| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Department of Pathology,* University of Texas Medical Branch, Galveston, Texas; and the Department of Pathology, Duke University Medical Center, Durham, North Carolina
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
The clinical features of rickettsial diseases correspond to damage to the main target cells, endothelial cells, particularly importantly those in the lungs and brain. The initial symptoms are those of a flu-like syndrome; however, if untreated, severe manifestations can develop including noncardiogenic pulmonary edema/adult respiratory distress syndrome, interstitial pneumonia, acute renal failure, hemorrhagic phenomena, peripheral edema, and hypovolemic hypotension because of leakage of intravascular fluid into the extravascular space. Central nervous system manifestations include delirium, focal neurological deficits, seizures, and coma.6,7
Studies in animal models have demonstrated that a successful anti-rickettsial immune response requires CD8+ T lymphocytes,8
and in terms of their effector functions both interferon (IFN)-
production and cytolytic mechanisms are required.9
In humans, the presence of perivascular infiltrating CD8+ T cells has also been demonstrated in skin lesions of patients with spotted fever group rickettsioses.10
Also, endothelial cells become activated on rickettsial infection as evidenced by the production of cytokines and expression of adhesion molecules.11-14
The stable adhesion between cytotoxic lymphocytes and target cells through integrin activation is a requirement for the focalization of their effector mechanisms.15
In the majority of the studies addressing lymphocyte adhesion to the endothelium, this step is viewed as a requirement for the transmigration of leukocytes to the areas of inflammation beneath the endothelium. Human rickettsioses and the rickettsial mouse model offer the unique opportunity to study CD8+ T lymphocyte effector mechanisms in a relevant infectious disease in which endothelial cells themselves are presenting the antigen and are therefore the targets. The role of chemokines, particularly inflammatory chemokines that specifically target activated T cells through the CXCR3 receptor such as CXCL9 (Mig) and CXCL10 (IP-10), has not been explored in this regard; however, there is suggestive evidence from other systems including in vitro flow studies,16
transplantation studies, and other infectious diseases caused by intracellular organisms that do not target endothelial cells,17-22
that chemokines might play an important role in the anti-rickettsial immune response. We undertook the following studies to investigate the possible role of CXCL9 and CXCL10 in the immune response against spotted fever group rickettsiae.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
R. conorii (Malish 7 strain) is a human isolate from South Africa with an unknown number of passages in the yolk sacs of embryonated chicken eggs. For all of the experiments described in this study, a stock of R. conorii was produced by cultivation in specific pathogen-free embryonated chicken eggs. Yolk sacs from infected eggs with dead embryos were homogenized in a Waring blender, diluted to a 10% suspension in sucrose-phosphate-glutamate buffer (0.218 mol/L sucrose, 3.8 mmol/L KH2PO4, 7.2 mmol/L K2HPO4, 4.9 mmol/L monosodium L-glutamic acid, pH 7.0) and aliquoted for storage at -80°C after discarding the pellet produced by low-speed centrifugation (200 x g, 10 minutes). Rickettsial content of this stock was quantified by plaque assay,23 and the LD50 was determined experimentally in C3H/HeN mice.
Human Samples
Fourteen brainstem specimens from nine fatal pediatric cases of RMSF collected at Duke University between 1935 and 1978 were used for immunohistochemical analysis. Four brainstem samples from trauma cases were used as negative controls. No other tissues were analyzed because they were not available, and one case reported as negative for rickettsial antigen (see case 4 in Table 2
) contained R. rickettsii in other tissues that were studied in the past.24,25
Two pathologists analyzed the samples and qualitatively graded each parameter according to the intensity of staining in a scale of 1+ to 4+.
