TLR2-Induced IL-10 Production Impairs Neutrophil Recruitment to Infected Tissues during Neonatal Bacterial Sepsis Free

Newborns are highly susceptible to a variety of infectious diseases. At birth, they undergo a dramatic transition from the sterile intrauterine environment to the nonsterile outside world. Given the limited exposure to Ags in utero and the defects in neonatal adaptive immunity (1), newborns must rely on their innate immune system to mount a response against invading pathogens (2). Neutrophils are one of the first blood cells to be recruited into the site of infection to eliminate the microorganisms. Immune cells are activated by sensing conserved molecular signatures associated with pathogens through a limited number of germline-encoded pattern-recognition receptors, including the TLRs (3). Engagement of different TLRs can induce overlapping, yet distinct, patterns of gene expression that contribute to an inflammatory response and to the pathophysiology of sepsis (4–8). Among mammalian TLRs, TLR2 recognizes the largest number of ligands (9–14). Thus, during infection in the early period of life, TLR2 is likely stimulated locally and systemically by a variety of microorganisms. Among these, the Gram-positive Group B Streptococcus (GBS) are responsible for severe forms of neonatal disease, such as pneumonia, sepsis, and meningitis (15–17). In addition to the well-characterized TLR2-mediated inflammation, data exist supporting the notion that TLR2 signaling can lead to the production of the anti-inflammatory cytokine IL-10 (18–20). We showed previously that the susceptibility of adult and neonatal mice to GBS infection is dependent on early host IL-10 production (21, 22). In the present work, we investigated whether TLR2-induced IL-10 production contributes to the sepsis process by preventing neutrophil recruitment into infected tissues. Our data clearly indicate that TLR2-induced IL-10 production is a key signaling pathway in the innate immune response to neonatal bacterial infections.

GBS strains NEM316 (ST23) and COH-1 (ST17), both belonging to the capsular serotype III, were cultured at 37°C in Todd–Hewitt broth or agar (Difco Laboratories) containing 5 μg/ml colistin sulfate and 0.5 μg/ml oxalinic acid (Streptococcus Selective Supplement; Oxoid).

Six- to eight-week-old male and female BALB/c, C57BL/6, and TLR2-deficient C57BL/B6.129-Tlr2^tm1Kir/J (TLR2^−/−) mice were purchased from The Jackson Laboratory. IL-10–deficient BALB/c (IL-10^−/−) mice were kindly provided by Dr. A. O’Garra (National Institute for Medical Research, London, U.K.). All animals were kept at the animal facilities of the Institute Abel Salazar during the experiments. All procedures were performed according to the European Convention for the Protection of Vertebrate Animals used for Experimental and Other Scientific Purposes (ETS 123) and 86/609/EEC Directive and Portuguese rules (DL 129/92). All animal experiments were planned to minimize mice suffering.

Neonatal (2-d-old) C57BL/6 wild-type (WT) or TLR2^−/− mice were infected i.p. with 10⁶ CFU GBS NEM316 or 10⁵ CFU GBS COH-1 in a maximum volume of 40 μl. Newborns were kept with their mothers during the experiment. Survival curves were determined over a 30-d experimental period. Liver, lungs, and brain of infected pups were aseptically removed at indicated time points and homogenized in PBS. Blood was collected in heparinized tubes and centrifuged to collect the sera. When possible, 10 μl blood was saved for CFU counts. The sera were stored at −80°C until analysis. To quantify the bacterial load, serial dilutions in sterile saline were plated on Todd–Hewitt agar and incubated overnight at 37°C.

Newborn or adult sera cytokines were quantified by ELISA (R&D Systems), according to the manufacturer’s instructions.

Total RNA was extracted from liver with TRIzol reagent (Invitrogen), and reverse transcription was performed using M-MLV Reverse Transcriptase (Invitrogen), as recommended by the manufacturer. Transcript products were amplified with platinum TaqDNA polymerase SuperMix (Invitrogen) on a GeneAmp PCR system (Applied Biosystems) using the following specific primer sets (Integrated DNA Technologies): IL-1α, sense, 5′-CTCTAGAGCACCATGCTACAG-3′ and anti-sense, 5′-TGGAATCCAGGGGAAACACTG-3′; IL-6, sense, 5′-CATCCAGTTGCCTTCTTGGGA-3′ and anti-sense, 5′-CATTGGGAAATTGGGGTAGGAAG-3′; IL-10, sense, 5′-ATGCAGGACTTTAAGGGTT-3′ and anti-sense, 5′-ATTTCGGAGAGAGGTACA-3′; TNF-α, sense, 5′-GGCAGGTCTACTTTGGAG-3′ and anti-sense, 5′-ACATTCGAGGCTCCAGTG-3′; and β-actin sense, 5′-GTGGGGCGCCCCAGGCACCA-3′ and anti-sense, 5′-CTCCTTAATGTCACGCACGATTTC-3′.

