Advertisement

Contribution of NK Cells to the Innate Phase of Host Protection Against an Intracellular Bacterium Targeting Systemic Endothelium

  • Rong Fang
    Affiliations
    Department of Pathology, University of Texas Medical Branch, Galveston, Texas
    Search for articles by this author
  • Nahed Ismail
    Affiliations
    Department of Pathology, University of Pittsburgh, Pittsburgh, Pennsylvania
    Search for articles by this author
  • David H. Walker
    Correspondence
    Address reprint requests to David H. Walker, M.D., Department of Pathology, Center for Biodefense and Emerging Infectious Diseases, University of Texas Medical Branch, 301 University Blvd., Galveston, TX 77555-0609
    Affiliations
    Department of Pathology, University of Texas Medical Branch, Galveston, Texas

    Center for Biodefense and Emerging Infectious Diseases, University of Texas Medical Branch, Galveston, Texas
    Search for articles by this author
Open AccessPublished:May 21, 2012DOI:https://doi.org/10.1016/j.ajpath.2012.03.020
      We investigated the mechanisms by which natural killer (NK) cells mediate innate host defense against infection with an endothelium-targeting intracellular bacterium, Rickettsia. We found that a robust Rickettsia-induced innate response in resistant mice cleared the bacteria early in the infection and was associated with significantly higher frequencies of splenic interferon (IFN)-γ (+) CD8+ T cells and cytotoxic NK cells compared with susceptible mice. More importantly, NK cell-deficient Rag−/−γc−/− animals displayed significantly increased susceptibility to Rickettsia infection compared with NK cell-sufficient Rag−/− mice, as evidenced by impaired bacterial clearance, early development of severe thrombosis in the liver, and a decreased serum level of IFN-γ. Furthermore, the lack of NK cells also impaired host resistance of CB-17 scid mice to Rickettsia, similar to what was observed in Rag−/−γc−/− mice. Interestingly, perforin deficiency in Rag−/−Prf1−/− mice resulted in greater thrombosis and insignificantly different systemic levels of IFN-γ compared with Rag−/− mice, suggesting that perforin, which is mainly produced by NK cells, is involved in the prevention of vascular damage. Together, these findings reveal that NK cells mediate the innate phase of host protection against infection with rickettsiae, most likely via IFN-γ production. Furthermore, NK cells are involved in preventing rickettsial infection-induced endothelial cell damage, possibly via perforin production.
      Rickettsiae are obligately intracellular α-proteobacteria that primarily target the microvascular endothelium.
      • Walker D.H.
      Rocky Mountain spotted fever: a seasonal alert.
      • Woods M.
      • Olano J.
      Host defenses to Rickettsia rickettsii infection contribute to increased microvascular permeability in human cerebral endothelial cells.
      The main pathogenic mechanism involved in rickettsial disease is increased systemic microvascular permeability leading to edema, hypovolemia, and hypotension. The severity of rickettsial infection in both human and animal models is dependent on bacterial virulence, host factors, and bacterial dose.
      • Walker D.H.
      • Hawkins H.K.
      • Hudson P.
      Fulminant Rocky Mountain spotted fever Its pathologic characteristics associated with glucose-6-phosphate dehydrogenase deficiency.
      • Walker D.H.
      The role of host factors in the severity of spotted fever and typhus rickettsioses.
      • Walker D.H.
      • Popov V.L.
      • Wen J.
      • Feng H.M.
      Rickettsia conorii infection of C3H/HeN mice A model of endothelial-target rickettsiosis.
      • Feng H.M.
      • Walker D.H.
      Cross-protection between distantly related spotted fever group rickettsiae.
      Rocky Mountain spotted fever caused by Rickettsia rickettsii and Mediterranean spotted fever caused by R. conorii are considered to be important due to their wide geographic distribution and a potentially fatal outcome in severe cases.
      • Weinberger M.
      • Keysary A.
      • Sandbank J.
      • Zaidenstein R.
      • Itzhaki A.
      • Strenger C.
      • Leitner M.
      • Paddock C.D.
      • Eremeeva M.E.
      Fatal Rickettsia conorii subsp. israelensis infection.
      • Aliaga L.
      • Sánchez-Blázquez P.
      • Rodríguez-Granger J.
      • Sampedro A.
      • Orozco M.
      • Pastor J.
      Mediterranean spotted fever with encephalitis.
      • Papa A.
      • Dalla V.
      • Petala A.
      • Maltezou H.C.
      • Maltezos E.
      Fatal Mediterranean spotted fever in Greece.
      • de Sousa R.
      • Nobrega S.D.
      • Bacellar F.
      • Torgal J.
      Mediterranean spotted fever in Portugal risk factors for fatal outcome in 105 hospitalized patients.
      On the other hand, other rickettsial diseases, such as African tick bite fever caused by R. africae, present as a mild disease. In addition, the severity of rickettsial diseases is in large part determined by host factors.
      • Walker D.H.
      • Hawkins H.K.
      • Hudson P.
      Fulminant Rocky Mountain spotted fever Its pathologic characteristics associated with glucose-6-phosphate dehydrogenase deficiency.
      • Walker D.H.
      The role of host factors in the severity of spotted fever and typhus rickettsioses.
      Fulminant Rocky Mountain spotted fever often occurs in African-American males with glucose-6-phosphate dehydrogenase deficiency associated with an overwhelming bacterial load, extensive endothelial damage, and thrombosis.
      • Walker D.H.
      • Hawkins H.K.
      • Hudson P.
      Fulminant Rocky Mountain spotted fever Its pathologic characteristics associated with glucose-6-phosphate dehydrogenase deficiency.
      • Walker D.H.
      The role of host factors in the severity of spotted fever and typhus rickettsioses.
      Furthermore, we have previously established several murine models of spotted fever rickettsiosis with different mouse strains and different bacterial inocula
      • Walker D.H.
      • Popov V.L.
      • Wen J.
      • Feng H.M.
      Rickettsia conorii infection of C3H/HeN mice A model of endothelial-target rickettsiosis.
      • Feng H.M.
      • Wen J.
      • Walker D.H.
      Rickettsia australis infection: a murine model of a highly invasive vasculopathic rickettsiosis.
      • Walker D.H.
      • Popov V.L.
      • Feng H.M.
      Establishment of a novel endothelial target mouse model of a typhus group rickettsiosis: evidence for critical roles for gamma interferon and CD8 T lymphocytes.
      . For example, C3H mice are genetically susceptible to a high, but not a low dose, of R. conorii, whereas C57BL/6 (B6) mice are highly resistant to both inocula.
      • Fang R.
      • Ismail N.
      • Shelite T.
      • Walker D.H.
      Differential interaction of dendritic cells with Rickettsia conorii: impact on host susceptibility to murine spotted fever rickettsiosis.
      Using these murine models of spotted fever rickettsiosis, we found that protective adaptive immunity during primary infection correlates with induction of strong cell-mediated immunity, including effector CD8+ CTLs, Th1 cells, and production of inflammatory cytokines, such as interferon (IFN)-γ and tumor necrosis factor-α.
      • Feng H.M.
      • Walker D.H.
      Interferon-gamma and tumor necrosis factor-alpha exert their antirickettsial effect via induction of synthesis of nitric oxide.
      • Feng H.M.
      • Popov V.L.
      • Walker D.H.
      Depletion of gamma interferon and tumor necrosis factor alpha in mice with Rickettsia conorii-infected endothelium: impairment of rickettsicidal nitric oxide production resulting in fatal, overwhelming rickettsial disease.
      • Feng H.
      • Popov V.L.
      • Yuoh G.
      • Walker D.H.
      Role of T lymphocyte subsets in immunity to spotted fever group Rickettsiae.
      • Valbuena G.
      • Feng H.M.
      • Walker D.H.
      Mechanisms of immunity against rickettsiae New perspectives and opportunities offered by unusual intracellular parasites.
      The mechanisms involved in the innate phase of host responses against rickettsial infection in resistant and susceptible murine hosts, however, remain ill-defined.
      Natural killer (NK) cells are essential effectors of the innate immune system against infections as they mediate elimination of a variety of pathogens through secretion of IFN-γ and perforin/granzyme-mediated killing.
      • Korbel D.S.
      • Finney O.C.
      • Riley E.M.
      Natural killer cells and innate immunity to protozoan pathogens.
      • Bancroft G.J.
      The role of natural killer cells in innate resistance to infection.
      • Scharton-Kersten T.M.
      • Sher A.
      Role of natural killer cells in innate resistance to protozoan infections.
      • Lodoen M.B.
      • Lanier L.L.
      Natural killer cells as an initial defense against pathogens.
      • Paust S.
      • Senman B.
      • von Andrian U.H.
      Adaptive immune responses mediated by natural killer cells.
      • French A.R.
      • Yokoyama W.M.
      Natural killer cells and viral infections.
      There is increasing evidence, however, that cytotoxic granules produced by NK or CD8+ T cells are involved in development of immunopathology after infections with certain pathogens, such as lymphocytic choriomeningitis virus and Epstein-Barr virus.
      • de Saint Basile G.F.A.
      The role of cytotoxicity in lymphocyte homeostasis.
      • Matloubian M.
      • Suresh M.
      • Glass A.
      • Galvan M.
      • Chow K.
      • Whitmire J.K.
      • Walsh C.M.
      • Clark W.R.
      • Ahmed R.
      A role for perforin in downregulating T-cell responses during chronic viral infection.
      • Jordan M.B.
      • Hildeman D.
      • Kappler J.
      • Marrack P.
      An animal model of hemophagocytic lymphohistiocytosis (HLH): CD8+ T cells and interferon gamma are essential for the disorder.
      Other studies suggest that infection-induced immunopathology can be restricted by perforin-dependent negative regulation of cytotoxic T lymphocytes responses.
      • Matloubian M.
      • Suresh M.
      • Glass A.
      • Galvan M.
      • Chow K.
      • Whitmire J.K.
      • Walsh C.M.
      • Clark W.R.
      • Ahmed R.
      A role for perforin in downregulating T-cell responses during chronic viral infection.
      • Kägi D.
      • Odermatt B.
      • Mak T.W.
      Homeostatic regulation of CD8+ T cells by bperforin.
      • Badovinac V.P.
      • Hamilton S.E.
      • Harty J.T.
      Viral infection results in massive CD8+ T cell expansion and mortality in vaccinated perforin-deficient mice.
      In the present work, we used various mouse strains with different susceptibilities to infection with R. conorii, and we studied the contributions of NK and T cells to host immunity, particularly during the early phase. We found that rapid bacterial clearance at the early phase of infection in the resistant host was associated with an increased production level of IFN-γ by CD8 T cells, and enhanced levels of activation and cytotoxic activity of NK cells. Further genetic manipulation of resistant hosts with a high rickettsial challenge dose proved our hypothesis that NK cells contribute greatly to the early phase of host protection, independent of acquired T-cell responses, through effective bacterial elimination, as well as preventing infection-induced pathology.

