Phagosomes Induced by Cytokines Function as anti-Listeria Vaccines

Background: The effectiveness of phagosomes as vaccines is unknown.

Results: Listericidal phagosomes contain a compartmentalized signaling pathway and a nontoxic listeriolysin form bound to immune molecules. As vaccines they activate effector T cells and recruit immune cells.

Conclusion: Protection with listericidal phagosomes requires recruitment of dendritic cells and T cell regulation.

Significance: Phagosomes are effective immunotherapies and are a new generation of vaccine tools.

Abstract

Phagosomes are critical compartments for innate immunity. However, their role in the protection against murine listeriosis has not been examined. We describe here that listericidal phago-receptosomes are induced by the function of IFN-γ or IL-6 as centralized compartments for innate and adaptive immunity because they are able to confer protection against murine listeriosis. These phago-receptosomes elicited LLO(91–99)/CD8⁺- and LLO(189–201)/CD4⁺-specific immune responses and recruited mature dendritic cells to the vaccination sites controlled by T cells. Moreover, they present exceptional features as efficient vaccine vectors. First, they compartmentalize a novel listericidal STAT-1-mediated signaling pathway that confines multiple innate immune components to the same environment. Second, they show features of MHC class II antigen-loading competent compartments for cathepsin-D-mediated LLO processing. Third, murine cathepsin-D deficiencies fail to develop protective immunity after vaccination with listericidal phago-receptosomes induced by IFN-γ or IL-6. Therefore, it appears that the connection of STAT-1 and cathepsin-D in a single compartment is relevant for protection against listeriosis.

Introduction

Macrophages (MØ)² are critical cells for the innate and adaptive immune responses against Listeria monocytogenes (1–3). The phagosomal compartments in MØ regulate all of these immune processes by undergoing a profound transformation to mediate efficient listericidal functions, high levels of oxidative burst and lysosomal proteases, and increased antigen processing capacity (4–6). Several pro-inflammatory cytokines such as TNF-α, IFN-γ, and IL-6 enhance the microbicidal mechanisms of MØs and restrict the intracellular growth of L. monocytogenes (7). It is unclear whether the microbicidal signaling of these cytokines is connected with phagosomal trafficking or with protection against infections. Two main listericidal mechanisms the oxidative and nonoxidative pathways operate within the phagosomal compartments. However, degradation of L. monocytogenes requires the action of nonoxidative mechanisms (8–11) that are mediated by lysosomal proteins as cathepsin-D (CTSD). In this regard, CTSD participates in innate immunity and inactivates the main phagosomal L. monocytogenes cytolysin, listeriolysin O (LLO) (12–14). CTSD-mediated degradation of the immunodominant antigen LLO occurs through a unique cleavage site between ⁴⁹¹WW⁴⁹² residues. This site also contains the phagosomal binding domain (15). Therefore, a connection might exist between listericidal components and L. monocytogenes immunity within the phagosomes. Here, we examine the hypothesis that a common listericidal route induced by pro-inflammatory cytokines may be compartmentalized in unique vesicles connecting STAT-mediated signaling, trafficking regulators, listericidal lysosomal enzymes such as CTSD, and immune phagosomal functions. We also examined the possibility that the compartmentalization of functions within phagosomes might be useful to confer protection against listeriosis. Our approach involved the use of differential gene expression methods combined with basic proteomic, functional analyses of L. monocytogenes phagosomes, and their use as vaccine vectors against listeriosis. All these studies were verified using MØs genetically deficient in putative upstream components of this signaling route such as STAT-1 and STAT-3 and the postulated downstream lysosomal component CTSD. Finally, we also evaluated the efficiency of phagosomes as vaccine vectors in wild type and experimental CTSD^low-deficient mice and explored the contribution of T cells in the potency of these vaccines using SCID mice.

In this study, we describe a novel phagosomal compartment, the listericidal phago-receptosomes induced by IFN-γ or IL-6, which may be important Listeria-induced immune vesicles that regulate IL-6 production and constitute effective vaccines to confer protection against listeriosis.

EXPERIMENTAL PROCEDURES

Cell Lines and Cytokine Treatment

The following MØ-like cell line was used through out the study, J-774 cells cultured in DMEM, 10% FCS, 1 mm glutamine, 1 mm nonessential amino acids, 50 μg/ml gentamicin, 30 μg/ml vancomycin. Bone marrow-derived macrophages (BM-DM) were obtained as follows: from femurs of 8–12-week-old female CBA/J, C57BL/6, SCID, or 129/Sv mice (Taconic Farms, Denmark) (for CTSD^low experiments, Ctsd^low mice features were previously reported, 15 or for SCID experiments); from 3-week-old CTSD^−/− and wild type littermate mice CTSD^+/+ (for CTSD^+/+ and CTSD^−/− experiments) (Biochemisches Institute Albrechts-Ludwig-Universität of Kiel, Kiel, Germany); from femurs of 6–10-week old Stat1^−/− and wild type littermate mice (Taconic Farms, Denmark); or from femurs of 6–10-week-old macrophage/neutrophil-specific STAT3^−/− (here called STAT3^−/−) and wild type littermate mice (named STAT3^+/+). These mice were obtained by inter-crossing of FLOXSTAT3^−/− mice from S. Akira with the macrophage/neutrophil-specific LysMcre mice and wild type littermate mice from I. Förster at Borstel animal facilities (Research Center Borstel, University of Lubeck, Borstel, Germany). Bone marrow-derived cells were cultured in DMEM, 20% FCS, 1 mm glutamine, 1 mm nonessential amino acids, 25 ng/ml M-CSF, 50 μg/ml gentamicin, 30 μg/ml vancomycin (D20) in bacteriological dishes for 7-days to differentiate into MØ (BM-DM). Murine recombinant IFN-γ, TNF-α, IL-6, IL-10, or IL-12 cytokines were obtained from Sigma. Cells were treated 72 h with 10–20 ng/ml with the different murine cytokines before infection kinetics or phagosome isolation.

Bacteria

L. monocytogenes 10403S strain) was obtained from D. A. Portnoy (University of California, Berkeley), and GFP-Listeria monocytogenes variant of the L. monocytogenes strain DH-L1039 (GFP-L. monocytogenes) was kindly provided by D. E. Higgins (Harvard Medical School, Boston).

Kinetic Infection Assays

MØ-like cell lines or BM-DM were cultured into 96-well plates at 1 × 10⁶ cells/ml in the presence or absence of the above-mentioned cytokines 72 h before infection. Cells were infected with L. monocytogenes at a ratio of 10:1 (bacteria/cell) as reported previously for different times (0, 4, 8, or 16 h). CFU ratios were performed as reported and represented the ratio of CFU at 8 h to CFU at 0 h ± S.D. of triplicates (4). Comparative kinetic infection assays were performed in J-774 cells and BM-DM from CBA/J cells pretreated or not with TNF-α, IL-6, or IFN-γ as reported previously (4, 10, 15).

Measurements of H₂O₂ and Nitrite Production

J-774 cells (2 × 10⁶ cells/ml) were cultured in microtiter plates. Cells were pretreated or not with cytokines for 72 h and next infected for 1 h with 2 × 10⁷ CFU/ml of L. monocytogenes. H₂O₂ and NO₂ production was measured as described previously (10). In brief, the H₂O₂ production was measured by the HRP-dependent conversion of phenol red by H₂O₂ into a compound with increased absorbance at 600 nm using a H₂O₂ standard curve. Results are expressed as nanomoles of H₂O₂ produced per cell. The nitrite production was determined with the Griess reagent using a sodium nitrite standard curve. Results were expressed as nanomoles of NO per cell. Samples were performed in triplicate, and results are the mean ± of three independent experiments.

Phagosome Isolation

J-774 cells or BM-DM cells were cultured at 1 × 10⁸ cells in the presence or absence of cytokines 72 h before infection. Cells were infected with L. monocytogenes at a ratio of 10:1 (bacteria/cell) for 20 min. Phagosome isolation was performed as described previously (4) and as detailed in supplemental material.