Amplified products (20 μl) were separated by electrophoresis on 1.5% agarose gels, and the intensity of each band was analyzed using ImageJ software and corrected relative to β-actin gene expression in the same sample.

Ab treatments were performed in newborn mice 12 h prior to GBS challenge, with 30 μg (i.p.) goat anti-mouse IL-1R, rat anti-mouse TNF-α, rat anti-mouse IL-6, or goat anti-mouse IL-10R mAbs (R&D Systems). Control animals received the same amount of control isotype IgGs. Pups from each litter were randomly assigned to control or to experimental groups, marked, and kept with their mother.

Neutrophil recruitment in the lungs of infected pups was evaluated by flow cytometry analysis. Briefly, 18 h after GBS infection, the organs were removed, gently homogenized in HBSS (Sigma), and passed through glass wool to remove cellular aggregates. FITC anti-mouse Ly6G Ab (clone 1A8; BD Pharmingen) was used for neutrophil detection. Bone marrow (BM) and blood neutrophils were stained with PerCP/Cy5.5 anti-mouse CXCR2 (clone TG11/CXCR2; BioLegend) together with FITC anti-mouse Ly6G Ab. Cells were washed, depleted of RBCs by hypotonic lysis, and fixed. Fluorescence was analyzed using an Epics XL cytometer (Beckman Coulter), and data were analyzed with FlowJo software (TreeStar).

Newborn mice were depleted of neutrophils by treatment with purified anti-Ly6G Abs (clone 1A8; BioLegend). Ab treatment was performed twice: 12 h before and immediately after GBS challenge. Each pup was injected with 80 μg anti-Ly6G Abs.

Spleens from neonatal mice (C57BL/6) were removed and mechanically dissociated. B cells were purified by magnetic cell sorting using magnetic CD19⁺ MicroBeads (Miltenyi Biotec). For macrophage differentiation, BM cells, removed from both the femurs and the tibias of adult mice, and fetal liver cells, isolated from newborns, were cultured in the presence of 10% L929 cell-conditioned medium as a source of M-CSF. Purified CD19⁺ cells, CD19⁻ cells, total splenocytes, and macrophages (5 × 10⁵/ml) were stimulated in vitro with Pam3CSK4 (1 μg/ml) for 18 h. For neutrophil isolation, BM cells were removed from both the femurs and the tibias of adult C57BL/6 mice and flushed on ice with HBSS containing BSA (0.1% w/v) and glucose (1% w/v). Cells were pelleted, and erythrocytes were removed by hypotonic lysis. The BM preparation was suspended in Dulbecco’s PBS, layered on a three-layer Percoll (GE-Healthcare) gradient (80, 65, and 55% in Dulbecco’s PBS), and centrifuged at 1200 × g for 30 min at 10°C. Mature neutrophils were recovered at the interface of the 65 and 80% fractions, and purity was ∼85%, as determined by FACS analysis, using anti-Ly6G Abs. Isolated neutrophils were plated on 96-well plates and stimulated for 4 h with IL-10 (4 ng/ml) or Pam3CSK4 (10 μg/ml).

Two-day-old or adult BALB/c mice were injected i.p. with 1 mg/kg Pam3CSK4 (InvivoGen). Blood was collected 1 and 3 h later for cytokine determination.

LTB₄ (Cayman Chemical) was diluted in a 0.9% NaCl aqueous solution (vehicle), filter sterilized, and administered as described (23), with slight modifications. Two-day-old BALB/c mice were injected with Pam3CSK4; 3 h later, they were exposed to LTB₄ or vehicle for 5 min in a whole-body exposure chamber. Pilot experiments were performed earlier to determine the most effective doses of LTB₄ for recruiting neutrophils into the lungs.