      Materials and Methods

      Rickettsia conorii and Plaque Assay

      Rickettsia conorii (Malish 7 strain) was obtained from the ATCC (VR 613; Manassas, VA). For animal inoculation, rickettsiae were cultivated in specific pathogen-free embryonated chicken eggs. After homogenization, rickettsiae were diluted in a 10% suspension of sucrose-phosphate-glutamate buffer (0.218 mmol/L sucrose, 3.8 mmol/L KH2PO4, 7.2 mmol/L K2HPO4, 4.9 mmol/L monosodium glutamic acid, pH 7.0). The concentration of rickettsiae from yolk sac was determined by plaque assay and quantitative real-time PCR, described as follows. The rickettsial stock was stored at −80°C until used. Plaque assay for testing the quantity of viable rickettsiae in the infected tissue was performed as previously described.
      • Walker D.H.
      • Popov V.L.
      • Wen J.
      • Feng H.M.
      Rickettsia conorii infection of C3H/HeN mice A model of endothelial-target rickettsiosis.

      Mice and Rickettsial Infection

      Wild-type (WT) female C3H/HeN mice, NK cell-deficient-scid mice on CB-17 background and scid mice on CB-17 background were purchased from Harlan Laboratories (Indianapolis, IN) and used at 6 to 10 weeks of age. Age- and sex-matched WT CB-17 mice, B6 mice, and T-cell- and B-cell-deficient Rag−/− mice, NK cell-deficient-Rag−/− mice (Rag−/−γc−/−), and perforin-deficient- Rag−/− mice (Rag−/−Prf1−/−) were purchased from Jackson Laboratories (Bar Harbor, Maine) and Taconic Farms Inc. (Hudson, NY). Mice were housed in a biosafety level 3 facility at the University of Texas Medical Branch, Galveston, TX. All experiments and procedures were approved by the University of Texas Medical Branch Animal Care and Use Committee, and mice were used according to the guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Different mouse strains were infected intravenously with R. conorii at different doses as follows: WT C3H and B6 mice were inoculated with a low dose (3 × 104 plaque forming units) and a high dose (3 × 105 plaque forming units); Rag−/−, NK cell-depleted Rag−/−, Rag−/−γc−/−, and Rag−/−Prf1−/− mice were inoculated with a high dose (3 × 105 plaque forming units). Negative control mice were inoculated with sucrose-phosphate-glutamate buffer or 10% uninfected yolk sac processed in the same way as infected yolk sac, as previously described. Mice were monitored daily for signs of illness.

      NK Cell Depletion

      For NK cell depletion, a nonactivating polyclonal antibody against asialo-GM1 (Wako Chemicals, Inc., Richmond, VA) was used as previously described.
      • Kang S.J.
      • Liang H.E.
      • Reizis B.
      • Locksley R.M.
      Regulation of hierarchical clustering and activation of innate immune cells by dendritic cells.
      Rag−/− mice on B6 background were inoculated i.v. with 50 μL of 1:80 dilution of anti-asialo-GM1 antibody on days 0, 2, and 5 of infection. Depletion of NK cells was approximately 93% efficient, as determined by flow cytometric analysis of the number of DX5 (+) CD3 (−) NK cells in the spleen of depleted mice compared to the sham-depleted mice.

      Measurement of Tissue Rickettsial Loads by Real-Time PCR

      Rickettsial burdens in the livers of infected mice were determined using an iCycler IQ from BioRad (Hercules, CA). Primers (Sigma-Genosys, St. Louis, MO) and probes (Biosearch Technologies, Novato, CA) targeting R. conorii ompB and mouse GAPDH genes were used as previously described.
      • Valbuena G.
      • Bradford W.
      • Walker D.H.
      Expression analysis of the T-cell-targeting chemokines CXCL9 and CXCL10 in mice and humans with endothelial infections caused by rickettsiae of the spotted fever group.
      The results were normalized to and expressed as ompB copy number per 106 copies of GAPDH.