Differential Microarrays

J-774 cells (1 × 10⁶ cell/well) were cultured in 6-well plates in the presence or absence of 10 ng/ml IFN-γ or 20 ng/ml IL-6 for 16 h. Cytokine-treated cells were infected with L. monocytogenes, 10403S wild type strain (10:1 bacteria: cell ratio), for 15 min at 37 °C, followed by a 45-min incubation in the presence of 20 μg/ml gentamicin followed by 4 h in D5 medium. Noninfected samples (NI) were cultured at same cell density and treated with the above cytokines. Basal level control corresponded with nontreated and noninfected samples (NT and NI samples). Later on, the total RNA was extracted from cells using the Qiagen kit (RNeasy total RNA isolation kit, Qiagen, product number 74104). The amount of total RNA extracted per sample varied between 25 and 30 μg. RNA integrity was analyzed in 1% agarose gels (28 S and 18 S ribosomal RNA forms were observed as nondegraded in a 2:1 proportion). RNA quality was estimated by a close to 2.0 value of the ratio A_{260/280 nm}, and concentration was calculated with the assumption that 1 OD corresponds with 40 μg of RNA measured at A_{260 nm}. Differential microarrays were performed with the Affymetrix GeneChip MOE430A2.0 that evaluated 22626 mouse genes with GCOS 1.3 Affymetrix® software (Progenika S. A., Spain). The fold changes of gene expression values are expressed as the signal log ratio that corresponds to the log₂ of fold change (FC) in a previous version of Affymetrix software. Therefore, signal log ratio values of ≥0.3 were induced genes as they corresponded to values ≥1.2 FC, and ≤ −0.3 were depressed genes as they corresponded with values ≤ −1.2 FC. All our final values were subtracted from the values of basal controls (NT and NI values). Other controls include L. monocytogenes infected versus NT and NI values and are shown in supplemental Table S1. Gene ontology information was derived from Progenika S.A. (Affymetrix NetAffx Analysis Center).

Recombinant Proteins and LLO Peptides

GST-PBD_PAK (kindly provided by G. Bockoch, UCLA), the p21-activated kinase-derived binding domain for activated RAC2 protein, and the His-EEA1₂₁₈ (a gift from D. Lambright. Texas University), the FYVE domain of the EEA1 protein containing the binding domain for RAB5A-GTP, were expressed in Escherichia coli BL-21 strain. Recombinant proteins were induced with 5 mm isopropyl β-d-thiogalactoside for 5 h at 37 °C and purified with glutathione-Sepharose or TALON resins, respectively, according to instructions provided by the manufacturer (Clontech). LLO(1–99) and LLO(189–201) were synthesized by F. Roncal (Centro Nacional de Biotecnologia, Consejo Superior de Investigaciones Cientificase, Madrid, Spain) with a purity higher than 95% after HPLC and mass spectrometry.

Westerns, Immunoprecipitations, and Overlay Assays

30 μg of isolated phagosomes were loaded per lane onto SDS-polyacrylamide gels. Gels were transferred onto nitrocellulose membranes. Primary antibodies were incubated overnight at 4 °C as follows: mouse anti-Stat1, mouse anti-inducible NOS (Pharmingen), rabbit anti-STAT1-YPhos, rabbit anti-STAT3-SPhos, rabbit anti-STAT3-YPhos (Vitro), rabbit polyclonal anti-LLO specific antibody (Diatheva), mouse anti-PLY5 monoclonal antibody that recognize LLO undecapeptide sequence (15), 4F11 (mouse monoclonal anti-Rab5a), rabbit anti-ASMase (a gift from O. Utermöhlen, Kölhn University, Kölhn, Germany), and rabbit anti-cathepsin-D and rabbit anti-Rac2 (kindly provided by G. Bockoch, UCLA). Thereafter, secondary antibodies were HRP-conjugated (The Jackson Laboratory) and developed by ECL (Amersham Biosciences). To detect RAC2-GTP forms, we used a previously reported overlay assay (4). In brief, phagosomes were lysed and immunoprecipitated with GST-PBD_PAK. Immunoprecipitates were run onto SDS-PAGE and transferred to NC membranes. Primary antibody (rabbit anti-Rac2) was incubated overnight at 4 °C, followed by incubation with HRP-conjugated secondary antibody. For His-EEA1(218) overlay assay to detect RAB5A-GTP, we immunoprecipitated Rab5a using 4F11 antibody, run onto an SDS-polyacrylamide gel, and transferred to NC membranes. Membranes were treated sequentially with re-naturalization buffer and binding buffer, followed by o/n incubation at 4 °C with 100 ng/ml of His-EEA1₂₁₈ diluted into binding buffer. Blots were developed with secondary rabbit anti-His₆ antibody followed by ECL. To detect LLO bound to MHC class II molecules, phagosomal lysates were immunoprecipitated with mouse anti-IA^k antibody (10.3.62) or anti-IA^b antibody (Y3P). Immunoprecipitates were run onto SDS-polyacrylamide gel and transferred to NC membranes. Primary antibody (rabbit anti-LLO) was incubated overnight at 4 °C, followed by incubation with HRP-conjugated secondary antibody. As internal control in all Western blots, we included the marker, RAB5C, and the expression level was not modified by the cytokine treatment (4). Western blots were developed by ECL.

Vaccination Protocols

Ctsd^+/+ mice were treated (Ctsd^low) or not (Ctsd^+/+) (n = 5) for 3 days with 0.5 mg/ml pepstatin-A previous to vaccination and during the vaccination protocol as described previously (15). L. monocytogenes phagosomes containing 500 CFU/∼30 μg of phagosomal proteins were obtained from BM-DM of Ctsd^+/+ or CBA/J mice, previously treated for 72 h with 10 ng/ml mouse recombinant IFN-γ (P-IFN), IL-6 (P-IL-6), or left untreated (P-NT). These phagosomes were inoculated in the peritoneal cavity (i.p.) of CTSD^+/+ or CTSD^low (n = 5) mice for 7 or 14 days or were nonvaccinated (NV). Next, all mice were inoculated with 10³ CFU of L. monocytogenes intraperitoneally for 3 additional days. The vaccination timing follows similar protocols previously reported for studies using L. monocytogenes as vaccine (16). Mice were bled before sacrifice and serum-stored at −80 °C to measure cytokines by FACS analysis. Spleens and livers were homogenized, and CFUs were counted in homogenates. Similar protocols were performed in CBA/J, SCID, or 129/Sv mice using P-NT, P-IFN, or P-IL-6 phagosomes obtained from BM-DM from the bone marrow cells of these mouse strains and vaccinations each 7 days (SCID and 129/Sv mice) or 14 days (CBA/J mice).

FACS Analysis of BM-DM, Peritoneal Cells, and Cytokine Measurements

Peritoneal exudate cells (PEC) were obtained after Hanks' wash of the peritoneal cavity and cell surface-labeled with antibodies against the following markers: CD11b (marker for MØs), CD11c (marker for DC), Ly6C (marker for DCi), Ly6G (also known as Gr-1, marker for PMNs), Dx5 (marker for NK cells), CD3 and CD8 (markers for cytotoxic T cells) (Tc), IA^k (MHC-II for CBA/J mice), IA^d (MHC-II for SCID and 129/Sv mice), or IA^b (MHC-II for CTSD^+/+, CTSD^−/−, or CTSD^low mice) and analyzed by FACS.

We also used FACS analysis for cell surface labeling of BM-DM (100.000 cells) treated or not for 72 h with 10 ng/ml IFN-γ as activating stimulus using monoclonal antibodies FITC or phycoerythrin-labeled (BD Biosciences) against the above-mentioned cell surface markers. The analysis of BM-DM oxidative burst using L. monocytogenes as stimulus (10 min, 37 °C) and using the Phagoburst® kit (Orphegen Pharma, Heidelberg, Germany) was also performed by FACS analysis. TNF-α and IL-6 production by BM-DM infected with wild type L. monocytogenes was also measured in culture supernatants by FACS analysis as well as in sera from different mice (CBA kit from BD Biosciences). BM-DM were incubated in microtiter plates at a density of 2 × 10⁶ cells/ml with medium alone or with 2 × 10⁷ CFU/ml of L. monocytogenes for 1 h without antibiotics, followed by 24 h of incubation in D5 complete medium. Cells were centrifuged, and half of the supernatants volume was harvested and stored at −80 °C until FACS analysis. Samples were performed in triplicate, and the results are the means ± S.D. of two separate experiments.

