| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Microbiology and Immunology, Göteborg University, Göteborg, Sweden
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
and IL-1
upon ex vivo culture, DC do not. In addition, although recruited monocytes internalize Salmonella in vitro and in vivo they did not induce the proliferation of OT-II CD4+ T cells after coincubation with Salmonella expressing OVA despite their ability to activate OT-II cells when pulsed with the OVA323–339 peptide. We also show that recruited monocytes enter the PP of infected mice independently of the mucosal addressin cell adhesion molecule-1 (MAdCAM-1). Finally, recruited but not resident monocytes increase in the blood of orally infected mice, and MHC-II up-regulation, but not TNF- or iNOS production, occur already in the blood. These studies are the first to describe the accumulation and function of monocyte subsets in the blood and GALT during oral Salmonella infection. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Monocytes circulating in the blood are composed of a heterogeneous population of cells that can be divided into distinct phenotypic and functional subsets in humans, rodents, and other species (3, 8, 14, 15). In murine blood, Gr-1lowCCR2lowCX3CR1high and Gr-1intCCR2highCX3CR1low (int,3intermediate) monocyte subpopulations have been described where the latter is more prone to migrate to inflammatory sites while the former migrates to tissues under steady-state conditions (3, 8, 14, 16). It is suggested that Gr-1intCCR2highCX3CR1low monocytes are more immature monocytes newly released from bone marrow that undergo a maturation process and become Gr-1lowCCR2lowCX3CR1high monocytes (8, 14). In addition to maturation, monocytes, under the influence of local stimuli, have the potential to differentiate into dendritic cells (DC), macrophages, and Langerhans cells (3, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25). Thus, recent in vivo studies of monocytes are beginning to elucidate the complex nature of this pivotal mononuclear phagocyte population in steady-state and inflammatory conditions. However, relatively little is known about the role of the recently identified monocyte subsets in bacterial infections, and no information is yet available on the function of these populations in the GALT after oral bacterial challenge.
Salmonella enterica is an intracellular bacterial species that contains several serovars of food and waterborne pathogens, such as the human pathogen serovar Typhi and its cousin serovar Typhimurium (Salmonella enterica serovar Typhimurium). Although S. enterica serovar Typhimurium causes a localized infection of the gastrointestinal tract in humans, it causes a systemic infection resembling typhoid fever in mice. After the oral infection of mice, S. enterica serovar Typhimurium (or Salmonella for brevity) penetrates the intestinal epithelium using M cell-dependent and independent pathways (26). The first organs targeted by orally acquired Salmonella are the Peyer’s patches (PP) and the mesenteric lymph nodes (MLN) (27, 28), the lymphoid organs that drain the intestine and initiate immune responses to gastrointestinal Ags. Although the PP are directly seeded by bacteria that cross the intestinal epithelium, bacteria arrive in the MLN via intestinal lymph, likely transported by DC (26, 29). The bacteria also spread to the spleen and liver via the blood (29). The control of Salmonella early during infection is dependent on IL-12, IFN-
, and TNF-, molecules that enhance the microbicidal capacity of phagocytes, and mice lacking any of these cytokines are more susceptible to infection (30). The production of reactive nitrogen intermediates via the inducible NO synthase (iNOS) is also important to controlling Salmonella infection (30).
Studies in mice made neutropenic with a depletion strategy using mAb RB6-8C5, which primarily recognizes Ly6G but also cross-reacts with Ly6C (31), have concluded that neutrophils are important for host survival to Salmonella (32, 33). Although this general conclusion is not debated, these studies may also have depleted monocytes, particularly inflammatory monocytes that express Ly6C and are recognized by RB6-8C5 mAb (3, 7, 8, 34). Similarly, studies addressing the role of macrophages during Salmonella infection obtained in studies using chemical depletion methods (35) should be considered with the caveat of monocyte depletion (8) and the possible effects on bystander phagocytes.
Although the importance of mononuclear phagocytes in host survival to Salmonella infection is not disputed, the related phenotypic and functional nature of neutrophils, monocytes, and tissue macrophages (5, 34, 36, 37, 38, 39), combined with the recent data revealing monocyte subsets with distinct roles (3, 7, 8), underscores the need for precise analysis of these populations during infection with this intracellular pathogen. This is particularly true for the organs that are the immediate targets of Salmonella after penetration of the intestinal epithelial barrier, the PP, and the MLN.
These issues are addressed here by performing direct ex vivo five- to seven-color flow cytometry analysis of cells accumulating in the PP and MLN during the first few days of oral Salmonella infection of mice. We characterize the influx of monocytes and neutrophils into the gut-draining lymphoid tissues and show that recruited monocytes are the major producers of iNOS and TNF-, most of which is made by uninfected bystander monocytes. We also examine the capacity of the different myeloid lineage populations recruited to infected PP and MLN to internalize and kill Salmonella in vitro and in vivo, produce TNF-, IL-1, and IL-12 upon ex vivo culture, and induce the proliferation of OT-II CD4+ T cells after coincubation with Salmonella expressing OVA or pulsed with OVA peptide. In addition, chemokine receptor expression on the recruited myeloid lineage populations is assessed and the role of mucosal addressin cell adhesion molecule-1 (MAdCAM-1) in recruiting monocytes to infected PP is analyzed. Finally, infection-induced recruitment and activation of monocyte subsets in the blood of orally infected mice is examined. These studies are the first to characterize recruited monocytes in PP and MLN after oral Salmonella infection and assess the number and function of their counterparts in the blood of orally infected mice.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
C57BL/6 mice were purchased from Charles River Laboratories. IFN-–/–, IL-12p40–/–, RAG–/– mice and OT-II mice, all backcrossed more than nine generations on the C57BL/6 background, were bred at the Experimental Biomedicine Animal Facility of Göteborg University, Göteborg, Sweden. Mice were used at 8 to 12 wk of age and provided with food and water ad libitum. All experiments were performed using protocols approved by the Swedish government’s animal ethical committee and followed institutional animal use and care guidelines.