      NK Cytotoxicity Assay

      For NK cell cytotoxicity assays, splenocytes were isolated from infected and uninfected C3H and B6 mice. Target YAC-1 cells (ATCC, Manassas, VA), mouse lymphoma cells which are the optimal target for mouse NK cells, were stained with 3, 3′-dioctadecyloxacarbocyanine using the LIVE/DEAD Cell-Mediated Cytotoxicity Kit (Molecular Probes, Inc., Eugene, OR). Effector spleen cells were isolated and co-cultured with YAC-1 cells (ATCC, Manassas, VA) at an effector: target cell ratio of 100:1. Cells were then collected, washed, and stained with Live/DEAD Fixable Violet Dead Cell Stain Kits (Life Technologies, Grand Island, NY), according to the manufacturer's instructions. The cells were analyzed by flow cytometry after 4 hours co-culture. The percent specific lysis was determined as follows: 100 × (experimental lysis–spontaneous lysis)/(maximum lysis–spontaneous lysis). The percent specific lysis was normalized to the number of NK cells in the spleen.

      In Vitro Splenocyte Culture and Determination of Cytokines and Chemokines

      Infected mice were sacrificed on day 2 postinfection (p.i.), and the spleen and serum were collected. Splenocytes were cultured in 96-well round bottom plates containing 5 × 105 cells/well, or 24-well plates containing 1.5 × 106 cells/well with or without rickettsial antigen stimulation. The culture supernatants were collected after 72 hours. The concentrations of cytokines and chemokines in the culture supernatant and sera were determined by quantitative ELISA kit (R&D Systems, Minneapolis, MN) or microsphere multiplexed cytokine immunoassays (Bio-Plex Cytokine Assay, Bio-Rad Laboratories, Hercules, CA) according to the manufacturer's instructions.

      Antibodies and Flow Cytometry

      Spleen cells were isolated and stimulated with or without rickettsial antigens or phorbol 12-myristate 13-acetate (10 ng/mL) and ionomycin (400 ng/mL) in the presence of Golgi stop (BD Bioscience, San Diego, CA). Then the cells were suspended in fluorescence activated cell sorter buffer (PBS containing 0.1% bovine serum albumin and 0.01% NaN3). Fc receptors were blocked with anti-CD16/32 (clone 2.4G2). The following fluorescein isothiocyanate-, phycoerythrin (PE)-, peridinin chlorophyll protein Cy5.5 (PerCP-Cy5.5)-, and allophycocyanin (APC)-conjugated antibodies were purchased from BD Bioscience unless indicated otherwise: APC- or PE–anti-CD3 (clone 145-2C11), PE–anti-IFN-γ (clone XMG1.2), PE–anti-IL-12R β1 chain (clone 114), PE–anti-CD69 (clone HI.2F3), PercP– or APC–anti-CD8 (clone 53–6.7), and PercP– or APC–anti-CD4 (clone RM4-5). Isotype control antibodies included fluorescein isothiocyanate-, PE-, PercP-Cy5.5- and APC-conjugated hamster IgG1 (clone A19-3), rat IgG1 (clone R3-34), and rat IgG2a (clone R35-95). Specific antibodies including PE–anti-granzyme B (clone 16G6), fluorescein isothiocyanate–anti- CD49b (clone DX5, Pan-NK cells), and isotype control antibodies, including rat IgG2b and rat IgM were purchased from eBioscience (San Diego, CA); 20,000 events were collected using the FACSCalibur or FACSCanto system (BD Biosciences, Franklin Lakes, NJ). Data were analyzed with FlowJo software version 7.6.1 (TreeStar Inc., Ashland, OR).

      Histopathology and Immunohistochemistry

      Formalin-fixed, paraffin-embedded liver and lung samples were sectioned and stained with H&E. The quantity of pathological foci in livers in 10 high-power fields was determined using MetaMorph for Olympus (Olympus America Inc., Center Valley, PA). Thrombi were confirmed by staining with anti-mouse fibrinogen (Abbiotec, LLC., San Diego, CA) polyclonal antibody using Vectastain ABC reagents and Vector Red substrate (Vector Laboratories Inc., Burlingame, CA).

      Statistical Analysis

      For comparison of mean values of different experimental groups, the one-way analysis of variance or paired t-test was determined using GraphPad Prism software version 5.01. Posthoc group pairwise comparisons were conducted using the Bonferroni procedure and overall α level of significance of 0.05. For testing the difference in survival between different mouse groups, data were analyzed by the product limit (Kaplan-Meier) method using GraphPad Prism software version 5.01. A difference in mean values was deemed significant when P < 0.05.

      Results

      R. conorii Establishes Infection in the Resistant Host and Induces Robust Innate Immunity that Clears Infection

      We first determined the bacterial loads in the liver of susceptible and resistant mice during the course of infection. Rickettsiae replicated progressively in susceptible mice and reached a peak before mice died, whereas bacteria were substantially cleared in fewer than 4 days in resistant B6 mice (Figure 1A). Using plaque assay, the decreased bacterial burden detected by real-time PCR in the liver of resistant B6 mice correlated with the presence of live rickettsiae in the lung of B6 mice at both 6 and 24 hours after infection with a high dose of R. conorii (Figure 1B). In addition, infected liver from B6 mice contained multiple inflammatory foci at 2 and 5 days p.i., even though rickettsiae were completely cleared at these time points (Figure 1C). Next we determined the acquired T-cell response in C3H and B6 mice infected with low and high doses of R. conorii. All B6 mice survived both doses of rickettsiae, whereas all C3H mice died on inoculation of the high dose after approximately 6 to 7 days, as shown in the previous studies.
      • Walker D.H.
      • Popov V.L.
      • Wen J.
      • Feng H.M.
      Rickettsia conorii infection of C3H/HeN mice A model of endothelial-target rickettsiosis.
      • Fang R.
      • Ismail N.
      • Shelite T.
      • Walker D.H.
      Differential interaction of dendritic cells with Rickettsia conorii: impact on host susceptibility to murine spotted fever rickettsiosis.
      Greater bacterial expansion in susceptible C3H mice resulted in a significantly higher number of IFN-γ-producing CD4+ T cells (Figure 2, A and B), but not IFN-γ-producing CD8+ T cells (Figure 2, C and D) on day 9 p.i., compared to resistant B6 mice with the same inocula. Collectively, these results suggested that R. conorii established infection and initiated adaptive immune responses in the resistant host. Furthermore, these data suggest that innate responses play an important role in mediating host control of rickettsial infection.
      Figure thumbnail gr1
      Figure 1Rickettsiae are completely cleared by the robust innate immune response in B6 mice but progressively propagate in susceptible mice. WT C3H and B6 mice were inoculated with a high dose of R. conorii (as described in Materials and Methods). A: On days 1, 2, 4, and 5 postinfection (p.i.), livers were collected to measure the bacterial burden kinetically (as described in Materials and Methods). B: The presence of viable bacteria in the lung of B6 mice at 6 and 24 hours p.i. was determined by plaque assay. Bars represent the mean ± SD of three to five mice in each group. C: H&E-stained liver sections from uninfected and infected B6 mice on days 2 and 5 p.i. Microscopic examination revealed cellular inflammatory infiltrates (arrows). The sections shown are representative of three mice in each group, and experiments shown are representative of two performed. Original magnification, ×10. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; PFU, plaque forming units.
      Figure thumbnail gr2
      Figure 2Adaptive immune response was initiated by infection with R. conorii in vivo in both resistant B6 mice and susceptible C3H mice. WT C3H and B6 mice were inoculated with a low dose and a high dose of R. conorii (as described in Materials and Methods). Mice were sacrificed on day 9 postinfection (p.i.). Splenocytes were stimulated with antigen in vitro (as described in Materials and Methods). The percentage and absolute number of IFN-γ- producing CD4+ CD3+ T cells (A, B) and CD8+ CD3+ T cells (C, D) were determined by flow cytometric analysis. Flow cytometry dot plots (A, C) reveal the percentage of IFN-γ producing cells, which are quantified in (B, D). Each group includes five mice, and the results shown represent two independent experiments. *P < 0.05; ns, not statistically significant.

