 in gp130^Y757F/Y757F peritoneal fluid at 3 h post-SES (Fig. 6,C). Similarly, flow cytometric analyses confirmed that the surface expression of CXCR2, the common chemokine receptor for KC/CXCL1 and other CXC chemokines containing the ELR (Glu-Leu-Arg) motif, was comparable between gp130^+/+ and gp130^Y757F/Y757F mice (Fig. 6 D). Thus, the above data collectively demonstrate that STAT3-mediated down-regulation of neutrophil-activating chemokine expression represents a primary mechanism by which IL-6 controls neutrophil trafficking.

Regulation of neutrophil trafficking during inflammation, which involves the initial recruitment of neutrophils and their subsequent controlled clearance, is critical for the successful resolution of acute inflammation. Impaired clearance of activated neutrophils can result in damage to the inflamed tissue and resulting immunopathology. To provide new molecular insights into the mechanisms governing this process, in this study we have used a unique series of gp130 knockin mutant mice displaying differential activation of specific gp130-signaling pathways in an established model of SES-induced acute peritonitis. We report here that alterations to specific IL-6/gp130-signaling pathways dramatically influence the rate of neutrophil trafficking (Table I) (27) and, in particular, that neutrophil clearance is positively regulated by gp130-mediated STAT3 signaling.

Apoptosis has been implicated in the clearance of neutrophils from sites of inflammation and is important for the resolution of the inflammatory process (10, 28). Although several lines of evidence suggest that the ERK MAPK and PI3K pathways promote the survival of neutrophils (29, 30, 31, 32), neutrophil apoptosis during the inflammatory response in gp130^Y757F/Y757F mice was comparable to gp130^+/+ mice. Similarly, the extent of apoptosis over 24 h was normal within the total leukocyte population of both gp130^Y757F/Y757F and gp130^Y757F/Y757F:IL-6^−/− mice genetically modified to lack gp130-mediated SHP2-ERK MAPK and -PI3K pathways (data not shown). Thus, these observations suggest that the accelerated removal of neutrophils in the peritoneal cavity of gp130^Y757F/Y757F mice is not due to increased apoptosis as a consequence of impaired ERK/PI3K signaling. We speculate that one possible explanation for the removal of these neutrophils from the peritoneal cavity which is currently under investigation in our laboratory is the increased (3- to 4-fold) numbers of monocytes/macrophages in the peritoneal inflammatory infiltrates of only gp130^Y757F/Y757F mice, which occurs in a STAT3-dependent manner (B. J. Jenkins, unpublished observations) and may lead to enhanced phagocytosis of dying neutrophils by macrophages as previously demonstrated (33).

A key and definitive finding of this current study was our genetic evidence that hyperactivation of STAT3, rather than STAT1, via gp130 was primarily responsible for the aberrant regulation of neutrophil trafficking in gp130^Y757F/Y757F mice. Specifically, the accelerated clearance of neutrophils and impaired production of the critical neutrophil-attracting chemokine CXCL1/KC in the peritoneal cavity of gp130^Y757F/Y757F mice were rescued in gp130^Y757F/Y757F:Stat3^+/− mice displaying reduced gp130-dependent STAT3 activity. This genetic observation identified that gp130-dependent STAT3 signaling influences neutrophil clearance by negatively regulating CXCL1/KC production. Because the SES-induced neutrophil trafficking phenotype of gp130^Y757F/Y757F mice was not completely corrected in gp130^Y757F/Y757F:Stat3^+/− mice, we interpret this finding as a further indication that the STAT3-signaling threshold for specific pathophysiological responses varies in a cell/organ-type specific manner (34). For instance, we have previously reported that genetically reducing the level of STAT3 activity in gp130^Y757F/Y757F:Stat3^+/− mice completely rescues the STAT3-driven splenomegaly and thrombocytosis observed in gp130^Y757F/Y757F mice, whereas lymphadenopathy was only partially ameliorated (18, 26). Accordingly, certain cell types/organs are more sensitive to the reduction in the net STAT3 signal output in gp130^Y757F/Y757F:Stat3^+/− mice than others. We therefore propose that this scenario also applies to the STAT3-driven molecular processes (e.g., CXCL1/KC regulation) facilitating neutrophil trafficking. Nonetheless, our data presented here are supported by other studies demonstrating that STAT3 activation negatively regulates production of a number of inflammatory genes, including chemokines (35, 36). It is possible that this is a selective effect on a specific subset of chemokines and may not be a global suppression of all chemokines. In support of this notion, inflammation-induced expression levels of the monocyte-attracting chemokine CCL2 (MCP-1) remained unchanged in gp130^Y757F/Y757F mice. Although the precise molecular mechanism for STAT3-negative regulation of proinflammatory gene expression remains controversial, a number of studies have reported that STAT3 negatively regulates NF-κB signaling or DNA binding, either through a direct interaction or via an intermediate protein (37, 38, 39, 40).

Although our data presented here suggested that IL-6 signaling has a major role in regulating the neutrophil trafficking component of peritonitis, it is still possible that other gp130-acting cytokines may augment this process either independently or in combination with IL-6. Oncostatin M, for example, displays the same capability as IL-6 to down-regulate neutrophil-activating chemokines in vitro (41), and a number of other IL-6 family cytokines (e.g., IL-11, IL-27 and IL-31) are produced during inflammation which may have distinct or overlapping roles in regulating acute inflammatory responses. Thus, the role of such cytokines in the aberrant neutrophil trafficking observed in gp130^Y757F/Y757F mice during acute peritoneal inflammation warrants further investigation.

In summary, this study has demonstrated a clear and major role for gp130-mediated STAT3 signaling in regulating neutrophil clearance during acute peritoneal inflammation. These effects were driven by IL-6, were independent of other gp130-signaling pathways (e.g., STAT1, ERK1/2), and involved specific down-regulation of neutrophil-attracting chemokine production. Importantly, these current findings expand upon our earlier observations identifying the gp130-STAT3 signaling axis as a key promoter of chemokine-directed T cell recruitment during acute inflammatory responses (25), and therefore further highlight the critical role gp130/STAT3 signaling plays in controlling both innate and adaptive arms of the immune system. Although the studies presented here are based on an acute model of peritoneal inflammation, the involvement of deregulated IL-6/STAT3 signaling in numerous chronic inflammatory conditions (e.g., inflammatory bowel disease and arthritis) (42, 43, 44) suggests a broader implication for the role of gp130/STAT3-mediated chemokine regulation during inflammatory episodes.

We thank Paul Hertzog (Monash Institute of Medical Research, Clayton, Victoria, Australia) for critical reading of this manuscript. We also thank Iain Campbell (School of Molecular and Microbial Biosciences, University of Sydney, New South Wales, Sydney, Australia) for generously providing the Stat1^−/− mice.

The authors have no financial conflict of interest.