Frozen lungs were thawed and homogenized in an iced solution of 0.5% hexadecyltrimethylammonium and 50 mM KPhos (Sigma) (pH = 5). After centrifugation, the supernatants were mixed in a solution of H₂O₂–sodium acetate and tetramethylbenzidine (Sigma). One unit of myeloperoxidase (MPO) activity was defined as the quantity of enzyme that degraded 1 μM peroxide/min at 25°C and was expressed as U/mg tissue sampled.

The Student unpaired t test was used to analyze the differences between groups. Survival studies were analyzed with the log-rank test, and bacterial counts were analyzed using the Mann–Whitney U test using Prism software (GraphPad). A p value < 0.05 was considered statistically significant.

To investigate the role of TLR2 in GBS-induced neonatal sepsis, newborn TLR2^−/− and WT C57BL/6 mice were infected 48 h after birth with 10⁶ CFU of GBS NEM316. Compared with WT mice, the survival of TLR2^−/− mice was markedly increased (Fig. 1A). To investigate whether the longer survival of TLR2^−/− mice was associated with an early control of GBS growth, we next determined organ CFU at 6, 18, and 24 h postinfection. Although no differences were detected at the first two time points, TLR2^−/− pups had significantly lower numbers of viable bacteria compared with WT mice at 24 h postinfection (Fig. 1B). Accordingly, more bacteria were found in the brain and blood of WT mice compared with TLR2^−/− mice (Fig. 1B). In addition, no viable bacteria were detected 7 d postinfection in the organs of TLR2-deficient mice (data not shown).

Local and systemic overactivation of the immune response is critical for the pathogenesis of sepsis, including organ dysfunction (24, 25). Likewise, we compared the serum concentrations of IL-1α, TNF-α, and IL-6 at 6, 18, and 24 h postinfection in WT and TLR2^−/− mice. TLR2-deficient mice presented decreased levels of pro-inflammatory cytokines as well as the anti-inflammatory IL-10 (Fig. 2). IL-1α is of relevance in this situation because it is released from dying cells, as observed in the case of severe pathological tissue necrosis (26). Significantly higher levels of serum IL-1α were observed in WT pups at 6 and 18 h postinfection and declined thereafter, reaching values similar to those in TLR2^−/− pups at 24 h after GBS challenge (Fig. 2A). TNF-α was never detected in appreciable levels and was found almost exclusively in the serum of WT mice (Fig. 2A). The serum levels of IL-6 were similar at 6 h postinfection in both mice lineages. However, in WT mice, IL-6 levels increased slightly to reach a maximum at 18 h and strongly decreased at 24 h, whereas, in TLR2^−/− mice, IL-6 was barely detectable at these time points (Fig. 2A). The relative mRNA expression of these cytokines in the liver of neonatal WT and TLR2^−/− mice was also evaluated by semiquantitative RT-PCR assay (Fig. 2B, 2C). This analysis revealed that the expression of IL-1α was only significantly different between WT and TLR2^−/− newborn mice at 24 h postinfection (i.e., when almost all WT pups had already died). This observation is consistent with the fact that IL-1α is an intracellular cytokine that is not released from healthy tissues in the absence of cell membrane breakdown (27). The transcription of IL-6 was also significantly lower in the liver of infected newborn TLR2^−/− mice compared with WT pups, concordant with what was observed in the serum (Fig. 2A, 2C). The levels of TNF-α mRNA were not significantly different in WT and TLR2^−/− mice, a finding consistent with the posttranscriptional regulation of this cytokine (28). With respect to IL-10, both its serum protein (Fig. 2A) and liver mRNA (Fig. 2B, 2C) levels were significantly higher in WT newborn mice than in TLR2^−/− newborn mice at all time points.

To determine the relative importance of proinflammatory cytokines or IL-10 in GBS-induced sepsis, the activity of individual cytokines was blocked in WT mice before infection. Thus, newborn WT mice were treated 12 h prior to the GBS challenge with anti–TNF-α, anti–IL-1R or anti–IL-6 or with IL-10R mAbs, and survival was analyzed for 30 d. Controls received the same amount of isotype-matched mAbs. Anti–TNF-α mAb treatment did not improve the survival of newborns challenged with GBS, despite a slight delay in mortality (Fig. 3A). Anti–IL-1R mAb treatment reduced the mortality to 60%, whereas anti–IL-6 mAb treatment did not significantly influence the survival of GBS-infected mice (Fig. 3A). Strikingly, nearly all anti–IL-10R mAb–treated mice survived (9/10), whereas all mice in the control and the anti–TNF-α mAb–treated groups died during the first 3 d postinfection (Fig. 3A). Moreover, pups treated with anti–IL-10R mAb exhibited an increased survival rate compared with those treated with anti–IL-1R mAb. To directly assess the influence of IL-10 on the outcome of a lethal challenge of GBS in TLR2^−/− pups, mouse rIL-10 (1 μg) or PBS was administered to TLR2^−/− newborns at 5 and 12 h postinfection. As expected, rIL-10–treated TLR2^−/− pups showed significantly increased mortality compared with PBS-treated controls (Fig. 3B).

