Overproduction of TNF-α by CD8+ Type 1 Cells and Down-Regulation of IFN- γ Production by CD4 + Th1 Cells Contribute to Toxic Shock-Like Syndrome in an Animal This information is current as Model of Fatal Monocytotropic Ehrlichiosis of September 29, 2021. Nahed Ismail, Lynn Soong, Jere W. McBride, Gustavo Valbuena, Juan P. Olano, Hui-Min Feng and David H. Walker

J Immunol 2004; 172:1786-1800; ; Downloaded from doi: 10.4049/jimmunol.172.3.1786 http://www.jimmunol.org/content/172/3/1786 http://www.jimmunol.org/ References This article cites 53 articles, 31 of which you can access for free at: http://www.jimmunol.org/content/172/3/1786.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

by guest on September 29, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2004 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Overproduction of TNF-␣ by CD8؉ Type 1 Cells and Down-Regulation of IFN-␥ Production by CD4؉ Th1 Cells Contribute to Toxic Shock-Like Syndrome in an Animal Model of Fatal Monocytotropic Ehrlichiosis1

Nahed Ismail,* Lynn Soong,*†‡§ Jere W. McBride,*‡ Gustavo Valbuena,* Juan P. Olano,*†‡ Hui-Min Feng,* and David H. Walker2*‡§

Human monocytotropic ehrlichiosis (HME) is an emerging, life-threatening, infectious disease caused by Ehrlichia chaffeensis,an obligate intracellular bacterium that lacks cell wall LPS. We have previously developed an animal model of severe HME using a

strain of Ehrlichia isolated from Ixodes ovatus ticks (IOE). To understand the basis of susceptibility to severe monocytotropic Downloaded from ehrlichiosis, we compared low and high doses of the highly virulent IOE strain and the less virulent Ehrlichia muris strain that are closely related to E. chaffeensis in C57BL/6 mice. Lethal infections caused by high or low doses of IOE were accompanied by extensive liver damage, extremely elevated levels of TNF-␣ in the serum, high frequency of Ehrlichia-specific, TNF-␣-producing CD8؉ T cells in the spleen, decreased Ehrlicha-specific CD4؉ T cell proliferation, low IL-12 levels in the spleen, and a 40-fold decrease in the number of IFN-␥-producing CD4؉ Th1 cells. All groups contained negligible numbers of IL-4-producing cells in ؉ ؉ the spleen. Transfer of Ehrlichia-specific polyclonal Abs and IFN-␥-producing Ehrlichia-specific CD4 and CD8 type 1 cells http://www.jimmunol.org/ protected naive mice against lethal IOE challenge. Interestingly, infection with high dose E. muris provided protection against rechallenge with a lethal dose of IOE. Cross-protection was associated with substantial expansion of IFN-␥-producing CD4؉ and ,CD8؉ cells, but not TNF-␣-producing CD8؉ T cells, a high titer of IgG2a, and a low serum level of TNF-␣. In conclusion -uncontrolled TNF-␣ production by CD8؉ T cells together with a weak CD4؉ Th1 cell response are associated with immunopa thology and failure to clear IOE in the fatal model of HME. The Journal of Immunology, 2004, 172: 1786–1800.

uman monocytotropic ehrlichiosis (HME)3 is an emerg- suggested that humoral immune responses play an important role ing, tick-borne, zoonotic disease caused by infection of during infections with E. chaffeensis and other related Ehrlichia H monocytes by the obligately intracellular bacterium, spp. (8). Although immunocompetent mice are generally resistant by guest on September 29, 2021 Ehrlichia chaffeensis (1–3). Infection of humans with E. chaffeen- to infection with E. chaffeensis, SCID mice are highly susceptible, sis results in a spectrum of clinical manifestations, ranging from developing persistent and fatal infection. Transfer of immune se- fatal to self-limited disease. In immunocompromised individuals, rum obtained from immunocompetent C57BL/6 mice as well as such as those infected with HIV (4) or treated with immunosup- Abs specific to the 28-kDa major outer membrane proteins of E. pressive drugs, E. chaffeensis infection can result in a severe, life- chaffeensis to C57BL/6 SCID mice provide significant, but tran- threatening disease. Immunocompetent individuals who do not re- sient, protection from disease (9). As infection of SCID mice with Ն ceive doxycycline treatment until 8 days after the onset of E. chaffeensis results in fatal disease with pathology that does not symptoms are also at increased risk of developing fatal HME. Se- mimic the histopathological findings in HME, SCID mice may not vere cases of HME are often comparable in severity to Rocky be a suitable model to elucidate the immunological basis of resis- Mountain spotted fever or toxic shock syndrome (5–7). tance and susceptibility to infection by monocytotropic ehrlichiae. Little is known about the host factors that influence susceptibil- The role of the cell-mediated response in the host defense ity and resistance to severe HME, although some studies have against Ehrlichia is supported by the observations that intracellular killing of E. chaffeensis requires CD4ϩ T cell-dependent cellular Departments of *Pathology and †Microbiology and Immunology, ‡Sealy Center for effector mechanisms, including NO production by IFN-␥-activated § Vaccine Development, and Center for Biodefense and Emerging Infectious Diseases, macrophages, and granulomatous inflammation (10). However, the University of Texas Medical Branch, Galveston, TX 77555 roles of CD4ϩ and CD8ϩ T cells and their cytokines in host de- Received for publication October 24, 2003. Accepted for publication November 18, 2003. fense and pathogenesis of the disease are not yet defined. The costs of publication of this article were defrayed in part by the payment of page In the present study we have analyzed the immunopathological charges. This article must therefore be hereby marked advertisement in accordance mechanisms associated with susceptibility or resistance to ehrli- with 18 U.S.C. Section 1734 solely to indicate this fact. chial infection, using two ehrlichial strains that are phylogeneti- 1 This work was supported by Grant AI31431 from the National Institute of Allergy cally related to E. chaffeensis. The first one is a highly virulent and Infectious Diseases. ehrlichial strain isolated from Ixodes ovatus ticks (IOE) native to 2 Address correspondence and reprint requests to Dr. David H. Walker, Department of Pathology, Center for Biodefense and Emerging Infectious Diseases, 301 Univer- Japan (11) that causes fatal disease in immunocompetent mice sity Boulevard, Galveston, TX 77555-0609. E-mail address: [email protected] (12). The second organism is a mildly virulent strain (Ehrlichia 3 Abbreviations used in this paper: HME, human monocytotropic ehrlichiosis; EM/ muris) that causes mild and self-limited disease in immunocom- IOE, mice primed with high dose E. muris and rechallenged with high dose IOE; IOE, Ixodes ovatus Ehrlichia; MOI, multiplicity of infection; OMP, outer membrane pro- petent mice (13). Phylogenetic analysis supports the close rela- tein; p.i., postinfection. tionships among monocytotropic Ehrlichia spp., including E.

Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00 The Journal of Immunology 1787 chaffeensis, E. muris, and IOE (14–17). Additionally, serological preparations was determined using a bicinchoninic acid protein assay kit cross-reactions occur between closely related monocytotropic Ehr- (Pierce, Rockford, IL) and was used as the Ag in ELISPOT assay and lichia species (18–20), attributed mainly to a molecularly charac- ELISA. The uninfected cell lysates were prepared similarly and used as a negative control (mock Ag). For preparation of host cell-free IOE Ag, terized major outer membrane proteins (p28) of Ehrlichia and ma- IOE-infected spleens and livers were harvested from day 7 infected mice, jor antigenic protein 2, the P28 orthologue identified in E. and cell-free IOE Ags were prepared as previously described (21). Spleen ruminantium (18–20). and liver of naive mice were prepared as the E. muris-infected P388D1 In this study we assessed the course of infection and immune cells and were used as a negative control in all experiments using cell-free IOE Ags (mock Ag). responses in C57BL/6 mice susceptible to fatal infection with IOE, but resistant to severe disease caused by E. muris infection. We ELISPOT assays for Ag-specific, cytokine-producing T cells also investigated whether a primary infection with E. muris can Single-cell suspensions were obtained from the spleen of control and in- protect mice against rechallenge with virulent IOE and analyzed fected mice. CD4ϩ and CD8ϩ T cells were isolated by negative selection the immunological correlates of this cross-protection. We provide using mouse CD4 or CD8 subset enrichment columns (R&D Systems, in this report compelling evidence that infection with virulent IOE Minneapolis, MN), and the purity ranged from 80–90% as determined by causes overproduction of TNF-␣ by Ag-specific CD8ϩ T cells, FACS analysis. Splenocytes and purified T cell subsets were assessed via ELISPOT for cytokine production, as described previously (22–24). which strongly correlates with pathology and a septic shock-like Briefly, 96-well nitrocellulose plates (Millipore, Bedford, MA) were coated syndrome. More importantly impaired CD4 T cell proliferation at 4°C overnight with mAbs (1.25 ␮g/ml; 100 ␮l/well) that are specific for and down-regulation of the Th1 response are other detrimental murine IFN-␥, IL-4, or TNF-␣ (BD PharMingen, San Diego, CA). Two- 6 5 factors that result in high mortality in the animal model of fold dilutions of spleen cells were added to wells starting at 10 to 2 ϫ 10 cells/well in the presence of an additional 1 ϫ 106 spleen cells from naive, fatal HME. Downloaded from unimmunized mice. The addition of normal spleen cells was necessary to ensure that the number of Ag-dependent spots observed was proportional Materials and Methods to the number of immune spleen cells plated and that the response was Bacterial stocks and experimental design linear. Immune spleen cells or purified T cells were stimulated with either the specific E. muris or IOE Ags at concentration of 10 ␮g/well or with Two monocytotropic ehrlichial strains were used in this study, highly vir- nonspecific Ag such as SRBC. Spleen cells from EM/IOE-infected mice ulent Ehrlichia spp. (designated IOE) isolated from Ixodes ovatus ticks (a were stimulated with cell-free E. muris or IOE Ags. Positive and negative gift from Dr. M. Kawahara, Nagoya City Public Health Research Institute, controls contained 5 ␮g/ml Con A or medium, respectively. After incuba- http://www.jimmunol.org/ Nagoya, Japan) and mildly virulent E. muris (provided by Dr. Y. Rikihisa, tion at 37°Cin5%CO2 for 16 h, the plates were washed and incubated Ohio State University, Columbus, OH). To produce infectious stocks for with 100 ␮l of the appropriate biotinylated secondary mAb (1.25 ␮g/ml; reproducible studies, C57BL/6 mice were inoculated i.p. with 1 ml of a BD PharMingen) at 37°C for 2 h. The spots were developed by the addition Ϫ1 10 dilution of the frozen stock. On day 7 after inoculation, the mice were of alkaline phosphatase-streptavidin (0.2 ␮g/ml; BD PharMingen), fol- sacrificed, the spleens and livers were harvested, and the homogenate was lowed by the phosphatase substrate (buffer containing 0.1 M Tris, 0.1 M suspended in sucrose-phosphate-glutamate buffer (0.218 mol/L sucrose, ␮ NaCl, 0.05 M MgCl2, 8.75 g/ml nitro blue tetrazolium chloride, and 9.4 0.0038 mol/L KH2PO4, 0.0072 mol/L K2HPO4, and 0.0049 mol/L monoso- mg/ml 5-bromo-4-chloro-3-indolyl phosphate, toluidine in 67% (v/v) dium glutamic acid, pH 7.0). Large particles of debris were removed by DMSO; Sigma-Aldrich, St. Louis, MO). The reaction was stopped by thor- ϫ centrifugation at 200 g for 3 min, and the supernatant was then aliquoted ough washing with deionized water. The spots were counted under a dis- Ϫ Ϫ1 and stored at 80°Casa10 stock of IOE or E. muris. The LD50 of the secting microscope. In all experiments, Ag-specific spots were determined Ϫ5 by guest on September 29, 2021 IOE stock was estimated to be a dilution of 10 , because i.p. inoculation by subtraction of the background spots (spots detected in the Ag-negative Ϫ4 with a 10 dilution killed 100% of the mice, whereas 100% of the mice wells) from spots detected in the Ag-positive wells. survived when a 10Ϫ6 dilution was used (data not shown). Sex-matched C57BL/6 mice were obtained from The Jackson Labora- Flow cytometry tory (Bar Harbor, ME) and were used at 6–8 wk of age in all experiments. Mice were infected i.p. with 1 ml of fresh inoculum at the following doses: Before staining, spleen cells were incubated with an anti-Fc III/II receptor 10Ϫ2 or 10Ϫ4 dilution of IOE stock or 10Ϫ1,10Ϫ2,or10Ϫ3 dilution of E. (BD PharMingen, San Diego, CA) mAb and 10% normal mouse serum in Ϫ2 muris stock. Quantitative real-time PCR determined that the 10 dilution PBS containing 0.1% BSA and 0.01% NaN3. The lymphocytes were iden- of IOE stock (high dose IOE) contained 4 ϫ 106 bacterial genomes, tified by characteristic size (forward light scatter) and granularity (side whereas the 10Ϫ1 dilution of E. muris stock (high dose E. muris) contained light scatter) in combination with anti-CD4 or anti-CD8 (FITC conjugated; 6 ϫ 108 bacterial genomes. For cross-protection experiments, mice were BD PharMingen) surface staining or FITC-conjugated isotype control rat infected i.p. with a 10Ϫ1 dilution of E. muris and then challenged i.p. 30 IgG2a or IgG2b (BD PharMingen). For each sample, between 200,000 and days later with a high dose (10Ϫ2 dilution) of IOE. Subsequently, we refer 400,000 cells were analyzed. The data were collected and analyzed using to these mice as EM/IOE. Control mice were given 1 ml of a 10Ϫ1 or 10Ϫ2 a FACS flow cytometer (BD Biosciences, Mountain View, CA). dilution of a spleen homogenate from naive C57BL/6 mice. On the indi- ϩ cated days of infection, mice were sacrificed, and immune responses were CD4 T cell proliferation assay assessed. Selected organs were harvested for histology, immunohistochem- Purified CD4ϩ T cells from individual mice in different groups were plated istry, and determination of bacterial load by real-time PCR and culture. in triplicate at (1.5 ϫ 106 cells/ml) together with irradiated APC (5 ϫ 105) Histology and immunohistochemistry from syngeneic naive mice in 96-well plates and were stimulated with the relevant cell-free E. muris or IOE Ags. The cultures were incubated for ␮ 3 Samples of liver, spleen, lung, and kidney were processed for histopatho- 72hat37°Cin5%CO2 and then pulsed with 50 Ci/ml [ H]thymidine for ϩ logical examination as described previously (12). For immunohistochem- 16 h. CD4 T cells from IOE- or EM/IOE-infected mice were stimulated istry, slides were incubated for 45 min at 37°C with canine anti-E. with cell-free E. muris, IOE Ags, or Con A. Control cultures excluded chaffeensis polyclonal Ab at dilution of 1/1000, which cross-reacts with E. either Ag or APC, or cells were stimulated with 5 ␮g/ml Con A. Plates muris and IOE Ags. Slides were then incubated for 30 min with a biotin- were harvested, and the incorporation of [3H]thymidine was measured us- ylated goat anti-canine IgG (HϩL) Ab used at a 1/800 dilution (Vector ing liquid scintillation spectroscopy (25). Laboratories, Burlingame, CA). The slides were then washed and incu- bated with avidin-HRP conjugate for 20 min at 37°C, followed by incu- Peritoneal macrophages and in vitro stimulation bation with substrate containing 3-amino-9-ethylcarbazole for 8 min at For in vitro stimulation, peritoneal exudate cells were used. To obtain 37°C (Vector Laboratories). Slides were counterstained with hematoxylin. peritoneal exudate cells, mice were injected with 1 ml of 10% thioglycolate Normal canine serum was used as a negative control. (Difco Laboratories, Detroit, MI) i.p., and peritoneal lavage was performed Preparation of host cell-free Ehrlichia with 10 ml of ice-cold PBS 18 h later. Macrophages were isolated by negative selection using mouse CD11b subset enrichment columns (R&D E. muris was cultivated in P388D1 cells with 5% bovine calf serum-sup- Systems, Minneapolis, MN), and the purity ranged from 85–90%, as deter- plemented MEM at 37°C. Ehrlichiae were harvested when ϳ90–100% of mined by FACS analysis. Purified macrophages were stimulated with medium the cells were infected, and cell-free ehrlichiae were prepared as previously or IOE or E. muris Ags for 6 and 24 h, and culture supernatants were examined described (21). The total protein concentration of the resulting bacterial for IL-12p40 production by ELISA as previously described (22). 1788 ROLE OF TNF-␣ AND IFN-␥ IN MURINE Ehrlichia INFECTION

Cytokine ELISA added to Ag-coated wells and incubated at 25°C for 2 h. HRP-conjugated goat anti-mouse IgG1 and IgG2a Abs were added at a dilution of 1/2000, ϫ 6 Single-cell suspensions of spleens were cultured at 5 10 cells/ml in and the color was developed using substrate 3,3Ј,5,5Ј-tetramethylbenzidine DMEM containing 10% FBS, 2 mM glutamine, 100 U/ml penicillin G (Cal-Biochem, San Diego, CA.). The color was measured using an ELISA ␮ ϫ Ϫ5 sodium, 100 g/ml streptomycin sulfate, and 5 10 M 2-ME in the plate reader at 450 nm. For determining the concentration of IgG sub- ␮ presence or the absence of 50 g/ml E. muris or IOE Ags (prepared as classes, serial dilutions of purified mouse IgG1 or IgG2a were added to described above). Supernatants were collected at 72 h and assayed for IL-4, plates coated with goat anti-mouse Ig. Standard curve obtained by incu- IL-10, IL-12p40, and IL-12p70 using a sandwich ELISA, as previously bating standards with goat anti-mouse IgGs, and substrate was used to described (22). Standard curves were generated using recombinant mouse determine the concentration of each isotype. All assays including the serial cytokines (BD PharMingen). The detection limits were 50 pg/ml for IL-4, dilution standards were performed in triplicate wells, and the average val- 150 pg/ml for IL-10, and 122 pg/ml for IL-12p70. In some cases, sera were ues were calculated for analysis. collected at different time points of infection and were assayed for IL- 12p40 and TNF-␣ levels using an immunoassay kit (Quantikine; R&D Systems). The detection limit for TNF-␣ was 5 pg/ml. Statistical analysis Ehrlichial load determination by quantitative real-time PCR For comparisons of the means of various groups, the pairwise t test was used. The ehrlichial load in tissues was determined by real-time PCR (with SYBR Green) of the Ehrlichia dsb gene, which encodes a thio-disulfide oxireductase or disulphide bond formation protein of E. muris and IOE Results (GenBank accession no. AY236484 and AY26485). Primer sequences are Course of infection in animal models of mild and fatal human as follow: IOE forward, GAATAGAAAATGAAGAAATGAG; IOE re- monocytotropic ehrlichiosis verse, CAATAGCCACAAGAATAGTCAAAGA; E. muris forward, GAACAGAGGGGTCATTAAAAGCTGTTC; E. muris reverse, GAT We compared the course of disease in mice infected with different Downloaded from TCAACGCTGCATGGTAA; mouse GAPDH forward, CAACTACATG doses of highly virulent IOE or mildly virulent E. muris. C57BL/6 GTCTACATGTTC; and GAPDH reverse, CTCGCTCCTGGAAGATG. mice given a high dose (10Ϫ2 dilution) of IOE developed progres- The substrate for amplification was DNA purified from frozen tissue sam- sive disease with weight loss and severe hypoglycemia (data not ples using the DNeasy Tissue kit (Qiagen, Valencia, CA). Quantitative real-time PCR was performed using the iCycler from Bio-Rad (Hercules, shown), and all died between days 8 and 10, presumably due to the CA) and SYBR Green iQ Supermix (Bio-Rad). The results were normal- lethal inflammatory multiorgan pathology previously reported (12)

ized to the levels of expression of the eukaryotic housekeeping gene in this mouse model of fatal HME (Fig. 1A). When a lower dose http://www.jimmunol.org/ 4 GAPDH in the same sample and expressed as copy number per 10 (10Ϫ4 dilution) of IOE was given, mice showed no major signs of GAPDH copies (standard curves for dsb and GAPDH with Ͼ94% effi- ciency and linear amplification across 1 to 106–107 copies were used to disease on day 7, but started to lose weight thereafter. All mice Ϫ4 obtain the copy number of the samples). PCR analyses were considered infected with low dose (10 dilution) IOE succumbed to infection