      T and/or B Cells Contribute to Early Elimination of Bacteria but Are Not the Determinants of Host Resistance to Rickettsial Infection

      Next we examined the response of T- and B-cell deficient (ie, Rag−/−) mice to an infection with a high dose of R. conorii. Similar to our previous data, all WT C3H mice succumbed to infection, whereas WT B6 and Rag−/− mice survived infection (Figure 3A). No Rag−/− mice succumbed to infection under observation until day 35 (data not shown). Interestingly, lack of T and B lymphocytes in Rag−/− mice resulted in a fourfold greater bacterial burden in the spleen compared to WT B6 mice on day 2 p.i. (Figure 3B), whereas no significant difference in bacterial burden between the two groups of mice was observed on day 5 p.i. These data suggested that although acquired immunity is dispensable for host survival after infection with Rickettsia in these animals, T and B lymphocytes contributed to early elimination of rickettsiae.
      Figure thumbnail gr3
      Figure 3T and/or B cells contribute to early elimination of bacteria but are not the determinants for host resistance to rickettsial infection. WT C3H, B6, and Rag−/− mice were inoculated with a high dose of R. conorii (as described in Materials and Methods). A: Mice were monitored daily for illness and survival until day 18 postinfection (p.i.). B: On days 2 and 5 p.i., bacterial burdens in liver were determined by real-time PCR. Bars represent the mean ± SD of three to five mice in each group. The results here represent two to four independent experiments with three to six mice/group. *P < 0.05. GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

      Innate Host Resistance to R. conorii Is Associated with an Elevated Level of Type 1 Response in the Spleen and a Greater Frequency of IFN-γ- Producing CD8+ T Cells

      Induction of IFN-γ producing Th1-type responses is critical for protection against infections with intracellular pathogens including rickettsiae.
      • Feng H.M.
      • Popov V.L.
      • Walker D.H.
      Depletion of gamma interferon and tumor necrosis factor alpha in mice with Rickettsia conorii-infected endothelium: impairment of rickettsicidal nitric oxide production resulting in fatal, overwhelming rickettsial disease.
      Thus, we examined whether the robust innate response in the resistant host is mediated by IFN-γ. At the single cell level, protective immunity in resistant B6 mice was associated with a significantly higher frequency of IFN-γ-producing splenocytes in B6 mice compared to C3H mice on day 2 p.i. (Figure 4A). Efficient bacterial clearance at the early stage of infection in B6 mice was associated with increased frequency of IFN-γ-producing splenic CD8+ T cells (Figure 4, B and C) compared to susceptible mice. These CD8+ T cells expressed memory phenotype markers (CD62Llow CD44high), suggesting effector memory CD8+ T cells (data not shown). These results suggested that IFN-γ produced by activated memory CD8+ T cells contributed to a host control of rickettsial infection at the early stage.
      Figure thumbnail gr4
      Figure 4Host resistance to R. conorii infection correlates with an elevated level of IFN-γ- producing spleen cells and a greater frequency of IFN-γ-producing CD8+ T cells. WT C3H and B6 mice were inoculated with a high dose of R. conorii. Splenocytes were collected on day 2 postinfection (p.i.), and then stimulated (as described in Materials and Methods). Cells were stained with the corresponding antibodies for detection of IFN-γ-producing spleen cells and counted (A). The percentage and absolute number of IFN-γ-producing-CD8+ T cells (B, C, respectively) were determined by using flow cytometric analysis. The dot plot (C) represents the mean ± SD of percentage of IFN-γ producing CD8+ T cells of five mice per group. Data shown are mean ± SD of five mice per group with similar results in two independent experiments. *P < 0.05.

      Enhanced Susceptibility to R. conorii Is Associated with Suppressed Local Production of IFN-γ, Enhanced IL-10, and Altered Systemic Proinflammatory Cytokines and Chemokines

      Excessive systemic or local production of IFN-γ and other proinflammatory cytokines can be detrimental to the host immune responses against infection.
      • Feng C.G.
      • Kaviratne M.
      • Rothfuchs A.G.
      • Cheever A.
      • Hieny S.
      • Young H.A.
      • Wynn T.A.
      • Sher A.
      NK cell-derived IFN-gamma differentially regulates innate resistance and neutrophil response in T cell-deficient hosts infected with Mycobacterium tuberculosis.
      • Stevenson H.L.
      • Estes M.D.
      • Thirumalapura N.R.
      • Walker D.H.
      • Ismail N.
      Natural killer cells promote tissue injury and systemic inflammatory responses during fatal Ehrlichia-induced toxic shock-like syndrome.
      Our data show that splenocytes from uninfected C3H mice produced a significantly higher level of IL-10 compared to resistant mice, but no difference in the levels of IFN-γ production (Figure 5A). In contrast, splenocytes of infected C3H mice produced a suppressed level of IFN-γ and an enhanced level of IL-10 after antigen stimulation compared to uninfected C3H mice and infected B6 mice (Figure 5A). Furthermore, infection of susceptible hosts with R. conorii increased the serum level of IFN-γ, but not IL-12p40, compared to resistant mice (Figure 5B). Resistance of B6 mice against fatal Rickettsia infection was associated with significantly lower serum levels of monocyte chemoattractant protein-1 (Figure 5B) compared to susceptible mice. These results suggested that enhanced susceptibility to rickettsial infection was associated with a suppressed type 1 response in the spleen but increased systemic production of monocyte chemoattractant protein-1.
      Figure thumbnail gr5
      Figure 5Greater host susceptibility to R. conorii infection is associated with a suppressed level of IFN-γ and a higher level of IL-10 production by spleen cells, as well as enhanced serum levels of monocyte chemoattractant protein-1 (MCP-1) and IFN-γ compared to the resistant host. WT C3H and B6 mice were inoculated with a high dose of R. conorii. On day 2 postinfection (p.i.), the spleen cells were stimulated in vitro (as described in Materials and Methods). A: The supernatant was collected, and the concentrations of IFN-γ and IL-10 were measured by ELISA. B: Concentrations of IFN-γ, IL-12p40, and MCP-1 in the sera were measured by Bio-plex assay (Bio-Rad Laboratories) or ELISA (R&D Systems). Data shown are mean ± SD of four mice per group with similar results in two independent experiments. *P < 0.05.

      Rapid Clearance of Bacteria in Resistant Mice Is Associated with Greater Activation and Cytotoxic Function of NK Cells

      Next we examined whether differential host susceptibility to Rickettsia infection is associated with altered activation and cytotoxic activity of NK cells. Our data showed that infection of B6 mice with R. conorii significantly enhanced activation of NK cells, as evidenced by greater expression of CD69 and IL-12p40 receptor β1 on NK cells when compared to susceptible C3H mice (Figure 6A).
      Figure thumbnail gr6
      Figure 6Host resistance to rickettsial infection is associated with greater activation and cytotoxicity of natural killer (NK) cells during the early stage of infection. WT C3H and B6 mice were inoculated with a high dose of R. conorii (as described in the Materials and Methods). On days 1 and 2 postinfection (p.i.), the frequencies of CD69- and IL-12p40 receptor β1-expressing DX5 (+) CD3 (-) NK cells (A, top and bottom panels, respectively) and granzyme B-expressing NK cells (B) were measured in the spleens of all mice by flow cytometry. C: Cytotoxicity was examined on day 2 p.i. (as described in the Materials and Methods). The percentage of rickettsial infection-associated lysis of YAC-1 cells by NK cells was normalized relative to the number of splenic NK cells in C3H and B6 mice. DF: The percentage and absolute number of IFN-γ-producing NK cells in spleen were analyzed on day 2 p.i. using flow cytometric analysis. The dot plot (D) reveals the mean ± SD of percentage of IFN-γ producing NK cells of five mice per group. The percentage and total number of IFN-γ producing NK cells in spleen, respectively, were assessed (E, F). Data shown are mean ± SD of five mice per group, and data are representative of two independent experiments. *P < 0.05.
      Furthermore, the frequency of granzyme-B-expressing splenic NK cells (Figure 6B) and NK cell cytotoxic activity, measured by the percentage of YAC-1 cell killing at an effector: target cell ratio of 100: 1 (Figure 6C), was significantly higher in infected resistant mice compared to susceptible mice on days 1 and/or 2 p.i. (Figure 6B). Interestingly, protective immunity in B6 mice was associated with significantly lower frequencies of IFN-γ-producing NK cells when compared to susceptible C3H mice on day 2 p.i. (Figure 6, D–F). These data do not exclude a potential protective role of IFN-γ produced by NK cells in host defense against Rickettsia, although it suggests that NK cell-mediated cytotoxic killing of infected target cells plays a pivotal role in effective rickettsial elimination.