Mice pups were infected 48 h after birth with 10⁵ CFU of COH-1, a hypervirulent strain of GBS. All WT pups died within 2 d, whereas only 2 of 19 TLR2^−/− infected pups died within 7 d postinfection (Supplemental Fig. 1A). These results show that TLR2 deficiency increases the survival of neonatal mice even upon infection with a hypervirulent strain of GBS. Moreover, 24 h postinfection, the resistance of TLR2^−/− mice to COH1 infection reflects a lower bacterial colonization of the liver, lungs, and brain compared with WT pups (Supplemental Fig. 1B). Resistance to a lethal sepsis induced by GBS strain COH-1 and associated with TLR2 deficiency was reported (29), and TNF-α was suggested as the molecular mediator of bacterial clearance and septic shock. Therefore, we evaluated serum levels of TNF-α and IL-10 in infected WT and TLR2-deficient pups 18 h postinfection (Supplemental Fig. 1C). Both cytokines were detected only in the serum of WT neonates and not in TLR2-deficient mice. Consequently, anti–TNF-α or anti–IL-10R mAbs were administered to WT pups before GBS challenge. The results showed that only neutralizing IL-10 signaling increased survival (Supplemental Fig. 1D). Thus, host susceptibility to GBS-induced sepsis is associated with IL-10 production through TLR2 signaling, irrespective of the bacterial strain used.

Neutropenia is a serious risk factor for neonatal GBS infections (2). Thus, we hypothesized that TLR2-mediated IL-10 production was inhibiting neutrophil migration into infected organs. No neutrophil infiltration was detected in the lungs of infected WT pups. Indeed, the frequency and total number of neutrophils were almost the same as in noninfected pups 18 h after GBS infection (Fig. 4A, 4B). On the contrary, an efficient recruitment of neutrophils into infected lungs was observed in TLR2^−/− pups (Fig. 4A, 4B), with an ∼4-fold increase in the numbers of neutrophils compared with WT pups (221.58 ± 39.51 × 10⁴ versus 57.56 ± 10.30 × 10⁴ cells/lung) (Fig. 4B). To further test our hypothesis, newborn WT mice were treated with anti–IL-10R mAb prior to GBS challenge, and the total number of neutrophils was evaluated 18 h postinfection. In agreement with our previous observations (21), blocking IL-10 signaling re-established neutrophil migration into the lungs (Fig. 4C). A significantly lower number of viable GBS cells was recovered at all time points postinfection in anti–IL-10R–treated mice compared with untreated mice (Fig. 4D). Moreover, because infection of TLR2-deficient mice also led to a discrete rise in IL-10 levels in sera (Fig. 2A), bacterial load was determined in the lungs (Fig. 4D), liver, and brain of TLR2^−/− mice treated with anti–IL-10R before infection (Supplemental Fig. 2). Interestingly, although TLR2^−/− pups were naturally able to control GBS infection, a lower bacterial load was observed at 6 and 18 h postinfection in neonates treated with anti–IL-10R mAb compared with those treated with isotype control. These results showed a strong association between the circulating level of IL-10 and the bacterial load in the organs of GBS-infected neonates.

To further investigate whether neutrophils are the effector cells responsible for GBS clearance, we next depleted neutrophils with anti-Ly6G mAbs prior to infection in newborn TLR2^−/− mice. Depletion of neutrophils abrogated the survival advantage of TLR2^−/− pups to GBS-induced sepsis (Fig. 4E). Taken together, these results strengthen the hypothesis that early TLR2-induced IL-10 production in WT mice impairs the clearance of a massive GBS challenge due to abrogated recruitment of neutrophils.