negative for ehrlichial DNA if the critical threshold (CT) values exceeded and died between days 15 and 17. These two models were used to 40 cycles. Expression of the ehrlichial load in terms of the number of evaluate the host immune response associated with fatal ehrlichio- copies of GAPDH is a valid approach in this case, because ehrlichiae are sis caused by the highly virulent ehrlichial strain. obligately intracellular bacteria. In contrast, C57BL/6 mice infected with the high dose (10Ϫ1 Adoptive transfer dilution) of E. muris developed mild disease, and all survived the by guest on September 29, 2021 T cells were isolated from the spleens of mice infected with E. muris (E. acute infection (Fig. 1A). Infection of BALB/c or C3H mice with Ϫ1 muris-specific T cells) or E. muris-primed mice rechallenged with a lethal the high dose of E. muris (10 dilution) or infection of C57BL/6 dose of IOE (EM/IOE-specific T cells), respectively, on day 7 postinfection mice with a higher inoculum resulted in a mild, self-limited disease (p.i.) or rechallenge. The Th1/Th2 phenotype of transferred cells was de- ϩ ϩ (our unpublished observations). Therefore, infection of C57BL/6 termined by ELISPOT assay. CD4 and/or CD8 T cells were purified and mice with low virulence E. muris was used as a model of mild, adoptively transferred i.p. (107 each subset) into naive mice with or without mouse polyclonal serum anti-ehrlichial Abs. Serum polyclonal Ab was self-limited ehrlichiosis. obtained from EM/IOE-infected mice on day 30 after rechallenge with IOE We then assessed whether an infection with E. muris could pro- and diluted 1/5, and 1 ml was transferred i.p. to each mouse. The titer and vide cross-protection against subsequent infection with a lethal isotype of the polyclonal Abs at this time point after infection, as deter- dose of IOE. Four weeks after primary infection with high (10Ϫ1) mined by immunofluorescence assay and ELISA, were 1/8192 and IgG2a Ϫ2 Ϫ3 isotype, respectively. Control group mice were adoptively transferred i.p. or low (10 and 10 ) doses of E. muris, mice were challenged with naive CD4ϩ and/or CD8ϩ T cells with or without 1 ml of normal with the high dose of IOE. Control groups were injected with PBS mouse serum. One day later, mice were challenged i.p. with a high dose and challenged with a high dose of IOE or a high dose (10Ϫ1 Ϫ2 (10 ) of IOE. The course of infection and cytokine responses were as- dilution) of E. muris. All mice primarily infected with a high dose sessed on day 7 after IOE challenge. of E. muris survived for Ͼ4 mo after lethal challenge with IOE. Immunofluorescence assay for detection of specific or cross- We used the term EM/IOE to refer to this cross-protected group. In reactive Abs to E. muris Ags contrast, survival was only prolonged in mice primarily infected with the lower doses of E. muris and then challenged with the high Serum samples from infected and control mice were measured for Ehrli- chia-specific IgG Abs by indirect immunofluorescence assay using E. dose of IOE, where mice succumbed between days 16 and 20 p.i. muris as a surrogate or homologous Ag as previously described (12). A (data not shown). serial 2-fold dilution of serum samples was applied to the Ag slides. After Histopathological examination of the liver on day 7 p.i. in mice incubation at 37°C for 30 min in a humid chamber, slides were stained with infected with high dose IOE revealed abundant, partially confluent FITC-labeled anti-mouse IgG (Kirkegaard & Perry Laboratories, Gaithers- foci of necrosis of contiguous hepatocytes with regions of cellular burg, MD) at a dilution of 1/100. Slides were examined under a photofluo- rescent microscope (Nikon, Tokyo, Japan). Serological titers were ex- infiltration (Fig. 2A). Apoptotic cells were observed diffusely pressed as the reciprocal of the highest dilution at which specific throughout the liver, which is consistent with the presence of a fluorescence was detected. large number of TUNEL-positive cells (12) (data not shown). Sim- Quantitative ELISA to measure IgG subclasses ilarly, histopathological evaluation of the lung tissue from these mice revealed severe interstitial pneumonitis marked by infiltration Quantitative ELISA was performed to measure the concentration of Ehr- of lymphohistiocytic cells throughout interstitial regions, necrosis, lichia-specific IgG subclass Abs as described previously (21). Briefly, the 96-well ELISA plates were coated with the purified E. muris or IOE Ags and the presence of inflammatory cell infiltration into interalveolar at a concentration of 25 ng/well using 50 mM sodium bicarbonate buffer, septa (Fig. 2B). The marginal zone of the spleen was prominent pH 9.6. Serum samples were diluted 1/50, and 100 ␮l of each sample was and consisted of activated, moderately enlarged lymphocytes and The Journal of Immunology 1789

FIGURE 1. Analysis of susceptibility, survival, and bacterial burden in C57BL/6 mice after challenge with high and low virulence ehrlichial strains. A, Sur- vival of B6 mice over 50 days after i.p. infection with a high (4 ϫ 106) or low (4 ϫ 104) dose of IOE (F), a high dose (6 ϫ 108)ofE. muris (f), or EM/IOE (E). EM/IOE are mice primarily infected with a high dose of E. muris and then challenged 4 wk later with a high dose of IOE. The data shown represent one of three independent experiments with a total of 30 mice/ group. B, Quantitative real-time PCR and relative bac- terial load. Ehrlichial DNA from different organs of Downloaded from the following mouse groups was prepared, and a Dsb gene of IOE or E. muris was amplified: group I, IOE amplification from control naive mice, E. muris am- plification from control mice was similar to that of IOE (data not shown); group II, IOE burden in mice infected with high dose IOE on day 7 p.i.; group III, E. http://www.jimmunol.org/ muris burden in mice infected with high dose E. muris on day 7 p.i.; group IV, IOE burden in EM/IOE-in- fected mice on day 7 after rechallenge; group V, E. muris burden in EM/IOE-infected mice on day 7 after rechallenge; and group VI, E. muris burden in mice infected only with E. muris on day 37 p.i. The levels of IOE or E. muris DNA in different tissues were nor- malized to the levels of GAPDH. Data represent the average and SD of triplicate amplifications with three mice per group. Similar results were observed in three by guest on September 29, 2021 independent experiments.

macrophages. The red pulp contained focal areas of moderately areas. The hepatocytes manifested no evidence of injury and were increased numbers of macrophages and plasma cells. similar to hepatocytes in control animals. Most notably, mice in- Of note, mice infected with low dose IOE had less pronounced fected with a high dose of E. muris developed well-defined gran- pathology on day 7 p.i., characterized by widespread moderate ulomas on day 30 p.i. (time point at which these mice were re- hepatic and pulmonary cellular infiltration. On day 14 p.i., shortly challenged with a high dose IOE) in the liver, mainly adjacent to before death, mice developed severe pathological changes similar blood vessels (Fig. 2C) in foci of infiltrates containing plasma cells to those observed on day 7 in mice infected with high dose IOE. and lymphocytes. Development of multiorgan inflammation in mice infected with Immunohistochemical staining of the liver revealed that mice low dose IOE on day 14 was also associated with weight loss and infected with high dose IOE (Fig. 2D), low dose IOE (not shown), hypoglycemia (data not shown). The presence of severe, diffuse or high dose E. muris (Fig. 2E) showed similar distributions of liver necrosis and apoptosis together with weight loss, hypoglyce- morulae containing ehrlichiae. Morulae were located mostly in mia, and elevated serum levels of liver enzymes previously re- cells lining the sinusoidal spaces (Kupffer cells and/or endothelial ported in these animals (12) suggest that the cause of death fol- cells) and in monocytes present in the vascular lumens. Ehrlichiae lowing lethal IOE infection is due to a toxic shock-like syndrome were seldom detected within hepatocytes, suggesting that the liver (7, 26), which is probably mediated by a soluble factor. pathology was independent of direct hepatocellular infection. The In contrast, histopathological examination of the liver in mice relative bacterial burden on day 7 p.i., as judged by immunohis- infected with a high dose of E. muris showed moderate focal ac- tochemical examination, was not markedly different among these cumulation of lymphocytes in the perisinusoidal and perivascular three groups (compare Fig. 2, D and E). 1790 ROLE OF TNF-␣ AND IFN-␥ IN MURINE Ehrlichia INFECTION Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 2. Hepatic and pulmonary histopathology and hepatic immunohistochemical staining. Mice were infected with high dose IOE, high dose E. muris, or EM/IOE, and paraffin sections of liver or lung tissues collected on day 7 p.i. were stained with H&E (AÐC). A, Mice infected with high dose IOE developed extensive hepatic necrosis and apoptosis (arrows). B, Lung pathology shows interstitial pneumonia characterized by extensive infiltration with inflammatory cells and thickened alveolar septa (arrow) compared with normal alveolar wall (arrowhead). C, EM/IOE-infected mice had well-formed granulomas (arrow), especially adjacent to the hepatic blood vessels, consisting of macrophages, lymphocytes, and plasma cells. D, Immunohistochemical staining of the liver of a mouse infected with high dose E. muris revealed numerous morulae in sinusoidal lining cells (arrows). There was moderate focal inflammation with no cell death of the hepatocytes. E, Immunohistochemical staining of the liver of a mouse infected with high dose IOE revealed ehrlichia-infected macrophages within a blood vessel lumen (black arrows). There were scattered foci of necrotic and apoptotic hepatocytes. Liver sections from EM/IOE mice contained few ehrlichiae (data not shown).