Bacterial strains
The S. enterica serovar Typhimurium 
3181, SR11 derivative 4666 (40) was used for the studies shown in Fig. 1, A and B, and Figs. 3, 5, and 8. S. enterica serovar Typhimurium SL1344 and its enhanced GFP (eGFP)-expressing derivative SMO22 were used where bacterial uptake in vivo was examined (see Fig. 4) (41). Strain 14028r (42) expressing OVA (43) was used for cytokine production and bacterial uptake experiments (Fig. 2, C–E, and Fig. 6B). Strain 4550 expressing OVA-GFP or GFP (44) was used for in vitro infection in the Ag presentation assays (see Fig. 7). Strain 8554 (SR11 
asdA16 rpsL hisG) was used for in vivo infection of mice in Fig. 2, C and D, and Figs. 6, 7, and 9. Bacteria were grown in Miller’s Luria-Bertani (LB) broth or Lennox (8554) broth overnight at 37°C. For SL1344 and SMO22 the medium was supplemented with antibiotics (SL1344, 100 µg/ml streptomycin; SMO22, 50 µg/ml kanamycin and 100 µg/ml streptomycin). 14028r/OVA was grown on agar supplemented with 50 µg/ml carbenicillin. The bacterial suspension was diluted and the OD was measured at 600 nm. After centrifugation, the bacteria were resuspended at the appropriate concentration in sterile PBS. The actual bacterial dose administered was determined by the serial plating of bacteria on LB agar plates.
|
|
|
|
|
|
|
|
|
Mice were given 0.1 ml of 1% NaHCO3 intragastrically 10 min before infection. C57BL/6 or RAG–/– mice were then infected intragastrically with 2–8 x 107 bacteria in 200 µl of PBS. IFN-–/– and IL-12–/– mice are more susceptible to Salmonella and were given 5–8 x 106 bacteria in 200 µl PBS intragastrically. Only IFN-–/– and IL-12–/– mice with a bacterial load similar to that of wild-type mice were included. In experiments where cell interactions with eGFP-expressing bacteria in vivo were studied (Fig. 4), mice were infected intragastrically with 109 SMO22 or SL1344 bacteria in 200 µl of PBS. In experiments using 8554 (Fig. 2, C–E, and Figs. 6, 7, and 9), mice were infected intragastrically with 8 x 108 bacteria in 200 µl of PBS. Mice were sacrificed after 2–5 days and the bacterial load in PP, MLN, spleen, and blood was determined by plating serial dilutions of single cell suspensions on LB or Lennox (8554) agar plates.
Cell preparation
Mice were sacrificed 2–5 days postinfection and the spleen, MLN, and PP were removed. Blood was collected from the heart by vascular perfusion with 5 ml of PBS containing 4 mM EDTA. The organs were digested with 0.45 mg/ml Liberase CI (Roche) in HBSS for 30 min at 37°C and pipetted into a single cell suspension or pressed through nylon mesh (TP filter). Splenocytes and blood were treated with 0.14 M NH4Cl for 5 min at room temperature once or twice, respectively, to lyse RBC, and debris was removed by filtration. Cells were then washed three times with HBSS and resuspended in RPMI 1640 (Invitrogen Life Technologies) with heat-inactivated 10% FBS (PAA Laboratories). The total number of viable cells per organ was determined by trypan blue exclusion. For intracellular cytokine staining, 2 x 106 cells/ml were resuspended in 5 ml of complete RPMI 1640 supplemented with 10% FBS and incubated at 37°C in the presence of 5 µg/ml brefeldin A (Sigma-Aldrich) for 4 h. For bacterial uptake studies in vivo, CD11b-expressing cells were enriched from PP and MLN pooled from 2–5 mice by magnetic separation using autoMACS (Milteny Biotech) after incubation with anti-CD11b beads (clone M1/70), (Milteny Biotech) according to manufacturer’s protocol.
Flow cytometry
Single cell suspensions were washed in FACS buffer (HBSS containing 3% FBS, 1 mM EDTA, and 10 mM HEPES) that was used throughout the surface staining procedure. Fc receptors on cells were blocked by incubating with the anti-FcRII/III mAb 2.4G2 for 15 min at 4°C. Cells were then stained for 20 min at 4°C with the following mAbs that were biotinylated or conjugated to FITC, PE, allophycocyanin, allophycocyanin-Cy7, PE-Cy7, Pacific Blue, or Qdot605: anti-CD11c (clone HL3), anti-Gr-1 (clone RB6-8C5), anti-CD11b (clone M1/70), anti-F4/80 (clone CI:A3-1), anti-MHC-II (clone M5/114.15.2), anti-CD80 (clone 16-10A1), anti-CD86 (clone GL1), anti-CD62L (clone MEL-14), anti-NK1.1 (clone PK136), anti-CD19 (clone 1D3), anti-B220 (clone RA3-6B2), anti-TCR -chain (H57-597), anti-CD4 (clone GK1.5), anti-Ly6C (clone AL-21), and anti-Ly6G (clone 1A8). The following anti-chemokine receptors were used: anti-CCR2 (clone MC21), anti-CCR6 (clone 1C12), anti-CXCR2 (clone 242216), and anti-CXCR3 (clone 220803). The Abs rat IgG1 (clone 8R3-34), IgG2a (clone R35-95), and IgG2b (clone A95-1) were used as isotype controls and biotinylated mouse anti-ratIgG2a as the secondary Ab. All mAb except anti-F4/80 (Caltag Laboratories), anti-CXCR2 and anti-CXCR3 (R&D Systems), and anti-CCR2 (generously provided by M. Mack, University of Regensburg, Regensburg, Germany), were purchased from BD Pharmingen. Anti-CD4 (clone GK1.5), purified in house, was conjugated to Qdot605 (Quantum Dot Corporation), and anti-Ly6G was conjugated to Pacific Blue (Molecular Probes) according to the manufacturer’s protocol. 7-Aminoactinomycin D (7-AAD; Sigma-Aldrich) was included in all stainings to define viable cells.