FACS Analysis of Spleens to Measure IFN-γ Intracellularly

We used FACS analysis for cell surface labeling of spleen cells (100,000 cells), using monoclonal antibodies FITC- or antigen-presenting cells labeled (BD Biosciences) against CD4 or CD8. Peptides used were LLO(1–99) or LLO(189–201). For in vitro culture, spleen cells were plated into 96-well round-bottom plates (5 × 10⁶ cells/ml) and stimulated with each of the LLO(91–99) or LLO(189–201) peptides independently (10⁻⁵ m each peptide) for 5 h in the presence of brefeldin A (intracellular cytokine staining) as described previously (16). Stimulated cells were surface-stained for CD4 and CD8 and then fixed and permeabilized using a cytofix/cytoperm kit (BD Biosciences). Cells were stained for IFN-γ with antibody anti-IFN-γ phycoerythrin-labeled. Samples were acquired using a FACSCanto flow cytometer (BD Biosciences). Data were gated to include exclusively CD4⁺ or CD8⁺ events, and the percentages of these cells expressing IFN-γ were determined according to the manufacturer's recommendations (BD Biosciences). Results of LLO peptide-stimulated splenocytes were corrected according to the percentages of total CD4⁺ and CD8⁺ cells, respectively. Data were analyzed using FlowJo software (Treestar, Ashland, OR).

Statistical Analysis

For statistical analysis, the Student's t test was applied.

Ethics Statement

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Spanish Ministry of Science, Research and Innovation. The Committee on the Ethics of Animal Experiments of the University of Cantabria approved this protocol (Permit Number 2009/12) that follows the Spanish legislation (RD 1201/2005). All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.

RESULTS

IFN-γ and IL-6 Trigger Similar Listericidal Mechanisms in MØ

J-774 MØ-like cells display basal levels of MØ-associated microbicidal abilities and can be transformed into highly bactericidal cells upon cytokine treatment (4, 11, 17, 18). Therefore, we evaluated the activation of these cells when exposed to pro-inflammatory cytokines that are involved in L. monocytogenes immune responses (i.e. TNF-α, IFN-γ, IL-6, IL-10, and IL-12). IFN-γ, TNF-α, and IL-6 induced killing of intracellular L. monocytogenes in J-774 cells as observed by the reduced replication ability of L. monocytogenes, although IL-10 and IL-12 had no effect (Fig. 1A). We also confirmed that these cytokines promoted L. monocytogenes killing comparing the infection kinetics of J-774 cells and real MØ such as BM-DM. These cytokines clearly enhanced the microbicidal abilities of BM-DM and also increased the listericidal abilities of J-774 cells (supplemental Fig. S1, A and B). Therefore, J-774 cells and BM-DM were comparative models to study the listericidal mechanisms induced by pro-inflammatory cytokines. Oxidative bactericidal mechanisms were examined by measuring hydrogen peroxide (H₂O₂) and nitric oxide (NO) production, both of which decrease intracellular L. monocytogenes viability. Only IFN-γ and IL-6 induced the production of both oxidative compounds in J-774 cells (Fig. 1B). We next determined the relevance of these cytokines at the phagosomal level by examining intraphagosomal L. monocytogenes viability as a measurement of L. monocytogenes degradation as reported previously (4, 13, 15). Only IFN-γ and IL-6 treatment of J-774 cells decreased bacterial viability, indicating their importance for L. monocytogenes degradation within the phagosomes (Fig. 1C, white bars). Lysosomal hydrolases are responsible for the nonoxidative listericidal mechanisms that degrade L. monocytogenes within the phagosomal compartment (13–15). Chloroquine is a general inhibitor of lysosomal hydrolases that increase the phagosomal pH and block the transport of lysosomal proteases to L. monocytogenes phagosomes (19). Treatment of IFN-γ or IL-6-treated J-774 cells with chloroquine increased L. monocytogenes intraphagosomal viability to control levels, demonstrating that cytokine-induced L. monocytogenes degradation within the phagosomes required the activity of lysosomal hydrolases (Fig. 1C, black bars). Similar results were obtained using a cathepsin-D inhibitor, pepstatin A (data in Fig. 1C, legend), further supporting the participation of CTSD in L. monocytogenes degradation (15). These results suggest that IFN-γ and IL-6 induce similar oxidative and nonoxidative listericidal mechanisms in J-774 cells and that TNF-α does not mediate phagosomal L. monocytogenes degradation controlled by lysosomal proteases (Fig. 1C).

FIGURE 1.

IFN-γ or IL-6 triggers similar listericidal mechanisms in MØs. J-774 cells (1 × 10⁶ cells/ml) were treated 72 h with 10 ng/ml recombinant cytokines (TNF-α, IL-6, IL-10, IL-12, or IFN-γ) before infection with L. monocytogenes (bacteria/cell ratio of 10:1). A, cells were infected with L. monocytogenes for 8 h, and results are expressed as the ratio of CFU at 8 h to CFU at 0 h ± S.D. of triplicates. Kinetic assays are shown in supplemental Fig. S1, A and B. B, NO₂ production (left plot) of J-774 cells infected with L. monocytogenes for 24 h and measured in cell supernatants using the Griess reagent and described under “Experimental Procedures.” Results were expressed as nanomoles of NO₂ per cell ± S.D. of triplicates. H₂O₂ production (right plot) was measured adding L. monocytogenes as a stimulus for 2 h as described under “Experimental Procedures,” and results are expressed as nanomoles of H₂O₂ per cell ± S.D. of triplicates (p < 0.05). C, L. monocytogenes phagosomes isolated from J-774 cells (1 × 10⁸ cells) pretreated or not with 1 mm chloroquine (CQ) or 100 mm pepstatin A (pepA). CFU of pepstatin A-treated cells corresponded to 3.6 × 10⁴ ± 10³ for control (Cont), 0.9 × 10⁴ ± 35 for IFN-γ, 0.8 × 10⁴ for IL-6, 3.5 × 10⁴ ± 120 for IL-10, and 3.6 × 10⁴ ± 115 for TNF-α (p < 0.01).

Listeria-specific Transcriptional Response Is Induced by IFN-γ or IL-6

To identify and select genes induced synergically by IFN-γ or IL-6, we performed a detailed transcriptional analysis. Therefore, we analyzed the differential expression of genes included on the Affymetrix GeneChip MOE430A2.0 (∼14,000 mouse genes) in J-774 cells using three strategies (Fig. 2A). These three strategies were performed to differentiate the genes specifically induced or repressed by IL-6 or IFN-γ in activated J-774 cells in response to L. monocytogenes infection (strategy 1, IFN + LM or IL-6 + LM in Fig. 2A) from those genes induced or repressed by L. monocytogenes infection itself (strategy 2 in Fig. 2A). Strategy 3 was developed based on the observation that the IFN-γ-associated transcriptional signal was greater in J-774 cells than IL-6 associated signal (Fig. 2B, DE numbers). The data from five independent experiments examining differential gene expression (Fig. 2A) were analyzed using two different approaches. In the first approach detailed in the supplemental material, we analyzed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional term enrichment in the sets of differentially expressed genes without any aprioristic assumptions (supplemental Tables S1 and S2). In the second approach, we performed a hypothesis-driven analysis focused on genes specific for phagosomal destruction (cluster I) and those involved in L. monocytogenes immunity (cluster II) to determine the response localized within the phagosomes because they may function as destructive immune vesicles (20). Cluster I was focused on genes reported to encode phagosomal trafficking regulators and lysosomal proteins involved in phagocytic particles destruction (4, 12, 19, 21, 22). Cluster II was focused on genes reported to encode proteins important in the anti-L. monocytogenes innate immune response, mainly type I and II IFN-regulated genes (23). The housekeeping genes used as a reference for basal transcriptional expression were cytosolic β-actin and the pyruvate carboxylase. The results of all strategies analyzed indicated that, in the intracellular trafficking subfamily of the cluster I genes, the gene encoding RAB5A was induced, and IFN-γ induction of RAB5A was slightly higher than IL-6-induced expression (Fig. 2C, 1.55-fold change (FC) versus 0.98 FC, respectively). LAMP-2, the regulatory A subunit of the H⁺-ATPase (Atp6), and LIMP-2 were the lysosomal component-associated genes also induced by both cytokines in most strategies (Fig. 2C and Table 1). Genes belonging to the type I IFN pathway in the cluster II gene set, such as the secreted factors C3F6 and IL-6, were induced in all strategies. However, genes belonging to the type II IFN pathway in the cluster II gene set that are involved in L. monocytogenes-specific innate immunity such as the JAK-STAT cytosolic component, STAT-1 (5, 24, 25), appeared to be L. monocytogenes-specific; they were only induced by treatment with either cytokine plus L. monocytogenes. Because IFN-γ induced a stronger signal than IL-6 in J-774 cells (Fig. 2B), several genes from both clusters were only induced by IFN-γ. Examples of IFN-γ-induced genes are detailed in supplemental Table S2. We also observed no induction of the classical STAT-1-mediated genes after L. monocytogenes infection such as those encoding the MHC II IAβ or IAα chains or the transactivator CIITA. These results were not surprising given that strong signals such as IFN-γ or IL-6 would induce fewer additional genes with an L. monocytogenes infection. Moreover, these findings suggest that IFN-γ- or IL-6 induced genes share a single signaling pathway.

