INFECTION AND IMMUNITY, May 2002, p. 2576–2582 Vol. 70, No. 50019-9567/02/$04.00�0 DOI: 10.1128/IAI.70.5.2576–2582.2002Copyright © 2002, American Society for Microbiology. All Rights Reserved.

Rickettsia-Macrophage Interactions: Host Cell Responses toRickettsia akari and Rickettsia typhi

S. Radulovic,1* P. W. Price,2 M. S. Beier,1 J. Gaywee,1 J. A. Macaluso,1 and A. Azad1

University of Maryland, School of Medicine, Department of Microbiology and Immunology, Baltimore,Maryland 21201,1 and Beckman Coulter, Inc., Miami, Florida 331162

Received 27 July 2001/Returned for modification 12 October 2001/Accepted 22 January 2002

The existence of intracellular rickettsiae requires entry, survival, and replication in the eukaryotic host cellsand exit to initiate new infection. While endothelial cells are the preferred target cells for most pathogenicrickettsiae, infection of monocytes/macrophages may also contribute to the establishment of rickettsial infec-tion and resulting pathogenesis. We initiated studies to characterize macrophage-Rickettsia akari and -Rick-ettsia typhi interactions and to determine how rickettsiae survive within phagocytic cells. Flow cytometry,microscopic analysis, and LDH release demonstrated that R. akari and R. typhi caused negligible cytotoxicityin mouse peritoneal macrophages as well as in macrophage-like cell line, P388D1. Host cells responded torickettsial infection with increased secretion of proinflammatory cytokines such as interleukin-1� (IL-1�) andIL-6. Furthermore, macrophage infection with R. akari and R. typhi resulted in differential synthesis andexpression of IL-� and IL-6, which may correlate with the existence of biological differences among these twoclosely related bacteria. In contrast, levels of gamma interferon (IFN-�), IL-10, and IL-12 in supernatants ofinfected P388D1 cells and mouse peritoneal macrophages did not change significantly during the course ofinfection and remained below the enzyme-linked immunosorbent assay cytokine detection limits. In addition,differential expression of cytokines was observed between R. akari- and R. typhi-infected macrophages, whichmay correlate with the biological differences among these closely related bacteria.

As an obligate intracellular pathogen, rickettsial survival isdependent upon entry into, growth, and replication within theeukaryotic host cell cytoplasm and then exit to initiate a newinfection cycle (1). Generally, the outcome of infection is thehost cell lysis and the release of rickettsial progeny. While theeukaryotic endothelial cells are the preferred target cells formost pathogenic rickettsial species, infection of monocytes/macrophages has also been observed (3, 24, 25). Rickettsial-host cell interaction is complex, and we are just beginning tounderstand the molecular mechanisms by which rickettsiaeinitiate infection and cause pathology. Rickettsial infectionseldom results in complete shutdown of the host machinery,and in most cases, vigorous host responses to infection clearthe rickettsial pathogens. The host’s immune responses to therickettsial infection lead either to clearance of the pathogen orto the persistence of a subclinical infection. As an example ofthe latter, Rickettsia prowazekii, the causative agent of louse-borne typhus, can persist in humans years after primary infec-tion and/or appropriate antibiotic treatment (22). Persistenceof the rickettsiae in otherwise healthy persons with subsequentreappearance of clinical symptoms (Brill-Zinsser disease) isone mechanism by which rickettsiae perpetuate in the absenceof an animal reservoir. How rickettsiae manage to survive inthe host cells has been a subject of several recent studies (4, 7,8). One possible means by which rickettsiae may survive in thehost cells is via the suppression of the antimicrobial activities ofeukaryotic target cells, specifically monocytes/macrophages(27).

This study was initiated to characterize macrophage-rickett-sia interactions and to determine how rickettsiae survive withinmacrophages. Rickettsia akari, the causative agent of rickettsi-alpox, was used, since it infects mammalian monocytes/macro-phages and thus differs significantly from other members of thespotted fever group rickettsiae and particularly Rickettsia rick-ettsii, the etiologic agent of Rocky Mountain spotted fever. Inaddition, for comparative analysis, a typhus group rickettsia, R.typhi, capable of surviving in mouse peritoneal macrophages(3) was included in this study. Here we report that R. akari andR. typhi infect the P388D1 cell line as well as mouse peritonealmacrophages. Furthermore, macrophage infection with theserickettsiae resulted in differential cytokine synthesis and ex-pression.

MATERIALS AND METHODS

Rickettsial strains and culture conditions. In this study, we used R. akari(Kaplan) and R. typhi (Wilmington) propagated and maintained in African greenmonkey kidney cells (Vero) at respective concentrations of 106 and 108 PFU/ml,as previously described (11). After 4 days postinfection, cells were assayed for thepresence of rickettsiae by indirect fluorescent antibody assay (IFA) (14, 17) withmouse monoclonal antibodies against the R. akari and R. typhi outer membraneproteins, rOmp (gifts from Centers for Disease Control and Prevention andDavid H.Walker, UTMB, Galveston, Tex.). In vitro viability of Rickettsia wasdetermined with tissue culture plaque assays (15).