Evaluation of Ehrlichia infection in different groups of mice On day 7 p.i., mice infected with high dose IOE had 2- to 3-fold We reasoned that the fatal infection with the IOE strain might be more ehrlichiae in the lung and liver than E. muris-infected mice. due to a greater ability to disseminate and/or replicate in vivo In contrast, the concentrations of IOE and E. muris were compa- compared with the nonlethal E. muris. To examine this possibility, rable in the spleen of infected mice. the number of bacterial genomes in different organs of mice in- Interestingly, compared with naive mice infected with high dose fected with a high dose of IOE, a high dose of E. muris, or EM/IOE IOE, cross-protected EM/IOE-infected mice were able to radically was determined on day 7 p.i./rechallenge using quantitative real- eliminate IOE from spleen, liver, and lung on day 7 after IOE time PCR. IOE and E. muris organisms were equally capable of challenge. Additionally, EM/IOE-infected mice harbored fewer E. dissemination in vivo, with marked tropism to the spleen, liver, muris organisms in all organs than naive mice infected with E. and lung (Fig. 1B). muris only on day 37 p.i. (Fig. 1B). Ehrlichiae were detected in The Journal of Immunology 1791 spleen cell cultures from E. muris-infected mice through day whether lethal IOE infections induced similar levels of IL-10 com- 37 p.i. (Table I). In contrast, Ehrlichia were not cultured from the pared with nonlethal E. muris infection. Lethal infection with high spleens of EM/IOE-infected mice at any time point after IOE chal- dose IOE induced IL-10 to an extent not significantly different lenge (Table I). These data are consistent with our demonstration from that induced by E. muris infection (Fig. 3C). by real-time PCR that the spleens of EM/IOE-infected mice con- tained fewer E. muris than were present in the spleens of naive ϩ mice infected with E. muris on day 37 p.i. (Fig. 1B). The EM/IOE Ag-specific CD4 T cell responses are impaired in susceptible model was used subsequently to define the immunological param- C57BL/6 mice infected with a high dose of IOE eters associated with cross-protection against monocytotropic The studies described above revealed diverse outcomes of disease Ehrlichia spp. in mice infected with different doses and combinations of IOE and E. muris and the relative contributions of TNF-␣ and CD8ϩ T cells ␣ ϩ Roles of TNF- in IOE-induced toxic shock-like syndrome in these infections. To further analyze the roles of CD4 and ϩ The systemic pathology during lethal IOE infection suggested that CD8 T cells and their cytokines in these infections, we first ex- ϩ ϩ tissue damage might be a consequence of overinduction of proin- amined the expansion of CD4 and CD8 cells quantitatively by flammatory cytokines (26). To determine whether infection with FACS. The expansion of total spleen cells (not shown) as well as ϩ ϩ high doses of IOE induce greater systemic levels of TNF-␣ than CD4 and CD8 T cells (Fig. 4A) in the spleens of E. muris-or nonlethal ehrlichial infections, serum levels of TNF-␣ were deter- EM/IOE-infected mice was significantly higher than that after infec- mined at different time points during acute infection. An extremely tion with high dose IOE. Interestingly, prolonged survival of mice high serum level (ϳ900 pg/ml) of TNF-␣ was detected on day infected with low dose IOE was associated with significantly higher Downloaded from ϩ ϩ 7 p.i. in mice infected with high dose IOE (Fig. 3A). During in- expansion of CD4 , but not CD8 , T cells on day 7 p.i. compared fection with the lower dose of IOE, serum TNF-␣ levels were only with that in mice infected with high dose IOE ( p Ͻ 0.005). ϩ moderately increased on day 7 (Fig. 3A). However, 2 days before To assess the Ehrlichia-specific CD4 T cell response, we stim- ϩ death, TNF-␣ was elevated to similar levels during both high and ulated purified CD4 splenic T cells from mice infected with high low dose infections. In contrast, serum levels of TNF-␣ in naive dose E. muris or high dose IOE with the relevant Ag in vitro, 3 mice infected with E. muris or EM/IOE were not significantly respectively, and measured [ H]thymidine incorporation. Prolifer- http://www.jimmunol.org/ ϩ different from those in uninfected mice. ation above background levels was not detected when CD4 T Next we examined whether TNF-␣ is produced by IOE-specific cells from naive control mice were stimulated with either E. muris ϩ CD4 or CD8ϩ T cells, as type 1 cells would be the major source or IOE Ag, whereas CD4 T cells from mice infected with high of TNF-␣ at later stages of infection. TNF-␣-producing, Ehrlichia- dose E. muris proliferated extensively in response to different con- ϩ specific total splenocytes and purified CD4ϩ or CD8ϩ T cells after centrations of E. muris Ags (Fig. 4B). Conversely, CD4 T cells in vitro Ag stimulation were quantified using an ELISPOT assay. from mice previously infected with high dose IOE proliferated The absolute total number of IOE-specific, TNF-␣-producing cells poorly in response to different concentrations of IOE Ags (Fig. ϩ ϩ in the spleen of naive mice infected with high dose IOE was sig- 4B). Naive CD4 T cells or E. muris-specific CD4 T cells ex- nificantly higher than that of E. muris or IOE-specific, TNF-␣- hibited maximal proliferation in vitro upon stimulation with Con A by guest on September 29, 2021 producing cells in naive mice infected with E. muris or in E. muris- or E. muris Ags, respectively, in the presence or the absence of primed mice rechallenged with IOE (EM/IOE; 5.7 ϫ 104, IOE Ags (data not shown). These data suggest that the decreased ϩ compared with 4.4 ϫ 103 and 7.2 ϫ 103), respectively (Fig. 3B). in vitro proliferation of IOE-specific CD4 T cells is not due to the Unstimulated spleen cells did not produce TNF-␣. More intrigu- ability of IOE Ags to suppress Ag nonspecific or specific prolif- ϩ ing, IOE-specific CD8ϩ T cells comprised ϳ60% of the TNF-␣- eration. CD4 T cells isolated from mice infected with low dose of producing cells, whereas no significant production of TNF-␣ was IOE or EM/IOE on day 7 after IOE infection proliferated strongly derived from IOE-specific CD4ϩ T cells (Fig. 3B). in response to in vitro stimulation with an optimum IOE Ag con- ϩ The high serum level of TNF-␣ induced during lethal IOE in- centration (Fig. 4C). The proliferation of CD4 T cells isolated fection could have been due to inadequate stimulation of down- from EM/IOE in response to IOE (Fig. 4C)orE. muris Ag stim- regulatory cytokines such as IL-10. Therefore, we determined ulation (data not shown) was comparable. These data exclude the possibility that the concentration of IOE Ags could be a limiting factor that may account for the impaired in vitro proliferation of CD4ϩ T cells isolated from mice infected with high dose IOE. Table I. Assessment of infectious bacteria in the liver and spleen of These studies demonstrated that a lethal IOE infection results in a a infected mice selective impairment of the CD4ϩ T cell proliferative response in vivo. Bacterial Culture (day)

Ehrlichial Inoculum 7143037 Down-regulation of type 1 response is associated with High dose IOE (10Ϫ3) ND bbb susceptibility to IOE infection Low dose IOE (10Ϫ5) ND ND bb High dose E. muris (10Ϫ1) ϩϩϩϩ To determine whether the Th1/Th2 cytokine profile predicts the EM/IOEc ϪϪϪϪ resistance or susceptibility of a host to infection with monocyto-

a tropic ehrlichiae, we quantified the number of Ehrlichia-specific, Mice (nine mice per group) were infected with the indicated doses of E. muris ␥ or IOE or with a high dose of E. muris before challenge with a high dose of IOE IFN- - and IL-4-producing cells in the spleen of different mice (EM/IOE) and were sacrificed on the indicated days. Splenocytes were cultured for 21 groups using ELISPOT assay. On day 7, splenocytes from mice days and were assessed for the presence of Ehrlichia by Diff-Quik staining and im- resistant to infection with high dose E. muris or cross-protected munofluorescence assay. Data were scored as negative or positive, respectively. ND, an effective method for the cultivation of IOE has yet to be developed. EM/IOE-infected mice contained substantially higher numbers of b All animals died on day 9. Ehrlichia-specific, IFN-␥-producing cells than mice susceptible to c Liver tissues from EM/IOE-infected mice were examined by tissue culture for the presence of both E. muris and IOE. These data are representative of three inde- infection with either a high or a low dose of IOE (Fig. 5A). A pendent experiments. substantial number of IFN-␥-producing spleen cells was detected 1792 ROLE OF TNF-␣ AND IFN-␥ IN MURINE Ehrlichia INFECTION

when splenocytes from all mouse groups, including IOE-infected mice, were stimulated with Con A in the ELISPOT assay (not shown). All groups contained negligible numbers of IOE-specific, IL-4-producing cells by ELISPOT assay (Fig. 5A), and we did not detect significant IL-4 production in the culture supernatant of splenocytes from any infected group by ELISA (data not shown). Next, we determined the expansion of Ag-specific, IFN-␥-pro- ducing T cells per spleen in different mouse groups (Fig. 5B). On day 7 p.i., E. muris-infected mice had ϳ7- and 3-fold more Ag- specific IFN-␥-producing cells (3 ϫ 105/spleen) than mice infected with either high or low dose IOE (4.7 ϫ 104 and 1.0 ϫ 105/spleen, respectively; Fig. 5B). On day 14 p.i. with low dose IOE, the total number of IFN-␥-producing cells in the spleen had decreased to 3.6 Ϯ 2 ϫ 104 cells (data not shown). This significant decrease in the number of IFN-␥-producing cells was concomitantly associ- ated with an increase in serum TNF-␣ on day 14 (Fig. 3A) and a decrease in survival. Similarly, cross-protection of EM/IOE-in- fected mice against lethal IOE rechallenge was associated with

substantial expansion of IOE- and E. muris-specific, IFN-␥-pro- Downloaded from ducing cells (Fig. 5A). Significant production of Ag-specific IFN-␥ by immune spleen cells derived from EM/IOE-infected mice on day 7 after IOE rechallenge was detected by ELISPOT assay in the presence of E. muris or IOE Ags, but not in the absence of specific ehrlichial Ags or in the presence of unrelated Ag such as SRBC

(data not shown). These data suggest that cross-protection of E. http://www.jimmunol.org/ muris-primed mice against lethal IOE challenge is mediated by an Ag-specific response. IFN-␥-producing cells were not detected in the spleen of naive mice stimulated in vitro with either E. muris or IOE. Together, these data suggest that a strong type 1 response is associated with resistance to disease caused by ehrlichiae, whereas a weak type 1 response, but not a Th2 response, is associated with susceptibility to severe ehrlichiosis.

ϩ ϩ CD4 and CD8 T cell responses to infection with different by guest on September 29, 2021 ehrlichial strains To further analyze the contributions of both CD4ϩ and CD8ϩ T cell responses to resistance or susceptibility to these pathogens, splenic CD4ϩ and CD8ϩ T cells were isolated on day 7 and as- sessed for Ag-specific IFN-␥ production by ELISPOT assay. Ap- proximately 9% of IFN-␥-producing cells in the spleens of mice infected with a high dose of IOE were IOE-specific, CD4ϩ T cells (Fig. 6A), whereas 94% of IFN-␥-producing cells in the spleens of E. muris-infected mice were E. muris-specific, CD4ϩ T cells (Fig. 6B). Infection of naive mice with high dose IOE was associated with a 40-fold lower number of IFN-␥-producing CD4ϩ T cells per spleen (2.7 ϫ 105 compared with 6.8 ϫ 103/spleen). In contrast, IOE rechallenge of E. muris-primed mice (EM/IOE) resulted in a high frequency of IOE-specific, IFN-␥-producing CD4ϩ T cells (Fig. 6D). ELISPOT analysis of IFN-␥ production by purified CD8ϩ T cells showed that 88% of IFN-␥-producing cells in the spleens of mice infected with high dose IOE are IOE-specific CD8ϩ T cells (Fig. 6A), whereas only 6% of IFN-␥-producing cells in the FIGURE 3. Serum TNF-␣ level, phenotype of TNF-␣-producing spleen ϩ cells, and IL-10 production by immune splenocytes during infection with spleens of E. muris-infected mice are E. muris-specific CD8 T lethal IOE and nonlethal E. muris. A, High levels of serum TNF-␣ were cells (Fig. 6B). Compared with nonlethal infection with E. muris, detected in mice primarily infected with high or low doses of IOE on days 7 and 14, respectively. No data were available for days 14 and 30 in mice ␣ infected with high dose IOE because no animals survived, and no TNF- a total of nine mice per group. C, Splenocytes from mice infected with high was detected on day 0 (not shown). B, ELISPOT analysis of the number of dose E. muris, high dose IOE, or EM/IOE were isolated on day 7 after ␣ ϩ ϩ TNF- -producing total splenocytes, and purified CD4 and CD8 T cells primary or secondary IOE infection and stimulated in vitro with 10 ␮g/ml from IOE-, E. muris-, and EM/IOE-infected mice. The number of Ag- of the corresponding Ag for 3 days. The levels of IL-10 in culture super- specific TNF-␣ spots was determined by subtracting the number of spots in natant were measured by ELISA. These results represent the mean Ϯ SD wells without Ag. Data represent the average and SD of three mice per of three experiments with nine mice per group. The IL-10 level was not group. Similar results were observed in three independent experiments with significantly different in the three groups ( p ϭ 0.130). The Journal of Immunology 1793

