ABSTRACT
Ehrlichia chaffeensis is an obligate intracellular bacterium that infects primarily monocytes and macrophages and causes potentially fatal human monocytic ehrlichiosis (HME) that mimics toxic-shock-like syndrome in immunocompetent hosts. Early recruitment of neutrophils to the sites of infection is critical for the control of bacterial infection and inflammatory responses. We recently observed rapid and sustained neutrophil recruitment at a primary site of infection (peritoneum) following lethal murine ehrlichial infection compared to innocuous ehrlichial infection. We examined here the contribution of neutrophils to protective immunity or immunopathology during infection with monocytic Ehrlichia. Unexpectedly, depletion of neutrophils from lethally infected mice enhanced bacterial elimination, decreased immune-mediated pathology, and prolonged survival. Furthermore, compared to lethally infected sham controls, neutrophil depletion in infected mice resulted in amelioration of pathogenic responses, as evidenced by a decreased number of tumor necrosis factor alpha (TNF-α)-producing CD8+ T cells, which is known to mediate immunopathology and toxic shock in a murine model of fatal ehrlichiosis. Although neutrophil depletion did not influence the number of CD4+ Th1 cells and NKT cells producing gamma interferon (IFN-γ), it increased the ratio of IFN-γ- to IL-10-producing NKT cells as well as the ratio of IFN-γ to interleukin 10 (IL-10) transcripts in the liver. This may ameliorate the net suppressive effect of IL-10 on IFN-γ-mediated activation of infected macrophages and thus may account for the enhanced bacterial elimination. Finally, transcriptional analysis of gene expression in the liver indicated that neutrophils contribute to overproduction of cytokines and chemokines during fatal ehrlichiosis. In conclusion, these results revealed an unexpected role of neutrophils in supporting bacterial replication indirectly and promoting immunopathology during severe infection with an intracellular bacterium.
INTRODUCTION
Human monocytotropic ehrlichiosis (HME) is a highly prevalent life-threatening tick-borne disease in North America caused by Ehrlichia chaffeensis (1, 2), an intracellular Gram-negative bacterium. Ehrlichia infection leads to a systemic inflammatory disease associated with immunopathology and a high mortality rate (3). Established murine models of mild and fatal ehrlichiosis using mildly virulent Ehrlichia muris and highly virulent Ixodes ovatus Ehrlichia (IOE) have greatly facilitated our understanding of the pathogenesis and mechanisms of host defenses during HME. E. muris and IOE are antigenically and genetically closely related to each other and to E. chaffeensis (4, 5). Mildly virulent E. muris causes mild, self-limited but persistent infection in immunocompetent mice and mimics E. chaffeensis infection in its natural host, white-tailed deer (6). In contrast, murine infections with high doses of IOE cause severe and fatal Ehrlichia-induced toxic shock that mimics the clinical and pathological findings in patients with potentially fatal HME (1, 2, 7, 8). Similar to toxic-shock-like syndrome caused by other bacterial pathogens, late antibiotic treatment of patients with HME is usually ineffective in preventing disease progression. Thus, an immune-based therapeutic approach may be effective in the management of HME patients. However, a significant hurdle in design of new therapy against Ehrlichia is the limited understanding of the immunopathogenesis of HME.
Neutrophils have long been viewed as short-lived effector cells of the innate immune system (9). Once recruited into tissues, neutrophils engage in complex bidirectional interactions with macrophages, dendritic cells, natural killer cells, and B and T cells (10). In response to different signals, neutrophils express a vast and diverse repertoire of molecules that are crucial to the development of innate and adaptive immune responses against several pathogens (11). For example, activated neutrophils release several cytokines that mediate the induction and recruitment of CD4+ Th1 cells, T helper 17 (Th17) cells, and CD8+ T cells (12). Neutrophils also can act as antigen-presenting cells, as they can cross-present exogenous antigens to CD8+ T cells and provide costimulatory signals necessary for T cell activation. In turn, T cells attract neutrophils via the release of CXCL8 (interleukin 8 [IL-8]) and activate neutrophils by secreting gamma interferon (IFN-γ). Finally, neutrophils, dendritic cells (DCs), and natural killer (NK) cells colocalize and enhance each other's activity via receptor-receptor interactions and soluble mediators (13–16). As a host defense mechanism during microbial infection, neutrophils are recruited to the site of infection, where they mediate effective bacterial clearance via different mechanisms, including release of lytic enzymes, such as defensins and cathepsin G (CG), production of reactive oxygen intermediates (ROI), and extruding neutrophil extracellular traps (NETs) that contain DNA and granules and function as pattern recognition molecules (PRMs) (8, 17–22). Although infection or inflammation increases the life span of neutrophils, these neutrophils undergo apoptosis as a mechanism to ameliorate inflammatory responses. However, failure to remove apoptotic neutrophils resulted in sustained inflammation (23–26). Studies have demonstrated that elevated levels of Ccr1 during bacterial peritonitis and sepsis were associated with neutrophil accumulation and severe pathology while Ccr1 deficiency decreased neutrophil accumulation and abrogated inflammation and pathology in a murine model of kidney injury and inflammation (27), suggesting that neutrophils play a pathogenic role in sepsis.
Recently, we showed that patients with fatal HME have a substantially higher level of serum IL-8 than patients with mild disease (28), a finding which further implies a role of neutrophil inflammation in pathogenesis of fatal ehrlichiosis. However, the exact contribution of neutrophils to protective immunity or immunopathology during Ehrlichia infection remains undefined. In this study, we examined the contribution of neutrophils to host defense against Ehrlichia and their roles in the pathogenesis of Ehrlichia-induced toxic shock. Neutrophils were depleted from the mice prior to and at the time of infection with a lethal dose of virulent Ehrlichia species. Unexpectedly, neutrophil depletion from lethally infected mice enhanced bacterial elimination, decreased immune-mediated pathology, and prolonged survival. Our results revealed an unexpected role of neutrophils in supporting bacterial replication indirectly and promoting immunopathology during severe infection with an intracellular bacterium.
