MyD88 and Type I Interferon Receptor-Mediated Chemokine Induction and Monocyte Recruitment during Listeria monocytogenes Infection¹

Mice and infections

All mice used in this study were bred and maintained at the Memorial Sloan-Kettering Research Animal Resources Center. IFNAR^−/− mice were purchased from B&K Universal, MyD88^−/− mice were obtained from Dr. Shizuo Akira (Osaka University, Osaka, Japan), and MCP-1^−/− mice were obtained from Dr. Barrett Rollins (Harvard Medical School, Boston, MA). IFNAR^−/− and MyD88^−/− mice strains were crossed to obtain MyD88^−/− IFNAR^−/− mice. MyD88^−/− and MCP-1^−/− mice were back-crossed over 10 generations onto the C57BL/6 background while IFNAR^−/− mice crossed for seven generations to the C57BL/6 background. Mice were infected i.v. with 3000 L. monocytogenes strain 10403S. At indicated times following infection, spleens were harvested and dissociated in PBS containing 0.05% Triton X-100 and bacterial CFUs were determined by plating on brain-heart infusion agar plates.

Cell culture

Macrophages were grown from bone marrow precursors in antibiotic-free DMEM medium supplemented with 30% supernatant from l-cell fibroblasts and 20% FBS. On day 5, cells were harvested in ice-cold PBS, and plated at 3 × 10⁵/well in 96-well flat-bottom plates. Bacteria were grown to log phase (A₆₀₀ of 0.1), pelleted at 10,000 rpm for 10 min, and resuspended in PBS before addition to cells. WT 10403S and LLO^ko DP-L2161 bacteria were added to the cells at a 5:1 bacteria:cell ratio. Total volume in the wells during infection was 200 μl in 96-well plates. Infection was allowed to proceed for 1 h at which time extracellular bacteria were washed away, and gentamicin-containing medium was added to each well to prevent extracellular bacterial growth (T = 0). Supernatants were collected 1, 2, 4, and 6 h postinfection to assay for chemokines. TRIzol reagent (In-vitrogen) was added at 15 min, 30 min, 1, 2, 4, and 6 h postinfection to extract RNA from macrophages.

Flow cytometry

At various times after infection, spleens were collected, dissociated, and digested with 0.3% collagenase type 4 (Worthington Biochemical). Bone marrow cells were collected from mouse femurs as described previously (3). The following Abs were purchased from BD Pharmingen: anti-CD11b-PerCP(M1/70), anti-Ly6C-FITC(AL-21). For FACS analysis, a large FSC/SSC gate was drawn to include lymphocyte/monocyte populations.

ELISA

Murine MCP-1 was quantified using an ELISA kit from BD Biosciences. To obtain lysates for chemokine assays, organs were harvested at indicated times following infection, macerated in ice-cold PBS containing 0.01% Triton X-100, and centrifuged at 10,000 × g.

Real-time PCR analysis

Total RNA was isolated from macrophages using the TRIzol reagent (In-vitrogen). DNase-treated RNA underwent randomly primed cDNA synthesis and real-time PCR analysis. SYBR Green-based real-time PCR was performed using the DyNAmo SYBR Green qPCR Kit (Finnzymes). MCP-1-specific primers were obtained from Qiagen. Signals were normalized to GAPDH RNA (forward: 5′-ACCACAGTCCATGCCATCAC-3′; reverse: 5′-TCCACCACCCTGTTGCTGTA-3′). Normalized data were used to quantify relative levels of MCP-1 using ΔΔCt analysis.