FIGURE 2.

Differential expression microarrays and functional clustering to identify genes common to IFN-γ or IL-6 signaling in MØ. J-774 cells nontreated and noninfected (NT and NI), treated with IFN-γ (IFN-γ and NI), or IL-6 (IL-6 and NI) (10 ng/ml) but noninfected or treated with IFN-γ (IFN-γ + LM) or IL-6 (IL-6 + LM) and L. monocytogenes infected were used for RNA isolation and microarrays. A, description of the differential expression strategies performed by microarrays (strategies 1–3). Next, functional clustering was applied to all genes induced ≥0.3 signal log ratio (corresponding to 0.5 FC of previous Affymetrix software version). Two gene clusters were analyzed as follows: genes involved in phagosomal destruction (cluster I) and genes involved in L. monocytogenes-specific immunity (cluster II). A total of 34 selected genes were analyzed further. B, representation of all genes from strategy 1 corresponding to differential expression (DE), either induced or decreased. Of note, the total differential expression numbers were significantly higher in IFN + L. monocytogenes compared with NT and NI samples than in IL-6 + L. monocytogenes compared with NT and NI samples. C, heat map presentation of selected genes (cluster I, cluster II, and housekeeping genes) and their expression profile in IFN-γ- or IL-6-activated J-774 cells infected with L. monocytogenes corresponding to all strategies. Rows colorimetrically represent expression ratios as described under “Experimental Procedures” from ≤ −4 FC repressed genes shown in green to ≥4 FC-induced genes shown in red. Black boxes correspond to no differential expression change genes.

TABLE 1.

Specific LLO immune response in spleens of vaccinated wild type mice

^a CBA/J mice were vaccinated i.p. or not (NV) with L. monocytogenes phagosomes (500 CFU/∼30 μg) (P-NT, P-IFN, or P-IL-6) or L. monocytogenes-infected BM-DM (10⁶ cells) (BM-DM/L. monocytogenes, BM-DM-IFN/L. monocytogenes, or BM-IL-6/L. monocytogenes) for 14 days (n = 5/vaccination type) and challenged i.p. with 10³ CFU/mice of L. monocytogenes for 3 days.

^b Splenocytes from vaccinated mice were LLO peptide-stimulated, and percentages of total CD4⁺ or CD8⁺ T cells are expressed as means ± S.D. (p < 0.05).

^c CD4 or CD8 cells and LLO peptide-stimulated were stained for intracellular IFN-γ as under “Experimental Procedures” and “Results” corrected from total CD4⁺ or CD8⁺ and expressed as means of CD4⁺/LLO(189–201) or CD8⁺/LLO(91–99) cells ± S.D. (p < 0.05).

^d PEC from vaccinated mice were stained for CD11b (MØ) or CD11c (DC). CD11c⁺ cells were double-stained with Ly6C or anti-IA^k antibodies to distinguish mature DC (DC_m) (CD11c⁺IA^k+) or immature DC (DC_i) (CD11c⁺Ly6C⁺), and results are expressed as the mean ± S.D. of triplicates (p < 0.05).

^e Cytokines measured in sera from vaccinated mice were expressed as picograms/ml TNF-α or IL-6. Pro-inflammatory cytokines show no variation among vaccines as follows: MCP-1, 1525 ± 20; IL-12, 52 ± 5; IL-10, 15 ± 2, and IFN-γ, 795 ± 11 (p < 0.01).

Phagosomal Components of the IFN-γ or IL-6 Listeriocidal Route

We confirmed the transcriptional results using a basic proteomic and functional analysis of isolated L. monocytogenes phagosomes from J-774 cells treated or not with IFN-γ or IL-6. Most of the induced genes, such as Rab5a and Rab5a-GTP forms, LAMP-2 and LIMP-2, belonged to cluster I and were also observed at high protein levels in L. monocytogenes phagosomes induced by IFN-γ or IL-6 (Fig. 3A). However, we also observed high RAC2-GTP and CTSD protein levels, although the corresponding genes were not induced or even repressed according to the differential expression analysis. Similarly, we observed high protein levels of p67^Phox and inducible NOS in IFN-γ- or IL-6-induced phagosomes (Fig. 3B), both of which belonged to the phagosomally related type II IFN subfamily of genes, despite a lack of detectable induction of gene expression (Fig. 2C). Therefore, not all proteins involved in microbe inactivation need to be induced at the gene level to demonstrate increased protein levels in IFN-γ- or IL-6-induced L. monocytogenes phagosomes. The most surprising findings corresponded to the expression of the JAK-STAT-associated cytosolic specific genes belonging to the type II IFN subfamily, which are important for the anti-L. monocytogenes innate immune response. We observed high protein levels of JAK1 and JAK2 and high total protein levels of STAT-1 (Stat1 lane in Fig. 3B) and active forms (P-Stat lane in Fig. 3B) in both IFN-γ- and IL-6-induced L. monocytogenes phagosomes. This is the first evidence demonstrating that the cytosolic components of the JAK-STAT signaling route associated with IFN-γ and IL-6 receptors can be found in phagosomes. The STAT-1-signaling route is connected to the induction of MHC II expression and transport to antigen-loading compartments (MIIC) (23). MIIC vesicles contain MHC II molecules loaded with peptides (26), and SDS-stable αβ MHC II dimers provide a valid measurement of peptide-loaded MHC II molecules (27–29). We found that the phagosomes (P-IFN, P-IL-6, and P-NT lanes in Fig. 3C) demonstrated significant levels of SDS-stable αβ MHC II dimers, especially the IFN-γ-induced L. monocytogenes phagosomes. However, endosomes only expressed low levels of SDS-stable αβ MHC II dimers (Endo lane in Fig. 3C). These results suggest that phagosomes are MIIC-competent, and endosomes are not. However, higher or lower levels of SDS-stable αβ MHC II dimers are not necessarily a measurement of the antigen processing ability. Therefore, using LLO, the most common L. monocytogenes antigen (2, 3), we evaluated the specific CTSD-mediated phagosomal degradation of LLO, which, as we have reported previously, occurs between residues ⁴⁹¹WW⁴⁹² (15). This experimental method uses two anti-LLO antibodies to distinguish phagosomes with adequate LLO processing function from phagosomes with low processing capability by using the levels of intact LLO and CTSD-degraded LLO as a read-out. IFN-γ and IL-6-induced L. monocytogenes phagosomes are equally good LLO-processing compartments; they each contained high levels of CTSD-degraded LLO with no significant amount of intact LLO (IFN or IL-6 lanes, Fig. 3D). On the contrary, nonlistericidal phagosomes were poor LLO-processing compartments; they contained high levels of intact LLO and low levels of CTSD-degraded LLO(1–491) forms (lanes labeled as NT in Fig. 3D). We next examined which LLO form co-precipitated with MHC class II molecules by performing an immunoprecipitation with a monoclonal anti-MHC class II antibody, followed by Western blot with specific rabbit anti-LLO antibody. Only the LLO(1–491) form that corresponds with the CTSD-degraded LLO form co-precipitated with MHC class II molecules. LLO intact forms did not co-precipitate (Fig. 3D, lower lanes in IP: anti-MHC and WB, Rb anti-LLO gel). We verified with this approach that P-IFN or P-IL-6 phagosomes were better LLO-processing compartments than P-NT as they presented higher levels of LLO(1–491) forms that co-precipitated with MHC class II molecules. These results suggest that CTSD-mediated phagosomal LLO processing was promoted in listericidal IFN-γ or IL-6 phagosomes facilitating the LLO-specific loading of MHC-II dimers. As a control of all the proteomic analyses performed with L. monocytogenes phagosomes, we used the RAB5C marker because we previously reported it was not modified by cytokine treatment (4). L. monocytogenes phagosomes obtained from BM-DM pretreated or not with IFN-γ or IL-6 showed similar proteomic patterns as L. monocytogenes phagosomes obtained from J-774 cells (supplemental Fig. S2), confirming that these cells were comparative models for L. monocytogenes infection.

FIGURE 3.