Isolation and culture of macrophages. The mouse macrophage-like cell lineP388D1 (106 per ml) was seeded into flat-bottom 96-well culture plates andcultured for 24 h in RPMI 1640 medium (Gibco BRL, Bethesda, Md.) supple-mented with 10% fetal bovine serum (FBS). P388D1 cells were infected with R.akari or R. typhi at a multiplicity of infection (MOI) of 50 rickettsiae/cell. Theculture medium was changed daily, and cells were then harvested 24, 48, and 96 hpostinfection and analyzed by fluorescence-activated cell sorter (FACS) analysisto measure Rickettsia-induced apoptosis and necrosis. Controls included un-treated P388D1 cells as well as 96-h-infected cells that have been treated withEscherichia coli lipopolysaccharide (LPS) (0.5 mg/ml).

Peritoneal macrophages were harvested from 3-week-old female C3H/HeN-

* Corresponding author. Mailing address: University of MarylandSchool of Medicine, 655 W. Baltimore St., Baltimore, MD 21201.Phone: (410) 706-1341. Fax: (410) 706-0282. E-mail: sradu001@umaryland.edu.

2576

on January 5, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

CrlBr mice purchased from Charles River Laboratories (Wilmington, Mass.).Mice were injected intraperitoneal (i.p.) with 10 ml of sterile ice-cold serum-freeRPMI medium supplemented with 0.2 mM EDTA (RPMI-EDTA). The abdo-men was gently agitated, and the medium was withdrawn. Harvested cells werewashed three times (50 � g, 30 min, 4°C) with serum-free RPMI-EDTA sup-plemented with addition of 100 �g of gentamicin (Gibco BRL) per ml andresuspended to 106 cells/ml in RPMI 1640 supplemented with 10% FBS. Thecells were cultured for 2 h at 37°C and washed gently to remove nonadherentcells. Afterward, the medium was replaced with RPMI supplemented with so-dium pyruvate (1 mM), gentamicin (100 �g/ml), and 15% FBS. After overnightincubation, the cell medium was discarded and replaced with gentamicin-freeRPMI and 10% FBS. The macrophages were then infected with R. akari and R.typhi. Untreated cells and LPS-treated and infected cells were also included inthis study. Cells were harvested 24, 48, and 96 h postinfection and analyzed byFACS.

Flow cytometry. Flow cytometry was performed with a FACSort flow cytom-eter having both argon and helium/neon lasers and with CellQuest software ona MacIntosh G4 computer (Becton Dickinson, Sunnyvale, Calif.). Analysis ofphosphatidylserine on the outer membrane of P388D1 apoptotic cells was per-formed with the Annexin-V-Fluos staining kit (Roche). Briefly, 106 infected cellsper ml (untreated and treated) were washed with phosphate-buffered saline andcentrifuged at 200 � g for 5 min. The pellet was resuspended in 100 �l of stainingsolution (20 �l of Annexin-V-fluorescein and 20 �l of propidium iodide (PI)diluted in 1,000 �l of HEPES buffer) incubated at room temperature for 15 minand analyzed by FACS. We used a 488-nm excitation and a 515-nm band passfilter for fluorescein detection and a �600-nm-pore-size filter for PI detection.Positive controls for apoptosis included P388D1 cells incubated for 4 h at 37°Cin the presence or absence of 4 mg of camptothecin per ml (Sigma) or anti-Fasmonoclonal antibody (CD95/APO-1; Boehringer Mannheim) at a concentrationof 1 �g/ml.

LDH assays. A colorimetric assay (Roche) was used to quantitate cell deathand cell lysis. This assay is based upon quantitation of lactate dehydrogenase(LDH) released from the cytosol of damaged cells. Supernatants from unin-fected control and infected cells were collected in tubes and spun for 5 min at12,000 rpm. Fifty-microliter portions of cleared supernatants were then used forthe LDH assay. The assay was performed at room temperature according to themanufacturer’s instructions (16). To quantify cytotoxicity, we calculated theaverage A490 of triplicate samples and compared the values to those of back-ground controls (cells incubated with camptothecin).