Studies in humans and mice suggested that, during sepsis, downregulation of CXCR2 expression on neutrophils impaired their migration to infected sites (5, 30, 31). Therefore, we investigated whether the enhanced neutrophil recruitment observed after anti–IL-10R mAb treatment was associated with an increased CXCR2 expression on circulating neutrophils. As shown in Fig. 5, the surface expression of CXCR2 decreased 6 h after GBS-induced sepsis compared with uninfected pups. However, circulating neutrophils from anti–IL-10R mAb–treated pups displayed a similar decrease in CXCR2 expression (Fig. 5A–C). At this time point, significantly higher neutrophil migration into the lungs was observed in pups treated with anti–IL-10R mAb compared with those treated with isotype mAb (Fig. 5D). Moreover, significant lower numbers of bacteria were recovered from the liver of anti–IL-10R mAb–treated mice compared with isotype mAb–treated pups (Fig. 5E). To further exclude a direct role for IL-10 on CXCR2 downregulation, neonatal BM neutrophils were isolated and cultured in the presence of rIL-10 or Pam3CSK4. As shown in Supplemental Fig. 3, Pam3CSK4 treatment reduced the surface expression of CXCR2 on neutrophils. In contrast, IL-10 treatment did not inhibit surface CXCR2 expression (Supplemental Fig. 3). The supernatants from Pam3CSK4-treated neutrophils were also tested for the presence of IL-10, but it was not detected at any time point tested (data not shown).

These results show that IL-10 is not directly involved in the downregulation of CXCR2 expression on neutrophils, although it plays a critical role in the impairment of neutrophil recruitment to the site of infection.

TLR2^−/− newborn pups are more resistant to GBS-induced sepsis as the result of an efficient recruitment of neutrophils. Based on our results, we propose that systemic activation of TLR2 leads to IL-10 production that inhibits neutrophil recruitment. Interestingly, as early as 3 h after injection, high concentrations of IL-10 could be detected in the sera of 2-d-old BALB/c mice treated i.p. with Pam3CSK4 (Fig. 6A). To substantiate our hypothesis, we developed an in vivo neutrophil-recruitment model in noninfected neonatal mice; pups were nebulized with the potent neutrophil chemoattractant LTB₄ to induce neutrophil infiltration into the lungs. In this model, MPO activity was increased 1 h after stimulation but decreased over time (Fig. 6B). Subsequently, we tested whether TLR2-dependent IL-10 production induced by injection of Pam3CSK4 could suppress the pulmonary influx of neutrophils induced by LTB₄ treatment. Pam3CSK4 or PBS (control) was administered i.p. (defined as time 0), LTB₄ was nebulized at 3 h, and MPO activity and total number of neutrophils in the lungs were evaluated at 4 h. A significant inhibition of LTB₄-induced neutrophil migration was observed after Pam3CSK4 treatment (Fig. 6C, 6D). To confirm that this inhibition was IL-10 dependent, pups were treated with anti–IL-10R mAb prior to Pam3CSK4 administration. Neonatal neutrophils were efficiently recruited into the lungs of anti–IL-10R mAb–treated pups compared with control mice (Fig. 6C, 6D). Moreover, when these experiments were carried out in IL-10–deficient mice (IL-10^−/−), we observed efficient neutrophil recruitment into the lungs after LTB₄ treatment that was not affected by Pam3CSK4 injection (Fig. 6E). These data clearly show that the migration of neutrophils to inflamed or infected neonatal organs is modulated by TLR2-induced IL-10.

We next investigated whether TLR2-mediated IL-10 production is restricted to the neonatal context. For that purpose, cultures of splenic cells derived from neonatal and adult mice were stimulated with Pam3CSK4, and the levels of IL-10 were quantified at 12 and 18 h after stimulation. Higher levels of IL-10 were detected in the supernatant of neonatal cells compared with those from adults (Fig. 7A). Moreover, because neonatal B cells produce high levels of IL-10 (32), we sorted this population (Supplemental Fig. 4) from neonatal and adult spleens and quantified IL-10 production after stimulation with Pam3CSK4. As shown in Fig. 7B, CD19⁺ cells derived from neonates produced higher amounts of IL-10 after TLR2 triggering than did those derived from adults. Moreover, we also investigated IL-10 production by non-B cells and found higher concentrations of this cytokine in the supernatants derived from neonates, albeit at lower levels than in the B cell population (Fig. 7C). The quantification of IL-10 in macrophage cultures showed that this cell population was not responsible for IL-10 production after TLR2 triggering because similar low levels were found in macrophages derived from newborns and adults (Fig. 7D).