MATERIALS AND METHODS
Mice and Ehrlichia infection.Female wild-type (WT) C57BL/6 mice (B6), 8 to 12 weeks of age, were obtained from Jackson Laboratories (Bar Harbor, ME). All animals were housed under specific-pathogen-free conditions at the University of Pittsburgh in accordance with the institutional guidelines for animal welfare. The highly virulent monocytotropic Ixodes ovatus Ehrlichia (IOE) and mildly virulent Ehrlichia muris were used in this study. The strains were provided by Y. Rikihisa (Ohio State University, Columbus, OH). IOE or E. muris stock was propagated by passage through WT C57BL/6 mice or DH82 (ATCC canine macrophage cell line), respectively. Single-cell suspensions from the spleens of IOE- or E. muris-infected mice on day 7 postinfection (p.i.) were stored in liquid nitrogen and used as stocks. Mice were infected intraperitoneally (i.p.) with 5 × 103 IOE organisms/mouse or 2 × 105 E. muris organisms/mouse and were monitored daily for signs of illness and survival. On the indicated day p.i., four to six mice/group were sacrificed, and selected organs were harvested for further analysis.
In vivo neutrophil depletion.For neutrophil depletion, mice were treated via an i.p. route with 300 μg of anti-Gr1 antibody (clone RB6-8C5, rat anti-mouse Ly-6G and Ly-6C) or isotype control antibody (isotype rat anti-mouse IgG2b, κ) on days 0, 1, and 3 p.i. with IOE. We refer to these mice as IOE/anti-Gr1 and IOE/sham control, respectively. Another control group included naïve mice that are treated with 300 μg of the anti-Gr1 antibody and is referred to as naïve/sham control. The number of neutrophils in the blood of IOE-infected anti-GR1 monoclonal antibody (MAb)-treated mice was examined by flow cytometry on days 2, 4, 6, and 8 after IOE infection and compared to the number of neutrophils in IOE-infected sham controls treated with isotype control MAb. Anti-Gr1-1 antibody depleted ≥90% of neutrophils on day 2 p.i., and the same efficiency of depletion was observed until day 8 p.i. compared to sham controls.
Isolation of peritoneal exudate cells.The peritoneal cavities of the infected mice were washed with 10 ml sterile phosphate-buffered saline (PBS), and the peritoneal wash/lavage fluid was collected at different time points after infections. Cells were centrifuged at 275 × g for 5 min to separate the peritoneal exudate cell fractions (PECs). The PECs were directly processed for flow cytometry analysis to determine the frequency of resting neutrophils (CD11b− Ly6G+), activated neutrophils (CD11b+ Ly6G+), and macrophages (CD11b+ Ly6G−).
Histology.Formalin-fixed, paraffin-embedded samples from livers, lungs, and spleens were sectioned and stained with hematoxylin and eosin (H&E). Histologic sections were evaluated qualitatively, and liver lesions were assessed by four parameters, including hepatocyte damage, frequency of lesions, size of inflammatory lesions, and extent of perivascular inflammation. Stained slides were viewed under an Olympus BX51 microscope, and the images were recorded by an RT Slider digital camera (Diagnostic Instruments Inc., Sterling Heights, MI).
Flow cytometry and intracellular cytokine staining/analysis.Splenocytes were harvested, counted, and resuspended in fluorescence-activated cell sorter staining buffer at a concentration of 106 cells/well. FcRs were blocked with a MAb (clone 2.4G2) against the mouse cell surface antigens (Ags) CD16 and CD32 for 15 min. The following fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-, PerCP-Cy5.5-, Alexa Fluor-, and allophycocyanin-conjugated Abs were purchased from BD Biosciences unless indicated otherwise: anti-CD45.2 (clone 69), anti-CD3 (clone 145-2C11), anti-CD11c (clone HL3), anti-F4/80 (clone 6F12), anti-CD4 (clone RM4-4), anti-CD8a (clone 53-6.7), anti-CD11b (clone M1/70), anti-NK1.1 (clone PK136), anti-IL-10, anti-tumor necrosis factor alpha (TNF-α) (clone MP6-XT22), anti-IFN-γ (clone XMG102), anti-CD95 (clone JO2), and anti-Ly6G (clone 1A8). Isotype control MAbs, including FITC-, PE-, or allophycocyanin-conjugated hamster IgG1 (A19-3), rat IgG1 (R3-34), rat IgG2a (R35-95), mouse IgG2a (X39), mouse IgG2b (MPC-11), mouse IgG1 (X40), and rat IgG2b (A95-1), were purchased from BioLegend (San Diego, CA). For intracellular cytokine staining, splenocytes were incubated at 37°C for 4 h in complete medium with the addition of BD Golgi Plug (BD Biosciences, CA) according to the manufacturer's recommendations. Lymphocyte and granulocyte populations were gated based on forward- and side-scatter parameters as well specific staining for CD3 and CD45 surface markers. Approximately 20,000 to 50,000 events were collected using BD-LSR or BD FACSCalibur (BD Immunocytometry Systems, San Jose, CA) flow cytometry, and data were analyzed using FlowJo software (TreeStar, Ashland, OR).