ELISA-based cytokine detection assays. Macrophages were infected with R.akari or R. typhi at an MOI of 50 rickettsiae/cell. The rickettsiae were allowed toadhere to cells for 1 h at room temperature and maintained at 34°C in 5% CO2.Afterward, the cell culture medium was changed daily, and the infection rate wasmonitored at selected time points. The cell culture medium was collected andprocessed for cytokine analysis. The cytokines interleukin-1� (IL-1�), IL-6, IL-10, IL-12, gamma interferon (IFN-�), transforming growth factor � (TGF-�),and tumor necrosis factor alpha (TNF-�) were measured as secreted proteinproducts of infected and uninfected macrophages at various time points by usingcytokine-specific enzyme-linked immunosorbent assays (ELISAs). Assays wereperformed according to the manufacturer’s guidelines at the Cytokine CoreLaboratory located at the University of Maryland, Baltimore. Briefly, mousecytokines were measured by two-antibody ELISA with a biotin-streptavidin-peroxidase detection kit and compared against a standard curve with SoftPro(Molecular Dynamics). Results were expressed as picograms of cytokine permilliliter. The lower limits of detection for the assays were as follows: IL-1�, 15pg/ml; IL-6, 12 pg/ml; IL-10 and IL-12, 31.2 pg/ml; IFN-� and TGF-�, 31 pg/ml;and TNF-�, 15.6 pg/ml. All data shown are from reproducible experiments.Values are expressed as means with standard deviations from triplicate samples.

RNA isolation and RT-PCR assays. For reverse transcription-PCR (RT-PCR),9 � 105 P388D1 cells and mouse peritoneal macrophages (C3H/HeNCrlBr) wereinoculated with R. akari and R. typhi at an MOI of 50 rickettsiae per cell andmonitored by Gimenez staining 24, 48, and 96 h postinfection. Cells were de-tached from the 150-cm2 flasks with a cell scraper, transferred to a 50-ml tube,and centrifuged at 13,000 � g for 10 min (26). The supernatant was discarded,and the cells were treated with 1 ml of Trizol reagent (Life Technologies,Gaithersburg, Md.). For subsequent RNA extractions, we followed the manu-facturer’s protocol (Gibco BRL). The RNA pellets were dissolved in diethylpyrocarbonate (DEPC) water, quantified spectrophotometrically, and stored at�80°C. RNAs derived from infected P388D1 cells, peritoneal macrophages, anduninfected controls were examined by formaldehyde agarose gel electrophoresisto confirm that RNA had not become degraded during the extraction procedure.Prior to RT-PCR, it was essential to treat all RNA samples with DNase, by using1 U of DNase/�g of RNA. Following DNase treatment, RNAs were precipitated

with 100% isopropanol, and pellets were washed with 70% ethanol and resus-pended in DEPC water. For RT, the Superscript RT protocol for randomhexamer-mediated cDNA synthesis was used.

We chose to amplify the 17-kDa cell surface common antigen gene as ourtarget (11). The RT-PCRs were performed with 6 �l of cDNAs as the template,8 mM (each) primer, 1� PCR buffer (20 mM Tris-HCl [pH 8.3], 50 mM KCl),1.5 mM MgCl2, 200 mM deoxynucleoside triphosphate (dNTP) mix, and 1 U ofTaq polymerase. Following denaturation of the initial cDNA-RNA hybrid (95°Cfor 4 min), an extra 1 U of Taq polymerase was added to each sample to improvethe yield of amplicons. Amplification conditions were 94°C for 30 s, 50°C for 45 s,and 72°C for 45 s over 30 cycles. An aliquot of 10 �l of each RT-PCR productwas electrophoresed on a 1% agarose gel and stained with ethidium bromide.

Semiquantitative RT-PCR of I�B� mRNA. Total cellular RNA was isolatedfrom R. akari- and R. typhi-infected P388D1 cells, infected peritoneal macro-phages, and uninfected controls as described above. For detection of IB�, 15 �lof the RT reaction mixture was amplified in 100 �l of volume. The primers usedwere as follows: IB� forward primer, 5-GCTCGGAGCCCTGGAAGC-3;IB� reverse primer, 5-GCCCTGGTAGGTAACTCT-3 (566-bp product). ThePCR cycle consisted of an initial incubation at 95°C for 105 s, followed by cyclingat 95°C for 30 s, 65°C for 30 s, and 72°C for 60 s, with a final incubation at 72°Cfor 7 min (19). Ten-microliter samples were subjected to electrophoresis on a 2%NuSieve 3:1 agarose gel (FMC Bioproducts, Rockland, Maine) in 1� TAEbuffer.

Statistical analysis. Appropriate transformations were performed for statisti-cal comparison of means for cytokine data by the Student t test. Nonparametricstatistics were also used with similar data. The significant differences between themeans at P � 0.05 are marked in the figures where appropriate.