|
The mouse model of endothelial-target spotted fever group rickettsioses consists of R. conorii infection of C3H/HeN mice and has been previously described in detail.26,27 All mice were housed in a biosafety level 3 facility and were infected intravenously (through the tail vein) with 0.25 LD50 of R. conorii for the sublethal dose model or 3 LD50 of R. conorii for the lethal dose model (in a volume of 300 µl in sucrose-phosphate-glutamate buffer). Animals infected with the higher dose usually died 6 to 7 days later. In three separate experiments, eight mice received a sublethal dose of R. conorii, and three mice received a lethal dose. One mouse from each group was euthanized every other day by cervical dislocation, and all organ tissues were harvested and divided for RNA preservation in RNAlater (Ambion, Austin, TX), and fixation in 10% buffered formalin or zinc fixative (BD Pharmingen, San Diego, CA). Tissues from mock-infected mice were similarly processed and evaluated as negative controls. Organs fixed in formalin were transferred to 70% ethanol after 24 hours to avoid overfixation.
Cell Line
The C3H/HeN mouse endothelial cell line SVEC4-10 was kindly provided by Dr. M. Edidin (Johns Hopkins University, Baltimore, MD).28 This cell line was cultivated in Dulbecco’s modified Eagle’s medium (Life Technologies, Inc., Carlsbad, CA) containing 2% fetal bovine serum (Hyclone Inc., Logan, UT). The cells were passaged twice weekly.
Chemokine Cloning
Standards for quantitative real-time polymerase chain reaction (PCR) were generated by cloning the chemokines and GAPDH cDNA into the vector pCR2 (Invitrogen, Carlsbad, CA) after PCR amplification using primers designed to encompass the entire reading frame. All of the inserts were confirmed by sequencing. The sequences of the primers are presented in Table 1
. The relevant mRNA was obtained from splenic tissue of C3H/HeN mice infected for 5 days with 3 LD50 of R. conorii. The mRNA was reverse-transcribed with Sensiscript (Qiagen, Valencia, CA) and oligo(dT) primers. cDNA for mouse CXCL9 and CXCL10 (cloned into pBluescript; Stratagene, La Jolla, CA) was a kind gift of P. Murphy (National Institutes of Health. Bethesda, MD).
|
Optimal quality RNA was obtained by combining an organic extraction method (Totally RNA, Ambion) with the RNaqueous 4PCR kit from Ambion. Briefly, RNAlater-preserved samples (Ambion) were homogenized in denaturation solution using a mixer mill (Qiagen, Valencia, CA) and subsequently extracted with phenol/chloroform. The upper aqueous phase was mixed with an equal volume of 70% ethanol. This mixture was loaded into the filter cartridges of the RNAqueous kit and the protocol followed according to the manufacturer’s instructions. The RNA concentration was adjusted to 0.2 mg/ml before DNase treatment (DNAfree, Ambion). In all cases, RNA quality was verified by gel electrophoresis before cDNA synthesis with Moloney murine leukemia virus reverse transcriptase (Superscript-II first strand synthesis system; Invitrogen, Life Technologies, Carlsbad, CA) and oligo(dT) primers. In addition, mRNA was PCR-amplified directly (with no reverse transcriptase) to verify that there was minimal or no DNA contamination of the cDNA samples.
Quantitative real-time PCR was performed using the iCycler from Bio-Rad (Hercules, CA) and Stratagene’s Brilliant quantitative PCR kit (La Jolla, CA). Two-step cycle parameters (95°C and 60°C) were used and the primers and probes used are presented in Table 1
. Probes for CXCL9 and CXCL10 were labeled with FAM and Black Hole Quencher 1 (BHQ1), and the probe for GAPDH was labeled with TAMRA and Black Hole Quencher 2 (BHQ2) (Biosearch Technologies, Novato, CA). The primers used for real-time PCR with SYBRgreen (IQ SYBR green supermix; Bio-Rad, Hercules, CA) are presented in Table 1
. Optimal magnesium concentrations obtaining efficiencies close to 100% were determined experimentally. In all cases 3 to 5 mmol/L Mg could be used. We used the delta delta method (comparative CT[
CT] method) for the quantitative analysis of gene expression.29
The rickettsial load was determined by real-time PCR (with TaqMan probes) of the Rickettsia-specific gene rompB. For these experiments, the substrate of amplification was DNA-purified from RNAlater-preserved samples using the Dneasy tissue kit from Qiagen. The results were normalized to GAPDH in the same sample and expressed as copy number per 104 (for the lethal model) or 105 (for the sublethal model) GAPDH copies (standard curves with more than 94% efficiency and linear amplification across 1 to 106 to 107 copies were used to obtain the copy number of the samples). The sequences of the primers and probes are presented in Table 1
. The rompB probe was labeled with FAM and Black Hole Quencher 1 (BHQ1) (Biosearch Technologies).