RNA extraction and gene profiling analysis.Total RNA was isolated from livers harvested from infected mice on days 3 and 7 p.i. Liver tissues from naïve mice were used as controls. cDNA was synthesized using an SA Biosciences reverse transcriptase (RT)2 first-strand kit (Qiagen, Valencia, CA). cDNA, along with RT2 SYBR green master mix, was used to characterize expression of ∼420 genes using a different Pathway Finder RT2 Profiler PCR array by following the manufacturer's recommendation (Qiagen, Valencia, CA). Data were analyzed using Applied Biosystems 7900 HT real-time PCR. The array plate contains five housekeeping genes, including GAPDH (glyceraldehyde-3-phosphate dehydrogenase) and β-actin, and one genomic DNA contamination as reference genes and a control. Comparative threshold cycle (CT) values were analyzed using SA Biosciences software, and fold regulations were plotted. Fold up- or downregulations were calculated by dividing the expression fold changes of the candidate genes by the expression fold changes of the reference genes using the CT method. Using cutoff criteria, a 5-fold upregulation or downregulation was considered to be of biologic importance.
Measurements of bacterial burden by real-time PCR.The number of bacteria in Ehrlichia stock and the bacterial burden in different organs (liver, lung, spleen, and peritoneum) were determined at different time points postinfection using an iCycler IQ multicolor real-time detection system (Bio-Rad, Hercules, CA), as previously described (8, 29). The results were normalized to the levels of expression of the GAPDH gene in the same samples and expressed as copy numbers per 104 GAPDH. PCR analyses were considered negative for ehrlichial DNA if the critical threshold values exceeded 40 cycles.
Statistical analyses.The two-tailed t test was used for a comparison of mean values for two experimental groups, and one-way analysis of variance (ANOVA) was used for comparisons of multiple experimental groups. Data were represented by the means and standard deviations (SD) or standard errors (SEM). P values of ≤0.01 were considered highly significant (**), and P values of ≤0.05 were considered moderately significant (*).
RESULTS
Changes in myeloid cell populations following lethal Ehrlichia infection.We recently showed that fatal ehrlichiosis that mimics toxic shock syndrome is associated with an expansion of neutrophils and inflammatory dendritic cells in the spleen during the early stages of infection (30). These changes correlated with tissue injury, expansion of cytotoxic CD8+ T cells that cause immunopathology, and suppression of protective CD4+ Th1 responses (8, 18). In this study, we examined further the sequence of local inflammatory responses in the peritoneum, the initial site of infection, following lethal or nonlethal intraperitoneal (i.p.) challenge of mice with IOE or E. muris, respectively. A prominent but transient infiltration with activated neutrophils (CD11b+ Ly6G+) was observed in the peritoneum of E. muris-/nonlethally infected mice at 4 h but dropped markedly between 1 and 7 days (Fig. 1A). The significant increase in the number of CD11b+ Ly6G− macrophages in E. muris-infected mice lagged behind the neutrophil responses and peaked dramatically on day 7 p.i. (Fig. 1B). In contrast, prominent neutrophil infiltrates were observed at 4 h in the peritoneum of IOE-/lethally infected mice, peaking at 24 h and then dropping between days 3 and 7 p.i. Peak expansion of neutrophils in IOE-infected mice at 4 h was accompanied by peak elevation of macrophages, followed by marked drops in their numbers by 24 h p.i. The number of macrophages remained relatively unchanged from steady state compared to naïve mice until day 7 p.i. Despite early expansion or recruitment of neutrophils to the peritoneum of lethally/IOE-infected mice (Fig. 1A), these mice failed to eliminate ehrlichiae in the peritoneum at 4 h postinfection (Fig. 1C). Further, this was associated with higher bacterial burden in the livers and spleens of IOE-infected mice (Fig. 2A and B) as well as with liver injury (Fig. 2C and D) at later time points postinfection compared to nonlethally infected mice. Consistent with our earlier reports (8, 28, 29), all lethally/IOE-infected mice succumbed to infection on days 8 to 10 p.i. while all nonlethally/E. muris-infected mice survived until day 60 p.i. (data not shown). Together, failure to eliminate ehrlichiae at the initial site of infection in lethally infected mice despite the marked expansion of neutrophils and macrophages suggested that IOE evade the microbicidal functions of these professional phagocytes in lethally infected mice.
Differential expansion of activated neutrophils and macrophages in the peritoneum of nonlethally and lethally infected mice. The peritoneal exudate cells (PECs) were collected from IOE-infected (lethal), E. muris-infected (nonlethal), and naïve mice at 4 h, 24 h, 72 h, and 7 days p.i., as described in Materials and Methods. Cells were analyzed by flow cytometry, and bacterial burden was measured by real-time PCR (RT-PCR). (A and B) Temporal changes in numbers of activated neutrophils and macrophages in lethally and nonlethally infected mice. (C) Data show higher bacterial burden in the peritoneum of lethally infected mice than in nonlethally infected mice. (D) Representative flow cytometry data showing the percentages of CD11b+ Ly6G+ neutrophils and CD11b+ Ly6G− macrophages at 4 h and 24 h p.i. in lethally (IOE) and nonlethally (E. muris) infected mice. The percentages of different cell subsets in naïve mice were similar at 4 and 24 h, and therefore we included one zebra dot for naïve mice at the 4-h time point. Data represent the means and SD of data for 3 mice per group from one representative experiment (*, P ≤ 0.05; **, P ≤ 0.01).
