The Interleukin-1β/CXCL1/2/Neutrophil Axis Mediates Host Protection against Group B Streptococcal Infection

Abstract

Previous studies have indicated that group B streptococcus (GBS), a frequent human pathogen, potently induces the release of interleukin-1β (IL-1β), an important mediator of inflammatory responses. Since little is known about the role of this cytokine in GBS disease, we analyzed the outcome of infection in IL-1β-deficient mice. These animals were markedly sensitive to GBS infection, with most of them dying under challenge conditions that caused no deaths in wild-type control mice. Lethality was due to the inability of the IL-1β-deficient mice to control local GBS replication and dissemination to target organs, such as the brain and the kidneys. Moreover, in a model of inflammation induced by the intraperitoneal injection of killed GBS, a lack of IL-1β was associated with selective impairment in the production of the neutrophil chemokines CXCL1 and CXCL2 and in neutrophil recruitment to the peritoneal cavity. Decreased blood neutrophil counts and impaired neutrophil recruitment to the brain and kidneys were also observed during GBS infection in IL-1β-deficient mice concomitantly with a reduction in CXCL1 and CXCL2 tissue levels. Notably, the hypersusceptibility to GBS infection observed in the immune-deficient animals was recapitulated by neutrophil depletion with anti-Gr1 antibodies. Collectively, our data identify a cytokine circuit that involves IL-1β-induced production of CXCL1 and CXCL2 and leads the recruitment of neutrophils to GBS infection sites. Moreover, our data point to an essential role of these cells in controlling the progression and outcome of GBS disease.

INTRODUCTION

Streptococcus agalactiae (also known as group B streptococcus [GBS]) is an encapsulated Gram-positive bacterium that can establish a complex series of relationships with the human host (1). Although it most commonly behaves as a commensal of the intestinal and genital tracts, GBS has the potential to cause severe infections, such as pneumonia, sepsis, and meningitis (2). For example, GBS persists as the leading cause of neonatal infections, despite the introduction of universal screening of pregnant women for GBS carriage and intrapartum antibiotic prophylaxis (3). Moreover, in recent years, there has been a steady increase in the number of invasive infections in elderly adults and in patients with underlying conditions, such as liver cirrhosis, diabetes, and malignancies (4).

The innate immune response has a central role in the outcome of GBS infections and is a major determinant in the transition from commensalism to pathogenicity (5–7). Initial recognition of GBS and other microorganisms is mediated by innate immunity receptors (also called pattern recognition receptors [PRRs]), which are capable of recognizing evolutionarily conserved products (i.e., pathogen-associated molecular patterns [PAMPs]) that are common to large groups of microbial species. Membrane-bound PRRs, such as Toll-like receptors (TLRs), and cytosolic NOD-like receptors (NLRs) sense bacterial molecules produced early during infection and orchestrate defensive responses consisting of the activation of NF-κB, interferon regulatory factors, and other transcription factors and of nontranscriptional activation phenomena (8). Prominent among the latter is the assembly of a multimeric complex, the inflammasome, culminating in the cleavage of procaspase-1 in its mature, active form (9). Activated caspase-1 is responsible for the processing and subsequent secretion of the proinflammatory cytokines interleukin-1β (IL-1β) and interleukin-18 (IL-18) (10) and for the development of a lytic form of cell death, called pyroptosis (11, 12), which contributes to the initiation and propagation of inflammation (13). The NLRs better known to be involved in caspase-1 activation include NLRC4 (or interleukin-converting enzyme protease-activating factor) and pyrin domain-containing NLRs (NLRPs), which form large complexes consisting of the adaptor molecule apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) and of caspase-1.