RESULTS

The underlying rationale for this study was R. akari survivalwithin macrophages that may contribute to the establishmentof rickettsial infection and resulting pathogenesis. Initially, wefocused our study on determining whether R. akari survivesand replicates within the P388D1 cell line and mouse perito-neal macrophages. Both R. akari and R. typhi were able toinfect and replicate inside P388D1 cells. The macrophage in-fection rates were determined by IFA and FACS with mono-clonal antibodies against R. akari and R. typhi outer membraneprotein (rOmpB). At 24 h postinfection, IFA indicated that24% of P388D1 cells were infected with R. akari, while only15% of the cells were infected with R. typhi. Infection rates forR. akari increased to 52% by 48 h and 85% at 96 h postinfec-tion, while those for R. typhi were 32 and 52% at the same timepoints. FACS analysis revealed slightly higher infection rates ofP388D1 cells for R. akari (25% at 24 h, 55% at 48 h, and 92%at 96 h) and R. typhi (17% at 24 h, 35% at 48 h, and 56% at96 h). Infection rates for C3H/HeNCrlBr peritoneal macro-phages as determined by IFA ranged from 18 to 78% for R.akari and 10 to 49% for R. typhi during the 96 h of observa-tions. FACS analysis revealed the same general pattern formouse peritoneal macrophage infection rates ranging from 19to 78% for R. akari and 13 to 51% for R. typhi (Fig. 1). Asshown in Fig. 1, the addition of 0.5 mg of E. coli LPS per ml to96-h-infected cells drastically reduced the infection rates. Thepresence and progress of infection with R. akari and R. typhiwere additionally confirmed by RT-PCR (Fig. 2).

To determine if rickettsial infection resulted in macrophagekilling, LDH release was measured as an indicator of R. akari-and R. typhi-induced cytotoxicity. Overall, low-level cytotoxic-ity was observed, ranging from 2.9 to 8.4% cell lysis for P388D1cells and 3.5 to 13% cell lysis for the peritoneal macrophagesinfected with rickettsiae during the 96-h observation period(Fig. 3). The background lysis of uninfected P388D1 andmouse peritoneal macrophages ranged from 1.4 to 2.8%. Host

VOL. 70, 2002 MACROPHAGE-RICKETTSIA INTERACTIONS 2577

on January 5, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

cell lysis rates for R. akari varied from 4.8% at 24 h to nearly9% over the next 72 h postinfection. In contrast, R. typhiexhibited even lower cytotoxicity, with only 2.8 to 4.5% of theP388D1 cells lysed during the 96 h of the infection cycle.Overall, infection with both rickettsia species resulted in com-parable low cell lysis. In order to determine whether lysis ofinfected macrophages would have been increased if the mac-

rophages were activated, E. coli LPS was added. Overnighttreatment of 96-h-infected and uninfected P388D1 cells with E.coli LPS (0.5 mg/ml) resulted in a significant increase in celllysis (Fig. 3). The addition of LPS increased the cell lysis overthreefold in both R. typhi (13.6%)- and R. akari (24.4%)-in-fected macrophages compared to the level in their untreatedcontrols (no LPS). However, the level of increase due to LPSin infected macrophages was substantially lower than thoseobserved in uninfected macrophages (62.8%).

Since rickettsial infections have been associated with controlof apoptosis (4, 18, 27), we tested whether the cell lysis inducedby R. akari and R. typhi was due to apoptosis or necrosis. Weaddressed this question by using FACS analysis of Annexin-V-and PI-stained R. akari- and R. typhi-infected P388D1 andmouse peritoneal macrophages at 96 h postinfection. Whilepositive staining for PI alone identified necrotic cells, stainingwith both PI and Annexin-V was counted as host cell apopto-sis. Analyses of data for R. akari-infected P388D1 revealed that7.4% of the cells were apoptotic and 1.4% of the cells werenecrotic. Similarly, 8.3% of R. akari-infected mouse peritonealmacrophages were apoptotic, while 4.5% were necrotic. Incontrast, all of the R. typhi-infected cells that lysed were pos-itive by PI and by Annexin-V, indicating apoptotic events.

Prevention of apoptosis prolongs rickettsial endothelial in-fection and thereby allows rickettsiae to multiply and reinfect

FIG. 1. The rate of rickettsial infection of P388D1 cells (A) andmouse macrophages (B) as determined by FACS analysis. E. coli LPS(0.5 mg/ml) was added at 96 h postinfection.

FIG. 2. RT-PCR amplification of the 17-kDa gene from R. typhi(lanes 1 to 3, both panels)- and R. akari (lanes 4 to 6, both panels)-infected P388D1 cells (A) and peritoneal macrophages (B) at 24, 48,and 96 h postinfection. M corresponds to the 100-bp DNA marker(MBI Fermentas).

FIG. 3. Percentage of cell lysis of R. akari- and R. typhi-infectedP388D1 cells (top) and peritoneal macrophages (bottom) as deter-mined by LDH assay during the course of infection. E. coli LPS (0.5mg/ml) was added at 96 h postinfection.

2578 RADULOVIC ET AL. INFECT. IMMUN.

on January 5, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

neighboring cells. It has been shown for R. rickettsii, the caus-ative agent of Rocky Mountain spotted fever, that activation ofhost NF-B prevents apoptosis of endothelial cells (20). In apilot study, we also looked for NF-B p65 expression in R.akari- and R. typhi-infected macrophages during the rickettsialgrowth cycle. Rickettsial infections of macrophages resulted inan increased level of IB� mRNA as determined by RT-PCRof IB� (Fig. 4).