Lethal ehrlichial infection was associated with higher bacterial burden and more-severe hepatic necrosis and apoptosis than nonlethal infection. C57BL/6 mice were infected via the i.p. route with a high dose of IOE (lethal) or E. muris (nonlethal). Livers (A) and spleens (B) were collected on days 1, 3, 5, and 7 p.i. Bacterial burden was determined by RT-PCR, and the copy numbers of IOE/E. muris were normalized to numbers of GADPH. (C and D) Liver pathology in IOE- (C) and E. muris-infected (D) mice. Lethally infected mice develop an extensive but focal necrosis (arrows) and apoptosis of hepatocytes and immune cells in the liver compared to localized granuloma-like formation (arrowhead) in nonlethally infected mice (magnification, ×40). Data are representative of three independent experiments with 9 to 12 mice/group.
Neutrophil depletion enhanced bacterial elimination, decreased pathology, and prolonged mouse survival following lethal Ehrlichia infection.To address whether neutrophils might modulate the immune responses during lethal IOE infection, we examined the effect of neutrophil depletion on host defense against IOE. Anti-Gr1 MAb treatment of IOE-infected mice depleted more than 90% of CD11b+ Ly6G+ neutrophils in peripheral blood during the course of infection until day 8 p.i. (Fig. 3A and B and data not shown) compared to naïve and IOE/sham control mice. A decreased number of neutrophils in anti-GR1-depleted, IOE-infected mice was not due to increased neutrophil recruitment to other sites of infection (e.g., the peritoneum), since our data demonstrate a similar decline of neutrophils in the peritoneum of IOE-infected wild-type mice starting from 24 h postinfection (Fig. 1A and D). Anti-Gr1 MAb treatment did not influence the total number of leukocytes (CD45+ cells) (Fig. 3A) or the number of Ly6G− cells, including monocytes (F4/80+ CD11b+), DCs (CD11c+), NKT (NK1.1+CD3+), NK (NK1.1+CD3−), CD8+ T (CD3+CD8+), and CD4+ T (CD3+CD4+) cells in the peripheral blood or spleens of treated mice (data not shown).
In vivo neutrophil depletion using anti-Gr1 antibody. (A) Flow cytometric analysis of the peripheral blood on day 3 revealed that administration of anti-Gr1 antibody significantly reduced the number of Ly6G+ CD11b+ neutrophils (P = 0.006) in Gr1-depleted mice infected with IOE (IOE/anti-Gr1) compared with that in IOE-infected mice treated with isotype control Ab (IOE/sham control) and naïve mice. Total numbers of F4/80+ CD11b+ macrophage and CD45+ leukocyte/monocyte cells were not significantly influenced upon administration of anti-Gr1 antibody. Mononuclear and granulocytic cells were gated based on forward and side scatter. (B) Dot plots show a reduced percentage of Ly6G+ CD11b+ neutrophils in IOE/anti-Gr1. Data are representative of three independent experiments with 9 mice/group (*, P ≤ 0.05; **, P ≤ 0.01).
Next, we measured bacterial burden in neutrophil-depleted mice and controls during the course of IOE infection. Similar to our previous studies, IOE-infected wild-type mice succumbed to infection on days 8 to 10 p.i. In contrast, a lack of neutrophils resulted in prolonged survival of IOE-infected mice until day 18 p.i. (Fig. 4A). Unexpectedly, neutrophil-depleted mice infected with IOE had significantly lower bacterial burden in the liver than that detected in IOE/sham controls on days 3 and 7 p.i. compared to infected, undepleted wild-type mice (Fig. 4B and C). Compared to day 7 p.i., bacterial burden in different organs from neutrophil-depleted mice did not decrease further on days 11, 13, and 15 p.i. (Fig. 4D), which may account for partial protection of neutrophil-depleted mice. Interestingly, enhanced survival and lower bacterial burden in neutrophil-depleted mice correlated with attenuated pathology, as marked by the presence of fewer scattered apoptotic cells in the liver and no evidence of necrosis (Fig. 4E) compared to naïve mice and infected controls. The latter group developed extensive foci of necrosis and apoptosis in the liver (Fig. 4E). Together, these data suggest that neutrophils support bacterial replication indirectly and promote immunopathology during lethal Ehrlichia infection.
Prolonged survival, enhanced bacterial elimination, and decreased Ehrlichia-induced inflammation and immunopathology in IOE-infected, neutrophil-depleted mice. C57BL/6J mice were infected with IOE via i.p. injection and either depleted of neutrophils (IOE/anti-Gr1) or treated with isotype control (IOE/sham control) as described in Materials and Methods. Tissues from lungs, liver, and spleen were collected, and bacterial burden was determined by real-time PCR. (A) Neutrophil-depleted mice (IOE/anti-Gr1) had prolonged survival compared to that of the IOE/sham control. (B and C) Bacterial burdens in different organs of IOE/anti-Gr1 mice were significantly lower than those detected in the IOE/sham control on days 3 (B) and 7 (C) p.i. (C and D) Neutrophil-depleted mice had similar bacterial burdens in different organs on days 11, 13, and 15 p.i. (D) compared to that detected on day 7 p.i. (C) in neutrophil-depleted mice. (E) H&E staining of liver sections shows necrotic foci (arrows) in IOE/sham control mice but not in IOE/anti-Gr1-mice. Magnification in upper panel, ×20; magnification in lower panel, ×40. Data represent the means and SEM of three mice/group and are representative of three independent experiments with similar results (*, P ≤ 0.05; **, P ≤ 0.01).