Rip2 and Trif are not required for MCP-1 expression following L. monocytogenes infection

L. monocytogenes infection induces inflammatory responses that are mediated by TLRs and MyD88. TLR-mediated recognition of microbial molecules is restricted to the cell surface or to membrane bound intracellular compartments and thus, because L. monocytogenes invades the host cell cytosol, additional innate immune sensing receptors are likely to contribute to the aggregate inflammatory response. Indeed, TLRs and Nod-like receptors (NLRs) mediate distinct transcriptional responses to L. monocytogenes (33). Nod1 and Nod2 are two cytosolic molecules that bind to peptidoglycan-derived molecules and contribute to in vivo defense against intestinal L. monocytogenes infection (34, 35) and lung Chlamydia pneumoniae infection (36). It is unclear, however, whether NLR signaling pathways promote MCP-1 production following L. monocytogenes infection. Rip2 is an intracellular kinase that mediates NLR-signaling and Rip2 deficiency abrogates Nod1 and Nod2 signaling (37, 38). Thus, we investigated in vivo induction of MCP-1 following L. monocytogenes infection in Rip2^−/− mice and in mice lacking MyD88 or Trif, two intracellular adaptor molecules that mediate TLR-signaling. Spleens of Trif^−/− and Rip2^−/− mice that had been infected 24 h earlier with WT L. monocytogenes expressed the same level of MCP-1 as infected WT C57BL/6 mice (Fig. 1A). In contrast, and as demonstrated in previous studies from our laboratory (11), MCP-1 levels were reduced in the spleens of L. monocytogenes infected MyD88^−/− mice (Fig. 1A). We next examined Tip-DC recruitment to the spleens of Rip2, Trif, and MyD88-deficient mice following L. monocytogenes infection. Following infection, the total cell numbers in the spleens of Trif^−/−, Rip2^−/−, and MyD88^−/− mice were similar to WT mice (Fig. 1B). Tip-DCs were identified by high-level expression of Ly6C, and intermediate to high-level expression of CD11b. As demonstrated in Fig. 1, C and D, recruitment of Tip-DCs 24 h following infection was similar in WT, Trif^−/−, Rip2^−/−, and MyD88^−/− mice. In Trif^−/− and Rip2^−/− mice, normal recruitment of Tip-DCs was correlated with normal splenic expression of MCP-1. However, in MyD88^−/− mice, Tip-DC recruitment was normal despite a roughly 70% reduction in splenic MCP-1 levels, suggesting that the level of MCP-1 detected in the spleens of infected WT mice exceeds the amount required for Tip-DC recruitment.

MyD88 and type I IFN signals optimize MCP-1 production following infection

In vivo induction of MCP-1 is correlated with the ability of L. monocytogenes to escape the phagosome and enter the host cell cytosol. In contrast, LLO-deficient L. monocytogenes, which remains confined to the phagosome, does not induce MCP-1 expression in vivo (11). Studies from the Portnoy laboratory have demonstrated that cytosol invasion by L. monocytogenes is accompanied by the induction of IFN-β (39), a type I IFN involved in antiviral, but generally not antibacterial defense. We hypothesized that type I IFN induction provides an important in vivo stimulus for MCP-1 production. To investigate this, we measured the relative contribution of MyD88-mediated and type I IFN-mediated signaling for MCP-1 expression during L. monocytogenes infection. As a first step, bone marrow-derived macrophages (BMMØs) from C57BL/6, IFNAR^−/−, MyD88^−/−, and MyD88^−/−IF-NAR^−/− mice were infected with WT L. monocytogenes and MCP-1 protein levels were determined by ELISA (Fig. 2A). Relative to WT BMMØs, MCP-1 production at 6 h following infection was reduced by 60% in MyD88^−/− BMMØs, 70% in IFNAR^−/− BMMØs, and 85% in MyD88^−/−IFNAR^−/− BMMØs. These results indicate IFNAR- and MyD88-mediated signals both contribute to the production of MCP-1. Next, we infected C57BL/6, IFNAR^−/−, MyD88^−/−, and MyD88^−/−IFNAR^−/− mice and measured MCP-1 levels in spleens and serum by ELISA 24 h following infection. MCP-1 production was markedly reduced in MyD88^−/−IFNAR^−/− mice following L. monocytogenes infection, while MCP-1 protein levels were modestly reduced in both IFNAR^−/− and MyD88^−/− mice (Fig. 2, B and C). Loss of MyD88-mediated signals resulted in a greater decrease in MCP-1 production in vivo compared with loss of IFNAR-mediated signaling. MCP-1 production in the spleen and serum of IFNAR^−/− mice was reduced by 35 and 50%, respectively (Fig. 2, B and C), while the corresponding decrease in MyD88^−/− mice was 75 and 95%, respectively. Consistent with the in vitro findings using BMMØ, MCP-1 protein levels were reduced by 80 and 100% in the spleens and serum of infected MyD88^−/−IFNAR^−/− mice (Fig. 2, B and C). Of note, in spleen and in BMMØ, low levels of MCP-1 were detectable in the absence of both MyD88 and IFNAR-mediated signals, indicating that additional, undefined signaling pathways induce chemokine production, albeit the contribution is relatively minor.