Phagosomal components of the IFN-γ or IL-6 listeriocidal route. Phagosomes were isolated from J-774 cells pretreated or not (NT lanes) with IFN-γ or IL-6 as described under “Experimental Procedures.” A and B, Western blots of different proteins using Rab5c levels as controls. Rab5a-GTP was detected after Rab5a immunoprecipitation and EEA1₂₁₈ overlay. Rac2-GTP was detected after immunoprecipitation with GST-PBD_PAK. C, SDS-stable αβ dimers were detected using a rabbit anti-IA α chain antibody (26). D, LLO was detected as reported previously (15, 29) using a rabbit polyclonal anti-LLO antibody that recognizes all LLO forms (Rb anti-LLO lanes) or a mouse anti-PLY5 antibody recognizing only intact LLO (Mo anti-PLY5 lanes). Binding of LLO to MHC class II was performed as described under “Experimental Procedures” (IP: MoαMHC-II, WB: RbαLLO lane). E, CBA/J mice were vaccinated i.p. or not (NV bars) with different vaccines as under “Experimental Procedures” and as follows: BM-DM (1 × 10⁶ cells) pretreated or not with cytokines and infected with L. monocytogenes (BM-DM, BM-DM-IFN or BM-DM-IL-6 bars) or different phagosomes containing 500 CFU (P-NT, P-IFN, or P-IL-6 bars) were vaccinated for 14 days (n = 5/vaccine type) and challenged i.p. with 10³ CFU of L. monocytogenes/mice for 3 days. CFU of NV mice: BM-DM-NT, 5.2 × 10⁴ ± 130; BM-DM-IFN, 5.1 × 10⁴ ± 120; BM-DM-IL-6, 5.2 × 10⁴ ± 125 and saline, 5 × 10⁴ ± 130. Results of spleens homogenates are the mean ± S.D. of three different experiments (p < 0.01). F, SCID or 129/Sv mice were vaccinated i.p. or not (NV) with different L. monocytogenes phagosomes as in E for 7 days (n = 5/vaccination type), and results also expressed as in E (CFU shown in supplemental Fig. S6). Pie chart corresponds to percentages of DC cells ± S.D. in PEC of vaccinated mice (p < 0.01).

Listericidal Signal Induced by IFN-γ or IL-6 within Phagosomes Is Connected with Specific Immunity and Protection against Listeriosis

Next, we examined whether this listericidal pathway induced by IFN-γ or IL-6 within L. monocytogenes phagosomes was also linked to specific immunity. The priming of listericidal CD8⁺ or CD4⁺ T cells by antigen-presenting cells is a requirement to develop protective anti-L. monocytogenes immunity that fully depends on the main virulence factor LLO (1–3, 7). Therefore, we analyzed as an approach protection against murine listeriosis after i.p. inoculation of L. monocytogenes phagosomes (500 CFU/∼30 μg of phagosomal proteins) from untreated BM-DM (P-NT in Fig. 3E and Table 1) or treated with IFN-γ (P-IFN) or IL-6 (P-IL-6) for 14 days. Next, all mice were challenged intraperitoneally for 3 days with L. monocytogenes (10³ CFU/mice). We used a vaccination protocol similar in timing to those reported to generate effector immune responses using L. monocytogenes as vaccine carrier (16). However, previous to in vivo vaccination, we examined the safety of the vaccines by measuring their hemolytic activity onto sheep red blood cells as an approach. None of the vaccines tested show any cytotoxicity at all (supplemental Table S3). All vaccines were stored in liquid nitrogen until used and showed similar protein yields of ∼1 mg/ml and insignificant contaminations of 2–3% of RNA (supplemental Fig. S3). Protection was examined in livers (data not shown) and spleens of vaccinated mice as above by examining the CFU ± S.D. of triplicates (n = 5 mice/vaccination type) (Fig. 3E). We also included as controls mice inoculated intraperitoneally for 14 days with noninfected BM-DM (10⁶ cells/mice) pretreated or not with IFN-γ or IL-6 or controls inoculated with saline (NV bars) and next, all mice intraperitoneally for 3 days with 10³ CFU of L. monocytogenes. All controls showed similar results (data in legend of Fig. 3E). In further vaccination experiments, we only used NV as controls. The approach to measure the specific LLO immunity of spleens from vaccinated mice examines the IFN-γ intracellularly of CD4⁺ or CD8⁺ T cells specific for the immunodominant LLO(91–99) or LLO(189–201) peptides and corrected for total CD4⁺ or CD8⁺, respectively (Table 1). The procedure is detailed in supplemental Fig. S4). Phagosomes from J-774 cells cannot be used for these studies because there is no available and genetically compatible murine model for in vivo experiments with these cells. The protection achieved after vaccination with IFN-γ or IL-6 L. monocytogenes phagosomes was in the same range of 85–95% than the protection obtained with L. monocytogenes-infected BM-DM pretreated or not with IFN-γ or IL-6 (data in legend of Fig. 3E). The protection achieved with L. monocytogenes phagosomes from nontreated BM-DM (P-NT) was lower in a range of 56%. These results indicated that P-IFN and P-IL-6 phagosomes were as good vaccine vectors as L. monocytogenes-infected BM-DM. Moreover, all vaccines showed good stability and integrity because lysis of phagosomal membranes previous to vaccination abolished the protection achieved with each vaccine type (supplemental Table S4). When we examined LLO-specific immunity elicited by these vaccination types, we observed higher percentages of spleen LLO-specific and IFN-γ producers LLO(189–201)/CD4⁺ and LLO(91–99)/CD8⁺ T cells after vaccination with P-IFN and P-IL-6 phagosomes than with P-NT phagosomes (Table 1, rows b and c). These values were even higher than the percentages of LLO-specific and IFN-γ producers LLO(189–201)/CD4⁺ and LLO(91–99)/CD8⁺ observed with BM-DM treated or not with IFN-γ or IL-6 and infected with L. monocytogenes (Table 1, BM-DM/LM, BM-DM-IFN/LM, or BM-DM-IL-6/LM rows). Therefore, the protection obtained with P-IFN or P-IL-6 against L. monocytogenes, correlates with LLO-specific immunity.

Analysis of Efficiency of the Vaccines Using Listericidal IFN-γ or IL-6 Phago-receptosomes

Recruitment of DC to the vaccination sites is related to vaccine efficiency (7, 30). Therefore, we explore the phenotypes of the PEC recovered after vaccination because we inoculated our vaccines intraperitoneally. We observed that P-NT recruited low amounts of mature and immature DC (DC_m and DC_i) in a range of 35% (Table 1, row d). In supplemental Fig. S5, we detailed the protocol followed to stain the double-positive CD11c⁺IAk⁺ (DC_m) or CD11c⁺Ly6C⁺ (DC_i) cells of PEC after vaccination with P-IFN. P-IFN and P-IL-6 recruited higher amounts of DC_i in a range of 46–47% but interestingly very high amounts of DC_m in a range of 79–83%. However, all vaccine types recruited MØ, NK, or PMN to the vaccination sites at a low range of 1–10%. These results indicate that our vaccination protocol elicits a secondary effector immune response and not a primary response as L. monocytogenes alone induces, recruiting MØ, NK, or PMN at high ranges of 10–75% (NV data in row d of Table 1).

Vaccination efficiency is also related to the production of several pro-inflammatory cytokines such as TNF-α (1, 3, 7, 25, 31) and down-regulation of IL-6 production (25). Therefore, we analyze the cytokine pattern after the vaccination of mice with the above-mentioned vaccine types. Although all vaccination types induced significant amounts of TNF-α, P-IFN or P-IL-6 showed low levels of IL-6 after 7 days of vaccination and 3 days with a second challenge of L. monocytogenes (10³ CFU/mice) (row e of Table 1). Therefore, the high protection obtained after P-IFN or P-IL-6 vaccination correlated to high peritoneal recruitment of matured dendritic cells (DC_m).

Vaccine efficiency has also been related to the ability of DC to expand the immune response stimulating different T cells (30). To explore the role of T cells in the efficiency of L. monocytogenes phagosomal vaccines, we used a widely reported lymphocyte-deficient model, such as the SCID mice that show a normal innate immune response against L. monocytogenes (1). Nonvaccinated SCID and control 129/Sv mice showed similar numbers of bacteria in spleens (Fig. 3F, NV bars). However, SCID- and 129/Sv-vaccinated mice with L. monocytogenes phagosomal vaccines, P-NT, P-IFN, or P-IL-6 for 7 days, showed significant differences in protection (supplemental Fig. S6 and P-IFN and SCID bars in Fig. 3F). In fact, 129/Sv control mice showed 200% protection after vaccination with effective vaccines (Fig. 3F and supplemental Fig. S6) and SCID mice no protection but a 100-fold increase in L. monocytogenes numbers compared with NV mice (Fig. 3F). Moreover, the lack of protection of P-IFN vaccines in SCID mice correlated to low recruitment of mature and immature DC to the vaccination sites in ranges of 20% compared with 66% in 129/Sv control mice (Pie Chart in Fig. 3F). We conclude that T cells controlled the recruitment of DC cells to vaccination sites and seemed to improve the vaccine efficiency.