LPS-activated macrophages are capable of producing a sub-stantial array of cytokines, chemokines, and toxic mediators

(12, 13). To determine whether interaction between patho-genic rickettsiae and macrophages resulted in increased cyto-kine synthesis and release by infected cells, we initially studiedR. akari- and R. typhi-induced cytokine profiles in P388D1 cellsin vitro. Cytokine profiles for R. akari- and R. typhi-infected aswell as uninfected P388D1 cells are presented in Fig. 5 and 6.Nonactivated P388D1 cells infected with either R. akari or R.typhi demonstrated increased levels of proinflammatory cyto-kines, such as TNF-�, IL-1�, and IL-6 (Fig. 5A, B, and C,respectively) compared to uninfected controls. In general,TNF-� production by R. akari-infected P388D1 cells displayeduniform elevation by 24 h and remained unchanged through-out the observed infection compared to that of uninfectedcontrols (P � 0.05; Fig. 5A). However, R. typhi-infected cellsdisplayed a significant increase in TNF-� at 24 h postinfectionand declined fourfold afterward (P � 0.05). In contrast, therewere no changes in TNF-� levels between infected and unin-fected peritoneal macrophages during the 96-h sampling.

We noted a significant difference in IL-1� production be-tween R. akari- and R. typhi-infected cells (Fig. 5B). At 96 hpostinfection, cells infected with R. typhi secreted fourfoldmore IL-1� (68 pg/ml) compared to cells infected with R. akari

FIG. 4. Expression of IB� mRNA in R. akari (lane 1)- and R. typhi(lane 2)-infected P388D1 cells and peritoneal macrophages at 48 hpostinfection (lane 3). Lane 4, uninfected macrophages (24 h); lane 5,uninfected P388D1 cells (48 h); lane 6, uninfected macrophages (48 h);M, corresponds to 100-bp DNA marker.

FIG. 5. (A) Kinetics of TNF-� production in P388D1 cells infected with R. akari and R. typhi. �, P � 0.05 when 24-h R. typhi infected-P388D1cells were compared with their counterparts. ��, significant difference at P � 0.05 between infected and uninfected cells. (B) Kinetics of IL-1�production in P388D1 cells infected with R. akari and R. typhi expressed as mean � standard error. �, statistical significance at the �0.05 level ofIL-1� production by R. typhi-infected P388D1 versus R. akari-infected P388D1. (C) Early (24 h) and late (48 to 96 h) IL-1� production in peritonealmacrophages infected with R. akari and R. typhi. �, statistical significance at �0.05 level of IL-1� production between R. typhi- and R. akari-infectedmacrophages.

VOL. 70, 2002 MACROPHAGE-RICKETTSIA INTERACTIONS 2579

on January 5, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

and uninfected P388D1 cells (P � 0.05). The production ofIL-1� by R. typhi-infected cells continued to increase over theculture period, while R. akari-infected cells failed to producemore than baseline levels of IL-1� (Fig. 5C). Peritoneal mac-rophages secreted much greater levels of IL-1� than didP388D1 cells. In cells infected with R. akari, elevated IL-1�increased further at 96 h postinfection (745 pg/ml at 24 h, 736pg/ml at 48 h, and 1,247 pg/ml at 96 h). In contrast, R. typhi-infected cells displayed IL-1� secretion that peaked by 48 hpostinfection and then began to decline. IL-1� production byR. akari and R. typhi during 24 to 96 h postinfection wassignificantly different from that of the uninfected controls (P �0.001).

Infection with both R. typhi and R. akari induced dramaticincreases in IL-6 production by P388D1 cells (Fig. 6). IL-6secretion by R. akari-infected cells peaked at 24 h postinfection(1,500 pg/ml), followed by a steady decline to baseline 96 hpostinfection. In contrast, R. typhi-infected cells produced uni-formly high levels of IL-6 through 96 h postinfection. Interest-ingly, primary mouse peritoneal macrophages infected with R.akari and R. typhi showed a very different, although still robust,pattern of IL-6 secretion. At 24 h postinfection, both infectedcultures had elevated levels of IL-6 (R. akari, 1,125 pg/ml; R.typhi, 1,755 pg/ml). However, IL-6 secretion by R. akari-in-fected macrophages continued to increase throughout thecourse of the experiment, while secretion from R. typhi-in-fected cells decreased slightly, but remained substantiallyhigher than that from uninfected controls.