Immunohistochemistry
Sections (4 to 6 µm) from paraffin-embedded tissues fixed in 10% formalin in phosphate-buffered saline (PBS) were antigen-retrieved in citrate pH 6.0, stained with relevant primary antibodies, and detected with appropriate biotinylated secondary antibodies, ABC-peroxidase and diaminobenzidine as a substrate (Vector Laboratories, Burlingame, CA). After digestion with Protease I (Ventana Medical Systems, Tucson, AZ) for 10 minutes at room temperature, rabbit anti-spotted fever group rickettsia antibody was detected with anti-rabbit alkaline phosphatase antibody (Jackson Immunoresearch Laboratories, West Grove, PA) and Vector Red substrate. The strong red fluorescence produced by this substrate localized the rickettsiae using a fluorescence microscope equipped with a wide band-pass fluorescein-isothiocyanate filter (Olympus Optical, Melville, NY).
In the mouse model, CXCL9 and CXCL10 were detected with affinity-purified goat polyclonal antibodies (R&D Systems, Minneapolis, MN). In human cases, CXCL9 and CXCL10 were detected with rat monoclonal antibodies (clones 6/D4/D6/G2 and B8-11, respectively; BD Pharmingen). The anti-human CXCR3 antibody was also obtained from BD Pharmingen (clone 1C6/CXCR3) and the anti-human CD8 mAb (clone 4B11) from Novocastra Laboratories (Burlingame, CA). Mouse zinc-fixed tissues were used for the detection of CD8+ T cells with a rat anti-mouse CD8 monoclonal antibody (clone 53-6.7, BD Pharmingen) and biotinylated F(ab')2 fragments of mouse-absorbed mouse anti-rat IgG (Jackson Immunoresearch) as a secondary reagent. The anti-human CXCR3 antibody served as isotype-negative control for anti-human CXCL9 (they are both IgG1), and the anti-human CD8 antibody served as a negative control for anti-human CXCL10 (they are both IgG2) because they produced a clearly distinguishable pattern of staining. In addition, in the absence of primary antibody, no signal was detected. For mouse samples, noninfected mice that do not express the antigen (rickettsia and chemokines) served as negative controls.
Intracellular Cytokine Staining
Confluent monolayers of the mouse endothelial cell line SVEC4-10 were infected with 4 x 103 pfu/cm2 of R. conorii, mock-infected as a negative control, or stimulated with 5000 U/ml human tumor necrosis factor (TNF)-
(known to stimulate mouse TNF- receptors) and 500 U/ml mouse IFN- (Roche Applied Science, Indianapolis, IN) for 12, 24, or 48 hours. Four hours before fixation with 4% paraformaldehyde in PBS for 20 minutes, the cells were treated with Brefeldin A (Golgi Stop, BD Pharmingen) to paralyze the cellular secretory mechanisms and in this way permit detection of CXCL9 and CXCL10 inside the cells. Goat polyclonal antibodies against mouse CXCL9 and CXCL10 were used (R&D Systems) for detection.