Because neonatal cells display a decreased ability to produce proinflammatory cytokine compared with adult cells (33, 34), we analyzed whether the increase in IL-10 production by neonatal cells is a feature specific for this cytokine or is due to a general hyperresponsiveness of the neonatal cells. Pam3CSK4 was administered to neonatal and adult mice, and the serum levels of several cytokines (IL-10, IL-6, IL-12, and TNF-α) were assessed 1 and 3 h later. At 1 h, the levels of all cytokines were similarly low in neonatal and adult mice (data not shown). However, 3 h after treatment, IL-10 was detected only in the serum of neonates (370.50 ± 32.93 pg/ml). At this time point, the serum level of the inflammatory cytokine IL-6 was significantly higher in neonates compared with adult mice (5349.23 ± 933.0 versus 1923.33 ± 148.70 pg/ml; p = 0.0336), whereas IL-12 was detected only in adult serum (474.27 ± 16.07 pg/ml). In contrast, no significant difference was observed in the serum TNF-α levels in neonates and adult mice (138.12 ± 27.67 versus 88.03 ± 4.05 pg/ml; p = 0.2247). Thus, IL-10 production in neonatal cells was not associated with the increased production of two of the three tested proinflammatory cytokines, which indicates that it is not due to a general hyperresponsiveness to stimulation.

Our results demonstrate that TLR2-induced IL-10 production is higher in neonatal cells than in adult cells, an important feature rendering neonates more susceptible to bacterial infection.

Neonatal sepsis remains among the leading causes of death in the world, with an incidence that is predicted to increase each year. Several reports showed a marked defect in neutrophil migration into infected organs during severe sepsis, followed by failure of local bacterial clearance that enables dissemination of infection and, sometimes, death (35, 36). Nonetheless, specific underlying mechanisms of lethal outcomes of bacterial infections in newborns remain poorly understood and, until now, therapies aiming at boosting the granulopoiesis during sepsis failed (37). Furthermore, endogenous G-CSF levels are already elevated in septic neonates, suggesting that end-organ unresponsiveness may account for this lack of effect (38). The basal expression of TLR is similar in neonates and in adults (39). However, the neonatal immune response is characterized by a reduced level of proinflammatory Th1 cytokines and is biased toward Th2/Th17-polarizing and anti-inflammatory cytokine production (2, 33, 40). In this regard, it was shown that, after LPS (TLR4 agonist) stimulation, macrophages from neonatal mice secrete much less proinflammatory cytokines but more IL-10 than do those from adult mice (34). In this study, when macrophages from neonatal IL-10–deficient mice were used, the concentrations of proinflammatory cytokines were similar to those observed in adult mice. Neonates apparently possess regulatory mechanisms that control the inflammatory response. Accordingly, we recently showed that GBS infection in neonates leads to a rapid increase in serum levels of the anti-inflammatory cytokine IL-10, which is responsible for their susceptibility to this pathogen (21). Moreover, in that study, we also showed that abrogation of either IL-10 or IL-10 signaling conferred protection to neonates against GBS strains (21). Several reports showed that leukocytes from newborn mice (32, 34, 41, 42) and human neonates (33, 43, 44) are highly committed to produce increased amounts of IL-10 upon infection. Using a model of GBS-induced neonatal sepsis, we showed in this study that inhibition of either TLR2 or IL-10 signaling leads to an increased survival of newborns due to an efficient neutrophil migration into infected tissue and subsequent bacterial clearance. The fact that TLR2-deficient mice are more resistant to GBS-induced sepsis was reported previously (29), but TNF-α was suggested as the molecular mediator of bacterial clearance and septic shock. Contrastingly, other studies reported that TNF-α levels were not increased in infants with sepsis (45) and that abrogation of circulating TNF-α by anti–TNF-α treatment had little effect on mortality of GBS-infected animals (46–48). We similarly showed in this study that anti–TNF-α mAb treatment did not modify the course of GBS-induced sepsis, whereas treatment with anti–IL-10R mAb significantly increased the survival of neonatal mice. Thus, in neonates, IL-10 produced through TLR2 triggering has a major role in the immunopathogenesis of neonatal GBS-induced sepsis. The role of IL-10 in inflammatory-driven pathologies has been widely studied, but its relevance to neonatal sepsis has been underestimated. Despite the beneficial effects of IL-10 in controlling the degree and the duration of inflammation, the ability of the pathogen to induce host IL-10 production soon after infection is detrimental to the host, because this cytokine contributes to its dissemination. Unrelated microorganisms can induce TLR2-mediated IL-10 production to evade the host defense by downregulating its microbicidal functions (41, 49), and several reports showed that TLR2 agonists are specialized in inducing IL-10 expression by APCs (18–20). In the current study, increased IL-10 serum levels were observed in neonates very soon after treatment with the TLR2 agonist Pam3CSK4, whereas LTB₄-induced lung neutrophil migration was significantly inhibited in these pups. Blocking IL-10 signaling was sufficient to reverse the inhibition induced by the TLR2 agonist. Decreased expression of the chemokine receptor CXCR2 on neutrophils isolated from septic patients (30) and from lethal septic mice (50) was observed previously, as was TLR2-dependent downregulation of CXCR2 expression on circulating neutrophils recovered from septic mice (5). However, the impaired neutrophil migration to infectious sites due to IL-10 production could not be related to their CXCR2 expression because this receptor was similarly downregulated in anti–IL-10R–treated pups that displayed normal neutrophil migration. Rather, our results suggest that IL-10 produced through TLR2 signaling is responsible for the abrogation of neutrophil recruitment. This is in agreement with a recent report showing a significant survival benefit in TLR2^−/− mice with polymicrobial sepsis and a downregulation of CXCR2 expression in peritoneal neutrophils similar to that observed in WT neutrophils (51). A prominent role for CXCL12, but not CXCR2, in polymicrobial sepsis was also reported recently (52), which suggests that other neutrophil chemoattractants could be involved.