Neutrophil depletion enhanced resistance to infection in mice infected with IOE, which was not dependent on NK cells and NKT cells.Our previous study indicated that NKT cells play a dual role during lethal Ehrlichia infection. While IFN-γ production by NKT cells is essential for effective intracellular bacterial clearance, they also promote infection-induced toxic shock syndrome (27, 30, 31). In contrast, NK cells play a detrimental role during fatal ehrlichiosis as they inhibit protective anti-Ehrlichia immunity and mediate tissue injury during fatal disease (29, 31, 32). Although the exact mechanism by which NK and NKT cells contribute to disease severity during lethal Ehrlichia infection is unknown, our previous studies suggested that this may be mediated via production of TNF-α and IL-10. We thus examined the effect of neutrophils on the NK and NKT responses during lethal IOE infection by flow cytometry. The number of NK cells and their cytokine production, including secretion of IFN-γ, TNF-α, and IL-10, were not significantly influenced in IOE-infected, neutrophil-depleted mice compared to infected sham controls (Fig. 5A and C). In contrast, neutrophil depletion resulted in a significant (P = 0.05) decrease in the percentage and absolute number of NKT cells (Fig. 5B and C) compared with those for IOE-infected sham controls. Within the NKT cell populations in neutrophil-depleted mice, there was no significant difference in the number of NKT cells producing IFN-γ (P = 0.06) (Fig. 5B). However, we detected a significant decrease in the number of NKT cells producing TNF-α (P = 0.05) and IL-10 (P = 0.0007) (Fig. 5B). Previous studies have shown that systemic and local overproduction of TNF-α and IL-10 is associated with disease severity and immunopathology (8, 27–29). Although the absolute number of IFN-γ-producing NKT cells was decreased in neutrophil-depleted mice, the ratio of IFN-γ- to IL-10-producing NKT cells was substantially higher in these mice than in IOE/sham mice controls (21.5 versus 3.62) (Fig. 5B). Collectively, these data suggested that neutrophils contribute to increased frequency of NKT cells producing TNF-α and IL-10. Thus, neutrophil-mediated induction of TNF-α- and IL-10-producing NKT cells may potentially contribute to the development of inflammation and pathological responses in a murine model of fatal ehrlichiosis.
Neutrophil depletion influenced the frequency and functions of NKT, but not NK, cells during lethal Ehrlichia infection. Splenocytes from infected and naïve mice were harvested on day 7 p.i. and analyzed by flow cytometry. Lymphocytes were gated on the forward and side scatter. (A) Data show no significant changes in the numbers of cytokines produced by NK cells between the IOE/anti-Gr1 mice and the IOE/sham control. (B) Data show a reduced number of NKT cells (CD3+ NK1.1+; P = 0.05) and their intracellular cytokines TNF-α (P = 0.05) and IL-10 (P = 0.007), but not IFN-γ, in IOE/anti-Gr1 mice compared with those for the IOE/sham control. (C) Dot plots show a reduced percentage of NKT cells in IOE/anti-Gr1 mice compared with those in IOE/sham control and naïve mice (*, P ≤ 0.05; **, P ≤ 0.01).
Neutrophils mediate expansion of pathogenic TNF-α-producing CD8+ T cells during fatal Ehrlichia infection.We have shown before that CD4+ Th1 cells are critical for host defense and elimination of intracellular ehrlichiae. However, CD4+ T cell proliferation is suppressed early during lethal ehrlichial infection and followed by apoptosis of these cells at late stages of infection (8, 28, 29). On the other hand, we previously demonstrated that Ehrlichia-induced toxic shock and tissue injury are mediated by CD8+ T cells (8, 17, 18). Thus, we examined here the contribution of neutrophils to CD4+ and CD8+ T cell responses during lethal infection. Notably, compared with infected sham control mice, neutrophil depletion in infected mice significantly reduced the total numbers of CD8+ T cells (P < 0.02) (Fig. 6A and B) as well as their production of IL-10 (P = 0.001) and TNF-α (P = 0.002) cytokines on day 7 p.i. (Fig. 6A and C). Production of TNF-α by CD8+ T cells is known to be a marker of pathogenic functions of CD8+ T cells during Ehrlichia-induced toxic shock (8, 17, 18). Neutrophil depletion did not significantly influence the total number of CD4+ T cells, IFN-γ production by Th1 cells, or expression of an apoptotic receptor (i.e., CD95/Fas) on CD4+ T cells in the spleen (Fig. 6D and E). However, absence of neutrophils restored the normal ratio of 2:1 of CD4 to CD8 T cells, similar to the ratio of CD4 to CD8 T cells in nonlethally infected mice, which suggests that a lack of neutrophils restores protective immunity mediated by CD4+ T cells in the absence of a pathogenic CD8+ T cell response (Fig. 6A and D). Taken all together, these data suggest that although neutrophils did not directly influence the frequency of protective CD4+ Th1 responses, they contribute to expansion of pathogenic CD8+ T cells and an inverted CD4/CD8 ratio, which in turn lead to suboptimal protective immunity mediated by CD4+ Th1 cells during lethal Ehrlichia-induced toxic shock.
Neutrophils promoted the induction of pathogenic TNF-α-producing CD8+ T cells during lethal Ehrlichia infection but did not influence CD4+ T cell responses. Splenocytes were harvested from different groups of mice and stimulated in vitro with IOE antigens. The frequency of cytokine production by CD8+ and CD4+ T cells was analyzed using flow cytometry. (A) Data show a significantly reduced total number of antigen-specific CD8+ T cells (P = 0.02) producing IL-10 (P = 0.001) and TNF-α (P = 0.002), but not IFN-γ, in IOE/anti-Gr1 mice compared with those in IOE/sham control and naïve mice. (B and C) Dot plots show a reduced percentage of CD8+ T cells with cytokine IL-10 in IOE/anti-Gr1 mice compared to those in IOE/sham control and naïve mice. (D and E) Data show no significant differences in the total numbers and expression of Fas (CD95) on CD4+ T cells or their production of IFN-γ, as measured by flow cytometry between neutrophil-depleted and undepleted mice. Data are representative of three experiments (*, P ≤ 0.05; **, P ≤ 0.01).