Selected TLRs and the NLRP3 inflammasome were shown to play a significant role in GBS recognition and infection (14). TLR2/TLR6 homodimers expressed in macrophages sense extracellularly released GBS lipoproteins by a mechanism that does not require cell-to-cell contact (15). Moreover, whole live or killed GBS cells can potently stimulate tumor necrosis factor alpha (TNF-α) and pro-interleukin-1 (pro-IL-1) production through the activation of as yet unidentified endosomal TLRs by a mechanism that requires phagocytosis and phagolysosomal processing (14). In dendritic cells and macrophages, intracellular pro-IL-1 and pro-IL-18 are cleaved by caspase-1, after activation of the NLRP3 inflammasome by the GBS cytolysin/beta-hemolysin, resulting in the release of the mature cytokine forms (14). Notably, NLRP3 inflammasome activation makes a nonredundant contribution to anti-GBS host defenses, as shown by the decreased ability of mice deficient in NLRP3, ASC, or caspase-1 to control GBS replication in vivo (14). However, the mechanisms underlying the protective activities of the NALP3 inflammasome during GBS infection, including the potential role of IL-1β or IL-18, have not been studied. In the present study, we found that IL-1β has a central role in host defenses against GBS and that this effect is mediated by the increased production of the neutrophil chemokines CXCL1/2 and the consequent recruitment of neutrophils to infection sites.

MATERIALS AND METHODS

Mice.

IL-1β^−/− (IL-1β-deficient or -knockout [KO]) mice (The Jackson Laboratory) and control wild-type (WT) C57BL/6 mice were purchased from Charles River Italia. The mice were housed and bred under pathogen-free conditions in the animal facilities of the SPGMB Department of the University of Messina.

Experimental models of GBS disease.

Six-week-old female mice were injected intraperitoneally (i.p.) or intravenously (i.v.) with the H36B GBS strain at the doses indicated below as previously described (16–18). Briefly, bacteria were grown to the mid-log phase in Todd-Hewitt broth (Oxoid) and were diluted to the appropriate concentration in phosphate-buffered saline (PBS; 0.01 M phosphate, 0.15 M NaCl, pH 7.2) before inoculation of animals. In each experiment the actual number of injected bacteria was determined by colony counts. Mice were observed every 12 h for 15 days after inoculation. Deaths, however, were never observed after 4 days. Where indicated, WT mice were pretreated i.v. with 100 μg of rat monoclonal anti-mouse Ly-6G antibody or rat IgG2a control (isotype control; BD Pharmingen) 24 h before i.p. inoculation with 2 × 10⁵ CFU of GBS. In other experiments, IL-1β-defective mice were injected i.p. with mature recombinant IL-1β (rIL-1β; 50 ng in PBS; R&D Systems) or vehicle at the time of i.p. challenge with 2 × 10⁵ CFU of GBS. In some experiments, mice were sacrificed at the times indicated below to measure the bacterial burden in the peritoneal lavage fluid, blood, kidney, and brain. Before obtaining these organs, anesthetized mice were transcardially perfused with 20 ml of PBS to clear the intravascular compartment of leukocytes, bacteria, and cytokines. The brain and kidneys were collected as described previously (19), placed in preweighed sterile tubes (M-tubes; Miltenyi Biotech) containing PBS, and homogenized in a gentleMACS system (Miltenyi Biotech). Serial dilutions were prepared in duplicate and plated on blood agar for colony counts. Cytokine and intracellular myeloperoxidase (MPO) levels (19) in the homogenates were determined as described below.

GBS-induced peritonitis model.

In the GBS-induced peritonitis model, killed GBS, killed Staphylococcus aureus (strain Newman), or zymosan (InvivoGen) was injected i.p. at a dose of 0.5 mg/animal in 0.2 ml of PBS using 6-week-old female mice, and peritoneal lavage fluid was collected at the indicated times to measure cell numbers by flow cytometry and cytokine concentrations (see below). Killed GBS and S. aureus were prepared by heat treatment (80°C for 45 min), followed by extensive washing with distilled water and lyophilization, as described previously (20). The endotoxin levels of these preparations were <0.06 endotoxin units/mg, as determined by Limulus amebocyte lysate assay (PBI).

Cytokine and MPO measurement.

Granulocyte-macrophage colony-stimulating factor (GM-CSF), keratinocyte-derived chemokine (KC; CXCL1), macrophage inflammatory protein 2 (MIP-2; CXCL2), TNF-α, IL-17, IL-1β, and MPO concentrations were determined in duplicate using the murine enzyme-linked immunosorbent assay (ELISA) kits GM-CSF Quantikine, CXCL1/KC Quantikine, CXCL2/MIP-2 DuoSet, TNF-α DuoSet, IL-17 Quantikine, IL-1 beta/IL-1F2 DuoSet, and myeloperoxidase DuoSet according to the manufacturer's recommendations (R&D Systems) and previously described methods (20, 21). The lower detection limits of these assays were, respectively, 7.8, 15.6, 15.6, 15.6, 10.9, 16, and 250 pg/ml. In preliminary experiments, a close correlation between MPO concentrations, as determined here, and neutrophil numbers in kidney and bone marrow cell suspensions was found.