      Deficiency of NK Cells Impairs Host Resistance and Promotes Severe Pathology after R. conorii Infection Regardless of Host Genetic Background

      To determine whether genetic background influences the effector function of NK cells against R. conorii, we examined the outcome of infection and host responses in NK cell competent or deficient CB-17 scid mice and Rag−/− mice. Similar to highly resistant B6 mice and Rag−/− mice, relatively resistant WT CB-17 mice and CB-17 scid mice survived the same dose of infection with R. conorii (Figure 7A). Interestingly, NK cell-deficient CB-17 scid mice were more susceptible to fatal disease, as demonstrated by ∼50% survival compared to 100% survival of CB-17 scid mice (Figure 7A). On day 2 p.i., R. conorii-infected NK cell-deficient CB-17 scid mice had a negligible percentage (ie, <1%) of IFN-γ- and granzyme B-expressing NK cells compared to CB-17 scid mice (data not shown). These data suggest that NK cells contribute greatly to host resistance against Rickettsia, independent of acquired immune responses.
      Figure thumbnail gr7
      Figure 7Deficiency of natural killer (NK) cells impairs host resistance to R. conorii infection in CB-17 scid mice. WT CB-17 mice, CB-17 scid mice, and NK cell-deficient CB-17 scid mice were infected (as described in the Materials and Methods). A: Mice were monitored for survival until day 18 postinfection (p.i.). B: On day 2 p.i., livers were harvested and processed for H&E staining. Arrows demonstrate infarcts, whereas arrowheads indicate cellular infiltration. Original magnification, ×40. Data represent two independent experiments with a total of eight mice in each group.
      Analysis of tissue pathology showed that combined lack of NK cells and T and B cells in NK cell-deficient CB-17 scid mice resulted in extensive tissue pathology and vasculitis, as evidenced by numerous hepatic infarcts and thrombi compared to WT CB-17 and CB-17 scid mice (Figure 7B). Similarly, deficiency of NK cells in infected Rag−/−γc−/− mice resulted in 100% mortality by day 32 (Figure 8A). The delayed mortality in infected NK cell-deficient- Rag−/− mice differs from an acute Rickettsia infection in humans that lasts from 10 to 14 days after the onset of infection.
      • Walker D.H.
      Rocky Mountain spotted fever: a seasonal alert.
      These data suggest that one of the mechanisms by which NK cells contribute to host protection to rickettsial infection could be attributed to prevention of bacteria-induced vascular injury.
      Figure thumbnail gr8
      Figure 8Natural killer (NK) cells mediate host resistance to rickettsial infection not only through enhancing bacterial clearance but also by regulating vascular injury-caused pathology. Rag−/− mice and Rag−/−γc−/−mice were infected (as described in the Materials and Methods). A; Mice were monitored for survival until day 36 postinfection (p.i.). B: H&E stained liver sections from uninfected and infected mice on day 2 p.i. showed more infarcts (arrows) and extensive cellular infiltrates (arrowheads) in Rag−/−γc−/−mice compared to Rag−/− mice. Original magnification, ×20. C: The percentage of pathological foci including cellular infiltration, infarcts and thrombosis on day 2 p.i. in 10 high power fields was determined using MetaMorph for Olympus (Olympus America Inc.). D: the bacterial loads in the infected livers were determined as described in the Materials and Methods. E: Serum levels of IFN-γ in all mouse groups were measured by ELISA. The levels of IFN-γ in the sera of uninfected mice were negligible. Data shown are mean ± SD of 5 mice per group and represent two independent experiments. *P < 0.05. GAPDH, glyceraldehyde 3-phosphate dehydrogenase.
      To further confirm the role of NK cells in preventing rickettsiae-induced vascular injury, we statistically analyzed the tissue pathology in NK cell-deficient Rag−/− mice. On day 2 p.i., R. conorii infection of NK-cell deficient Rag−/− γc−/− mice resulted in more severe hepatic pathology, marked by significantly increased cellular infiltration, infarction, and thrombosis compared to infected Rag−/− mice (Figure 8, B and C). In addition, deficiency of NK cells in Rag−/−γc−/− mice significantly impaired rickettsial clearance on day 5 p.i. when compared to Rag−/− mice (Figure 8D). To ensure that impaired protection against Rickettsia infection in Rag−/−γc−/− mice was not due to compensatory mechanisms, infected Rag−/− mice were treated with anti-asialo-GM1 antibody to deplete NK cells, which resulted in ∼90% depletion of NK cells (see Supplemental Figure S1A at http://ajp.amjpathol.org). Consistent with Rag−/−γc−/− mice, NK cell-depleted Rag−/− mice contained significantly greater bacterial burdens in livers compared to sham-depleted controls (see Supplemental Figure S1B at http://ajp.amjpathol.org). Impaired host control of rickettsial replication in NK-deficient Rag−/− mice was associated with reduced systemic levels of IFN-γ (Figure 8E), suggesting that NK cells contribute to innate immunity against rickettsiae, most likely via production of IFN-γ.

      Perforin Produced by NK Cells Is Involved in Protection from Vascular Injury Induced by Rickettsial Infection

      Next we directly examined the contribution of the cytotoxic molecule, perforin, produced by NK cells to host protection against Rickettsia in Rag−/−Prf1−/− mice. Surprisingly, rickettsiae-infected Rag−/−Prf1−/− mice had a significantly lower bacterial load in the liver on day 2 p.i. compared to that in Rag−/− mice (Figure 9A). Effective bacterial elimination on day 2 p.i. in Rag−/−Prf1−/− mice was associated with a significantly higher level of IFN-γ in the spleen (∼150 pg/mL) compared to Rag−/− mice (Figure 9B). Systemic levels of IFN-γ, however, were comparable in the different mouse groups (Figure 9C). Lack of perforin in infected Rag−/−Prf1−/− mice also resulted in severe pathology characterized by significantly greater frequency of cellular infiltration, infarction, and thrombosis in the liver on days 2 and day 5 p.i. (Figure 9, D and E). The greater pathology in infected Rag−/−Prf1−/− mice was similar to what we found in infected Rag−/−γc−/− mice resulting from lack of NK cells (Figure 8B). Immunohistochemical staining of fibrinogen in thrombi and infarcts in infected livers from Rag−/−Prf1−/− mice confirmed the fibrin content of thrombi induced by rickettsiae in vivo (Figure 9F). Collectively, these results further confirm that IFN-γ production by NK cells contributes to early bacterial clearance and perforin production by NK cells is associated with prevention of pathology during rickettsial infection.
      Figure thumbnail gr9
      Figure 9Natural killer (NK) cells mediate inhibition of vascular injury-caused pathology via a mechanism involving perforin. Rag−/− and Rag−/−Prf1−/− mice were infected (as described in Materials and Methods). On days 2 and 5 postinfection (p.i.), spleen and liver were collected. A: Bacterial burden in liver was determined by real-time PCR. The concentrations of IFN-γ in the supernatant of spleen cell culture (B) and sera (C) were determined by ELISA. The levels of IFN-γ in uninfected spleen cell culture supernatant and the sera of uninfected mice were negligible. D: The liver lesions included cellular infiltration (arrowheads) and thrombi or infarcts (arrows). Original magnification, ×20. E: The percentage of tissues containing pathological lesions on days 2 and 5 p.i. including cellular infiltration, infarcts and thrombosis in 10 high-power fields was determined using MetaMorph for Olympus (Olympus America Inc.). F: Fibrin was identified by immunohistochemical staining in the thrombi in livers of Rag−/−Prf1−/− mice (shown as brown) on day 2 p.i. Data shown are mean ± SD of 5 mice per group. ns, not significant. *P < 0.05. GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