Also of interest was the production of TGF-� by infectedmouse peritoneal macrophages, although none was detectedfrom infected P388D1 cells. Secretion of TGF-� by R. typhi-infected mouse peritoneal macrophages dramatically increasedafter 48 h in culture and remained high through the 96 h ofobservation (P � 0.05) compared to that in controls. In con-trast, TGF-� in R. akari-infected macrophages remained atbackground levels (�30 pg/ml for uninfected cells) (Fig. 7).Levels of other cytokines, such as IFN-�, IL-10, and IL-12 in

supernatants of R. akari- and R. typhi-infected P388D1 andmouse peritoneal macrophages, were below the ELISA detec-tion limits.

DISCUSSION

We have shown by RT-PCR that both R. akari and R. typhiwere capable of infecting and surviving within P388D1 cellsand mouse peritoneal macrophages. R. akari attained a greaterrickettsial load in infected P388D1 cells and mouse macro-phages during 96 h than did R. typhi. This observation is con-sistent with rickettsial growth observed previously in endothe-lial cells, which are the main mammalian target for mostrickettsial species (5–8). R. typhi, after entry into the hostendothelial cells, multiplies to achieve and maintains a highrickettsial load before lysing the host cells. R. akari and R.rickettsii, on the other hand, maintain a low-level infection, buteventually lyse their target cells. Both R. typhi, and R. akari canfreely leave the endothelial cells before host cell lysis occurs (3,23). However, the utilization of host cell actin assembly variesbetween R. akari and R. typhi (10). Presumably, this escapemechanism allows the rickettsiae to leave the host cell’s hostileenvironment before being destroyed (9).

A combination of intracellular growth and lysis is consideredto be the basis for rickettsial pathogenesis (1, 7, 8). While themolecular mechanism of rickettsia-induced endothelial cell ly-sis is not known, the lysis mechanism or mechanisms allowrickettsiae to be released and to infect the neighboring cells (8,9, 21, 22). Our data indicated that infection of unactivatedmacrophages with R. akari and R. typhi did not result in effi-cient cytolysis through the 96-h time point of evaluation.

Consideration of the NF-B data permits an alternative ex-planation. Activation of host cell NF-B has been demon-strated in response to infection by numerous human patho-gens, and an increase in the level of IB� is a sensitiveindicator of NF-B activation (4). In the present study, weobserved the presence of NF-B and an increased level ofIB� mRNA in R. akari- and R. typhi-infected P388D1 cells.IB� mRNA was detected in R. typhi- but not in R. akari-infected peritoneal macrophages. Sporn et al. have demon-strated that R. rickettsii induced activation of the transcrip-

FIG. 6. IL-6 response of P388D1 cells and peritoneal macrophagesinfected with R. akari and R. typhi. �, statistical significance at P � 0.05for IL-6 production between R. akari- and R. typhi-infected P388D1cells. Overall, there were no statistical differences between R. akari-and R. typhi-infected mouse macrophages, but the patterns of the IL-6production during the time course of infection were different. IL-6production in infected cells was significantly different from that inuninfected controls (�0.001) with the exception of P388D1 cells in-fected with R. akari at 96 h, which did not differ from the baseline data.

FIG. 7. TGF-� response of peritoneal macrophages infected withR. akari and R. typhi at 24, 48, and 96 h postinfection. �, statisticalsignificance at P � 0.05 for TGF-� production between R. akari- andR. typhi-infected macrophages 48 h postinfection.

2580 RADULOVIC ET AL. INFECT. IMMUN.

on January 5, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

tional factor NF-B in vitro and in vivo (18, 19). R. rickettsiiinfection also was associated with an increased level of IB�mRNA. Furthermore, they suggested the involvement of theTNFR1 receptor, but not TNF-� or apoptosis, in R. rickettsii-induced signal transduction, leading to NF-B activation (20).Recently, Rikihisha et al. provided convincing evidence thathuman granulocytic Ehrlichia delayed apoptosis of infectedgranulocytes (27). These few studies suggested that delayedhost cell apoptosis is a general phenomenon used amongpathogenic rickettsiae to prolong their survival within the eu-karyotic cells. However, it would be premature to assume auniversal feature among rickettsiae in terms of intracellularsurvival, considering biological and genetic differences amongrickettsiae in terms of host cell preference, growth, and intra-cellular location. The inhibition of apoptosis at two levels, bothin primary infection and after induction of proapoptotic im-munological processes of acquired immunity, could set theconditions for persistent rickettsial infection. Surviving rickett-siae could then replicate within the macrophages remaining ata site of rickettsia-infected endothelium. The activation ofNF-B induced by R. akari and R. typhi thus could preventapoptotic signals triggered by the host defense mechanism,such as interaction with immune cells or inflammatory cyto-kines. Our ongoing experiments may address these questionsin the near future.