For staining with anti-CCR2 and anti-CCR6, cells were first blocked with 10% normal mouse serum (Sigma-Aldrich) for 15 min and stained with the primary Ab for 45 min. After washing, cells were incubated with a secondary biotin-conjugated Ab for 20 min, washed, and blocked with the FcRII/III mAb 2.4G2 for 15 min. Finally, the cells were incubated with the additional mAbs to stain surface molecules and with streptavidin-allophycocyanin for 20 min.
When intracellular staining was performed, the samples were washed after surface staining and fixed in 2% formaldehyde in PBS for 20 min at room temperature. After the cells were washed in permeabilization buffer (HBSS containing 0.5% saponin and 0.5% BSA (Sigma-Aldrich)), which was subsequently used throughout the staining procedure, they were incubated for 30 min at room temperature with one or more of the following mAbs: anti-TNF- (clone MP6-XT22), anti-IL12p40 (clone C15.6), anti-IFN- (clone 37895.11) (all purchased from BD PharMingen), anti-CD68 (clone FA/11; Serotec), S. enterica serovar Typhimurium O-4 (clone 1E6; Abcam), and the appropriate isotype control or iNOS Ab (M-19) (Santa Cruz Biotechnology). Staining with rabbit anti-iNOS was followed by incubation with allophycocyanin-conjugated anti-rabbit IgG (Santa Cruz Biotechnology) for 30 min at room temperature. An anti-Salmonella O-4 mAb was conjugated to Pacific Blue according to the manufacturer’s protocol.
Cells or bacteria (strains SMO22 and SL1344) were acquired on an LSR-II flow cytometer (BD Biosciences) using DIVA software (BD Biosciences) and analyzed using Flow Jo software (Tree Star). Whereas the absolute number of the cell populations is reported for the MLN and the spleen, the number and size of PP differs in individual mice and thus the percentage rather than the absolute number of cells in a given population was calculated.
Bacterial survival assay and Ag processing and presentation assays
Cell suspensions were made from the MLN, spleen, and blood pooled from 10 mice at day 4 postinfection as described above. Cells from the spleen and blood were subsequently pooled (to get enough cells to work with) and depleted of B, T, and NK cells by magnetic separation using autoMACS (Milteny Biotech) after incubation with a mixture of anti-CD19 beads, anti-CD90 (Thy1.2) beads, and anti-NK (DX5) beads according to the manufacturer’s protocol (Milteny Biotech). MLN cells were likewise depleted of T, B, and NK cells. After staining for flow cytometry, cells were sorted into CD11chigh DC, Ly6Chigh monocytes, Ly6Clow monocytes, and Ly6Ghigh neutrophils using a FACSAria flow cytometer (BD Biosciences) and DIVA software (BD Biosciences). The different cell populations were seeded at 1.5 x 105 cells/well in round-bottom 96-well plates in complete medium (RPMI 1640 Glutamax-1 containing 10% FBS (PAA Laboratories), 2 mM MEM sodium pyruvate, 20 mM HEPES, 0.05 mM 2-ME. and 2.5 µg/ml fungizone (all from Invitrogen Life Technologies)). For bacterial survival assays, sorted Ly6Ghigh neutrophils and Ly6Chigh monocytes were infected with strain 14028r, centrifuged at 270 x g for 4 min, and incubated for 2 h at 37°C. The bacteria-to-cell ratio was between 5:1 and 18:1. After washing three times in complete medium containing 80 µg/ml gentamicin, the cells were resuspended in the medium and incubated at 37°C for an additional 1 or 18 h. Cells were lysed with 0.2% Triton X-100 with vigorous pipetting and incubated at room temperature for 10 min. The supernatant was then serially diluted and plated onto LB agar plates. Bacterial colonies were counted after overnight incubation at 37° C and the total number of bacteria recovered was calculated. As controls for determining the number of bacteria not internalized by the cells but viable in the supernatant, cells were preincubated with 20 µg/ml CCD (cytochalasin D) 1 h before the addition of bacteria and present for the initial 2 h of bacterial infection.
For Ag-presenting assays, titrated numbers of 4550-OVA-GFP or 4550-GFP (no OVA) were added to the cells. Alternatively, the OVA323–339 peptide was added at a concentration of 0.1, 1, or 10 µg/ml. Cells were then centrifuged for 270 x g for 4 min and incubated at 37° C. After 2 h, the cells were washed three times in complete medium containing 80 µg/ml gentamicin and resuspended in 100 µl of medium with gentamycin. Finally, 2 x 105 MACS-purified, CFSE-labeled OVA323–339-specific CD4+ T cells from OT-II mice were added. The OT-II CD4+ T cells were negatively purified from pooled spleens plus MLN from naive mice using a CD4 isolation kit (Milteny Biotech) according to the manufacturer’s protocol and were labeled with CFSE. For CFSE labeling, cells were incubated with 5 µM CFSE (Molecular Probes) in PBS for 8 min at room temperature. After 3.5 days of coculture at 37°C, cells were harvested, stained with anti-CD4 and anti-TCR, and the percentage of divided CD4+ OT-II T cells was determined by flow cytometry.