Neutrophil depletion abrogated cytokine and chemokine gene expression during early lethal Ehrlichia infection.Studies by us and other investigators indicated that severe tissue damage and multiorgan failure following lethal Ehrlichia infection or infections with other pathogens are associated with local and systemic overproduction of several cytokines and chemokines (8, 33). To test the hypothesis that neutrophils contribute to induction of pathogenic immune responses during lethal ehrlichiosis via cytokine and chemokine production, we examine the changes in transcriptional responses in the livers during early IOE infection in neutrophil-depleted and IOE/sham controls on days 3 and 7 p.i. We chose 5-fold regulations compared to gene expression in naïve mice as cutoffs to identify several differentially upregulated or downregulated cytokine and chemokine genes in the two groups of infected mice. Overall, the extent of fold changes of gene transcripts in the neutrophil-depleted, IOE-infected mice was dramatically different than that detected in IOE-infected control mice. In particular, we found that neutrophil depletion significantly downregulated several genes of proinflammatory cytokines and their receptors on days 3 and 7 p.i., including IL-1α (P = 0.05), IL-1β (P = 0.003), and TNF (P = 0.05), as well as IL-6 (P = 0.005) only on day 7 p.i. (Fig. 7A and B). Consistent with flow cytometry data, neutrophil depletion decreased expression of IL-10 (P = 0.00001) (Fig. 7C). Interestingly, neutrophil depletion increased expression of IL-1 antagonist (IL-1ra) (also called IL-1rn) on day 3 p.i. (Fig. 7C) but decreased IL-1ra expression on day 7 p.i. (Fig. 7D). IL-1ra is an anti-inflammatory molecule, as it competes with IL-1β for binding to IL-1 receptor and thus inhibits IL-1β function. The expression level of IFN-γ was not significantly (P = 0.2) different from that detected in IOE-infected wild-type controls (Fig. 7D). However, the ratio of IFN-γ to IL-10 mRNA expression in the livers of neutrophil-depleted mice was higher than that in the livers of IOE/sham controls on days 3 (8.7 versus 1.1, respectively) and 7 (24.7 versus 2.4, respectively) p.i. (Fig. 7C and D), a finding which is consistent with the above-mentioned results showing an insignificant effect of neutrophil depletion on IFN-γ production by NKT and CD4+ T cells (Fig. 5B and 6D). Surprisingly, neutrophil depletion significantly decreased expression of type I interferon (IFN-β) (P = 0.01) on day 7 but not day 3 p.i. (Fig. 7D). IFN-β is shown to stimulate production of several chemokines and expression of many chemokine receptors that support migration of monocytes, NK, and T cells to sites of infection (34). Indeed, decreased type I interferon in neutrophil-depleted mice correlated with significant downregulation of several chemokines and their receptors (Table 1), in particular, CCL2, CCL7, CCL8, and CCL12 on day 3 p.i. as well as CCR1, CCR7, and CXCL10 at day 3 and day 7 p.i. (Fig. 8). Finally, neutrophil depletion decreased expression of several T cell costimulatory molecules, including CD80, CD86, CD40, and CD40Ig (Table 2), suggesting that neutrophils may contribute to T cell activation during fatal ehrlichiosis. Taken all together, these data suggest that neutrophils contribute to pathogenesis of fatal ehrlichiosis via several mechanisms, such as enhancing the dysregulated overproduction of pro- and anti-inflammatory cytokines and chemokines, which in turn lead to uncontrolled activation of T cells and innate cells and their migration to sites of infection where they cause tissue damage, multiorgan dysfunction or failure, and, finally, Ehrlichia-induced shock.
Neutrophil depletion reduced expression of proinflammatory cytokines, anti-inflammatory cytokines, and Th1 cytokines in IOE/anti-Gr1 mice compared with that in the IOE/sham control. Expression levels of several proinflammatory cytokines, IL-1α, IL-1β, IL-6, and TNF-α (A and B), anti-inflammatory cytokines, such as IL-10 and IL-1ra, Th1 cytokines, such as IFN-γ, and type I interferon (C and D) were measured in the livers of IOE-infected/neutrophil-depleted and IOE-infected/sham controls on days 3 and 7 p.i. Data shown are the means and SEM of 3 mice/group and are representative of two independent experiments (*, P ≤ 0.05; **, P ≤ 0.01).
Induction of genes associated with inflammatory chemokines and receptors
Neutrophil depletion reduced expression of chemokines and their receptors in IOE/anti-Gr1 mice compared to that of IOE/sham control mice. The expression levels of several chemokines (A and B) and their receptors (C and D) in IOE-infected/anti-Gr1-depleted and IOE/sham controls were measured on days 3 and 7 p.i. Neutrophil depletion in IOE-infected/anti-Gr1-depleted mice reduced the expression levels of Ccl2, Ccl7, Ccl12, and Cxcl10 as well as the chemokine receptors Ccr1, Ccr2, and Ccr7 compared to those of the IOE/sham controls. Data represent the means (and SEM) of 2 to 3 RNA samples (*, P ≤ 0.05; **, P ≤ 0.01).
Induction of genes associated with T cell activation
DISCUSSION
Inflammation constitutes a major host response in many microbial infections. Invasive infections by pathogenic microbes can result in an uncontrolled hyperinflammatory response, leading to severe host damage and sepsis. Neutrophils constitute a hallmark of protective innate immunity in bacterial infections but at the same time have been notoriously known for their inflammation and sepsis-promoting effects (23–25, 35–37). Our previous studies attributed severity of HME that mimics toxic shock syndrome to immunopathology. In this study, we showed that neutrophils contribute to immunopathology and substantial cytokine and chemokine overproduction. Unexpectedly, we found that neutrophils indirectly impaired bacterial clearance and enhanced pathogenic response marked by expansion of TNF-α-producing CD8+ T cells, which has been shown to mediate fatal ehrlichiosis.