Among the roles macrophages play in the immune system isthe production of an array of cytokines that are important indirecting the immune response to host cell damage and patho-gen removal (27). The inactivation of IL-6 release in infectedmacrophages by pathogen bacteria such as E. coli and Porphy-romonas gingivalis has been reported by bacterial secretion ofspecific proteinases (2). The response of macrophages to rick-ettsial infection included a differential increase in proinflam-matory cytokines, such as TNF-�, IL-1�, and IL-6, while theproduction of other cytokines (IFN-�, IL-10, and IL-12) re-mained undetectable. P388D1 cells and mouse peritoneal mac-rophages infected with R. akari failed to induce remarkableTNF-� or TGF-� secretion. IL-1� secretion was robust fromperitoneal macrophages, but not from the P388D1cell line, andIL-6 secretion was remarkable from both cultures. While theincreased production of IL-1� and IL-6 by P388D1 cells andmouse peritoneal macrophages infected with R. typhi was no-ticeable, the expression of TNF-� and TGF-� differed betweenthese two host cells. The relationship between TNF-�, thenitric oxide synthase pathway, and LPS to the killing of IFN-�-treated macrophagelike RAW264.7 cells by R. prowazekiiwas studied previously (21). Furthermore, depletion of IFN-�and TNF in mice infected with Rickettsia conorii resulted infatal and overwhelming rickettsial infection (7). Differencesobserved in cytokine profiles between R. akari and R. typhi maycontribute to their survival within the confines of their eukary-otic target cells, despite their relationship to host defensemechanisms.

The studies reported here were carried out in vitro in theabsence of many cellular components of host immune re-sponses. In vivo, vigorous host responses occur that are depen-dent upon the recruitment and activation of macrophages andneutrophils, which are required to effectively clear rickettsialinfection. In most cases, infection with pathogenic rickettsiaein natural hosts (e.g., house mice and rats for R. akari and R.

typhi, respectively) is resolved with rickettsial clearance. Incontrast, human infection is characterized by severe pathologyand some mortality in untreated cases. Low rickettsial cytotox-icity toward the host cells observed in this study not onlyprovides ample opportunity for rickettsial spread to neighbor-ing cells, but may also result in rickettsial persistence. Indeed,we have been successful in isolating R. typhi from spleen andkidneys of wild-caught rats exhibiting high antirickettsial anti-bodies (unpublished data), which further confirms rickettsialpersistence.

ACKNOWLEDGMENT

This research was supported by grants from the National Institute ofAllergy and Infectious Diseases (AI 17828).

REFERENCES

1. Azad, A. F., and C. B. Beard. 1998. Rickettsial pathogens and their arthropodvectors. Emerg. Infect. Dis. 4:179–186.

2. Banbula, A., M. Bugno, A. Kuster, P. C. Heinrich, J. Travis, and J. Potempa.1999. Rapid and efficient inactivation of IL-6 gingipains, lysine- and arginine-specific proteinases from Porphyromonas gingivalis. Biochem. Biophys. Res.Commun. 261:598–602.

3. Beaman, L., and C. L. Wisseman, Jr. 1976. Mechanism of immunity intyphus infections. V. Demonstration of Rickettsia mooseri-specific antibodiesin convalescent mouse and human serum cytophilic for mouse peritonealmacrophages. Infect. Immun. 14:1065–1070.

4. Clifton, D. R., R. A. Goss, S. K. Sahni, D. Van Antwerp, R. B. Baggs, V. J.Marder, D. J. Silverman, and L. A. Sporn. 1998. NF-kappa B-dependentinhibition of apoptosis is essential for host cell survival during Rickettsiarickettsii infection. Proc. Natl. Acad. Sci. USA 95:4646–4651.

5. Collins, T. 1993. Endothelial nuclear factor-kB and the initiation of theatherosclerotic lesion. Lab. Investig. 68:499–508.

6. Feng, H.-M., and D. H. Walker. 2000. Mechanisms of intracellular killing ofRickettsia conorii in infected human endothelial cells, hepatocytes, and mac-rophages. Infect. Immun. 68:6729–6736.

7. Feng, H. M., V. L. Popov, and D. H. Walker. 1994. Depletion of gammainterferon and tumor necrosis factor alpha in mice with Rickettsia conorii-infected endothelium: impairment of rickettsicidal nitric oxide productionresulting in fatal, overwhelming rickettsial disease. Infect. Immun. 62:1952–1960.

8. Feng, H. M., J. Wen, and D. H. Walker. 1993. Rickettsia australis infection: amurine model of a highly invasive vasculopathic rickettsiosis. Am. J. Pathol.142:1471–1482.

9. Heinzen, R. A., S. S. Grieshaber, L. S. Van Kirk, and C. J. Devin. 1999.Dynamics of actin-based movement by Rickettsia rickettsii in Vero cells.Infect. Immun. 67:4201–4207.

10. Heinzen, R. A., S. F. Hayes, M. G. Peacock, and T. Hackstadt. 1993. Direc-tional actin polymerization associated with spotted fever group rickettsiainfection of Vero cells. Infect. Immun. 61:1926–1935.