Flow cytometry.

Absolute counts of blood leukocytes in the blood and in peritoneal lavage fluid samples were determined using a BD TruCount system (BD Biosciences). Briefly, 50 μl of freshly collected EDTA-anticoagulated blood or peritoneal lavage fluid was dispensed into TruCount tubes and stained for 20 min at 4°C with antibodies directed against F4/80 (macrophages), CD3 (T lymphocytes), CD19 (B lymphocytes), or Ly-6G (neutrophils) (all from BD), using the respective isotype antibodies as controls. Following 20 min incubation at room temperature in the dark, erythrocytes were lysed for 15 min using a fluorescence-activated cell sorter lysing solution (BD). Samples were analyzed on a FACSCanto II flow cytometer using FlowJo software (both from BD).

Data expression and statistical significance.

Differences in cytokine levels and numbers of CFU in the organs were assessed by one-way analysis of variance and the Student's-Keuls-Newman test. Survival data were analyzed with Kaplan-Meier survival plots, followed by the log rank test (JMP software; SAS Institute). When P values were less than 0.05, differences were considered statistically significant.

Ethics statement.

All in vivo experiments were conducted at the animal facilities of the SPGMB Department of the University of Messina according to European Union guidelines for the handling of laboratory animals and were approved by the local animal experimentation committee (Comitato Etico per la Sperimentazione Animale permit no. 18052010).

RESULTS

IL-1β^−/− mice are hypersusceptible to GBS infection.

Our previous studies indicated that caspase-1 activation, a major consequence of which is the secretion of IL-1β and IL-18, plays a significant role in anti-GBS host defenses. To investigate whether the protective effects of caspase-1 could be accounted for by IL-1β release, we tested IL-1β^−/− mice for susceptibility to GBS infection. After i.p. inoculation with 2 × 10⁵ CFU of GBS strain H36B, all the WT control mice survived the challenge, while 50% of IL-1β-defective mice died (P < 0.05; Fig. 1A). To further assess the ability of the immune-deficient animals to control GBS infection, in subsequent experiments mice were infected as outlined above and counting of the colonies in peritoneal lavage fluid and blood samples obtained at 24 h after challenge was performed. Bacteria were rarely detected in samples obtained from WT animals but were always present, often in high numbers, in samples obtained from IL-1β-deficient mice (P < 0.05; Fig. 1B and C). These data indicate a significant impairment in the ability of the host to control localized GBS infection and bacterial spread into the blood. Next, we specifically analyzed the involvement of IL-1β in the ability of the host to clear GBS from the bloodstream and prevent hematogenous colonization of target organs, such as the brain, in which GBS can cause severe pathology. To this end, bacteria were injected directly into the bloodstream of WT and IL-1β-deficient mice. In these experiments, animals were inoculated with five times as many bacteria as those previously used for i.p. infection, on the basis of our previous findings that mice are less susceptible to i.v. than i.p. infection (18, 22). In a first series of studies, the numbers of CFU in the organs of surviving animals were measured at 24 and 48 h after i.v. challenge. At 24 and 48 h after infection, 21 and 36% of the immune-deficient animals, respectively, had died, while all of the WT mice were alive and showed no signs of disease. GBS counts were consistently higher in blood, brain, and kidney samples from the surviving immune-defective mice than samples from WT animals (Fig. 2A to C). To observe signs of disease and lethality over a 15-day period, we performed further experiments, in which groups of WT and IL-1β-KO mice were also inoculated i.v. Under these conditions, 10 out of 16 animals (62%) developed neurological signs, including lethargy and paralysis, and eventually died, while all control mice survived with no sign of disease (Fig. 2D). These data indicate that IL-1β plays a crucial role in the clearance of GBS from the bloodstream and in preventing hematogenous colonization of the brain and kidneys. Thus, together, this first set of studies conclusively indicated that IL-1β promotes the clearance of infection and has a major impact on GBS disease.