Western Blot
To confirm the identity of the proteins detected with the goat polyclonal antibodies against CXCL9 and CXCL10, confluent monolayers of SVEC4-10 cells treated with TNF- and IFN- for 24 hours were lysed with a buffer consisting of 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate in PBS supplemented with protease inhibitor cocktail (Roche Applied Science). Total protein was quantified with the bicinchinononic acid (BCA) reagent from Pierce (Bradford, IL). Thirty µg of protein per sample were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electrotransferred to nitrocellulose membranes. Chemokines were detected with the same antibodies used for immunohistochemistry and intracellular cytokine staining and were detected with an alkaline phosphatase system (Kierkegaard & Perry Laboratories, Gaithersburg, MD).
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
We initially screened for ligands of chemokine receptors present in effector and memory CD8+ T lymphocytes because these cells are essential in the immune clearance of rickettsiae from the infected vasculature of the host.8,9
The chemokines investigated were CCL2 (MCP-1), CCL12 (MCP-5), CCL3 (Mip-1), CCL4 (Mip-1ß), CXCL9 (Mig), CXCL10 (IP-10), and CX3CL1 (Fractalkine). The entire open reading frame of each one of these chemokines was PCR-amplified and cloned into the TOPO TA vector pCR2. The identity of the cloned products was verified by sequencing. These plasmid DNA samples were used to optimize primers for SYBRgreen real-time PCR, and in all cases the efficiency of the reaction was more than 95% with linear amplification across 6 log10 of target concentration (data not shown). We subsequently analyzed the expression of all these chemokines by real-time PCR of cDNA from the lungs of C3H/HeN mice infected with lethal (3 LD50) and sublethal (0.25 LD50) inocula of R. conorii (Figures 1 and 3
; CX3CL1 data not shown). Our study focused on CXCL9 and CXCL10 because they had the highest level of up-regulation after rickettsial infection among all chemokines analyzed and because they are known to target activated T cells.30
|
Before further characterizing the expression of these chemokines in the animal model of spotted fever group rickettsioses, we wanted to determine the relevance of these findings to human disease by characterizing the expression of CXCL9 and CXCL10 in nine fatal cases of RMSF. The expression of these chemokines, their receptor (CXCR3), CD8 antigen, and rickettsial antigen was analyzed by immunohistochemistry of samples taken from the brainstem (Figure 2)
. They are all from pediatric autopsies collected between 1935 and 1978. The results are summarized in Table 2
. CXCL9 and CXCL10 were detected in all cases, and their levels of expression almost paralleled each other. More importantly, the majority of the signal was found in endothelial cells (Figure 2; a, b, and d)
, although occasional astrocytes and infiltrating leukocytes did express CXCL9 and CXCL10 (Figure 2c
, and data not shown). Despite the fact that many examples of expression of these chemokines in rickettsia-infected endothelial cells were found (Figure 2d)
, many vascular beds expressing CXCL9 and CXCL10 had no detectable rickettsial antigen (Figure 2, a and b)
. This may be explained by immune- and/or antibiotic-mediated clearance of the rickettsiae. Infiltrating CD8+ and CXCR3+ cells around vessels with no detectable rickettsiae might also reflect this effect, and the converse situation, rickettsia-infected endothelium with perivascular CD8+ and CXCR3+ cells, was frequently found (Figure 2, e and f)
. In all cases many microvessels that did not express CXCL9 or CXCL10 served as an internal negative control. In addition, brainstem samples from noninfectious disease cases were negative throughout. It is also worth pointing out that the use of a fluorescent substrate for the detection of rickettsial antigen enhanced our ability to identify areas with very few rickettsiae (Figure 2, g and h)
.
|
The findings in human cases of RMSF prompted us to characterize the animal model of spotted fever group rickettsioses with respect to the expression of CXCL9 and CXCL10. Both the mRNA and protein expression of these chemokines were analyzed in the most important target organs, namely brain and lungs, as well as liver, by real-time PCR with TaqMan probes (to ensure specificity) and by immunohistochemistry. For real-time PCR, the levels of expression of CXCL9 and CXCL10 were normalized against the levels of expression of GAPDH in each organ, and the samples were compared by the delta delta method assigning a value of 1 to the control (Figure 3)
. In all cases the immunohistochemical results closely paralleled the real-time PCR results.