11. Higgins, J. A., S. Radulovic, B. H. Noden, J. M. Troyer, and A. F. Azad. 1998.Reverse transcriptase PCR amplification of Rickettsia typhi from infectedmammalian cells and insect vectors. J. Clin. Microbiol. 36:1793–1794.

12. Klotz, F. W., L. F. Scheller, M. C. Seguin, N. Kumar, M. A. Marletta, S. J.Green, and A. F. Azad. 1995. Co-localization of inducible nitric oxide syn-thase and Plasmodium berghei in hepatocytes from rats immunized withirradiated sporozoites. J. Immunol. 154:3391–3395.

13. Moore, K. J., L. P. Andersson, R. R. Ingalls, B. G. Monks, R. Li, M. A.Arnaout, D. T. Golenbock, and M. W. Freeman. 2000. Divergent response toLPS and bacteria in CD14-deficient murine macrophages. J. Immunol. 165:4272–4280.

14. Philip, R. N., E. A. Casper, W. Burgdorfer, R. K. Gerloff, L. E. Hughes, andE. J. Bell. 1978. Serologic typing of rickettsiae of the spotted fever group bymicroimmunofluorescence. J. Immunol. 121:1961–1978.

15. Policastro, P. F., M. G. Peacock, and T. Hackstadt. 1996. Improved plaqueassays for Rickettsia prowazekii in Vero76 cells. J. Clin. Microbiol. 34:1944–1948

16. Racher, A. J., D. Looby, and J. B. Griffiths. 1990. Use of lactate dehydro-genase release to assess changes in culture viability. Cytotechnology 3:301–307.

17. Roux, V., and D. Raoult. 2000. Phylogenetic analysis of members of thegenus Rickettsia using the gene encoding the outer-membrane proteinrOmpB. Int. J. Syst. Evol. Microbiol. 50:1449–1455.

18. Sahni, S. K., L. C. Turpin, T. L. Brown, and L. A. Sporn. 1999. Involvementof protein kinase C in Rickettsia rickettsii-induced transcriptional activationof the host endothelial cell. Infect. Immun. 67:6418–6423.

19. Shi, R.-J., P. J. Simpson-Haidaris, N. B. Lerner, V. J. Marder, D. J. Silver-

VOL. 70, 2002 MACROPHAGE-RICKETTSIA INTERACTIONS 2581

on January 5, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

man, and L. A. Sporn. 1998. Transcriptional regulation of endothelial celltissue factor expression during Rickettsia rickettsii infection: involvement ofthe transcription factor NF-B. Infect. Immun. 66:1070–1075.

20. Sporn, L. A., S. K. Sahni, N. B. Lerner, V. J. Marder, D. J. Silverman, L. C.Turpin, and A. L. Schwab. 1997. Rickettsia rickettsii infection of culturedhuman endothelial cells induces NF-B activation. Infect. Immun. 65:2786–2791.

21. Turco, J., and H. H. Winkler. 1994. Relationship of tumor necrosis factoralpha, the nitric oxide synthase pathway, and lipopolysaccharide to the killingof gamma interferon-treated macrophagelike RAW264.7 cells by Rickettsiaprowazekii. Infect. Immun. 62:2568–2574.

22. Turcinov, D., I. Kuzman, and B. Herendı́c. 2000. Failure of azithromycin intreatment of Brill-Zinsser disease. Antimicrob. Agents Chemother. 44:1737–1738.

23. Van Kirk, L. S., S. F. Hayes, and R. A. Heinzen. 2000. Ultrastructure of

Rickettsia rickettsii actin tails and localization of cytoskeletal proteins. Infect.Immun. 68:4706–4713.

24. Walker, D. H., S. D. Hudnall, W. K. Szaniawski, and H. M. Feng. 1999.Monoclonal antibody-based immunohistochemical diagnosis of rickettsial-pox: the macrophage is the principal target. Mod. Pathol. 12:529–533.

25. Walker, D. H., H. M. Feng, S. Ladner, A. N. Billings, S. R. Zaki, D. J. Wear,and B. Hightower. 1997. Immunohistochemical diagnosis of typhus rickett-sioses using an anti-lipopolysaccharide monoclonal antibody. Mod. Pathol.10:1038–1042.

26. Weiss, E., J. C. Coolbaugh, and J. C. Williams. 1975. Separation of viableRickettsia typhi from yolk sac and L cell host components by renografindensity gradient centrifugation. Appl. Microbiol. 30:456–463.

27. Yoshiie, K., H.-Y. Kim, J. Mott, and Y. Rikihisa. 2000. Intracellular infectionby the human granulocytic ehrlichiosis agent inhibits human neutrophil apo-ptosis. Infect. Immun. 68:1125–1133.

Editor: R. N. Moore

2582 RADULOVIC ET AL. INFECT. IMMUN.

on January 5, 2020 by guesthttp://iai.asm

.org/D

ownloaded from