Although the exact mechanisms by which neutrophils contribute to bacterial clearance are not directly examined, several possible mechanisms may account for the consequences or responses. First, although the primary target of Ehrlichia is macrophages, it is possible that early recruitment of neutrophils to the initial site of infection protects these obligate intracellular bacteria from being degraded or killed in the extracellular environment by complement or other immune factors (38). Neutrophils may also function as a “Trojan horse” by which uptake of infected neutrophils by macrophages may be a mechanism of silencing of macrophage microbicidal functions during severe Ehrlichia infection. Although our data do not directly support the Trojan horse mechanism, the finding that IOE infections resulted in higher bacterial burden in the peritoneum despite early expansion of neutrophils and macrophages compared to E. muris-infected mice suggests that IOE, but not E. muris, evade microbicidal functions of these phagocytic cells. Although E. muris and IOE are genetically and antigenically related, recent ultrastructure analysis suggests that IOE may differentially express specific subsets of genes that are essential for survival in phagocytic cells, similar to what has been shown in related pathogens (39–42). Nevertheless, direct interactions between these two Ehrlichia species and neutrophils and the impact of this interaction on macrophage microbicidal functions will be examined in future studies.
Our previous studies have shown that a lack of CD8+ T cells protected mice from lethal infection, enhanced bacterial elimination, and restored the number of CD4+ Th1 cells. In addition, adoptive transfer of cytotoxic CD8+ T cells from lethally infected wild-type mice into nonlethally infected mice caused tissue injury and inhibited bacterial clearance (8, 30). Thus, since neutrophil depletion reduced the number of TNF-α- and IL-10-producing CD8+ T cells in lethally infected mice (Fig. 6A, B, and C), it is conceivable that neutrophils indirectly inhibit bacterial elimination via supporting the induction and/or expansion of pathogenic CD8+ T cells during Ehrlichia-induced shock.
How neutrophils induce CD8+ T cells in ehrlichiosis is unclear. In vivo and in vitro studies have shown that under inflammatory conditions, neutrophils act as antigen-presenting cells (APC) as they migrate to lymph nodes, express major histocompatibility complex (MHC) class II and costimulatory molecules, and thus are able to cross-present phagocytosed antigens to CD8+ T cells (11, 13, 22, 25, 43, 44). Interestingly, in these lymph nodes, neutrophils were found to impede DC functions and optimal activation of CD4+ cells (14, 44), which implies that neutrophil depletion should enhance DC antigen presentation, leading to activation and expansion of CD4+ T cells (39). Our data showed that neutrophil depletion slightly increases the frequency of CD4+ Th1 cells compared with the IOE/sham control, but it is not known whether this effect is due to decreased CD8+ T cell responses or enhanced activation and differentiation of CD4+ Th1 cells.
The finding that neutrophil depletion enhanced bacterial elimination despite minimal changes in IFN-γ production at the single-cell level or at the transcription level is paradoxical since IFN-γ is known to be critical for host defense against intracellular bacteria. We argue that a protective anti-Ehrlichia effect is determined by two interrelated factors: total IFN-γ concentration and the ratio of IFN-γ to IL-10. It is well established that IL-10 impairs the stimulatory effect of IFN-γ on phagocytic cells, which influence intracellular bacterial elimination. IL-10 also inhibits the differentiation of protective CD4+ Th1 cells, which are critical for protection against intracellular bacteria (27, 28, 40). Thus, it is possible that the phenotype of immune responses against Ehrlichia is expected to be determined by the IFN-γ/IL-10 ratio and not just the absolute level of each cytokine. IL-10 expression is significantly decreased in IOE-infected, neutrophil-depleted mice (Fig. 7C and D and Table 1). The ratio of IFN-γ to IL-10 mRNA expression in the livers of neutrophil-depleted mice was higher than that in the livers of IOE/sham controls on day 3 (8.67 versus 1.03, respectively) and on day 7 (24.7 versus 2.44, respectively) p.i. Consistent with changes observed in IL-10 expression in the liver of infected mice, the ratio of IFN-γ- to IL-10-producing NKT cells (21.5 versus 3.62) is higher in IOE-infected, neutrophil-depleted mice than in IOE-infected/sham controls. In conclusion, our data showed an enhanced IFN-γ/IL-10 ratio, which is essential for effective clearance of intracellular ehrlichiae.
Our previous studies also showed a correlation between production of IL-10 and development of pathogenic CD8+ T cells which in turn causes apoptosis of CD4+ T cells and tissue injury. We showed in this study that neutrophil depletion not only resulted in decreased IL-10 production (at a single-cell level and also in total mRNA) but restored the normal ratio of CD4/CD8 (2:1) compared to sham-infected controls. Interestingly, an inverted CD4/CD8 ratio has also been observed in patients with fatal HME who develop toxic shock. Nevertheless, it is unclear whether IL-10 contributes directly or indirectly to the induction of pathogenic CD8+ T cells. We envisage that IL-10 promotes bacterial replication and host cell apoptosis and thus supports cross-presentation of Ehrlichia antigens within apoptotic cells to CD8+ T cells.