      Discussion

      This study was undertaken to examine the contribution of the innate immune response and its effector components to host protection against infection with Rickettsia. The present study demonstrated that NK cells contribute greatly to host defense during an intracellular infection with Rickettsia that targets the endothelial cells. In addition, our study suggests that innate immune effectors against rickettsiae also involve IFN-γ-producing CD8+ T cells.
      The results in this study have broad significance to our understanding of the interaction of NK cells with endothelium in vivo during bacterial infection. Rickettsial infection promoted endothelial injury-caused pathology in NK cell-deficient CB-17 scid mice and NK cell-deficient Rag−/− mice. These results suggest that NK cells effectively inhibit rickettsiae-induced severe vascular injury. Studies have shown that rickettsiae cause injury of endothelial cells, which leads to increased procoagulant activity.
      • Olano J.P.
      Rickettsial infections.
      Tthrombosis and infarction, however, do not occur in the fatal murine model of spotted fever group rickettsiosis in susceptible C3H mice or at autopsy of the vast majority of patients who die from infection with Rickettsia.
      • Davì G.
      • Giammarresi C.
      • Vigneri S.
      • Ganci A.
      • Ferri C.
      • Di Francesco L.
      • Vitale G.
      • Mansueto S.
      Demonstration of Rickettsia conorii-induced coagulative and platelet activation in vivo in patients with Mediterranean spotted fever.
      • Davidson M.G.
      • Breitschwerdt E.B.
      • Walker D.H.
      • Levy M.G.
      • Carlson C.S.
      • Hardie E.M.
      • Grindem C.A.
      • Nasisse M.P.
      Vascular permeability and coagulation during Rickettsia rickettsii infection in dogs.
      The role of NK cells in preventing rickettsiae-induced endothelial injury may explain the minimal ischemic necrosis observed in fatal rickettsiosis in immunocompetent humans and mice. Thus, our results provide valuable clues to understanding the pathogenesis of rickettsial diseases and crosstalk between NK cells and endothelium, the main target cells for Rickettsia.
      Perforin is primarily involved in NK cell-mediated host control of microbes, but the evidence for the contribution of perforin to regulating pathology induced by infectious agents is increasing.
      • Vankayalapati R.
      • Garg A.
      • Porgador A.
      • Griffith D.E.
      • Klucar P.
      • Safi H.
      • Girard W.M.
      • Cosman D.
      • Spies T.
      • Barnes P.F.
      Role of NK cell-activating receptors and their ligands in the lysis of mononuclear phagocytes infected with an intracellular bacterium.
      • Maroof A.
      • Beattie L.
      • Zubairi S.
      • Svensson M.
      • Stager S.
      • Kaye P.M.
      Posttranscriptional regulation of IL-10 gene expression allows natural killer cells to express immunoregulatory function.
      Our data showed that resistance to rickettsial infection in B6 mice is associated with increased cytotoxic activity of NK cells (Figure 6). Surprisingly, we found that perforin production by NK cells in Rag−/− mice did not contribute to early rickettsial elimination (Figure 9A). It has been shown that perforin and granzymes play distinct roles in protective immunity and immunopathology during murine cytomegalovirus infection with infected Gzma−/−Gzmb−/− mice having only increased viral titers, whereas infected perforin-deficient hosts develop fatal immunopathology.
      • van Dommelen S.L.
      • Sumaria N.
      • Schreiber R.D.
      • Scalzo A.A.
      • Smyth M.J.
      • Degli-Esposti M.A.
      Perforin and granzymes have distinct roles in defensive immunity and immunopathology.
      Therefore, it is possible that enhanced bacterial clearance in Rag−/−Prf1−/− mice might be mediated by granzyme produced by cytotoxic NK cells independent of perforin or enhanced local production levels of IFN-γ (Figure 9B).
      • Kägi D.
      • Odermatt B.
      • Mak T.W.
      Homeostatic regulation of CD8+ T cells by bperforin.
      • Trapani J.A.
      • Smyth M.J.
      Functional significance of the perforin/granzyme cell death pathway.
      The presence of severe tissue damage in the absence of overwhelming infection in infected Rag−/−Prf1−/− mice (Figure 9, D and E) is less due to immunopathology secondary to a compensatory increase in production of IFN- γ for the following reasons: i) our previous studies indicated that IFN- γ is essential for bacterial clearance in rickettsial infection in vitro and in vivo;
      • Feng H.M.
      • Walker D.H.
      Interferon-gamma and tumor necrosis factor-alpha exert their antirickettsial effect via induction of synthesis of nitric oxide.
      • Feng H.M.
      • Popov V.L.
      • Walker D.H.
      Depletion of gamma interferon and tumor necrosis factor alpha in mice with Rickettsia conorii-infected endothelium: impairment of rickettsicidal nitric oxide production resulting in fatal, overwhelming rickettsial disease.
      • Feng H.
      • Popov V.L.
      • Yuoh G.
      • Walker D.H.
      Role of T lymphocyte subsets in immunity to spotted fever group Rickettsiae.
      • Valbuena G.
      • Feng H.M.
      • Walker D.H.
      Mechanisms of immunity against rickettsiae New perspectives and opportunities offered by unusual intracellular parasites.
      ii) the amount of local IFN-γ production detected in Rag−/−Prf1−/− (∼150 pg/mL) mice is unlikely able to have caused tissue damage; in addition, the systemic levels of IFN- γ in infected Rag−/−Prf1−/− mice were comparable to those in Rag−/− mice (Figure 9C), which further suggests that the compensatory increase in IFN-γ is less likely to be a mechanism that would account for pathology; and iii) the lack of correlation between severe tissue damage and increased cellular infiltration on day 2 p.i. (Figure 9, D and E) further suggests that immunopathology is less likely.
      For the first time, this study defines the contribution of NK cells to early host control of rickettsial infection independent of T-cell-mediated immunity. A previous study reported that C3H mice depleted of NK cells are more susceptible to R. conorii infection, as demonstrated by increased bacterial burden in the presence of T cells.
      • Billings A.N.
      • Feng H.M.
      • Olano J.P.
      • Walker D.H.
      Rickettsial infection in murine models activates an early anti-rickettsial effect mediated by NK cells and associated with production of gamma interferon.
      Our data showed that the absence of NK cells in CB-17 scid mice or Rag−/− mice significantly enhanced the mortality of these infected mice, as well as impaired bacterial control (Figures 7A, 8, A and D). The importance of IFN-γ production by NK cells in early host control of rickettsiae was supported by the evidence that deficiency of NK cells in Rag−/− mice resulted in reduced systemic IFN-γ. These data indicate that in the absence of T and B cells the innate response against rickettsial infection is, at least in part, mediated by NK cells, and it is independent of the host genetic background. Enhanced mortality of Rag−/−γc−/− mice at the late stage of infection was also associated with disseminated infection and development of meningitis (data not shown), which further suggested that NK cells are essential for controlling bacterial elimination and/or replication. Our results showed that the frequency of IFN-γ-producing NK cells in WT C3H and B6 mice did not correlate with host resistance to rickettsial infection, although it should not be interpreted that IFN-γ production by NK cells is less important in innate immunity against Rickettsia. First, less IFN-γ production by NK cells in resistant B6 mice may result from much lower quantity of bacteria in tissues compared to susceptible mice, which also might explain the significantly reduced serum levels of IFN-γ (Figure 5B). Second, the synergistic effect of IFN-γ- producing NK cells and CD8 T cells on bacterial clearance may confer better protection against Rickettsia than only that from NK cells. Biron and colleagues
      • Lee S.H.
      • Kim K.S.
      • Fodil-Cornu N.
      • Vidal S.M.
      • Biron C.A.
      Activating receptors promote NK cell expansion for maintenance IL-10 production, and CD8 T cell regulation during viral infection.
      reported that NK cells regulate the balance of NK cell and CD8 T-cell responses in murine cytomegalovirus infection. The robust early host control of rickettsial infection in resistant mice may result from an optimally balanced response of CD8 T cells and NK cells, which is regulated by NK cells.
      Comparing the immune responses against Rickettsia in susceptible and resistant hosts revealed that a higher level of IFN-γ-producing CD8+ T cells is associated with early bacterial elimination, although deficiency of CD8+ T cells did not alter the innate resistance to R. conorii infection in B6 mice. The contribution of CD8+ T cells to innate resistance has also been demonstrated with a variety of other pathogens, such as murine cytomegalovirus and Listeria monocytogenes.
      • Berg R.E.
      • Crossley E.
      • Murray S.
      • Forman J.
      Relative contributions of NK and CD8 T cells to IFN-gamma mediated innate immune protection against Listeria monocytogenes.
      • D'Orazio S.E.
      • Troese M.J.
      • Starnbach M.N.
      Cytosolic localization of Listeria monocytogenes triggers an early IFN-gamma response by CD8+ T cells that correlates with innate resistance to infection.
      • Sumaria N.
      • van Dommelen S.L.
      • Andoniou C.E.
      • Smyth M.J.
      • Scalzo A.A.
      • Degli-Esposti M.A.
      The roles of interferon-gamma and perforin in antiviral immunity in mice that differ in genetically determined NK-cell-mediated antiviral activity.
      Furthermore, the important role of the synergistic effect of CD8+ T cells and NK cells in the protective immune response has been demonstrated in these infections. These results provide more clues on the importance of these early IFN-γ-producing CD8 T cells in the development of a vaccine against rickettsial infection.
      In conclusion, the current studies provide important and previously unrecognized insight into the roles of NK cells and CD8+ T cells in the innate response to R. conorii infection. Our study has shown that NK cells are an essential component of the machinery involved in controlling rickettsial replication and in mediating the innate phase of host protection to R. conorii infection in different mouse strains. NK cells participate in the regulation of the immune response to rickettsiae, more particularly through preventing early endothelial injury that resulted in tissue damage. These results highlight the potential role played by NK cells in controlling or regulating the response of other immune cells such as T cells and endothelial cells.