with low dose IOE contained a substantial number of IOE-specific, IFN-␥-producing CD4ϩ T cells on day 7 p.i (8.5 ϫ 104/spleen, 82% of total IFN-␥-producing cells/spleen; Fig. 6C). However, on day 14 p.i. (2–3 days before the death of the animals), the number of IFN-␥-producing CD4ϩ T cells in the low dose IOE-infected mice declined dramatically (5.6 ϫ 103/spleen and 18% of total IFN-␥-producing cells/spleen; Fig. 6C) and was accompanied by significant increase in the total number of Ag-specific, IFN-␥-pro- ducing CD8ϩ type-1 cells (78% of total IFN-␥-producing cells/ spleen) and very high serum level of TNF-␣ (Fig. 3A). The total number of IOE-specific, IFN-␥ producing CD8 T cells per spleen in these mice was 2.9 ϫ 104 on day 7 p.i., which increased to 5.3 ϫ 104/spleen on day 14 p.i. No significant production of IFN-␥ by macrophages was observed in any group of infected mice (data not shown).

Weak Th1 response and IOE lethality are associated with low IL-12 production in IOE-infected mice Downloaded from As the development of a protective IFN-␥ Th1 response after in- fection with several intracellular pathogens is dependent on IL-12 production (27, 28), we asked whether high and low virulence ehrlichial strains differ in their IL-12-inducing capability. IOE

(high virulence) or E. muris (low virulence) Ags were used to http://www.jimmunol.org/ stimulate 18-h thioglycolate-elicited peritoneal macrophages in vitro. As shown in Fig. 7A, the high virulence IOE strain produced a significantly lower level of IL-12p40 in culture supernatant col- lected 6 and 24 h after culture (data not shown) than that produced by low virulence E. muris. A minimal amount of IL-12p40 (360 pg/ml) was produced in supernatants from macrophages cultured with medium alone. To determine the effect of in vivo infection with high or low virulence ehrlichial strain on IL-12 production in the spleen, we by guest on September 29, 2021 analyzed infection-induced IL-12p40 and bioactive IL-12p70 lev- els in IOE-infected (high virulence) or E. muris-infected (low vir- ulence) mice on days 3 and 7 p.i., respectively. IL-12p40 and IL-12 p70 production was significantly impaired in the spleens of mice infected with a high dose of IOE compared with those in- fected with E. muris ( p Ͻ 0.05; Fig. 7, B and C). In addition, cross-protected, EM/IOE-infected mice produced a high level of IL-12 on day 7 after IOE rechallenge (Fig. 7, B and C). Interestingly, in vivo the serum IL-12p40 response on day 7 p.i.

ϩ after i.p. infection with IOE was significantly lower than that de- FIGURE 4. CD4 T cell responses to IOE and E. muris. A, The total tected in the serum of mice infected i.p. with E. muris or LPS (LPS number of CD4ϩ and CD8ϩ T cells in the spleens of different groups of stimulation, 118 Ϯ 1.8 ng/ml; E. muris infection, 106 Ϯ 8.4 ng/ml; mice on day 7 p.i., as measured by flow cytometry. The absolute number Ϯ of CD4ϩ and CD8ϩ T cells determined by FACS is based on comparison IOE infection, 28 6.6 ng/ml). Taken together, these results sug- with isotype control and cells present in the lymphocyte gate. The total gested that the defective development of Ag-specific, IFN-␥-pro- ϩ number of CD4ϩ and CD8ϩ T cells per spleen was calculated by compar- ducing CD4 Th1 cells in IOE-infected mice could be due to a ing the number of cells measured by FACS to the total number of spleen deficiency in IL-12 production. cells determined by trypan blue exclusion. B, CD4 T cells isolated from mice infected with a high dose of E. muris or IOE on day 7 p.i. were stimulated in vitro with different concentrations of the corresponding Ag. Substantial production of Ehrlichia-specific IgG2a is associated ϩ C, Ag-specific CD4 T cell proliferation. CD4 T cells isolated from the with cross-protection against a lethal dose of IOE in EM/IOE- indicated groups of mice were cultured in vitro in the absence or the pres- infected mice ence of the relevant Ag (25 ␮g/ml). CD4 cells from EM/IOE mice were stimulated with E. muris (not shown) or IOE. Data represent the mean Ϯ To determine the role of Abs in the control of Ehrlichia infection, SD of three mice per group in one experiment. Similar results were ob- we examined the titer and isotypes of serum anti-Ehrlichia Abs in served from three independent experiments. all infected groups of mice. Primary infection with either nonlethal E. muris or lethal IOE induced a high Ag-specific IgM titer by day 7 p.i., which declined in E. muris-infected mice on day 14 and was infection with high dose IOE was associated with a 2-fold greater undetectable by day 30 (Table II). The E. muris-specific IgG re- total number of IFN-␥-producing CD8ϩ T cells per spleen (2.6 ϫ sponse appeared on day 30 and had increased on day 60. All the 104 compared with 4.2 ϫ 104/spleen). Moreover, mice infected EM/IOE-infected mice developed a very high titer (1/4096) of 1794 ROLE OF TNF-␣ AND IFN-␥ IN MURINE Ehrlichia INFECTION

FIGURE 5. Frequency of Ag-specific IFN-␥-or IL-4-producing cells in infected mice. A, On day 7 after primary infection or rechallenge, splenocytes were collected from mice infected with high dose E. muris (group I), high dose IOE (group II), low dose IOE (group III), EM/IOE splenocytes stimulated in Downloaded from vitro with IOE Ags (group IV), or EM/IOE spleno- cytes stimulated in vitro with E. muris Ags (group V). Splenocytes from different mouse groups were stimu- lated in the ELISPOT assay with the relevant Ag (20 ␮g/ml) or were incubated with medium alone in the presence of naive syngeneic splenocytes (1 ϫ 106/ ␥ well). The number of Ag-specific, IFN- -producing http://www.jimmunol.org/ cells or IL-4-producing cells per 106 cells was deter- mined by subtracting the number of spots in wells without Ag. B, Total number of Ehrlichia-specific, IFN-␥-producing cells per spleen. Data represent the average and SD of four mice per group. Similar results were observed in three independent experiments, with a total of 12 mice/group. by guest on September 29, 2021

Ehrlichia-specific IgG on day 7 after IOE challenge. The level of -infected donors (data not shown) or naive CD4ϩ and CD8ϩ T IgG Abs increased further thereafter. Analysis of the IgG isotype cells plus polyclonal anti-Ehrlichia Abs (group II) succumbed to response showed that EM/IOE-infected mice have a high level of infection (Fig. 8A). All these groups of mice mounted a weak type-I Ehrlichia-specific IgG2a and a low level of Ehrlichia-specific response, which was indistinguishable from that induced in control IgG1 on days 7, 14, 30, and 60 after challenge with IOE (Fig. 7C). mice that received either PBS (group I; Fig. 8B), or naive CD4 and ϩ ϩ ϩ CD8 T cells and normal serum (data not shown). Transfer of E. Adoptive transfer of E. muris-primed CD4 and CD8 T cells muris-specific CD4ϩ and CD8ϩ T cells without (group III) or with provided partial protection against lethal IOE infection (group IV) polyclonal serum or EM/IOE-specific CD4ϩ and CD8ϩ To further determine what components of the immune response T cells without polyclonal serum (group V) resulted in the gener- induced during ehrlichial infection contribute to protection against ation of a significantly higher number ( p Ͻ 0.005) of IFN-␥-pro- lethal IOE infection, we adoptively transferred Ehrlichia-specific, ducing splenic Th1 cells compared with that in controls (Fig. 8B). IFN-␥-producing, CD4ϩ or CD8ϩ type 1 cells derived from either Although these mice survived longer than controls, they suc- E. muris (E. muris-specific) or EM/IOE-infected mice (EM/IOE- cumbed to infection at later time points (Fig. 8A). In contrast, mice specific) with or without polyclonal anti-Ehrlichia serum. Control that received both EM/IOE-specific CD4ϩ and CD8ϩ T cells iso- mice were transferred with naive CD4ϩ cells, CD8ϩ T cells, lated from EM/IOE-infected donors with anti-Ehrlichia polyclonal and/or normal mouse serum. Recipients of Ehrlichia-specific Abs (group VI) 1 day before infection generated a significantly CD4ϩ or CD8ϩ T cells isolated from day 7 E. muris or EM/IOE- higher number ( p Ͻ 0.001) of IFN-␥-producing cells than mice The Journal of Immunology 1795 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 6. Ehrlichia-specific IFN-␥ production by different T cell subsets. Mice were infected with high dose IOE (A), high dose E. muris (B), low dose IOE (C), or EM/IOE (D). On day 7, splenocytes were prepared for isolation of CD4ϩ, CD8ϩ, and CD11bϩ macrophages. Cells were cultured with the relevant Ag preparations. The number of Ag-specific, IFN-␥-producing cells was determined by subtracting the number of spots in wells without Ag. C, The number of IOE-specific, IFN-␥ spots produced by splenic CD4ϩ, CD8ϩ, and CD11bϩ cells from mice infected with a low dose of IOE was determined on days 7 and 14 p.i. These results represent the average Ϯ SD of three mice per group. Similar results were observed in three independent experiments. Approximately 10–30 spots were detected in CD11bϩ macrophages from all groups (not shown). transferred with EM/IOE-specific CD4 and CD8 T cells only which was followed by an increasing titer of anti-Ehrlichia IgG on (group V; p Ͻ 0.005; Fig. 8B) and survived until day 30 p.i. (Fig. day 14 (Table II). These results suggest that both EM/IOE-specific 8A). The previous group of mice (group VI) also developed a high CD4ϩ and CD8ϩ T cells and EM/IOE-specific Abs provide effec- titer of anti-Ehrlichia IgM Abs at early time points after infection, tive components for the protective responses.