FIG 1.

Mice lacking IL-1β are highly susceptible to GBS infection. (A) Survival of WT or IL-1β-deficient mice after i.p. challenge with 2 × 10⁵ CFU of GBS. Shown are cumulative data from two experiments, each involving 9 animals per group. *, P < 0.05 versus WT mice determined with Kaplan-Meier survival plots. (B and C) Numbers of bacterial CFU/ml in peritoneal lavage fluid (B) and blood (C) of WT or IL-1β-deficient mice at 24 h after i.p. challenge with 2 × 10⁵ CFU of GBS. Horizontal bars indicate mean values. Each determination was conducted on a different animal in the course of two experiments, each involving 7 animals per group. *, P < 0.05 versus WT mice determined by one-way analysis of variance and the Student's-Keuls-Newman test.

FIG 2.

Effects of a lack of IL-1β on hematogenous colonization of target organs by GBS. (A to C) Colony counts in the blood, brain, and kidney at 24 and 48 h after i.v. infection with 1 × 10⁶ CFU of GBS. Horizontal bars indicate mean values. Each determination was conducted on a different animal in the course of two experiments, each involving 7 animals per group. *, P < 0.05 versus WT mice determined by one-way analysis of variance and the Student's-Keuls-Newman test. (D) Survival of WT C57BL/6 and IL-1β-deficient mice after i.v. challenge with 1 × 10⁶ CFU of GBS. Shown are the cumulative results of two experiments, each involving 8 animals per group. *, P < 0.05 versus WT mice determined with Kaplan-Meier survival plot.

IL-1β-deficient mice have defective systemic and local neutrophil mobilization.

In further experiments, we sought to analyze the mechanisms underlying the increased susceptibility to GBS infection observed in the absence of IL-1β. Since the application of IL-1β to tissues or body cavities can result in localized neutrophil influx (23), we hypothesized that IL-1β-KO mice would display reduced neutrophil recruitment at sites of GBS infection. To test this, we set out to develop an experimental model whereby we could measure the influx of inflammatory cells in the peritoneal cavity after GBS challenge. We first determined the number and types of cells present under basal conditions in the peritoneal cavity and found that they did not differ between IL-1β-KO and wild-type mice (Fig. 3A). Next, mice were inoculated i.p. with killed GBS and peritoneal lavage fluid was obtained at different times after challenge. The use of live GBS was avoided because of the potentially interfering effects of uncontrolled bacterial growth in mice lacking IL-1β, as indicated by our previous experiments (Fig. 1B). As shown in Fig. 3A, cell counts in the peritoneal cavity rapidly increased after i.p. inoculation with killed GBS, peaking at 3 h and remaining elevated over the baseline values for up to 48 h. The early increase in cell counts was accounted for by neutrophils, while a moderate macrophage influx was observed at 24 and 48 h. GBS-induced cell recruitment was significantly blunted in IL-1β-defective mice (Fig. 3A), with a selective decrease in neutrophil counts being detected at 3 and 6 h after challenge (Fig. 3E). In contrast, macrophage influx was not decreased in the IL-1β-defective mice (Fig. 3B). These data indicate that IL-1β deficiency is associated with a marked selective decrease in neutrophil numbers in GBS-induced peritoneal exudates.

FIG 3.

Impaired neutrophil recruitment in IL-1β-deficient mice during GBS-induced peritonitis. (A to E) Cell counts in peritoneal lavage fluid samples after i.p. challenge with heat-killed GBS (0.5 mg/mouse). (A) Kinetics of total cell influx in WT and IL-1β-defective mice; (B to E) kinetics of recruitment of cells positive for F4/80 (macrophages; B), CD3 (T lymphocytes; C), CD19 (B lymphocytes; D), and GR1 (granulocytes; E) in WT and IL-1β-defective mice. (F) Blood neutrophil counts in WT and IL-1β-defective mice before and after i.p. challenge with 2 × 10⁵ GBS. Data are expressed as the means + SDs of three determinations, each conducted in a different animal, during the course of one of three experiments producing similar results. *, P < 0.05 relative to WT mice determined by one-way analysis of variance and the Student's-Keuls-Newman test.