Furthermore, we previously showed that tissue injury and ineffective bacterial elimination during fatal Ehrlichia infection are also mediated by cytotoxic NK cells producing IFN-γ, TNF-α, and IL-10; on the other hand, NKT cells may also contribute to immunopathology via production of TNF-α and IL-10. In this study, we demonstrated that neutrophil depletion significantly decreased numbers of NKT, but not NK, cells producing TNF-α and IL-10 compared to the IOE-infected sham control (Fig. 5). The differential effect of neutrophil depletion on NKT, but not NK, cell responses in the spleen cannot exclude the possibility that neutrophil depletion impairs the migration to and/or expansion of NK cells at the peripheral sites of infection, such as liver and lung. Increased migration of NK cells to the liver and their expansion were correlated with tissue injury and vice versa. In support of this possibility, we have found that a lack of neutrophils in depleted mice significantly decreased production of several cytokines and chemokines, including NK cell chemokines (Fig. 7 and 8 and Tables 1 and 2). Our data demonstrated that neutrophil depletion substantially reduced the expression of several C-C chemokines (mainly CCL2 [MCP-1], CCL3 [MIP-1α], CCL5 [RANTES], CCL7 [MCP3], CCL8 [MCP2], and CCL12 [MCP5]), CXC chemokines (mainly CXCL9 [Mig], CXCL10 [IP-10], and CXCL11 [I-TAC]), and their receptors (mainly CCR1, CCR3, and CCR7) that are essential for monocyte, neutrophil, NK cell, and CD8+ T cell migration to sites of infection (Fig. 8 and Table 2). CCR1 is a receptor for a number of C-C chemokines, including MIP-1α, MCP-3, and RANTES, and is expressed primarily on CD34+ cells, including lymphocytes and monocytes, but not on granulocytes. CCR1 binds to several MCP chemokines (MCP-3, MCP-4, and MCP-5) and is expressed primarily on epithelial cells, basophils, and eosinophils as well as on Th1 and Th2 cells. This receptor has been shown to be highly expressed during inflammatory allergic and parasitic infections as it enhances migration of lymphocytes and eosinophils. CCR7 binds to CCL19 and CCL21 and is expressed primarily on DCs and lymphocytes. CCR7 allows both tissue-resident and monocyte-derived DCs in peripheral sites to enter afferent lymphatics to present antigens to T cells (41). Thus, decreased expression of CCR7 may also explain defective induction of CD8+ T cells in infected neutrophil-depleted mice. Interestingly, among all chemokine receptors, CXCR2 (IL-8 receptor) was elevated in neutrophil-depleted mice. Studies showed that CXCR2 signaling regulates neutrophil release from the bone marrow. Absence of CXCR2 resulted in abnormal retention of neutrophils in the bone marrow (45, 46). Thus, overexpression of CXCR2 is most likely a compensatory mechanism due to neutrophil depletion (Fig. 8C).
Similar to sepsis, studies by us and other investigators have demonstrated a strong link between Ehrlichia-induced shock and development of cytokine storm. These studies highlight neutrophils as a major factor contributing to this immunopathogenic mechanism. Neutrophils, either spontaneously or following appropriate stimulation, produce numerous proinflammatory cytokines, anti-inflammatory cytokines, and chemokines (11). Our gene array analysis demonstrated that neutrophil depletion reduced the expression of several proinflammatory cytokines, including IL-1α, IL-1β, IL-6, TNF-α, and IFN-γ, and anti-inflammatory cytokines, such as IL-ra (IL-rn), IL-1r2, and IL-10 (Fig. 7 and Table 1). Among proinflammatory cytokines, IL-1β was found to be deleterious in melioidosis, and the detrimental role of IL-1β during melioidosis was due, in part, to excessive recruitment of neutrophils to the infection site (47). On the other hand, IL-1ra and IL-1r2 are important mechanisms that prevent IL-1β-mediated inflammation. Restoring IL-1r2 expression in IL-1r2-deficient cells resulted in significant cellular protection from IL-1β-mediated damage (48–51). Further, treatment of mice with IL-1ra reduced inflammation in collagen-induced arthritis (52). Consistent with these studies, we observed decreased IL-1β expression and elevated IL-1r2 expression at early stages of infection (day 3 p.i.) in lethally infected neutrophil-depleted mice (Fig. 7A and C and Table 1). This profile was similar to the expression profile of these genes in nonlethally infected mice that survived E. muris infection (P. Chattoraj, Q. Yang, A. Khandai, O. Al-Hendy, and N. Ismail, unpublished data). In contrast, lethal infection in IOE-infected WT mice was associated with reversed kinetics with increased Il-1r2 and IL-1ra only at late stages of infection, suggesting proactive but not productive immune responses to excess inflammation in this model.
In summary, our report describes enhancement of pathogenic T cell-mediated immune responses by neutrophils in vivo during lethal infection, partly via production of cytokines and chemokines. Further, our study suggests that neutrophils may function as a Trojan horse in our murine model of fatal ehrlichiosis, which is an intriguing area of further investigation. Modulation of neutrophil interaction with other immune cells or recruitment may thus be a potential novel immunotherapeutic strategy in the management of not only monocytic ehrlichiosis but also other intracellular bacterial infections that cause septic or toxic shock.
ACKNOWLEDGMENTS
We thank Partho Chattoraj and Ankita Khandia for their help with measurement of bacterial burden and flow cytometry analysis. We also thank the Genomics and Proteomics Core Laboratories (GPCL), University of Pittsburgh, for their help with an mRNA array.
This work is supported by a National Institutes of Health grant (NIAID-R56AI097679-01A to N.I.).
FOOTNOTES
- Received 18 December 2012.
- Returned for modification 7 January 2013.
- Accepted 1 March 2013.
- Accepted manuscript posted online 11 March 2013.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.