STAT-1 Participates Exclusively Upstream in the IFN-γ or IL-6 Listericidal Route

Next, we tempted to decipher the sequential action of the components of this listericidal IFN-γ or IL-6 signaling route. We envision upstream and downstream participants of this pathway. STAT-1 or STAT-3 could be upstream mediators of a putative IFN-γ or IL-6 listericidal route. CTSD may be a downstream effector given that it participates directly in phagosomal L. monocytogenes degradation (13–15), and RAB5A could also participate because the CTSD-mediated degradation of L. monocytogenes is linked to this GTPase (15). Therefore, we examined L. monocytogenes phagosomes from STAT-1^−/− and ^floxSTAT-3^−/− (STAT-3^−/−) BM-DM treated with IFN-γ, IL-6, or left untreated. We observed a 3-fold increase in L. monocytogenes intraphagosomal viability in STAT-1^−/− L. monocytogenes phagosomes (black bars in Fig. 4A) and replication indices of BM-DM (legend of Fig. 4A) compared with STAT-3^−/− or wild type controls. However, L. monocytogenes intraphagosomal viability and the replication index in the IFN-γ or IL-6-treated STAT-1^−/− BM-DM was similar to untreated BM-DM, indicating that no listericidal activity was induced by either cytokine. In contrast, L. monocytogenes intraphagosomal viability and the replication index in IFN-γ- or IL-6-treated STAT-3^−/− BM-DM was 5-fold reduced and was similar to control STAT-1^+/+ or STAT-3^+/+ cells. The lack of an IFN-γ effect was expected for the STAT-1^−/− BM-DM because the microbicidal effects of IFN-γ are STAT-1-mediated (23). However, the lack of an IL-6 effect on L. monocytogenes intraphagosomal viability in the STAT-1^−/− BM-DM was unexpected and indicated that STAT-1, and not STAT-3, was responsible for the listericidal effect induced by IFN-γ or IL-6. When we analyzed the levels of RAB5A and CTSD in STAT-1^−/− and STAT-3^−/− L. monocytogenes phagosomes, we detected very low levels of RAB5A and CTSD in STAT-1^−/− and levels similar to controls in STAT-3^−/− L. monocytogenes phagosomes (Fig. 4B). STAT-1^−/− also shows low MIIC competence because L. monocytogenes phagosomes do not co-precipitate CTSD-processed LLO(1–491) forms with MHC-class II molecules (IP: MoαMHCII/WB and RbαLLO lanes in Fig. 4B). RAB5C levels are also used as controls. These results suggest that the competence of RAB5A, CTSD, and MIIC is linked and acted on downstream of STAT-1.

FIGURE 4.

STAT-1 participates exclusively upstream in the IFN-γ or IL-6 listericidal route. BM-DM (1 × 10⁸ cells) from STAT-1^−/−, STAT-1^+/+, STAT3^−/−, or STAT3^+/+ mice were treated or not with IFN-γ or IL-6 and infected with L. monocytogenes. A, phagosomes were solubilized before CFU analysis. Results were expressed as CFU values× 10 ± S.D. of triplicates as under “Experimental Procedures.” Replication indices were expressed as CFU ± S.D. as follows: 10 × 10⁴ ± 120 for STAT-1^+/+, 40 × 10⁴ ± 140 for STAT-1^−/−, 11 × 10⁴ ± 112 for STAT3^+/+, and 10 × 10⁴ ± 102 for STAT3^−/− in nontreated cells; 3 × 10² ± 12 for STAT-1^+/+, 43 × 10⁴ ± 10³ for STAT-1^−/−, 4 × 10² ± 9 for STAT3^+/+, and 3.5 × 10² ± 13 for STAT3^−/− in IFN-γ-treated cells; 4.1 × 10² ± 11 for STAT-1^+/+, 45 × 10⁴ ± 112 for STAT-1^−/−, 3.2 × 10² ± 13 for STAT3^+/+, and 3.6 × 10² ± 11 for STAT3^−/− in IL-6-treated cells (p < 0.05). B, detection of RAB5A and Ctsd in L. monocytogenes phagosomes from to STAT-1^−/−, STAT-1^+/+, STAT3^−/−, or STAT3^+/+ BM-DM. LLO binding to MHC-class II was detected as in Fig. 3D (IP: MoαMHC-II, WB: RbαLLO lane). Rab5c levels were used as controls.

CTSD Participates Downstream in the IFN-γ or IL-6 Listericidal Route

Next, we examined if CTSD was the downstream component of the IFN-γ or IL-6 listericidal route using CTSD^−/− BM-DM. L. monocytogenes phagosomes in CTSD^−/− BM-DM displayed a 4-fold increase in L. monocytogenes intraphagosomal viability compared with Ctsd^+/+ phagosomes (Fig. 5A, black bars). IFN-γ or IL-6 failed to induce any listericidal activity in the CTSD^−/− phagosomes in contrast to CTSD^+/+ BM-DM (Fig. 5A, gray and white bars). IFN-γ or IL-6 increased Rab5a levels in L. monocytogenes phagosomes; however, L. monocytogenes phagosomes from CTSD^−/− BM-DM expressed only very low levels of RAB5A, and this was not modified by IFN-γ or IL-6 treatment (Western blot in Fig. 5A). This confirms that CTSD and RAB5A are linked in this cytokine pathway. We verified the defect in cytokine signaling by visualizing GFP-L. monocytogenes infection using conventional fluorescence (black and white images in Fig. 5B). The fluorescence signal observed in nontreated CTSD^+/+ BM-DM that corresponded to viable intracellular L. monocytogenes was clearly diminished in IFN-γ- and IL-6-treated BM-DM. However, the higher fluorescence signal observed in nontreated CTSD^−/− BM-DM was barely unmodified in IFN-γ- or IL-6-treated BM-DM. The intracellular replication indices after 8 h of L. monocytogenes infection indicated similar L. monocytogenes replication in the BM-DM (Fig. 5B, legend). We also evaluated other functions of activated MØ that might have also been affected by a defect in cytokine signaling, such as the oxidative burst, using flow cytometry or the transformation of phagosomes into antigen loading competent compartments (MIIC) using confocal fluorescence images (29). Ctsd^−/− BM-DM displayed a defect in the oxidative burst (Table 2, row b) and in the transformation of L. monocytogenes phagosomes into MIIC (Fig. 5B). MØ markers not involved in activation such as CD11b or CD11c were similar in CTSD^−/− and CTSD^+/+ BM-DM (Table 2, row c). CTSD^−/− BM-DM presented decreased co-localization of GFP-L. monocytogenes with anti-MHC II antibodies after IFN-γ or IL-6 treatment, although GFP-L. monocytogenes intracellular numbers were much higher (color images in Fig. 5B). These results suggested that CTSD^−/− L. monocytogenes phagosomes might display impaired antigen processing abilities. We used the CTSD-mediated LLO phagosomal processing approach to confirm a defect of CTSD^−/− BM-DM in antigen processing. CTSD^−/− L. monocytogenes phagosomes contained only high levels of intact LLO that were detected with the anti-PLY-5 antibody in all conditions as follows: nontreated, IFN-γ-, or IL-6-treated BM-DM (Fig. 5C). We also confirmed that CTSD^−/− L. monocytogenes phagosomes show no LLO form able to co-precipitate with MHC class II molecules (IP: Mo α MHC-II/WB, Rb α LLO lanes in Fig. 5C). RAB5C was also used as loading control. These results verified that CTSD^−/− L. monocytogenes phagosomes showed a clear impairment in LLO phagosomal processing and a severe defect in the phagosomal listericidal mechanisms connected to STAT-1. Next, we confirmed that CTSD was the downstream component of the STAT-1 listericidal pathway linked to L. monocytogenes-specific immunity. Following the same reasoning as in Fig. 3, we examined protection against listeriosis after i.p. inoculation of L. monocytogenes phagosomes (500 CFU/∼30 μg of phagosomal proteins) (P-NT, P-IFN, or P-IL-6 from BM-DM of CTSD^+/+ mice) into CTSD^low or CTSD^+/+ mice. We used an inducible mouse model of CTSD deficiency previously used and described for in vivo experiments (CTSD^low) (15) because CTSD^−/−-deficient mice do not live long enough to allow vaccination protocols. We observed that 7 days of vaccination to CTSD^+/+ mice with P-NT phagosomes conferred a protection in a range of 20%. Vaccination with P-IFN or P-IL-6 phagosomes conferred protection in a higher range of 90–95% compared with nonvaccinated mice (Fig. 5D, black bars). The protection achieved with IFN-γ or IL-6 L. monocytogenes phagosomes was in the same range of 95% of protection obtained after vaccination for 14 days with L. monocytogenes-infected BM-DM pretreated with IFN-γ or IL-6, whereas no protection was obtained after vaccination for 7 days (data in legend of Fig. 5D). Protection requires L. monocytogenes phagosomal integrity because lysis of phagosomes before mice inoculation causes no protection (Fig. 5D, gray bars). Vaccination of CTSD^low mice with P-NT, P-IFN-γ, or P-IL-6 L. monocytogenes phagosomes confers no protection with exacerbated numbers of L. monocytogenes (Fig. 5E), high levels of IL-6 (Table 3, row a), no LLO-specific immune response in spleens (Table 3, row b), and an unusual peritoneal recruitment of PMN but a lack of recruitment of DC_m (Table 3, row c). These results confirmed that CTSD was the downstream component of this STAT-1 listericidal pathway connected with L. monocytogenes-specific immunity and protection.