      Acknowledgments

      We thank Kerry Graves and Dr. Bin Gong for their preparation of histopathology slides and immunohistochemical staining. Dr. Leoncio A. Vergara assisted with the image analysis of the histopathology.

      Supplementary data

      • Supplemental Figure S1

        Depletion of natural killer (NK) cells impairs bacterial clearance in R. conorii-infected Rag−/− mice. Rag−/− mice were treated with anti-asialo-GM1 antibody and then infected as described in the Materials and Methods. A: The efficiency of depletion of NK cells was determined by flow cytometric analysis. Numbers in the dot plot represent the absolute number and percentage of DX5 (+) CD3 (-) NK cells per mouse spleen after antibody treatment. Uninfected mice served as controls. More than 90% of NK cells were depleted in both uninfected and infected mice compared to the sham controls. B: On day 2 postinfection (p.i.), the bacterial burden in liver was determined in all groups of mice by real-time PCR. Each group includes three mice. *P < 0.05; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

      References

        • Walker D.H.
        Rocky Mountain spotted fever: a seasonal alert.
        Clin Infect Dis. 1995; 20: 1111-1117
        • Woods M.
        • Olano J.
        Host defenses to Rickettsia rickettsii infection contribute to increased microvascular permeability in human cerebral endothelial cells.
        J Clin Immunol. 2008; 28: 174-185
        • Walker D.H.
        • Hawkins H.K.
        • Hudson P.
        Fulminant Rocky Mountain spotted fever.
        Arch Pathol Lab Med. 1983; 107: 121-125
        • Walker D.H.
        The role of host factors in the severity of spotted fever and typhus rickettsioses.
        Ann NY Acad Sci. 1990; 590: 10-19
        • Walker D.H.
        • Popov V.L.
        • Wen J.
        • Feng H.M.
        Rickettsia conorii infection of C3H/HeN mice.
        Lab Invest. 1994; 70: 358-368
        • Feng H.M.
        • Walker D.H.
        Cross-protection between distantly related spotted fever group rickettsiae.
        Vaccine. 2003; 21: 3901-3905
        • Weinberger M.
        • Keysary A.
        • Sandbank J.
        • Zaidenstein R.
        • Itzhaki A.
        • Strenger C.
        • Leitner M.
        • Paddock C.D.
        • Eremeeva M.E.
        Fatal Rickettsia conorii subsp. israelensis infection.
        Israel Emerg Infect Dis. 2008; 14: 821-824
        • Aliaga L.
        • Sánchez-Blázquez P.
        • Rodríguez-Granger J.
        • Sampedro A.
        • Orozco M.
        • Pastor J.
        Mediterranean spotted fever with encephalitis.
        J Med Microbiol. 2009; 58: 521-525
        • Papa A.
        • Dalla V.
        • Petala A.
        • Maltezou H.C.
        • Maltezos E.
        Fatal Mediterranean spotted fever in Greece.
        Clin Microbiol Infect. 2010; 16: 589-592
        • de Sousa R.
        • Nobrega S.D.
        • Bacellar F.
        • Torgal J.
        Mediterranean spotted fever in Portugal risk factors for fatal outcome in 105 hospitalized patients.
        Ann NY Acad Sci. 2003; 990: 285-294
        • Feng H.M.
        • Wen J.
        • Walker D.H.
        Rickettsia australis infection: a murine model of a highly invasive vasculopathic rickettsiosis.
        Am J Pathol. 1993; 142: 1471-1482
        • Walker D.H.
        • Popov V.L.
        • Feng H.M.
        Establishment of a novel endothelial target mouse model of a typhus group rickettsiosis: evidence for critical roles for gamma interferon and CD8 T lymphocytes.
        Lab Invest. 2000; 80: 1361-1372
        • Fang R.
        • Ismail N.
        • Shelite T.
        • Walker D.H.
        Differential interaction of dendritic cells with Rickettsia conorii: impact on host susceptibility to murine spotted fever rickettsiosis.
        Infect Immun. 2007; 75: 3112-3123
        • Feng H.M.
        • Walker D.H.
        Interferon-gamma and tumor necrosis factor-alpha exert their antirickettsial effect via induction of synthesis of nitric oxide.
        Am J Pathol. 1993; 143: 1016-1023
        • Feng H.M.
        • Popov V.L.
        • Walker D.H.
        Depletion of gamma interferon and tumor necrosis factor alpha in mice with Rickettsia conorii-infected endothelium: impairment of rickettsicidal nitric oxide production resulting in fatal, overwhelming rickettsial disease.
        Infect Immun. 1994; 62: 1952-1960
        • Feng H.
        • Popov V.L.
        • Yuoh G.
        • Walker D.H.
        Role of T lymphocyte subsets in immunity to spotted fever group Rickettsiae.
        J Immunol. 1997; 158: 5314-5320
        • Valbuena G.
        • Feng H.M.
        • Walker D.H.
        Mechanisms of immunity against rickettsiae.
        Microbes Infect. 2002; 4: 625-633
        • Korbel D.S.
        • Finney O.C.
        • Riley E.M.
        Natural killer cells and innate immunity to protozoan pathogens.
        Int J Parasitol. 2004; 34: 1517-1528
        • Bancroft G.J.
        The role of natural killer cells in innate resistance to infection.
        Curr Opin Immunol. 1993; 5: 503-510
        • Scharton-Kersten T.M.
        • Sher A.
        Role of natural killer cells in innate resistance to protozoan infections.
        Curr Opin Immunol. 1997; 9: 44-51
        • Lodoen M.B.
        • Lanier L.L.
        Natural killer cells as an initial defense against pathogens.
        Curr Opin Immunol. 2006; 18: 391-398
        • Paust S.
        • Senman B.
        • von Andrian U.H.
        Adaptive immune responses mediated by natural killer cells.
        Immunol Rev. 2010; 235: 286-296
        • French A.R.
        • Yokoyama W.M.