To assess whether this effect is specific for GBS, we also studied neutrophil recruitment in zymosan-induced peritonitis, a classical inflammation model, in the absence of IL-1β. The effects of IL-1β deficiency were also analyzed using Staphylococcus aureus, a widely studied extracellular pathogen, to induce peritonitis. Figure S1 in the supplemental material shows that neutrophil recruitment in the peritoneal cavity was significantly reduced in IL-1β-defective mice relative to WT ones in which peritonitis was induced by heat-killed S. aureus but not in mice in which peritonitis was induced by zymosan. The lack of an effect of IL-1β deficiency in zymosan-induced peritonitis that we observed here is in agreement with similar observations previously conducted in IL-1 receptor (IL-1R)-defective mice (24). These data suggest the involvement of IL-1β in promoting neutrophil recruitment after inoculation with cells obtained from two different species of bacteria, such as GBS and S. aureus. Interestingly, however, this effect of IL-1β seemed quite specific, since it was not observed using zymosan, a purified fungal cell wall component which has been widely used as a proinflammatory stimulus.

Because administration of IL-1α/β can induce the mobilization of neutrophils from the bone marrow into the circulation (25), we also assessed whether a lack of IL-1β influences blood neutrophil counts during GBS infection. No differences in blood neutrophil counts between wild-type and IL-1β-defective mice were detected under resting conditions (Fig. 3F). During GBS infection, however, neutrophil counts were significantly lower in IL-1β-deficient mice, consistent with the defective mobilization of these cells from the bone marrow. Next, since neutrophil recruitment in the blood and at infection sites is orchestrated by chemotactic cytokines/chemokines, such as GM-CSF, CXCL1 (KC), CXCL2 (MIP-2), TNF-α, and IL-17, whose expression may be increased by IL-1β, in further experiments we determined whether production of these mediators is reduced in IL-1β-defective mice. IL-17 levels in the exudates obtained from wild-type mice after inoculation with heat-killed GBS were below the limits of detection of the assay at all tested times (0, 3, 6, 18, and 24 h; data not shown). In contrast, elevations in GM-CSF, CXCL1, CXCL2, and TNF-α concentrations were measured early after challenge, peaking at 3 to 6 h and declining thereafter (Fig. 4A to D). Notably, IL-1β-defective mice showed significantly decreased levels of CXCL1 and CXCL2, but not TNF-α or GM-CSF, compared with those in WT mice, particularly at early time points. Taken together, these data suggest that IL-1β is involved in the acute inflammatory response to GBS and is required for the production of the neutrophil chemokines CXCL1/2.

FIG 4.

Kinetics of cytokine production in peritoneal lavage fluid after challenge with killed GBS. (A to D) GM-CSF, CXCL1 (KC), CXCL2 (MIP-2), and TNF-α protein levels in peritoneal lavage fluid from WT and IL-1β-deficient mice were measured at the indicated times after i.p. injection of killed GBS (0.5 mg/mouse). Data are expressed as the means ± SDs of three observations, each conducted on a different animal, during the course of one of two experiments producing similar results. *, P < 0.05 versus WT mice determined by one-way analysis of variance and the Student's-Keuls-Newman test.

Defective chemokine production and neutrophil infiltration at infection sites.

In further experiments, we sought to ascertain whether neutrophil chemokine levels were reduced in the tissues of IL-1β-deficient mice inoculated with live GBS. To this end, we infected mice by the i.v. route and obtained, after different times, the kidneys and brains, which are the main targets of hematogenous GBS dissemination. Protein levels of CXCL1, CXCL2, and IL-1β in organ homogenates obtained at 0, 6, and 18 h after i.v. inoculation with GBS were measured by ELISA (Fig. 5A to C). Elevated IL-1β levels were measured in samples from WT mice, while, as expected, this cytokine was undetectable in the organs of IL-1β-deficient mice. These organs also had significantly decreased CXCL1 and CXCL2 protein levels compared with those of WT mice early after challenge. Notably, significantly lower levels of MPO, a quantitative marker of neutrophil presence, were also measured in the organs of IL-1β-defective mice, despite the presence of higher bacterial numbers (Fig. 5D to F). To study in more detail the kinetics of early chemokine induction, in further experiments we measured chemokine, IL-1β, and MPO levels at 1, 3, and 6 h after i.v. infection with the same bacterial dose. Elevations of all of these proteins were measured as early as 1 h after infection and progressively increased over time in the organs of GBS-infected mice. Chemokine and MPO levels were significantly reduced at 1, 3, and 6 h in the IL-1β-deficient mice, while IL-1β was undetectable, as expected (see Fig. S2 in the supplemental material). Taken together these findings corroborate and extend the results previously observed with the model of challenge with killed GBS and indicate that IL-1β-deficient mice have a severe defect in CXCL1/2 chemokine production and in the subsequent recruitment of neutrophils to GBS infection sites.