Table II. Production of Ehrlichia-specific Abs during the course of infectiona

IgM IgG

Mice Day 7 Day 14 Day 30 Day 7 Day 14 Day 30 Day 60

High dose E. muris 1,024 64 Ͻ32 Ͻ32 Ͻ32 32 256 High dose IOE 1,024 bbϽ32 bbb Low dose IOE 128 64 b Ͻ32 Ͻ32 bb Adoptively transferred 1,024 1,024 Ͻ32 Ͻ32 64 256 b EM/IOE 64 64 64 4,096 8,192 8,192 16,384

a Mice were infected with the indicated doses of E. muris or IOE, as described in Table I, adoptively transferred with E. muris-immune CD4ϩ and CD8ϩ cells before challenge with high dose of IOE, or primed with high dose E. muris before challenge with a high dose of IOE (EM/IOE). Mice were sacrificed on the indicated days, and Ehrlichia-specific Abs were determined by immunofluorescence assay. P388D1 cells infected with E. muris were used as homologous or surrogate Ag for E. muris and IOE, respectively. Data represent the mean of Ab titers from individual mice. These data are representative of two independent experiments with nine mice per group. b Not determined due to the death of the animals before the time point. 1796 ROLE OF TNF-␣ AND IFN-␥ IN MURINE Ehrlichia INFECTION Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 7. A, Production of IL-12 by 18 h thioglycolate-elicited peritoneal macrophages and splenocytes in response to different ehrlichial strains. Thioglycolate-elicited peritoneal macrophages were stimulated with the indicated concentration of IOE (high virulence) or E. muris (low virulence) ehrlichial Ags. Cell-free supernatants were harvested for IL-12 cytokine measurement 6 h after culture initiation. Results are expressed as the mean Ϯ SD and are representative of three independent experiments. A minimal amount of IL-12 was detected upon culture of peritoneal macrophages with medium only (not shown). Levels of IL-12 were statistically significantly different between groups (p Ͻ 0.005). B, Mice were i.p. infected with high dose IOE or high dose E. muris. On days 3 and 7 p.i., spleen cells were isolated and cultured in medium (med), IOE, or E. muris Ag preparations for 72 h. Supernatant were assayed for IL-12 p40 (B) and Il-12p70 (C) production by ELISA. Values represent the mean Ϯ SD of three experiments with six mice per group. D, Ab responses in cross-protected, EM/IOE-infected mice. Mice were infected with high dose E. muris and then challenged with high dose IOE. Sera were collected at the indicated time points and pooled (three mice per group). The levels of Ag-specific IgG1 and IgG2a isotypes were measured by ELISA. These results are expressed as the mean ϮSD of triplicate wells. The data are representative of three independent experiments. The Journal of Immunology 1797

FIGURE 8. Enhanced survival and Th1/Th2 cyto- kine responses of mice adoptively transferred with CD4ϩ and CD8ϩ T cells isolated from EM/IOE-in- fected mice with polyclonal anti-Ehrlichia serum after challenge with a lethal dose of IOE. Groups of mice were adoptively transferred i.p. with CD4ϩ and CD8ϩ T cells isolated from mice infected with high dose E. muris (E. muris-specific T cells) on day 7 p.i., without (group III) or with (group IV) 1 ml of polyclonal anti- ϩ ϩ Ehrlichia serum, or with CD4 and CD8 T cells iso- Downloaded from lated from cross-protected EM/IOE (EM/IOE-specific T cells) on day 7 after IOE rechallenge, without (group V) or with (group VI) 1 ml of polyclonal anti- Ehrlichia serum. Control mice were injected with PBS (group I) or with naive CD4ϩ and/or CD8ϩ T cells with polyclonal anti-Ehrlichia serum (group II). All groups were infected 1 day later with high dose IOE. http://www.jimmunol.org/ A, Survival was monitored for 30 days. B, Splenocytes collected on day 7 after infections were assessed for IFN-␥- and IL-4-producing cells by ELISPOT assay. The results are expressed as the mean Ϯ SD of nine mice per group. Data are representative of two inde- pendent experiments. by guest on September 29, 2021

Discussion cells suggests that IFN-␥ production must come from CD4 T cells In this study we examined the immunological basis for suscepti- to be effective, and that IFN-␥ production by CD8 T cells is bility and resistance to Ehrlichia infection using two closely re- insufficient to control the infection. ϩ lated Ehrlichia spp.: the highly virulent IOE and the mildly viru- Second, overproduction of TNF-␣ by Ag-specific splenic CD8 lent E. muris, representing fatal and mild models of ehrlichiosis, T cells and very high levels of serum TNF-␣ are strongly associ- respectively. Most notably, we showed that priming with E. muris ated with severe and fatal disease. After primary infection, E. protected mice against lethal disease caused by rechallenge with a muris and IOE did not differ substantially in their ability to dis- high dose of IOE. Our study has, for the first time, revealed two seminate in vivo or to reach similar tissues (Fig. 1B). Therefore, critical factors determining susceptibility to severe ehrlichiosis. the ability of IOE to cause lethal infections in mice was not due First, marked impairment of the CD4ϩ T cell response, including merely to the direct effects of the bacteria. These data are consis- reduced expansion and Ag-specific proliferation, and a low num- tent with previous studies showing a disparity between ehrlichial ber of IFN-␥-producing Th1 cells are strongly associated with quantity and the severity of disease in immunocompetent patients susceptibility to fatal disease. Mice infected with a high dose of infected with E. chaffeensis (2, 3). However, a key property of the nonlethal E. muris produced a 75-fold greater number of total Ag- IOE Ehrlichia infection is the dramatic increase in serum TNF-␣ specific, IFN-␥-producing CD4ϩ T cells than mice infected with a levels and the number of Ag-specific, TNF-␣-producing CD8 T high dose of lethal IOE. Therefore, the contribution of CD4 T cells cells in the spleen after high or low dose challenge. to IFN-␥ production was dramatically decreased in IOE-infected Minimal production of TNF-␣ by CD8 T cells in mice infected mice. IL-12 production by macrophages, which is characteristic of only with nonlethal E. muris or cross-protected EM/IOE mice was a Th1 response (27), was also deficient on days 3 and 7 p.i. in associated with mild disease and long term survival, further sub- IOE-infected mice. The fact that mice infected with high dose IOE stantiating the contribution of TNF-␣ production by CD8 T cells to succumbed to infection despite production of IFN-␥ by CD8ϩ T disease pathogenesis (29–32). In support of our results, a previous 1798 ROLE OF TNF-␣ AND IFN-␥ IN MURINE Ehrlichia INFECTION study showed that CD8ϩ T cell knockout mice were less suscep- ficient stimulation of IL-12p40 and IL-12p70 in vivo. Moreover, tible than wild-type or CD4ϩ T cell knockout mice to infection thioglycolate-elicited macrophages were impaired in their ability with Ehrlichia ruminantum (another tick-transmitted Ehrlichia to produce IL-12 upon in vitro stimulation with IOE Ags compared strain that is very closely related to other monocytotropic Ehrli- with stimulation with E. muris Ags (Fig. 7A). As the host target ϩ chia, including IOE). In this study 50% of the CD8 T cell knock- cells for monocytotropic ehrlichiae, including IOE and E. muris, out mice survived infection, whereas the other half died after a are macrophages and monocytes, our data suggest that decreased prolonged period (33). IL-12 production by IOE-infected macrophages in vivo could be Decreased TNF-␣ production in EM/IOE cross-protected mice responsible for inefficient generation of IFN-␥-producing CD4ϩ could be attributed to efficient containment of IOE at early stages Th1 cells. Nevertheless, on day 7 p.i., spleen cells from susceptible of infection by an existing immune response against cross-reactive IOE-infected mice synthesized less IL-12p40 and IL-12p70 than E. muris Ags. The efficient decrease in bacterial burden in EM/ cells from resistant E. muris-infected mice. The latter observation IOE-infected mice at an early time point after infection may ac- could be due to weak IFN-␥ production by CD4ϩ Th1 cells, where count for the substantial generation of IOE-specific, IFN-␥-pro- IFN-␥ plays an important role in enhancing the induction of IL-12 ϩ ϩ ducing CD4 Th1 cells (Fig. 6D). Efficient generation of CD4 during infection with some intracellular pathogens (42). Th1 lymphocytes may play a regulatory role in controlling the CD4ϩ T cells have been believed to be important in protection ϩ generation of Ag-specific TNF-␣-producing CD8 T cells. against other infections, such as Mycobacterium tuberculosis (42), The elevated levels of systemic TNF-␣ preceding death together Plasmodium spp. (43), and Toxoplasma gondii (44). Ehrlichiae are with weight loss and hypoglycemia in lethal ehrlichiosis in IOE- obligately intracellular pathogens that replicate within host mac- infected mice resemble toxic shock-like syndrome caused by in- rophages. Therefore, macrophages activated by IFN-␥-producing Downloaded from fections with Gram-negative bacterial pathogens. Although genetic CD4 T cells are essential for limiting ehrlichial infection. In sup- and molecular characterization of E. muris and IOE has shown that port of this conclusion, the presence of granulomas in the liver these strains are very closely related (13–17), the basis for the (Fig. 2C) of EM/IOE-infected mice was associated with substantial dramatic difference in virulence between IOE and E. muris is pres- elimination of both IOE and E. muris organisms from tissues (Fig. ently unknown. Ehrlichia spp. differ from these prototypical 1B). Macrophages activated by IFN-␥ produce reactive nitrogen Gram-negative bacterial pathogens in the absence of LPS (34, 35). intermediates and kill intracellular pathogens (45–47). In vitro http://www.jimmunol.org/ Nevertheless, the massive inflammatory response that occurs dur- studies also suggest that IFN-␥-mediated activation of human mac- ing IOE infections could be due to host responses to a substance rophages controls Ehrlichia by limiting available cytoplasmic iron produced by the organism, analogous to the inflammatory response (48, 49). Therefore, an obvious conclusion from our studies is that generated by bacterial LPS (yet clearly distinct) (36, 37), or un- the primary role of CD4ϩ T cells in protection against ehrlichiosis ␣ ϩ controlled induction and production of TNF- -producing CD8 T is early production of IFN-␥. However, it is likely that early IFN-␥ cells. production is not the only role for CD4 T cells in protection against Our data exclude the possibility that overproduction of TNF-␣ is ϩ ehrlichiosis. due to a strong/uncontrolled CD4 Th1 response (38), because the ϩ ϩ