Measurement of cytokines by ELISA
Sorted Ly6Ghigh neutrophils, Ly6Chigh monocytes, and CD11chigh DC were resuspended in medium or infected for 2 h with 4550-OVA or 14028r, washed, and treated as described for bacterial survival assays. After 20–36 h, as indicated in the appropriate figure legends, the plates were centrifuged and the supernatants were frozen at –20°C until assayed for cytokine content. The presence of TNF-, IL-1, and IL-12p70 in supernatants was determined by ELISA following the manufacturer’s protocol (BD Biosciences). To determine naive cytokine levels, CD11b-expressing cells were enriched from the spleen of naive mice by autoMACS after incubation with anti-CD11b beads (clone M1/70). The CD11b+ cells were seeded at 2 x 105 cells/well in 96-well plates and incubated for 48 h before supernatants were frozen at –20°C.
Adoptive transfer and blocking experiments
Donor monocytes and neutrophils were negatively enriched using autoMACS and anti-CD49b (clone DX5) beads (Milteny Biotech) from pooled splenocytes and the blood of 10–15 RAG–/– mice infected 3 days earlier with Salmonella 4666. Donor T cells were negatively enriched from the pooled spleen and MLN of naive C57BL/6 mice using autoMACS and a CD4 isolation kit (Milteny Biotech). The donor monocytes, neutrophils, and T cells were then pooled and CFSE labeled as described above. Cells (10–15 x 106) containing 
35% T cells and 39% CD11b+NK1.1– cells were injected i.v. into C57BL/6 mice that were infected 3 days earlier with 4666. Some of the recipient mice were given 150 µg of the anti-MAdCAM-1 mAb MECA-367 (45) (BD Pharmingen) i.v. 6–8 h before cell transfer. Control animals received either 150 µg of the isotype control mAb 9B5 (anti-human CD44; kindly provided by A. Hänninen (University of Turku, Turku, Finland) or nothing. Twelve to 16 h after adoptive transfer, PP, MLN, spleen, and blood were removed from recipient mice and stained for flow cytometry as described above.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
The gut-draining lymphoid tissues PP and MLN are the first organs seeded by Salmonella after oral infection (27). However, little is known about the changes in the number, phenotype, and function of phagocytic cell populations, particularly the recently described monocyte subsets, responding to this oral pathogen in these organs. To address these issues, mice were orally infected with Salmonella and cells from PP and MLN were analyzed by five- to seven-color flow cytometry early during infection. Populations with differential expression of CD68 and Gr-1 responding to oral infection were apparent (Fig. 1). In particular, two populations that were relatively rare in naive mice, CD68intGr-1high (where int stands for intermediate) and CD68highGr-1int cells (gate R1 and R2, respectively, in Fig. 1), increased dramatically in the MLN and PP of infected mice starting at day 3 postinfection (Fig. 1, A and C and D). Although the increase in both populations was large, CD68highGr-1int (R2) cells remained more abundant than gated R1 cells and comprised 1–3% of total cells per organ at day 5 postinfection.
To further characterize the two populations responding to oral Salmonella infection, the expression of a number of surface and costimulatory molecules was examined (Fig. 1E). The gated CD68intGr-1high (R1) cells in infected mice lacked F4/80 and CD11c expression but had a very high level of CD11b. Few cells within this gate were positive for MHC-II, CD80, or CD86 at day 5 postinfection and the positive cells likely represent some spillover of R2 cells, as the junction between R1 and R2 gates in infected tissues becomes less absolute than in naive animals (Fig. 1A). A fraction of R1 cells were positive for CD62L. Together, these data suggest that cells in the R1 gate are neutrophils (34, 38).
Although most cells within the R2 gate of infected MLN were CD11bhigh, they had several features that distinguished them from R1 cells. This included the expression of F4/80 and a significant level of MHC-II by the majority of the gated cells. R2 cells from infected mice had increased CD80 and CD86 expression compared with naive mice, a fraction of the cells were CD62L+, and some expressed a low level of CD11c (Fig. 1E). Thus, infection-induced R2 cells are phenotypically distinct from R1 cells.
Although the very low level of CD11c on gated R2 cells suggested that they were not conventional DC, which in the mouse are characterized by high surface expression of CD11c, MHC-II, and costimulatory molecules (46), the phenotype of gated R2 cells was compared with that of the CD11chigh population in the same infected individual (Fig. 1, B and E). Consistent with previous reports, the CD11chigh cells expressed high levels of MHC-II, CD80, and CD86. Both myeloid and lymphoid DC are included in this gate as evidenced by the biphasic expression pattern for CD11b (Fig. 1E), which corresponds to the CD8+ and CD8– subsets (Ref. 46 and references therein). Moreover, additional studies showed that 20% of the CD11chighCD11blow DC express CD8, which is consistent with previous data (46, 47, 48). Although most of the DC expressed CD68 (data not shown), their low Gr-1 expression and the nonoverlapping gating for Gr-1 on R2 vs R3 cells exclude DC from the R2 gate (Fig. 1, A and B). Instead, most CD68highGr-1low cells in the PP and MLN of both naive and infected mice (Fig. 1A, cells below the R2 gate) were DC expressing high levels of CD11c and MHC-II (data not shown). Furthermore, very few plasmacytoid DC identified as TCR–CD19–NK1.1–CD11b–CD11cintB220+Ly6C+ (48, 49) were detected in naive MLN (0.06%), and they decreased to 0.01% of the total population in infected MLN. In addition, the CD11b–CD11cint phenotype of plasmacytoid DC (48, 49) excludes them from the R1–R3 gates. Less than 10% of R1, R2, and R3 cells at day 5 postinfection were positive for the lymphoid markers CD19, TCR, and NK1.1 (data not shown). B cells (CD19+) and T cells (TCR+ increased slightly in MLN and PP early during oral infection. In MLN, the absolute number of B and T cells increased 1.5-fold and 1.2-fold, respectively, at day 5 postinfection while the number of NK cells (NK1.1+TCR–) remained constant (data not shown).