FIG 5.

Effect of a lack of IL-1β on chemokine production and neutrophil influx during GBS infection. CXCL1 (KC), CXCL2 (MIP-2), IL-1β, and MPO protein levels (A to D) or numbers of CFU (E and F) in organ homogenates from WT and IL-1β-deficient mice were measured at the indicated times after i.v. infection with 1 × 10⁶ CFU of GBS. Data are expressed as the means ± SDs of three observations, each conducted on a different animal, during the course of one experiment. Horizontal bars indicate mean values. *, P < 0.05 versus WT mice determined by one-way analysis of variance and the Student's-Keuls-Newman test.

Neutrophil depletion impairs anti-GBS defenses.

The above-described experiments raised the possibility that the hypersusceptibility to GBS infection observed in IL-1β-deficient mice is linked to their decreased ability to recruit neutrophils to infection sites. However, little is known of the role of neutrophils in GBS disease. To investigate this, we analyzed the effects of neutrophil depletion on the infection process. We first determined that the i.v. inoculation of mice with 100 μg of rat monoclonal anti-Ly-6G antibody was sufficient to reduce the percentage of neutrophils from 6.3% ± 0.8% to 1% ± 1% of total peripheral blood leukocytes by 24 h. Low neutrophil counts persisted for at least 72 h, while other leukocyte populations were not affected (data not shown). Moreover, anti-Ly-6G antibody treatment severely reduced neutrophil influx into the peritoneal cavity, which was measured at 3 and 6 h after inoculation with heat-killed bacteria (Fig. 6A and B). Next, we assessed the impact of neutrophil depletion on the severity of GBS infection. After a low-dose GBS challenge, anti-Ly-6G-pretreated mice suffered from severe illness, which resulted in 75% mortality on day 3. In contrast, all control mice that had been pretreated with normal rat IgG survived the same challenge (Fig. 6C). Blood, kidney, and peritoneal lavage fluid cultures performed at 24 after GBS challenge confirmed the inability of granulocyte-depleted animals to control bacterial growth (Fig. 6D to F). Collectively, these data indicate that neutrophil depletion results in a severe impairment of the host's ability to control the growth of GBS and to prevent the systemic spread of bacteria from the initial site of infection. Thus, neutrophil depletion recapitulated the phenotype of GBS-infected IL-1β-deficient mice.

FIG 6.

Effects of neutrophil depletion on anti-GBS defenses. (A and B) Number of peritoneal cells positive for F4/80 (macrophages), CD3 (T lymphocytes), CD19 (B lymphocytes), and GR1 (granulocytes) in WT mice pretreated with anti-Ly-6G antibody or the isotype control after challenge with heat-killed GBS (0.5 mg/mouse). Data are expressed as the means + SDs of three independent experiments. *, P < 0.05 relative to isotype control-pretreated mice by one-way analysis of variance and the Student's-Keuls-Newman test. (C) Survival of WT mice pretreated with anti-Ly-6G antibody or the isotype control after i.p. challenge with 2 × 10⁵ CFU of GBS. Shown are the cumulative results of two experiments, each involving 8 animals per group. *, P < 0.05 relative to isotype control-pretreated mice determined with Kaplan-Meier survival plots. (D to F) Blood, kidney, and peritoneal lavage fluid bacterial numbers in WT mice pretreated with anti-Ly-6G antibody or the isotype control at 24 h after i.p. challenge with 2 × 10⁵ CFU of GBS. Each determination was conducted on a different animal in the course of one experiment involving 8 animals per group. *, P < 0.05 relative to isotype control-pretreated mice by one-way analysis of variance and the Student's-Keuls-Newman test.