FIGURE 5.

CTSD participates downstream in the IFN-γ or IL-6 listericidal route. A, L. monocytogenes phagosomes from CTSD^−/− or CTSD^+/+ BM-DM pretreated or not with IFN-γ or IL-6. Results are expressed as CFU values × 10 ± S.D. of triplicates and RAB5A detected as in Fig. 3A. B, untreated (NT), IFN-γ-, or IL-6-treated CTSD^−/− or CTSD^+/+ BM-DM were infected with GFP-L. monocytogenes for 4 h (black-white images corresponding to conventional fluorescence) or 1 h (color confocal images). MIIC compartments containing GFP-L. monocytogenes (yellow fluorescence) were labeled with biotinylated anti-IA^b antibody. Scale bars in black and white images: 6 μm (CTSD^+/+ NT, CTSD^−/− NT, CTSD^−/− IFN, and CTSD^−/− IL-6) and 10 μm (CTSD^+/+ IFN and CTSD^+/+ IL-6). CFU values at 8 h for CTSD^+/+ were 8 × 10⁴ (NT), 2.5 × 10² (IFN-γ), 2.3 × 10² (IL-6), and for CTSD^−/− were 45 × 10⁴ (NT), 40 × 10⁴ (IFN-γ) and 41 × 10⁴ (IL-6) (p < 0.05). Scale bars in color images: 10 μm. C, LLO in L. monocytogenes phagosomes from Ctsd^−/− or Ctsd^+/+ BM-DM detected as in Fig. 3D (RbαLLO lane shows all LLO forms, MoαPLY5 lane shows only LLO intact forms, and IP: MoαMHC-II, WB: RbαLLO lane shows LLO bound to MHC class II). Controls were RAB5C levels. D, vaccination of CTSD^+/+ mice i.p. (n = 5) or not (NV) with 500 CFU/∼30 μg of L. monocytogenes intact (CTSD^+/+) (black bars) or lysed phagosomes (CTSD^+/+-lysed) (gray bars) (P-NT, P-IFN, or P-IL-6) for 7 days and challenged with 10³ CFU of L. monocytogenes for 3 days. CFU in homogenized spleens are expressed as CFU × 100 ± S.D. CFU of L. monocytogenes infected BM-DM vaccines show no protection: 48 × 10³ ± 12 for nontreated, 45 × 10³ ± 9 for IFN-γ, and 49 × 10³ ± 11 for IL-6 similar to NV mice (58 × 10³ ± 6) (p < 0.01). E, CTSD^low mice (n = 5) vaccinated as in D and CFU quantified in homogenized spleens (p < 0.01).

TABLE 2.

Activation parameters of BM-DM from CTSD^−/− and CTSD^+/+ mice after L. monocytogenes infection

^a BM-DM were untreated (NT) and treated with IFN-γ or IL-6 for 72 h before L. monocytogenes infection for different times according to the experiment.

^b Oxidative burst capacity of BM-DM using the Phagoburst fluorescent reagent after 10 min of incubation with L. monocytogenes as the stimulus and according to the instructions of the supplier. Results correspond with the fluorescent intensity mean (MFI) of 100,000 cells using the CellQuest software program of data acquisition and analysis. Results are the mean ± S.D. of triplicate samples (p < 0.05).

^c BM-DM were infected with L. monocytogenes for 1 h and surface-labeled with phycoerythrin-conjugated specific monoclonal antibodies. Percentages reflect the amount of positive cells for a given marker or stimulus. Samples were performed in triplicate, and results express the mean ± S.D. (p < 0.01).

TABLE 3.

Vaccination parameters in CTSD^+/+ and CTSD^low mice

^a CTSD^+/+ and CTSD^low mice were vaccinated i.p. or not (NV) with different L. monocytogenes phagosomes (500 CFU/∼30 μg) (P-NT, P-IFN, or P-IL-6) for 7 or 14 days (n = 5/vaccination type) (results correspond to 14 days vaccinations) and challenged i.p. with 10³ CFU of L. monocytogenes/mice for 3 days. Cytokines were measured in sera by FACS analysis, and results are expressed as IL-6 concentrations of picograms/ml. Other pro-inflammatory cytokines were TNF-α, 1305 ± 17; MCP-1, 1520 ± 20; IL-12, 50 ± 5; IL-10, 15.6 ± 2, and IFN-γ, 695 ± 11 (p < 0.05). Results shown in Fig. 5, D and E (p < 0.01).

^b Splenocytes from vaccinated mice were stimulated with LLO(91–99) or LLO(189–201) peptides as under “Experimental Procedures.” Percentages of CD4⁺ or CD8⁺ T cells are expressed as the mean ± S.D. (p < 0.05). CD4⁺ or CD8⁺ cells and LLO peptide stimulated were stained for intracellular IFN-γ, and results are corrected from total CD4⁺ or CD8⁺, performed in triplicate, and expressed as the mean ± S.D. (p < 0.05).

^c PEC were surface-stained for CD11b (MØ) or CD11c (DC). CD11c⁺ cells were double-stained for Ly6C or anti-IA^b antibodies to distinguish for mature DC (DC_m) (CD11c⁺IA^k+) or immature DC (DC_i) (CD11c⁺Ly6C⁺) (45). Other phenotypes were as follows: MØ corresponds to CD11b⁺ cells, NK to Dx5⁺ cells, and neutrophils (PMN) to Ly6G⁺ cells (known as Gr-1). Results are expressed as the mean ± S.D. of triplicates (p < 0.05).

DISCUSSION

Cell-free membrane vesicles are a good alternative as safer vaccine vectors against infectious diseases. Current efforts are focused to implement their efficiency. In the case of listeriosis, the search for live vaccine delivery vectors is mainly focused to refinements to enhance their potency or reduce their cytotoxicity to get the approval for their use in humans (7, 16, 32–33).

In this study, we show that MØ are activated with IFN-γ or IL-6 to elicit a common listericidal signal mediated by STAT-1. This signal mediates the degradation of L. monocytogenes within compartmentalized phago-receptosomes induced by IFN-γ or IL-6. This functional compartmentalization serves to confine all innate immune elements required for L. monocytogenes destruction within the same environment and to mediate the onset of L. monocytogenes-specific immunity. Here, we present that the phago-receptosomes are new vaccine vectors inducing protection against murine listeriosis. In fact, they postulate as safe vaccines showing no cytotoxicity at all and higher efficiencies with shorter vaccination protocols than other cell-based vaccine carriers such as MØ infected with pathogens (34) or endosomes loaded with LLO (35). The induction of specific L. monocytogenes immunity is associated with their high competences as MIIC, binding of LLO-processed forms to MHC class II molecules, and down-regulation of IL-6. Protection appears linked to the recruitment of mature DC to the vaccination sites and induction of LLO-specific T cells in the spleens of vaccinated mice. What are the features of these listericidal compartments that make them unique as efficient vaccines against listeriosis (Fig. 6)?

FIGURE 6.