        Natural killer cells and viral infections.
        Curr Opin Immunol. 2003; 15: 45-51
        • de Saint Basile G.F.A.
        The role of cytotoxicity in lymphocyte homeostasis.
        Curr Opin Immunol. 2001; 13: 549-554
        • Matloubian M.
        • Suresh M.
        • Glass A.
        • Galvan M.
        • Chow K.
        • Whitmire J.K.
        • Walsh C.M.
        • Clark W.R.
        • Ahmed R.
        A role for perforin in downregulating T-cell responses during chronic viral infection.
        J Virol. 1999; 73: 2527-2536
        • Jordan M.B.
        • Hildeman D.
        • Kappler J.
        • Marrack P.
        An animal model of hemophagocytic lymphohistiocytosis (HLH): CD8+ T cells and interferon gamma are essential for the disorder.
        Blood. 2004; 104: 735-743
        • Kägi D.
        • Odermatt B.
        • Mak T.W.
        Homeostatic regulation of CD8+ T cells by bperforin.
        Eur J Immunol. 1999; 29: 3262-3272
        • Badovinac V.P.
        • Hamilton S.E.
        • Harty J.T.
        Viral infection results in massive CD8+ T cell expansion and mortality in vaccinated perforin-deficient mice.
        Immunity. 2003; 18: 463-474
        • Kang S.J.
        • Liang H.E.
        • Reizis B.
        • Locksley R.M.
        Regulation of hierarchical clustering and activation of innate immune cells by dendritic cells.
        Immunity. 2008; 29: 819-833
        • Valbuena G.
        • Bradford W.
        • Walker D.H.
        Expression analysis of the T-cell-targeting chemokines CXCL9 and CXCL10 in mice and humans with endothelial infections caused by rickettsiae of the spotted fever group.
        Am J Pathol. 2003; 163: 1357-1369
        • Feng C.G.
        • Kaviratne M.
        • Rothfuchs A.G.
        • Cheever A.
        • Hieny S.
        • Young H.A.
        • Wynn T.A.
        • Sher A.
        NK cell-derived IFN-gamma differentially regulates innate resistance and neutrophil response in T cell-deficient hosts infected with Mycobacterium tuberculosis.
        J Immunol. 2006; 177: 7086-7093
        • Stevenson H.L.
        • Estes M.D.
        • Thirumalapura N.R.
        • Walker D.H.
        • Ismail N.
        Natural killer cells promote tissue injury and systemic inflammatory responses during fatal Ehrlichia-induced toxic shock-like syndrome.
        Am J Pathol. 2010; 177: 766-776
        • Olano J.P.
        Rickettsial infections.
        Ann NY Acad Sci. 2005; 1063: 187-196
        • Davì G.
        • Giammarresi C.
        • Vigneri S.
        • Ganci A.
        • Ferri C.
        • Di Francesco L.
        • Vitale G.
        • Mansueto S.
        Demonstration of Rickettsia conorii-induced coagulative and platelet activation in vivo in patients with Mediterranean spotted fever.
        Thromb Haemost. 1995; 74: 631-634
        • Davidson M.G.
        • Breitschwerdt E.B.
        • Walker D.H.
        • Levy M.G.
        • Carlson C.S.
        • Hardie E.M.
        • Grindem C.A.
        • Nasisse M.P.
        Vascular permeability and coagulation during Rickettsia rickettsii infection in dogs.
        Am J Vet Res. 1990; 51: 165-170
        • Vankayalapati R.
        • Garg A.
        • Porgador A.
        • Griffith D.E.
        • Klucar P.
        • Safi H.
        • Girard W.M.
        • Cosman D.
        • Spies T.
        • Barnes P.F.
        Role of NK cell-activating receptors and their ligands in the lysis of mononuclear phagocytes infected with an intracellular bacterium.
        J Immunol. 2005; 175: 4611-4617
        • Maroof A.
        • Beattie L.
        • Zubairi S.
        • Svensson M.
        • Stager S.
        • Kaye P.M.
        Posttranscriptional regulation of IL-10 gene expression allows natural killer cells to express immunoregulatory function.
        Immunity. 2008; 29: 295-305
        • van Dommelen S.L.
        • Sumaria N.
        • Schreiber R.D.
        • Scalzo A.A.
        • Smyth M.J.
        • Degli-Esposti M.A.
        Perforin and granzymes have distinct roles in defensive immunity and immunopathology.
        Immunity. 2006; 25: 835-848
        • Trapani J.A.
        • Smyth M.J.
        Functional significance of the perforin/granzyme cell death pathway.
        Nat Rev Immunol. 2002; 2: 735-747
        • Billings A.N.
        • Feng H.M.
        • Olano J.P.
        • Walker D.H.
        Rickettsial infection in murine models activates an early anti-rickettsial effect mediated by NK cells and associated with production of gamma interferon.
        Am J Trop Med Hyg. 2001; 65: 52-56
        • Lee S.H.
        • Kim K.S.
        • Fodil-Cornu N.
        • Vidal S.M.
        • Biron C.A.
        Activating receptors promote NK cell expansion for maintenance.
        J Exp Med. 2009; 206: 2235-2251
        • Berg R.E.
        • Crossley E.
        • Murray S.
        • Forman J.
        Relative contributions of NK and CD8 T cells to IFN-gamma mediated innate immune protection against Listeria monocytogenes.
        J Immunol. 2005; 175: 1751-1757
        • D'Orazio S.E.
        • Troese M.J.
        • Starnbach M.N.
        Cytosolic localization of Listeria monocytogenes triggers an early IFN-gamma response by CD8+ T cells that correlates with innate resistance to infection.
        J Immunol. 2006; 177: 7146-7154
        • Sumaria N.
        • van Dommelen S.L.
        • Andoniou C.E.
        • Smyth M.J.
        • Scalzo A.A.
        • Degli-Esposti M.A.
        The roles of interferon-gamma and perforin in antiviral immunity in mice that differ in genetically determined NK-cell-mediated antiviral activity.
        Immunol Cell Biol. 2009; 87: 559-566