Administration of rIL-1β with the GBS inoculum rescues IL-1β-deficient mice.

In view of the essential role of IL-1β in promoting neutrophil recruitment and host defense against GBS, we evaluated whether the administration of active recombinant p17 IL-1β (rIL-1β) to IL-1β-defective mice could rescue their phenotype. To this end, we injected IL-1β-defective mice with rIL-1β (50 ng) at the time of i.p. challenge with 2 × 10⁵ CFU of GBS. Under these conditions, administration of rIL-1β resulted in 100% survival, while 8 out of 16 (50%) of the vehicle-treated immune-defective mice died (Fig. 7A). Moreover, administration of one dose of rIL-1β (50 ng) given i.p. with the GBS inoculum, as in the lethality experiments described above, resulted in blood, peritoneal lavage fluid, and kidney counts that were significantly lower than those for IL-1β-deficient mice inoculated with GBS plus vehicle (Fig. 7B to D). Thus, administration of active rIL-1β rescued the phenotype of IL-1β-deficient mice, demonstrating that the active form of IL-1β is sufficient for GBS clearance in vivo.

FIG 7.

Administration of rIL-1β restores host defenses in IL-1β-deficient mice. (A) Survival of IL-1β-deficient mice treated with rIL-1β or vehicle after i.p. challenge with 2 × 10⁵ CFU of GBS. *, P < 0.05 versus IL-1β-deficient mice treated with rIL-1β determined with Kaplan-Meier survival plots. Shown are cumulative results from two experiments, each involving 8 animals per group. (B to D) Blood, peritoneal lavage fluid, and kidney colony counts in IL-1β-deficient mice treated with rIL-1β or vehicle at 24 h after i.p. challenge with 2 × 10⁵ CFU of GBS. *, P < 0.05 versus IL-1β-deficient mice treated with rIL-1β by one-way analysis of variance and the Student's-Keuls-Newman test. Each determination was conducted on a different animal in the course of one experiment involving 8 animals per group.

DISCUSSION

Since we previously showed that GBS potently stimulates IL-1β secretion in macrophages and dendritic cells (14), in the present study we sought to investigate the in vivo functions of this cytokine during GBS infection. We initially found that IL-1β is a fundamental, nonredundant factor for host resistance against GBS. These data extend to this pathogen the results of previous work on the importance of IL-1β or IL-1R in host defenses against other species of extracellular bacteria, such as Streptococcus pneumoniae and S. aureus (26–30). Next, we went on to study the mechanisms underlying the effects of IL-1β and focused on neutrophilic polymorphonuclear leukocytes for three reasons. First, we preliminarily observed that blood neutrophil counts were selectively reduced in IL-1β-defective mice during GBS infection. Second, it is well established that IL-1β can contribute to the recruitment of neutrophils into the bloodstream and to infected body sites by a variety of mechanisms (31, 32). Third, little is known of the functions and dynamics of neutrophils in the course of GBS infection, despite clinical observations that neonates with decreased numbers of these cells are more susceptible to GBS sepsis (33, 34).