Together, the data in Fig. 1 show that the predominant cellular change that occurs in the GALT after oral Salmonella infection is the rapid accumulation of cells with a phenotype distinct from neutrophils and DC. Moreover, the phenotype of R2 cells suggests they may be monocytes recruited from the blood, as has been shown in peritoneal or i.v. inflammation models and in lymph nodes draining inflamed skin (1, 2, 3, 4, 6, 7, 8, 37). Because monocyte refers to a population circulating in the blood and macrophage indicates a tissue resident differentiated cell, we refer to the cells that accumulate in infected PP and MLN as monocyte/macrophage (mono/mac).
The mono/mac that accumulate in PP and MLN are the main producers of TNF- and iNOS during infection
To assess the function of the myeloid cells recruited to PP and MLN during oral Salmonella infection, their capacity to produce TNF- and iNOS, effector molecules important in controlling Salmonella (30), were assessed by direct ex vivo double intracellular staining. Mono/mac producing TNF- alone, iNOS alone, or both simultaneously were induced by Salmonella infection (Fig. 2A and Table I). Mono/mac were the predominant producers of these anti-bacterial molecules, accounting for 50–70% of all TNF-+ and 80–90% of all iNOS+ cells in MLN. Moreover, 20–25% of the mono/mac that stain positive for iNOS were also positive for TNF-.
|
was apparent among neutrophils and DC in infected mice (Table I).
We also determined the ability of sorted monocytes, neutrophils, and DC from Salmonella-infected mice to secrete TNF-, IL-1, and IL-12p70 after 20 h (TNF-) or 36 h of ex vivo culture. TNF- and IL-1 were detected in the supernatant of cultured monocytes and neutrophils, whereas DC production of these cytokines was not apparent (Fig. 2, C–E). Little if any increase in the amount of TNF- was detected upon the restimulation of sorted monocytes with bacteria, and no increase was detected upon DC restimulation (data not shown). In contrast, the restimulation of monocytes or neutrophils with bacteria during the ex vivo culture resulted in enhanced IL-1 production (Fig. 2, D and E). IL-12p70 was not detected in the supernatant of any of the sorted cell types cultured ex vivo with or without Salmonella (data not shown). Likewise, an increase in monocytes from the MLN or spleen of mice infected 3 or 5 days earlier with Salmonella that stained positive for intracellular IL-12p40 above the level detected in naive mice was not detected. Despite the fact that IL-12p70/IL-12p40 has a role during Salmonella infection (30), the cellular source(s) of this cytokine remain elusive.
Thus, mono/mac, defined as CD68highGr-1int cells, express MHC-II and costimulatory molecules (Fig. 1E) and are the main sources of TNF- and iNOS early during Salmonella infection. Moreover, the majority of mono/mac in infected tissues produce TNF- or iNOS rather than both simultaneously.
Impaired expression of iNOS and MHC-II but not TNF- in infected IFN-–/– and IL12p40–/– mice
We next asked whether iNOS and TNF- produced by recruited mono/mac early during infection depended on IFN- and IL-12. First, the accumulation of mono/mac and neutrophils in the PP, MLN, and spleen of infected IFN-–/– and IL-12p40–/– mice with a bacterial burden similar to that of infected wild-type mice 4 days postinfection was not compromised (data not shown). Likewise, the frequency of mono/mac producing TNF- in PP, MLN, and spleen was the same in wild-type, IFN-–/–, and IL-12p40–/– mice with a similar bacterial burden (Fig. 3, A–C). In contrast, iNOS production by mono/mac in all three organs in infected IFN-–/– and IL-12p40–/– mice was nearly absent (Fig. 3, A and B).
We also wished to determine whether the infection-induced increase in MHC-II on the mono/mac of infected wild-type mice was IFN--dependent. Indeed, Salmonella-induced MHC-II up-regulation on mono/mac was severely compromised in IFN-–/– and IL-12p40–/– mice (Fig. 3D). This is in contrast to MHC-II expression on DC, which was similar and high on CD11chigh (R3) cells in infected wild-type and knockout mice. Similar results were seen in the spleen (data not shown). We next examined the cellular source(s) of IFN- in the PP and MLN responsible for infection-induced mono/mac activation. These data revealed that NK and NKT cells (NK1.1+TCR+) and some T cells (TCR+NK1.1–) are the main producers of IFN- 4 days postinfection (Fig. 3E). Thus, the induction of iNOS and MHC-II expression but not TNF- production by mono/mac during oral Salmonella infection is dependent on IL-12 and IFN-, and NK/NKT cells and T cells are the predominant sources of IFN- in PP and MLN early during infection.
Uptake of eGFP-expressing Salmonella by phagocytic cells
To examine the relative uptake of bacteria by mono/mac and neutrophils in the same individual, mice were orally infected with Salmonella expressing eGFP and bacteria-associated cells were identified by flow cytometry. To get enough events to accurately assess the phenotype of eGFP+ cells, CD11b+ cells were first enriched by positive selection using MACS. In addition, an analysis of eGFP fluorescence negated using FITC-conjugated anti-CD68 (Fig. 1A), so an alternate strategy using CD11b, Gr-1, and F4/80 was used to identify the phagocytic cells (Fig. 4A) (7). To conclude that this strategy allowed identification of the neutrophil and mono/mac cells studied above, CD11b+Gr-1highF4/80– and CD11b+Gr-1intF4/80+ were backgated into a CD68 vs Gr-1 plot (Fig. 4B). This showed that the F1 and F2 populations defined by CD11b, Gr-1, and F4/80 identified the majority (63–95% in three independent experiments) of mono/mac (R1) and neutrophils (R2), respectively, with little cross-contamination between the two populations (Fig. 4, A and B) (7).