The effect of CD4 T cells on CD8 T cell development and by guest on September 29, 2021 level of IL-12 in serum and spleen and the number of IFN-␥ pro- ϩ ϩ function must also be taken into account. One role for CD4 T ducing CD4 Th1 cells were lower in IOE-infected mice than in ϩ cells in eliciting an effective and protective CD8 T cell response E. muris-infected mice. The finding that mice infected with high ϩ is the production of IL-2 (50). Additionally, CD4 T cells help dose IOE have very high serum levels of TNF-␣, but low IL-12 ϩ priming of effective cytotoxic CD8 T cells. Several studies have levels may in part be explained by the inhibitory effect of TNF-␣ ϩ classified the CD8 response to acute infections with intracellular on IL-12 production (39). Although lethal IOE infection was not pathogens such as M. tuberculosis (42), T. gondii (44), and Plas- associated with significant decrease in IL-10 production compared modium spp. (43) as CD4ϩ helper-dependent. CD4ϩ T cells are with that produced by E. muris-infected mice (Fig. 3C), one cannot required to activate APCs through CD40 signaling, licensing the exclude the possibility that the level of IL-10 in IOE-infected mice APC to stimulate a full-blown CD8 response. In contrast, other may be inadequate to down-regulate the pathological overproduc- infectious agents, such as the intracellular Gram-positive pathogen tion of TNF-␣ (40, 41). ϩ Listeria monocytogenes, which carries a plethora of immunostimu- This study shows that a weak CD4 Th1 response is associated latory signals (such as cell wall components, flagellin, and CpG with susceptibility to infection after infection with high dose IOE. DNA), activate APCs directly, thereby bypassing the need for CD4 As mice infected with a low dose of lethal IOE contained a sub- ϩ stantially higher number of IFN-␥-producing CD4ϩ T cells in the help (51, 52). Taken together, one can envisage that if CD4 Th1 spleen on day 7 p.i. than the high dose IOE recipients, and these cell differentiation were impaired early in the response as a result IFN-␥-producing CD4ϩ cells dramatically diminished on day of several factors such as cytokine environment, costimulatory sig- 14 p.i. (2 days before death), we hypothesize that the presence of nals from APCs, the amount of Ag, and the affinity of the epitope- ␥ ϩ TCR interactions (53), this could subsequently affect the induction the lower number of IFN- -producing CD4 T cells in the spleen ϩ of mice infected with high dose IOE on day 7 p.i. might be due to of effective CD8 CTL by dendritic cells or other APCs. This apoptotic cell death of the splenic CD4ϩ Th1 cells after infection situation might explain the failure of mice infected with high dose ␥ with high dose IOE. In support of this conclusion, we have dem- IOE to control infection despite significant generation of IFN- - producing CD8ϩ T cells. onstrated previously that the splenic pathology in mice infected ϩ with a high dose of IOE was associated with extensive apoptosis, The generation of a strong CD4 Th1 cell response in the pres- ␥ ϩ as detected by the TUNEL assay (12). ence of few IFN- -producing CD8 T cells in mice infected with ϩ The synthesis of IL-12 and IFN-␥ is often coordinately regu- E. muris suggests that the CD4 T cells themselves are sufficient lated, a finding that may relate to the ability of IFN-␥ to enhance to contain infection with the less virulent Ehrlichia spp. In con- IL-12 production and vice versa (42). Our data suggest that the trast, our adoptive transfer experiments and protection of E. muris- weak CD4ϩ Th1 response after lethal IOE challenge could be due primed animals against lethal rechallenge with IOE suggest that to an early decrease in IL-12 production. As shown in Fig. 7, B and both CD4ϩ and CD8ϩ type 1 cells producing high levels of IFN-␥, C, infection of C57BL/6 mice with high dose IOE results in inef- but not TNF-␣, contribute to protection against lethal ehrlichiosis The Journal of Immunology 1799 caused by IOE. These observations led us to hypothesize that the 7. Fichtenbaum, C. J., L. R. Peterson, and G. J. Weil. 1993. Ehrlichiosis presenting ϩ as a life-threatening illness with features of the toxic shock syndrome. absence of a strong CD4 Th1 response could result in not only Am. J. Med. 95:351. inefficient control of lethal IOE infection, but also failure in con- 8. Winslow, G. M., E. Yager, K. Shilo, D. N. Collins, and F. K. Chu. 1998. Infection trolling the induction of pathogenic TNF-␣-producing CD8ϩ T of the laboratory mouse with the intracellular pathogen Ehrlichia chaffeensis. Infect. Immun. 66:3892. cells. 9. Li, J. S., E. Yager, M. Reilly, C. Freeman, G. R. Reddy, A. A. Reilly, F. K. Chu, and M. Winslow. 2001. Outer membrane protein-specific monoclonal antibodies Role of Ab response against infection with an obligatory protect SCID mice from fatal infection by the obligate intracellular bacterial intracellular pathogen pathogen Ehrlichia chaffeensis. J. Immunol. 166:1855. 10. Ganta, R. R., M. J. Wilkerson, C. Cheng, A. M. Rokey, and S. K. Chapes. 2002. It is usually thought that Abs play a minor role in host defense Persistent Ehrlichia chaffeensis infection occurs in the absence of functional ma- against obligately intracellular pathogens because their lifestyle jor histocompatibility complex class II genes. Infect. Immun. 70:380. 11. Shibata, S., M. Kawahara, Y. Rikihisa, H. Fujita, Y. Watanabe, C. Suto, and renders them generally inaccessible to Abs. However, our data T. Ito. 2000. New Ehrlichia species closely related to Ehrlichia chaffeensis iso- show that complete protection of E. muris-primed mice against lated from Ixodes ovatus ticks in Japan. J. Clin. Microbiol. 38:1331. lethal challenge with high dose IOE is positively associated with 12. Sotomayor, E. A., V. L. Popov, H-M. Feng, D. H. Walker, and J. P. Olano. 2001. Animal model of fatal human monocytotropic ehrlichiosis. Am. J. Pathol. production of substantially high levels of Ehrlichia-specific Th1 158:757. isotype (IgG2a) Abs at all time points after infection. Furthermore, 13. Kawahara, M., T. Ito, C. Suto, S. Shibata, Y. Rikihisa, K. Hata, and K. Hirai. in these experiments adoptive transfer of E. muris-specific Abs 1999. Comparison of Ehrlichia muris strains isolated from wild mice and ticks ␥ ϩ ϩ and serologic survey of humans and animals with E. muris as antigen. J. Clin. together with IFN- -producing CD4 and CD8 T cells, but not Microbiol. 37:1123. alone, significantly increased the survival of naive mice challenged 14. Sumner, J. W., G. A. Storch, R. S. Buller, A. M. Liddell, S. L. Stockham, with high dose IOE (Table II). The mechanism by which IgG2a Y. Rikihisa, S. Messenger, and C. D. Paddock. 2000. PCR amplification and Downloaded from phylogenetic analysis of groESL operon sequences from Ehrlichia ewingii and mediates protection is most likely via FcR-enhanced opsonization Ehrlichia muris. J. Clin. Microbiol. 38:2746. (54), because the generation of a substantial Ag-specific IgM re- 15. Shibata, S., M. Kawahara, Y. Rikihisa, H. Fujita, Y. Watanabe, C. Suto, and sponse in IOE-infected mice, which is an effective component of T. Ito. 2000. New Ehrlichia species closely related to Ehrlichia chaffeensis iso- lated from Ixodes ovatus ticks in Japan. J. Clin. Microbiol. 38:1331. complement-mediated lysis, was not protective. These data are 16. Yu, X. J., X. F. Zhang, J. W. McBride, Y. Zhang, and D. H. Walker. 2001. consistent with previous studies showing that anti-E. chaffeensis Phylogenetic relationships of Anaplasma marginale and ‘Ehrlichia platys’ to other Ehrlichia species determined by GroEL amino acid sequences. Int. J. Syst. Abs confer a level of protective immunity in SCID mice (9, 55). Evol. Microbiol. 51:1143. http://www.jimmunol.org/ In summary, we have shown that susceptibility to severe ehrli- 17. McBride, J. W., X. J. Yu, and D. H. Walker. 2000. A conserved, transcriptionally chiosis is strongly correlated with two interrelated immunopatho- active p28 multigene locus of Ehrlichia canis. Gene 254:245. logical factors: first, overproduction of TNF-␣ production by IOE- 18. Ohashi, N., Y. Rikihisa, and A. Unver. 2001. Analysis of transcriptionally active ϩ gene clusters of major outer membrane protein multigene family in Ehrlichia specific CD8 T cells and a systemic TNF-␣-mediated canis and E. chaffeensis. Infect. Immun. 69:2083. inflammatory response; and second, down-regulation of Ag-spe- 19. Alleman, A. R., A. F. Barbet, M. V. Bowie, H. L. Sorenson, S. J. Wong, and M. Belanger. 2000. Expression of a gene encoding the major antigenic protein 2 cific CD4 T cell proliferation and Th1 differentiation. More im- homolog of Ehrlichia chaffeensis and potential application for serodiagnosis. portant, an IFN-␥, but not TNF-␣-producing, CD4 and CD8 type J. Clin. Microbiol. 38:3705. 1 response and a Th1-Ab response are critical for protection 20. Alleman, A. R., L. J. McSherry, A. F. Barbet, E. B. Breitschwerdt,