Rapid MyD88-dependent and delayed type I IFN-dependent MCP-1 induction

To further define the relative contributions of IFNAR and MyD88-mediated signaling in MCP-1 expression, we measured induction of MCP-1 mRNA in WT, MyD88^−/−, IFNAR^−/−, and MyD88^−/−IFNAR^−/− BMMØs infected for 15 min, 30 min, 1 h, 2 h, 4 h, and 6 h with WT L. monocytogenes (Fig. 3A). In WT BMMØ, MCP-1 mRNA was induced in two phases. The first phase occurred rapidly, with measurable increases in MCP-1 mRNA 15 min following infection, and reached a plateau 30 min after infection and then declined. During this phase, MCP-1 mRNA levels in infected BMMØs were increased up to 30-fold compared with uninfected BMMØs. The second phase of MCP-1 mRNA induction began 1 h following infection and continued during the course of examination for 5 h, at which time a 50-fold increase in mRNA was measured. Analysis of MCP-1 mRNA levels in infected MyD88^−/− BMMØs demonstrated that the first phase of MCP-1 mRNA induction is MyD88 dependent. The second phase of MCP-1 induction, however, is intact in MyD88^−/− BMMØs and follows similar kinetics as WT BMMØs. In MyD88^−/− BMMØ, MCP-1 mRNA levels remained low until 30 min following infection and then gradually increased. In contrast, in the absence of IFNAR signaling, the second phase of MCP-1 mRNA expression was abrogated while the early, MyD88-dependent phase remained intact. In MyD88^−/−IFNAR^−/− BMMØs, both phases are abrogated and MCP-1 mRNA remained at base level.

To determine the role of cytosol invasion by L. monocytogenes in MCP-1 induction, we used LLO-deficient L. monocytogenes to infect BMMØ. As shown in Fig. 3B, LLO^ko L. monocytogenes elicited MyD88-mediated signals and induced the first phase of MCP-1 mRNA expression. However, because type I IFN production is strictly dependent on cytosol invasion, LLO^ko L. monocytogenes did not induce type I IFN and also did not activate the delayed phase of MCP-1 mRNA induction. To determine whether the second phase was directly induced by type I IFNs, we added IFN-β (1 ng/ml) to uninfected BMMØs or LLO^ko-infected BMMØs. As shown in Fig. 3B, addition of IFN-β activated the second phase of MCP-1 mRNA expression in uninfected BMMØs. When IFN-β was added to macrophages infected with LLO^ko bacteria, both early and delayed phases of MCP-1 mRNA induction were activated (Fig. 3B) and resulted in MCP-1 protein secretion (Fig. 3C) in quantities that approximated those detected in BMMØs infected with WT L. monocytogenes.

Normal homeostatic Ly6C^high monocyte trafficking in MyD88^−/−IFNAR^−/− mice

Our previous studies demonstrated that MCP-1 is required to maintain normal levels of circulating Ly6C^high monocytes under homeostatic conditions (5). Because MyD88- and IFNAR-mediated signals can contribute to MCP-1 expression, we decided to determine whether these signaling pathways contribute to homeostatic monocyte trafficking. Thus, we determined the frequencies of Ly6C^high monocytes in the spleen and bone marrow of naive C57BL/6, IFNAR^−/−, MyD88^−/−, and MyD88^−/−IFNAR^−/− mice. Fig. 4 demonstrated that the frequency differences of Ly6C^high cells in the bone marrow (Fig. 4, A and C) and spleen (Fig. 4, B and D) of naive C57BL/6, IFNAR^−/−, MyD88^−/−, and MyD88^−/−IFNAR^−/− mice were not significant. These results indicated that MyD88- and IFNAR-mediated signaling pathways are not required for maintenance of circulating monocyte frequencies in the absence of infection and suggested that MyD88- and IFNAR-independent signaling pathways stimulate homeostatic production of MCP-1.