Features of the L. monocytogenes phago-receptosomes induced by IFN-γ or IL-6. IFN-γ or IL-6 binds to receptors and elicits a complete STAT-1 pathway, including phosphorylation of STAT-1 (P-STAT-1) bound to the receptor-associated kinases, JAK1 and JAK2 (feature 1). L. monocytogenes engulfed in phagosomes amplifies the STAT-1 signal and promotes Rab5a activation (RAB5A-GTP), triggering PHOX, inducible NOS, and CTSD-associated bactericidal mechanisms to assist L. monocytogenes killing and degradation, transforming the compartments in active innate immune vesicles with microbicidal abilities (feature 2). L. monocytogenes phago-receptosomes are transformed into competent MIIC immune vesicles by STAT-1-dependent mechanisms that stimulate CTSD-mediated antigen processing of LLO to its degraded LLO(1–491) form able to bind to MHC-II dimers. STAT-1-independent mechanisms as LIMP-2 also contribute transporting components of MIIC as LAMP-2 and αβ MHC II stable dimers (feature 3). Features 1–3 of L. monocytogenes phago-receptosomes reflect that they are active innate immune vesicles. Validation of these vesicles as vaccine vectors confirmed their participation in specific immunity (features 4–6). First, they confer protection against listeriosis and induce specific LLO(91–99)/CD8⁺ and LLO(189–201)/CD4⁺ effector T cells (feature 4). Second, their vaccine efficiency is related to stimulation of innate elements expanding the signals to T cells through the recruitment of DC_m to the vaccination sites and production of TNF-α with down-regulation of IL-6 (feature 5). Finally, protection requires intact innate and T cell-specific immune mechanisms in vaccine receptors (feature 6) because specific and protective immunity is achieved in CTSD^+/+ or 129/Sv control mice but CTSD-deficient (CTSD^low) or lymphocyte-deficient (SCID) mice fail to acquire specific immunity and protection against listeriosis.

The main characteristic of L. monocytogenes-specific immunity that makes this pathogen an attractive microbial vector is the ability to stimulate both arms of the immune system; it induces multiple innate immune pathways and an antigen-specific L. monocytogenes response that requires LLO to trigger a Th1 cytokine-based immune response (7, 32, 36). In this regard, IFN-γ or IL-6 initiates a complete listericidal pathway in MØ (Fig. 6, feature 1) via the activation of STAT-1 connected with the receptor-associated Janus tyrosine kinases, Jak1 and Jak2 (23). We obtained the first indication of this signaling pathway evaluating the global transcriptional response elicited in MØ by IFN-γ or IL-6 after infection with pathogenic L. monocytogenes. A general analysis of the enrichment of GO and KEGG terms indicated two profiles, cytokine and L. monocytogenes-specific patterns, and two main components, organelle regulation and L. monocytogenes immunity. The applied functional clustering based on L. monocytogenes phagosomal degradation (cluster I) and immunity (cluster II) restricted the common IFN-γ or IL-6 transcriptional responses to seven selected genes. Four genes corresponded to cluster I, and all of them played a role in phagocytosis, such as the trafficking regulator RAB5A (4, 19, 22, 37–39) and the lysosomal components LAMP-2 (21), LIMP-2 (4, 6, 29), and the regulatory subunit A of H⁺-ATPase (4, 12, 19, 21, 22). Three genes belong to cluster II and correspond to major components of L. monocytogenes-specific innate immunity, including the type I IFN-response genes C3F6 and IL-6 and the type II IFN response gene STAT-1 (1, 3, 5, 23). Therefore, it appears that this listericidal route connects signaling, trafficking regulators, listericidal lysosomal effectors, and cytokine production.

Next, L. monocytogenes internalized in phagosomes induces distinct phagosomal processes and amplifies the STAT-1-dependent listericidal signal to other innate immune elements (Fig. 6, feature 2). In this regard, this amplified signal promotes RAB5A translocation to phagosomes and activation of RAB5A to RAB5A-GTP (4) and the induction of the phagosomal oxidase (phox), H⁺-ATPase, inducible nitric oxidase, and the lysosomal protease, CTSD (4, 13, 19). Our basic proteomic analysis of purified phagosomes from IFN-γ or IL-6 activated MØ provided the second indication that phagosomes were signaling platforms for the onset on innate immunity. For the first time, we detected a complete JAK-STAT signaling pathway in L. monocytogenes phagosomes. The presence of high levels of JAK1, JAK2, and phosphorylated STAT-1 in these listericidal compartments appears to differentiate them from other reported phagosomal proteomes that contained latex beads but lacked these innate immune components (6). All together, these listericidal components, found in other phagosomal proteomes (6, 21, 22), contribute to the degradation of L. monocytogenes within the phagosomes. However, our functional validation of these platforms as powerful microbicidal vesicles revealed their singularity as innate immune vesicles.

Their transformation into MIIC-competent immune vesicles (Fig. 6, feature 3) is dependent on the STAT-1-mediated stimulation of CTSD-induced antigen processing of LLO, degrading it to the LLO(1–491) form (29). Moreover, MHC-II dimers become loaded with these LLO(1–491) forms (this study). However, other STAT-1 independent elements such as LIMP-2 (29) might also contribute transporting MIIC markers such as LAMP-2 and SDS-stable αβ MHC-dimers and inducing TNF-α. This transformation into MIIC competent vesicles might trigger the onset of specific immunity. Collectively, the results from experiments using the STAT-1^−/− and CTSD^−/− BM-DM revealed the sequence of elements involved in this listericidal route. STAT-1 acts upstream of RAB5A and CTSD in this pathway. STAT-1 is linked to the loading of MHC-class II molecules with CTSD-processed LLO(1–491) forms lacking the phagosomal binding domain (15).

The validation of these immune vesicles as effective vaccine vectors provided the fourth indication that L. monocytogenes phago-receptosomes induced by IFN-γ or IL-6 participated in specific immunity (Fig. 6, feature 4). In fact, they show high protection capacities against listeriosis and good abilities to induce specific LLO(91–99)/CD8⁺ and LLO(189–201)/CD4⁺ T cells (40).

Their vaccine efficiency appears related to their ability to trigger a secondary LLO-specific T cell immune response that recruits mature DC cells to the vaccination sites and triggers the production of Th1 cytokines (i.e. TNF-α, MCP-1, and IFN-γ) (Fig. 6, feature 5). In fact, SCID mice lacking T lymphocytes after vaccination with listericidal phagosomal vaccines only induce primary responses and recruit high levels of MØs instead of DC_m to the vaccination sites that causes a failure in protection against listeriosis.

Protection with L. monocytogenes phago-receptosomes also requires intact innate immune mechanisms in vaccine receptors (Fig. 6, feature 6). In fact, mice with severe deficiencies in innate immune responses (CTSD^−/−) (13, 15, 29) displayed a clear MØ defect with poor bactericidal abilities, oxidative burst capacities, CTSD-mediated LLO phagosomal processing, and impaired transformation of L. monocytogenes phagosomes into MIIC vesicles due to the absence of MHC-II dimers loaded with LLO(1–491). Moreover, vaccination of experimental CTSD-deficient mice (CTSD^low) previously described (15, 40) failed to confer protection against listeriosis because they do not recruit DC to the vaccination sites, showed uncontrolled neutrophilia and IL-6 exacerbated levels (25), and do not elicit LLO-specific T cells. Therefore, CTSD appears as the downstream component of the listericidal route that connects innate with specific immunity in listeriosis.

In brief, L. monocytogenes phago-receptosomes induced by IFN-γ or IL-6 present all the requirements to induce protective immunity, multiple innate immune elements, enhanced antigen presentation, stimulation of effector T cells, and recruitment of cells involved in the expansion of the signals (7, 30, 32, 34, 36, 41, 42).

We envisage that these L. monocytogenes phago-receptosomes induced by IFN-γ or IL-6 appear to function as naturally occurring vaccine carriers that might be liberated after pyroptosis of MØ at the inflammation sites (43, 44) and taken up by DC to confer protection (34, 45). Ongoing studies will evaluate their ability to induce long lasting T cell memory, the mechanisms of phago-receptosomes natural formation, and the refinements of doses and potency. Currently, they open up a field of possibilities for studies evaluating the relevance of natural vaccines that are similar to exosome release because of apoptosis (46). The lack of efficiency of this vaccination procedure in experimental deficiencies of CTSD or SCID mice lacking lymphocytes highlighted the convenience of testing bacterial vaccines in healthy donors as well as in immunodeficiencies with high susceptibility against certain bacterial groups (15, 40, 47).