To investigate the role of IL-1β in neutrophil mobilization, we set up an experimental model whereby leukocytes are recruited to the peritoneal cavity after challenge with GBS. We initially observed that, within few hours after the i.p. injection of bacteria, there was a marked influx of neutrophils, followed, several hours later, by a more moderate macrophage recruitment. Notably, neutrophil influx, but not macrophage influx, was severely reduced in IL-1β-defective mice. Since the dynamics and mechanisms of leukocyte recruitment can vary in different tissues according to the predominant resident cell types (31), we sought to confirm the effects of IL-1β on neutrophil recruitment in different body sites. It was found that, during infection, IL-1β promotes neutrophil infiltration into the brain and into the kidneys, which are important targets of the hematogenous dissemination of GBS. These findings indicated that IL-1β-mediated neutrophil recruitment occurs in different organs during GBS infection and raised the possibility that the decreased antibacterial resistance of IL-1β-defective mice is due to the decreased influx of these cells into infection sites. To verify this hypothesis, we analyzed the effects of neutrophil depletion in the context of GBS disease. Administration of an antibody directed against Gr1 (Ly-6G) recapitulated the phenotype of IL-1β deficiency in two ways. First, it severely reduced the neutrophil counts in GBS-induced exudates. Second, it resulted in overwhelming septicemia and death under conditions whereby control animals restricted bacterial growth and eventually cleared the infection. These findings point to a crucial role of neutrophils during GBS infection and indicate that the antimicrobial activities of these cells are critical to overcome the disease. Moreover, since the numbers of macrophages in the exudates of anti-Ly-6G-treated mice were not reduced, it is likely that these cells alone are not sufficiently effective to clear the invading GBS. Thus, our results confirm and extend previous observations on the importance of neutrophils in controlling the replication of extracellular bacteria, such as S. aureus and Vibrio cholerae, during experimental infection (35, 36). Having detected a link between neutrophils and IL-1β, we next investigated whether reduced neutrophil recruitment to GBS infection sites could be related to the IL-1β-dependent production of secondary mediators involved in neutrophil recruitment, including TNF-α, IL-17, and GM-CSF. Moreover, we focused on the ERL CXC chemokines CXCL1/2, which are characterized by potent, CXCR2 receptor-dependent neutrophil chemoattractant activity, whereas the CXC chemokines lacking the ELR motif are inactive toward neutrophils (37). We found that the absence of IL-1β resulted in a selective decrease in the early production of CXCL1 and CXCL1 after GBS stimulation using both the i.p. inflammation and the i.v. infection models. Our data are in general agreement with those from earlier studies showing that IL-1β promotes the production of neutrophil ERL CXC chemokines during acute inflammatory arthritis (38–40) and S. aureus-induced dermatitis (27). All together, the present study identifies a cytokine circuit whereby GBS-induced secretion of IL-1β stimulates resident tissue cells to produce ERL CXC chemokines. Since both CXCL1 and MIP-2 promote the migration of neutrophils from the bone marrow into the circulation (41) and the recruitment of these cells to local inflammation sites (31), impaired production of these chemokines can account not only for the reduced neutrophil influx into peripheral tissues but also for the neutropenia that we observed in IL-1β-deficient mice during GBS infection. Of course, our data do not exclude the possibility that other IL-1β-inducible factors, such as lipid mediators and endothelial adhesion molecules, in addition to CXC chemokines, may also participate in neutrophil mobilization during GBS infection. In addition, it should be pointed out that cytokines other than IL-1β can induce CXCL1/2-dependent neutrophil mobilization. For example, similar to IL-1β, TNF-α can induce rapid, CXCL1/2-dependent neutrophil recruitment after i.p. injection in mice (42). In addition, i.v. injection of granulocyte colony-stimulating factor causes CXCL1 release and rapid and selective neutrophil release from the bone marrow. Therefore, our data cannot rule out the possibility of and actually suggest the involvement of mediators other than IL-1β in GBS-induced leukocyte recruitment. It is important to note in this respect that, in the present study, CXCL1/2 levels were normal in the tissues of GBS-infected IL-1β^−/− mice late during infection, a time at which neutrophil infiltration was only partially reduced (Fig. 5), suggesting that, under some circumstances, the absence of IL-1β can be compensated for by other endogenous factors.

It should also be noted that, in a previous study, the release of IL-1β in macrophage and dendritic cell cultures critically required GBS hemolysin-induced caspase-1 activation (14). In contrast, in the present study, IL-1β could be released not only in response to live bacteria but also after the i.p. injection of heat-killed GBS, i.e., in the absence of active hemolysin. These features are likely related to differences in the IL-1β maturation/secretion mechanisms operating in vivo and those operating in vitro. In fact, recent studies using different inflammation models indicate that in vivo, but not in vitro, processing of IL-1β is at least partially caspase-1 independent (42–45). Preliminary data from our laboratory suggest that caspase-1-independent IL-1β production also occurs during GBS infections.

In conclusion, our data show that IL-1β-dependent neutrophil recruitment to sites of GBS infection is absolutely required for bacterial clearance. The critical role of neutrophils in anti-GBS defenses that we observed here in mice may also be seen in humans, since neonates with defective neutrophil number or function are particularly susceptible to infection by this pathogen (33, 34). Therefore, our finding may be useful in devising alternative strategies, based on immune-stimulatory interventions on the IL-1/CXL1/2/neutrophil axis, to treat GBS infections.