H. L. Sorenson, M. V. Bowie, and M. Belanger. 2001. Recombinant major an- by guest on September 29, 2021 against rechallenge with the lethal ehrlichial strain. The ability to tigenic protein 2 of Ehrlichia canis: a potential diagnostic tool. J. Clin. Microbiol. induce cross-protective immunity against heterogeneous ehrlichial 39:2494. strains might be useful as a vaccine strategy for the control of 21. McBride. J. W., R. E. Corstvet, S. D. Gaunt SD, C. Boudreaux, T. Guedry, and D. H. Walker. 2003. Kinetics of antibody response to Ehrlichia canis immuno- Ehrlichia infection in humans and animals. reactive proteins. Infect. Immun. 71:2516. Our data strongly indicate that CD4 and CD8 T cells are per- 22. Gunzer, M., C. Weishaupt, L. Planelles, and S. Grabbe. 2001. Two-step negative ϩ ϩ forming at least some different functions during ehrlichial infec- enrichment of CD4 and CD8 T cells from murine spleen via nylon wool adherence and an optimized antibody cocktail. J. Immunol. Methods 258:55. tion, and studies to further define these functions are essential to 23. Favre, N., G. Bordmann, and W. Rudin. 1997. Comparison of cytokine measure- our understanding of this disease and the most effective means for ments using ELISA, ELISPOT and semi-quantitative RT-PCR. J. Immunol. vaccination. A better understanding of why IOE and E. muris in- Methods 204:57. 24. Power, C. A., C. L. Grand, N. Ismail, N. C. Peters, D. P. Yurkowski, and fections trigger totally different types of responses in the same host P. A. Bretscher. 1999. A valid ELISPOT assay for enumeration of ex vivo, and how these divergent responses are regulated would lead to antigen-specific, IFN-␥-producing T cells. J. Immunol. Methods 227:99. more efficient control of the infection and treatment of severely 25. Huong, P. L., A. H. Kolk, T. A. Eggelte, C. P. Verstijnen, H. Gilis, and J. T. Hendriks. 1991. Measurement of antigen specific lymphocyte proliferation infected patients. using 5-bromo-deoxyuridine incorporation: an easy and low cost alternative to radioactive thymidine incorporation. J. Immunol. Methods 140:243. Acknowledgments 26. Hack, C. E., L. A. Aarden, and L. G. Thijs. 1997. Role of cytokine in sepsis. Adv. Immunol. 66:101. We express our gratitude to Kelly Cassity and Susan Butler for secretarial 27. Park, A. Y., B. D. Hondowicz, and P. Scott. 2000. IL-12 is required to maintain expertise in the preparation of this manuscript. a Th1 response during Leishmania major infection. J. Immunol. 165:896. 28. Sypek, J. P., C. L. Chung, S. E. Mayor, J. M. Subramanyam, S. J. Goldman, References D. S. Sieburth, S. F. Wolf, and R. G. Schaub. 1993. Resolution of cutaneous 1. Olano, J. P., and D. H. Walker. 2002. Human ehrlichioses. Med. Clin. North Am. leishmaniasis: interleukin-12 initiates a protective T helper type 1 immune re- 86:375. sponse. J. Exp. Med. 177:1797. 2. Walker, D. H., and J. S. Dumler. 1997. Human monocytotic and granulocytic 29. Leist, M., F. Gantner, I. Bohlinger, G. Tiegs, P. Germann, and A. Wendel. 1995. ehrlichioses: discovery and diagnosis of emerging tick-borne infections and the Tumor necrosis factor-induced hepatocyte apoptosis precedes liver failure in ex- critical role of the pathologist. Arch. Pathol. Lab. Med. 121:785. perimental murine shock models. Am. J. Pathol. 146:1220. 3. Dumler, J. S., and D. H. Walker. 2001. Tick-borne ehrlichiosis. Lancet 21:478. 30. Leist, M., F. Gantner, I. Bohlinger, P. G. Germann, G. Tiegs, and A. Wendel. 4. Paddock, C. D., D. P. Suchard, K. L. Grumbach, W. K. Hadley, 1994. Murine hepatocyte apoptosis induced in vitro and in vivo by TNF-␣ re- R. L. Kerschmann, N. W. Abbey, J. E. Dawson, B. E. Anderson, K. G. Sims, quires transcriptional arrest. J. Immunol. 153:1778. J. S. Dumler, et al. 1993. Brief report: fatal seronegative ehrlichiosis in a patient 31. Bradham, C., J. Plumpe, M. Manns, D. Brenner, and C. Trautwein. 1998. Mech- with HIV infection. N. Engl. J. Med. 329:1164. anisms of hepatic toxicity: TNF-induced liver injury. Am. J. Physiol. 275:G387. 5. Fishbein, D. B., J. E. Dawson, and L. E. Robinson. 1994. Human ehrlichiosis in 32. Anthony, J. M., B. L. Rosa, Y. V. Gessel, A. Alcaraz, S. K. Bliss, E. J. Pearce, the United States, 1985 to 1990. Ann. Intern. Med. 120:736. and E. Y. Denkers. 1999. Toxoplasma gondii and Schistosoma mansoni synergize 6. Childs, J. E., J. McQuiston, J. W. Summer, J. A. Comer, R. E. Massung, to promote hepatocyte dysfunction associated with high levels of plasma TNF-␣ S. M. Standaert, and C. D. Paddock. 1994. Human monocytic ehrlichiosis due to and early death in C57BL/6 mice. J. Immunol. 163:2089. Ehrlichia chaffeensis: how do we count the cases? In Rickettsiae and Rickettsial 33. Byrom, B., A. F. Barbet, M. Obwolo, and S. M. Mahan. 2000. CD8ϩ T cell Diseases at the Turn of the Third Millenium. D. Raoult and P. Brouqui, eds. knockout mice are less susceptible to Cowdria ruminantium infection than athy- Elsevier, Paris, p. 287. mic, CD4ϩ T cell knockout, and normal C57BL/6 mice. Vet. Parasitol. 93:159. 1800 ROLE OF TNF-␣ AND IFN-␥ IN MURINE Ehrlichia INFECTION

34. Rikihisa, Y. 1991. The tribe Ehrlichieae and Ehrlichial diseases. Clin. Microbiol. 45. Ismail, N., P. Olano, H. M. Feng, and D. H. Walker. 2002. Current status of Rev. 4:286. immune mechanisms of killing of intracellular microorganisms. FEMS Microb. 35. Dasch G. A., E. Weiss, J. C. Williams. 1990. Antigenic properties of the ehrli- Lett. 207:111. chiae and other rickettsiae. In Ehrlichiosis: A Vector Borne Illness of Animals 46. Unanue, E. R. 1997. Inter-relationships among macrophages, natural killer cells And Humans. J. C. Williams and I. Kakoma, eds. Kluwer Academic, Boston, and neutrophils in early stages of Listeria resistance. Curr. Opin. Immunol. 9:35. p. 323. 47. Feng, H. M., V. L. Popov, and D. H. Walker. 1994. Depletion of ␥ interferon and 36. Miethke, T., C. Wahl, K. Heeg, B. Echtenacher, P. H. Krammer, and H. Wagner. ␣ 1992. T cell-mediated lethal shock triggered in mice by the superantigen staph- tumor necrosis factor in mice with Rickettsia conorii-infected endothelium: ylococcal enterotoxin B: critical role of tumor necrosis factor. J. Exp. Med. impairment of rickettsicidal nitric oxide production resulting in fatal, overwhelm- 175:91. ing rickettsial disease. Infect. Immun. 62:1952. 37. Henderson, B., S. Poole, and M. Wilson. 1996. Bacterial modulins: a novel class 48. Barnewall, R. E., and Y. Rikihisa. 1994. Abrogation of ␥ interferon-induced of virulence factors which cause host tissue pathology by inducing cytokine syn- inhibition of Ehrlichia chaffeensis infection in human monocytes with iron-trans- thesis. Microbiol. Rev. 60:316. ferrin. Infect. Immun. 62:4804. 38. Mordue, D. G., F. Monroy, M. La Regina, C. A. Dinarello, and L. D. Sibley. 49. Barnewall, R. E., N. Ohashi, and Y. Rikihisa. 1999. Ehrlichia chaffeensis and 2001. Acute toxoplasmosis leads to lethal overproduction of Th1 cytokines. E. sennetsu, but not the human granulocytic ehrlichiosis agent, colocalize with J. Immunol. 167:4574. transferrin receptor and up-regulate transferrin receptor mRNA by activating 39. Hodge-Defour, J., M. W. Marino, M. R. Horton, A. Jungbluth, M. D. Burdick, iron-responsive protein. Infect. Immun. 67:2258. R. M. Strieter, P. W. Noble, C. A. Hunter, and E. Pure. 1998. Inhibition of 50. Rahemtulla, A., W. P. Fung-Leung, M. W. Schilham, T. M. Kundig, interferon ␥ induced interleukin 12 production: a potential mechanism for the S. R. Sambhara, A. Narendran, A. Arabian, A. Wakeham, C. J. Paige, and anti-inflammatory activities of tumor necrosis factor. Proc. Natl. Acad. Sci. USA ϩ R. M. Zinkernagel. 1991. Normal development and function of CD8 cells but 95:13806. markedly decreased helper cell activity in mice lacking CD4. Nature 353:180. 40. Gazzinelli, R. T., M. Wysocka, S. Hieny, T. Scharton-Kersten, A. Cheever, R. Kuhn, W. Muller, G. Trinchieri, and A. Sher. 1996. In the absence of endog- 51. Shedlock, D. J., J. K. Whitmire, J. Tan, A. S. MacDonald, R. Ahmed, and enous IL-10, mice acutely infected with Toxoplasma gondii succumb to a lethal H. Shen. 2003. Role of CD4 T cell help and costimulation in CD8 T cell re- immune response dependent on CD4ϩ T cells and accompanied by overproduc- sponses during Listeria monocytogenes infection. J. Immunol. 170:2053. tion of IL-12, IFN-␥ and TNF-␣. J. Immunol. 157:798. 52. Hamilton, S. E., A. R. Tvinnereim, and J. I. Harty. 2001. Listeria monocytogenes Downloaded from 41. Schopf, L. R., K. F. Hoffmann, A. W. Cheever, J. F. Urban, and T. A. Wynn. infection overcomes the requirement for CD40 ligand in exogenous antigen pre- 2002. IL-10 is critical for host resistance and survival during gastrointestinal sentation to CD8ϩ T cells. J. Immunol. 167:560.3. helminth infection. J. Immunol. 168:2383. ϩ ϩ 53. Constant, S. L., and K. Bottomly. 1997. Induction of Th1 and Th2 CD4 T cell 42. Serbina, N. V., V. Lazarevic, and J. L. Flynn. 2001. CD4 T cells are required responses: the alternative approaches. Annu. Rev. Immunol. 15:297. for the development of cytotoxic CD8ϩ T cells during Mycobacterium tubercu- losis infection. J. Immunol. 167:699. 54. Erbe, D. V., E. R. Pfefferkorn, and M. W. Fanger. 1991. Functions of the various 43. Carvalho, L. H., G. Sano, J. C. Hafalla, A. Morrot, M. A. Curotto de Lafaille, and IgG Fc receptors in mediating killing of Toxoplasma gondii. J. Immunol. ϩ 146:3145. F. Zavala. 2002. IL-4-secreting CD4 T cells are crucial to the development of http://www.jimmunol.org/ CD8ϩ T-cell responses against malaria liver stages. Nat. Med. 8:166. 55. Li, J. S., F. Chu, A. Reilly, and G. M. Winslow. 2002. Antibodies highly effective 44. Casciotti, L., K. H. Ely, M. E. Williams, and I. A. Khan. 2002. CD8ϩ-T-cell in SCID mice during infection by the intracellular bacterium Ehrlichia chaffeen- immunity against Toxoplasma gondii can be induced but not maintained in mice sis are of picomolar affinity and exhibit preferential epitope and isotype utiliza- lacking conventional CD4ϩ T cells. Infect. Immun. 70:434. tion. J. Immunol. 169:1419. by guest on September 29, 2021