The results in Fig. 4C show a slightly higher fraction of eGFP+ neutrophils (F1 cells) compared with mono/mac (F2) (1.6-fold difference) in the PP and MLN of infected wild-type mice. In infected IFN-–/– mice, very similar fractions of neutrophils and mono/mac were eGFP+ in both PP and MLN (1.2-fold difference). The somewhat higher fraction of eGFP+ cells in the MLN of infected IFN-–/– mice compared with wild-type mice reflects the higher number of bacteria recovered from IFN-–/– mice in these experiments (data not shown) due to the compromised capacity to kill Salmonella in the absence of IFN- (30).
To investigate the accuracy of detecting cell-associated bacteria in wild-type and IFN-–/– mice using eGFP fluorescence as the readout, two types of experiments were performed. First, the stability of eGFP in CD11b+ cells from infected wild-type and IFN-–/– mice was assessed. These data showed similar eGFP stability, as 94 ± 6.3% (n = 7) and 94 ± 2.2% (n = 6) of bacterial colonies recovered from CD11b+ cells isolated from infected wild-type and IFN-–/– mice, respectively, were eGFP+. Second, the detection bacteria by eGFP fluorescence was compared with the use of an anti-Salmonella O-4 LPS Ab. These studies revealed that 1.3–4.9% of CD11b+ cells from PP and MLN pooled from infected wild-type mice stained positive for O-4 reactivity, while eGFP fluorescence was detected in 0.4–2% of the cells in the same infected individuals. Identical percentages were obtained for infected IFN-–/– mice. Although the down-regulation of eGFP fluorescence in vivo cannot be eliminated as a contributor to this difference, the greater reactivity of an anti-LPS Ab is not unexpected because it will recognize LPS on intact bacteria as well as bacterial degradation products and shed LPS. In contrast, the detection of bacteria using eGFP fluorescence relies on intact protein emitting fluorescence, a method that will not detect bacteria degraded to a point that negates eGFP fluorescence.
Very few eGFP+ events were found among TCR+ or B220+ cells in infected tissues of the same animals, demonstrating that >80% of all bacteria-associated cells were either neutrophils or mono/mac (data not shown). The relationship between bacterial uptake and effector molecule production was also assessed directly ex vivo (Fig. 4D). Approximately 30–40% of eGFP+ CD11b+Gr-1int F2 cells were also iNOS+ (data not shown), demonstrating the tendency of bacteria-associated cells to produce this effector molecule. However, while 30–40% of all CD11b+Gr-1int F2 cells expressed iNOS (Fig. 3B), only a minor fraction of iNOS+ or TNF-+ CD11b+Gr-1+ cells were eGFP+ (Fig. 4D). This suggests that most iNOS and TNF--producing cells are bystander cells, as has been observed with other pathogens (5, 7, 12). In addition, bacterial uptake did not appear to influence the infection-induced up-regulation of MHC-II on mono/mac either in wild-type or IFN-–/– hosts (see Fig. 4E for PP; MLN data not shown). Similarly, bacterial uptake did not restore the ability of mono/mac from infected IFN-–/– mice to produce iNOS (Fig. 3A and data not shown).
Accumulation of cells in the blood of orally infected mice
Having established that neutrophils and cells with a phenotype suggesting a relationship to monocytes (3, 9, 13, 34, 50) (our so-called mono/mac population) accumulated and exerted effector functions in the gut-associated lymphoid tissues of orally infected mice, we asked when these cells appeared in the blood during infection and whether they were activated in the circulation. To this end, the populations defined in PP and MLN based on CD68 and Gr-1 expression were characterized in the blood of naive and Salmonella-infected mice. CD68 vs Gr-1 staining on blood cells revealed abundant R1 and R2 populations similar to those found in the tissues (Figs. 5A and 1A). In naive mice, R1 and R2 were 8 and 4%, respectively, of all leukocytes in blood. Both R1 and R2 populations accumulated in the blood as the infection progressed, increasing 2-fold and 4-fold, respectively, at day 3 postinfection and 2.5-fold and 4.5-fold, respectively, at day 5 (Fig. 5, A and C).
To further characterize the cell populations in blood during infection and compare them with the ones found in infected tissues, the expression of myeloid markers and costimulatory molecules on the blood populations was analyzed (Fig. 5D). The phenotype of gated Gr-1highCD68int (R1) cells in the blood was very homogenous and similar in both infected and naive mice (Fig. 5D). Essentially, all blood R1 cells expressed CD11b and CD62L but lacked F4/80, CD11c, MHC-II, CCR2, and the costimulatory molecules CD80 and CD86. In addition, gated R1 cells (as well as R2 and R4 cells) were negative for CD19, TCR, and NK1.1 (data not shown).
Like their tissue counterparts, blood CD68highGr-1int (R2) cells from infected mice were positive for CD11b, F4/80, and CCR2 (Fig. 5D and data not shown). The adhesion molecule CD62L was expressed on some blood R2 cells in naive mice, and the fraction of CD62L+ cells increased in the blood of Salmonella-infected mice. Similar to R2 cells in the MLN of infected mice, R2 cells in the blood expressed much higher MHC-II in infected compared with naive animals. In contrast, however, the up-regulation of CD86 and particularly CD80 on R2 cells in the blood was modest compared with that seen in infected MLN (Figs. 1E and 5D).