Defective inflammatory monocyte recruitment in MyD88^−/−IFNAR^−/− mice but not MyD88^−/− or IFNAR^−/− mice

Previously, our laboratory demonstrated that MyD88-mediated signals are required for activation and differentiation of Tip-DCs, but that recruitment is MyD88 independent, despite reduced MCP-1 production following infection (11). It remains unclear, however, how much MCP-1 is required to trigger monocyte emigration from bone marrow, and whether multiple or a single signaling pathway is essential for this process. When we examined monocyte recruitment in infected IFNAR^−/− mice, we observed similar results as in MyD88^−/− mice, with reduced MCP-1 production but normal numbers of Ly6C^high cells remaining in the bone marrow (Fig. 5, A and C) and trafficking to the spleen (Fig. 5, B and D). However, in the absence of both MyD88- and IFNAR-mediated signals, Tip-DC recruitment following infection was significantly decreased and similar to that detected in MCP-1^−/− mice. As shown in Fig. 5C, ∼3% of Ly6C^high monocytes emigrated from bone marrow in WT, MyD88^−/−, and IFNAR^−/− mice upon infection, while only 1.2% of monocytes emigrated from the bone marrow of MyD88^−/−IFNAR^−/− and MCP-1^−/− mice. Along similar lines, frequencies of Ly6C^high monocytes increased by nearly 2% in spleens of infected WT, MyD88^−/−, and IFNAR^−/− (Fig. 5D), while the increase in MyD88^−/−IFNAR^−/− mice and MCP-1^−/− mice was only 1%. The total numbers of splenocytes in different mouse strains were similar following infection (data not shown). Thus, deficiency of both MyD88 and IFNAR signaling pathways results in a 60% reduction in monocyte egress from bone marrow and a 50% reduction of monocyte trafficking to the spleen relative to deficiency of either of these pathways alone. These results demonstrate that the MyD88-mediated and type I IFN-mediated signaling pathways play redundant roles in monocyte recruitment during infection and that either of these signaling pathways can, in the absence of the other, maintain normal inflammatory monocyte recruitment following L. monocytogenes infection, but that absence of both compromises monocyte recruitment.

Beneficial effect of type I IFN in host defense against L. monocytogenes in the absence of MyD88 signaling

IFNAR^−/− mice are more resistant to L. monocytogenes infection, indicating the type I IFN production is deleterious in mice following L. monocytogenes infection (21–23). Recent studies have demonstrated that type I IFN production during infection enhances lymphocyte apoptosis and leads to IL-10 production, which hampers innate and adaptive immune mechanisms of bacterial clearance (21, 40). In our study, IFNAR signaling complements MyD88 deficiency by maintaining monocyte recruitment. This result suggested that IFNAR deficiency might render MyD88-deficient mice more susceptible to L. monocytogenes infection. To test this hypothesis we infected WT, IFNAR^−/−, MyD88^−/−, and IFNAR^−/− MyD88^−/− mice with virulent L. monocytogenes and measured in vivo bacterial growth in spleen and liver 24 h postinfection (Fig. 6). Consistent with previously published studies, IFNAR^−/− mice harbored fewer viable bacteria compared with WT mice, while MyD88^−/− mice had more severe infections. In contrast to IFNAR deficiency on the WT background, IFNAR deficiency in MyD88-deficient mice resulted in greater susceptibility to L. monocytogenes infection, with increased numbers of viable bacteria in spleen (Fig. 6A) and liver (Fig. 6B). These findings indicated that the detrimental effects of IFNAR-mediated signaling following L. monocytogenes infection are MyD88 dependent. In the absence of MyD88-mediated signaling, however, IFNAR-mediated signals enhance resistance to L. monocytogenes infection.