Pediatric and adult sepsis possess unique developmental features; however, in both cases, IL-10 can contribute to an increased susceptibility to sepsis. Among the different TLRs, TLR2, TLR4, and TLR9 have been associated with the pathophysiology of experimental adult sepsis (7, 8, 53, 54), and their roles in the upregulation of proinflammatory cytokines, such as IL-6, TNF-α, and IL-1β, is well documented. Interestingly, all of these TLRs also can induce IL-10 production (55). These data further support that IL-10 may also have a central role in the susceptibility of adults to sepsis. Accordingly, the development of an aberrant, prolonged, and massive release of IL-10 may lead to the development of an immunosuppressive state, called sepsis-induced immunoparalysis (56, 57), resulting in the inability to eradicate the primary infection and/or the development of new secondary infections in septic adults (58).

Different subsets of B cells display variable and unequal levels of TLR2 expression, which produce different amounts of IL-10 (59). The ability of B1 cells to produce IL-10 is well established (60). As opposed to adults, newborns display abundant levels of B1 cells (42), which may account for their ability to produce higher amounts of IL-10 (32). We show in this study that stimulation of splenic cell cultures derived from neonates with a TLR2 agonist lead to higher levels of IL-10, secreted almost exclusively by the B cell population; moreover, ∼30% of these B cells are CD5⁺ (a B1 cell marker) contrasting with ∼5% in adults. However, it is worth mentioning that the CD19⁻ cell (non-B) population also produces IL-10, which indicates that the propensity of neonates to produce IL-10 reflects at least the size of the CD5⁺ B cell population and the intrinsic differences in triggering TLR2-mediated IL-10 production. Therefore, we hypothesize that neonates are at risk of rapidly developing an immunoparalysis state after bacterial-induced sepsis, because they are committed to produce IL-10. Accordingly, we show in this study that blocking the inflammatory cytokines had no effect on the infection outcome, whereas blocking IL-10 signaling rendered neonatal mice resistant to GBS-induced sepsis. In the absence of TLR2, GBS is still recognized by other pattern-recognition receptors, such as the NOD-like receptors (61), TLR7, and TLR9 (62), and TLR/NOD/RIG-like receptor-independent pathway (63). It is likely that these pathways synergistically activate an efficient TLR2-independent innate immune response. TLR2 recognition leading to IL-10 production will prevent the development of this efficient immune response by preventing neutrophil recruitment, the effector cells that control bacterial clearance.

In summary, our results provide evidence that, in the context of a systemic neonatal infection, exposure to high levels of bacterial products triggers TLR2-mediated IL-10 production that prevents neutrophil migration to infected organs and bacterial clearance. Thus, IL-10 should be considered a key molecule in the pathophysiology of neonatal sepsis.

We thank Joana Palha for critical reading of the manuscript.

The authors have no financial conflicts of interest.