Cells consistent with the phenotype of murine DC present in the tissues of naive and infected mice (CD11chighMHC-II+ cells) were not found in the blood (Fig. 5B and data not shown) (51). However, two populations of monocytes, Gr-1+CD62L+CCR2+ inflammatory and Gr-1–CD62L–CCR2– resident monocytes, have been described in mouse blood (3). The Gr-1+CD62L+CCR2+ phenotype of gated R2 cells is consistent with the phenotype of inflammatory monocytes, while Gr-1–CD62L– cells in R4 share features with resident monocytes (Fig. 5, A and D) (3). R2 cells increased in the blood of infected mice while R4 cells did not (Fig. 5C). Moreover, most of the cells within the R4 gate, putative resident monocytes, express an intermediate level of CD11c and up-regulate MHC-II during infection. This suggests a possible relationship to, or capacity to become, DC. Cells within gate R4 appear to be the best candidate for resident monocytes, as the CD11cint cells in gate R5 in Fig. 5B were mostly (>80%) NK1.1+CD11bint NK cells (data not shown), consistent with other reports (51, 52). The remaining NK1.1– cells in R5 were CD68highGr-1lowCD11bint and fall within gate R4 in Fig. 5A.
Given the efficient activation of mono/mac to produce large amounts of iNOS and TNF- and to up-regulate MHC-II in the PP, MLN, and spleen during infection, we asked whether this activation already occurs in the blood of infected mice. As seen in Fig. 5D, MHC-II was readily up-regulated on R2 cells, putative inflammatory monocytes in blood. However, no Salmonella-induced iNOS and little TNF- above the level seen in naive mice were detected in this population (Fig. 5E). Together, the data in Fig. 5 show that two subsets of monocytes resembling resident (R4) and inflammatory monocytes (R2) can be identified in the blood of Salmonella-infected mice. R2 but not R4 monocytes increase in the blood during Salmonella infection and both have increased MHC-II expression but do not produce anti-bacterial effector molecules. Moreover, neutrophils increase in the blood of Salmonella-infected mice, while cells with the features of tissue DC are not readily apparent in the blood of naive or infected animals.
Bacterial uptake and Ag presentation capacity
To examine the phagocytic activity and killing capacity of monocytes and neutrophils, these cells were isolated from blood and spleen at day 4 postinfection, pooled, and depleted of T, B, and NK cells by MACS. The remaining cells were sorted into Ly6Chigh monocytes (CD11b+Ly6GlowLy6Chigh) and neutrophils (CD11b+Ly6GhighLy6Cint) (Fig. 6A). Because CD68 is expressed intracellularly and requires a staining procedure using fixation and permeabilization that kills the cells, we changed the gating strategy and used CD11b, Ly6C, and Ly6G to identify mono/mac and neutrophils (7, 8, 31, 34, 37, 53). Anti- TCR, CD19, NK1.1, and CD11c were also included in the staining to assure that no T cells, B cells, NK cells, or DC were included in the gates. To conclude that this strategy allowed identification of the mono/mac and neutrophils previously defined by CD68 and Gr-1, CD11b+Ly6GlowLy6Chigh monocytes and CD11b+Ly6GhighLy6Cint neutrophils were backgated into a CD68 vs Ly6G plot (Fig. 6A). This showed that 79% of the CD11b+Ly6GlowLy6Chigh cells fell within the mono/mac (R2) gate and 97% of the CD11b+Ly6GhighLy6Cint cells fell into the neutrophil (R1) gate, with little cross-contamination between the two populations (Fig. 6A).
The sorted monocytes and neutrophils were pulsed with Salmonella for 2 h, washed extensively, cultured for an additional 1 or 18 h in medium containing gentamicin, and the intracellular survival of bacteria was examined. After 3 h, both monocytes and neutrophils had phagocytosed bacteria but the neutrophils were more effective (Fig. 6B), which is consistent with the in vivo data (Fig. 4C). After 20 h, few viable bacteria were recovered from the two populations, showing that most of the phagocytosed bacteria had been killed. The use of CCD in parallel wells showed that the majority of bacteria were actively internalized rather than being attached to the cell surface (Fig. 6B).
Having established that Ly6Chigh monocytes were able to phagocytose bacteria in vitro, we wanted to examine whether they or Ly6Clow monocytes from infected mice were able to induce the proliferation of OT-II CD4+ T cells. To examine this, Ly6Chigh monocytes, Ly6Clow monocytes, and DC from infected mice were sorted based on expression of CD11b, Ly6C, Ly6G, and CD11c (Fig. 7A), pulsed ex vivo with Salmonella expressing OVA, and subsequently cocultured with CFSE-labeled OT-II CD4+ T cells. DC but not Ly6Chigh or Ly6Clow monocytes induced proliferation of OT-II cells (Fig. 7B). Bacterial titration showed that the T cell response peaked when DC were stimulated using a 5 to 1 bacteria to cell ratio, and that Ly6Chigh monocytes were unable to induce a T cell response at all bacteria to cell ratios tested (Fig. 7C). Furthermore, DC titration with a fixed bacteria to cell ratio (5 to 1) showed that proliferation of OT-II cells decreased as DC number declined (Fig. 7D). The observed proliferation of DC was epitope-specific, because a lack of proliferation was observed when the cells were pulsed with Salmonella not expressing OVA (Fig. 7D). Finally, pulsing with the OVA323–339 peptide showed that Ly6Chigh monocytes induced a modest proliferation of OT-II cells, while both Ly6Clow monocytes and DC efficiently induced proliferation (Fig. 7B).
Together, these results show that Ly6Chigh monocytes and neutrophils phagocytose and kill Salmonella in vitro. Moreover, and in contrast to DC, neither Ly6Chigh or Ly6Clow monocytes induce OT-II T cell proliferation after pulsing with Salmonella expressing OVA although Ly6Clow monocytes induce OT-II proliferation after pulsing with OVA peptide.
Monocytes and neutrophils are recruited to PP despite blocking of MAdCAM-1
Monocytes and neutrophils in blood express the selectin CD62L (Fig. 5D) and the 4 integrin (37, 54, 55). CD62L and 4 partnered with 7, in particular, are ligands for MAdCAM-1. MAdCAM-1 is expressed on the high endothelial venules of PP and is critical for lymphocyte homing to this organ (45, 56, 57). To investigate whether MAdCAM-1 is also involved in monocyte and neutrophil recruitment to PP during Salmonella infection, blocking studies were performed. Although T cell recruitment to the PP of Salmonella-infected mice was abrogated in animals treated with the anti-MAdCAM-1 mAb MECA-367, monocytes and neutrophils accumulated in PP equally well in MECA-367-treated and untreated mice (Fig. 8). MECA-367 blocked T cell recruitment to PP but not MLN (Fig. 8. B and C) or spleen (data not shown) as predicted (45). Thus, unlike T cells, monocytes and neutrophils do not require the addressin MAdCAM to enter into PP in mice orally infected with Salmonella.
Chemokine receptor expression on Ly6Chigh monocytes, Ly6Clow monocytes, neutrophils and DC in blood, PP, and MLN of Salmonella-infected mice
To gain further insight into entry of myeloid cells into the gut lymphoid tissues during oral Salmonella infection, we next studied chemokine receptor expression because these receptor/ligand interactions provide important information that direct cell entry into lymphoid tissues (58). Thus, CXCR2, CXCR3, and CCR6 expression on myeloid cells from blood, PP and MLN were analyzed by flow cytometry at day 4 postinfection. Almost all blood neutrophils express CXCR2, while <20% of neutrophils in PP and MLN express this receptor (Fig. 9). In contrast, few if any Ly6Chigh monocytes in any of the immune compartments examined were CXCR2+. The fraction of cells positive for CXCR3 was relatively low for all cell types in all tissues examined, except for neutrophils in MLN where 13% were positive. Finally, CCR6 expression was only found on DC in PP and MLN. Thus, the high expression of CCR2 on Ly6Chigh monocytes in blood (R2 cells; Fig. 5D) is consistent with the role of this receptor and its ligand MCP-1 (CCL2) in the release of these inflammatory monocytes into the blood for subsequent entry into infected tissues, as described previously (3, 7, 12, 59). Moreover, the data suggest a potential role for CXCR2 in neutrophil migration and CCR6 in the migration of DC.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
.
CD68highGr-1int monocytes recruited to PP and MLN during infection had increased MHC-II, CD80, and CD86 but yet had phenotypic and functional features distinct from DC in the same tissue. For example, low CD11c expression was detected on a small fraction of recruited monocytes compared with high CD11c expression on DC (40, 46, 47). Furthermore, MHC-II expression on Gr-1int monocytes was greatly increased during infection in an IL-12- and IFN--dependent fashion. In contrast, MHC-II expression on CD11chigh DC was very high in both steady-state and infected organs and was unaffected in the absence of IL-12 or IFN-. Functionally, recruited monocytes produced iNOS, TNF-, and IL-1 while DC produced little if any of these molecules. Double intracellular staining revealed that the majority of monocytes produced either TNF- or iNOS, with relatively few cells staining positive for both. The production of TNF- and/or iNOS by monocytes recruited to the PP and MLN of orally infected mice is consistent with studies of monocytes recruited to the spleen after the i.v. administration of Listeria monocytogenes (7, 12, 48). The capacity of neutrophils from Salmonella-infected tissues to release some cytokines is also consistent with the complex activities of these cells observed with other pathogens (60, 61).
Our studies using eGFP-expressing Salmonella in vivo revealed that a relatively small fraction of monocytes producing iNOS or TNF- were eGFP+. The findings that relatively few cells associate with Salmonella early during infection (1%; Fig. 4 and Ref. 47) and a large fraction of all mono/mac produce iNOS or TNF- (30%) make it unlikely that bacterial uptake and destruction (which negates eGFP detection) accounts for the predominance of eGFP– cells making iNOS or TNF-. The data rather suggest the production of TNF- and iNOS by bystander cells, an observation consistent with other infection models (5, 12). Cells containing bacteria can respond differently than bystander cells exposed only to factors in the environment (47). Bacterial association did not, however, overcome the IFN-–/– requirement for recruited monocytes to produce iNOS or up-regulate MHC-II. Thus, the IFN-–/–-mediated pathways inducing iNOS and MHC-II up-regulation during infection cannot be overcome by alternate pathways induced in bacteria-containing cells, as can occur for the up-regulation of costimulatory molecules on DC that cannot respond to TNF- (47).
The Ag presentation capacity of the myeloid lineage cells differed. For example, while recruited (Ly6Chigh) monocytes induced only marginal proliferation of OT-II CD4+ T cells after stimulation with OVA peptide, Ly6Clow monocytes induced robust proliferation. The reason for this differential capacity of Ly6Clow and Ly6Chigh monocytes is not clear, as both subsets express MHC-II and costimulatory molecules, albeit at much lower levels than DC. Effectively inducing T cell proliferation may be a property acquired during monocyte differentiation and is performed only by the mature Ly6Clow population. Alternatively, iNOS produced by monocytes could inhibit T cell proliferation (62).
In addition and in contrast to DC, neither Ly6Clow nor Ly6Chigh monocytes caused OT-II proliferation after coincubation with Salmonella expressing OVA. This was somewhat surprising given the capacity of the recruited Ly6Chigh monocytes to phagocytose Salmonella and up-regulate MHC-II and costimulatory molecules during infection. It could be that the Ag-processing machinery in monocytes is not capable of generating the OVA peptide from internalized Salmonella expressing OVA, at least within the time frame examined here. It has also recently been shown that a subset of DC in PP, those expressing CCR6, are required for CD4+ T cell priming du
0.05 was obtained for all groups. Samples contained 34–2550 eGFP+ events. D, CD11b+Gr-1+ cells from the MLN and PP of mice infected 4 days earlier were gated as in the left dot plot in A. eGFP expression on gated iNOS+ (left) or TNF-