|
. At the protein level, both CXCL9 and CXCL10 were detected in multiple foci in the sinusoids of the liver 24 hours after the infection (Figure 4a)
, and subsequently, on day 3, the signals for both chemokines were found globally throughout the sinusoids of the liver (Figure 4b)
. Between days 5 and 7, the expression of CXCL9 and CXCL10 became focal again (Figure 4c)
and eventually became undetectable between days 13 and 21. In the lungs, CXCL9 and CXCL10 proteins were also detectable in multiple foci between days 1 and 3 (Figure 4d)
; however, the peak of protein expression was not globally distributed as in the liver although it clearly involved the vasculature as well as infiltrating leukocytes (Figure 4, e and f)
.
|
.
CXCL10 is found in the brain, both at the protein and mRNA levels, but only beginning on day 3 after a sublethal rickettsial challenge, which is the peak of expression (Figure 3)
. In the case of CXCL9, the peak of expression is on day 7. Both chemokines are subsequently down-regulated in a similar manner as in the lungs and liver (Figure 3)
. In the lethal model of spotted fever group rickettsiosis, both CXCL9 and CXCL10 were detected 24 hours after the infection, and the levels had increased on day 3. At the protein level, CXCL9 and CXCL10 were detected in multiple foci of the vasculature, but the protein expression was not generalized at any point as it was in the liver sinusoids (data not shown).
Interestingly CXCL9 and CXCL10 were detected on the luminal side of kidney tubular epithelial cells at very early time points after rickettsial infection (Figure 4g)
. This distribution might reflect tubular reabsorption of these small proteins that are likely filtered through the glomerulus. Therefore, the analysis of the kidney for small cytokines and chemokines such as CXCL9 and CXCL10 might be useful as a reflection of systemic production.
Kinetics of CD8+ T-Cell Infiltration in C3H/HeN Mice Infected with R. conorii
The peak of expression of CXCL9 and CXCL10 on day 3 was surprising because these chemokines have been shown to preferentially target mature effector and memory T cells and these cells usually infiltrate infected tissues at later times. For this reason we decided to investigate the kinetics of T-cell infiltration into rickettsia-infected tissues by immunohistochemistry. CD8+ T lymphocytes are first observed in the parenchyma of all organs analyzed on day 6, and they are not found in infected tissues at earlier time points (Figure 4, h and i)
.
Kinetics of Rickettsial Growth in C3H/HeN Mice Infected with R. conorii
We investigated the rickettsial load in lungs, liver and brain of the same mice used for the chemokine expression studies by both real-time PCR of the rickettsial gene rompB (using TaqMan probes and immunohistochemistry. The PCR in these experiments was performed using DNA as a target, and the data were normalized against the eukaryotic gene encoding GAPDH. For these experiments, the results were expressed as number of rompB copies per 104 (lethal model) or 105 (sublethal model) GAPDH copies. Mice receiving either a lethal or a sublethal dose of R. conorii had a progressive increase in rickettsial load peaking on day 5 (Figure 5
; lung and brain results not shown). The animals receiving a lethal inoculum had 
10-fold higher rickettsial loads than mice receiving sublethal inocula, which is consistent with the 10-fold difference in the size of the inocula. In the sublethal model, the rickettsial load progressively decreased after day 5 with a dramatic decline between days 5 and 7. This is consistent with the time when infiltrating CD8+ T cells are first detected in infected tissues (day 6). The immunohistochemical results validated the real-time PCR results (data not shown); however, we noticed a difference in our capacity to detect rickettsial antigen in paraffin-embedded tissue sections depending on the type of fixative used. Rickettsial antigen can be detected at lower concentrations (2 to 3 days earlier) in tissues fixed with a zinc-based preparation as compared to tissues fixed in 10% formalin despite the fact that in both cases the antigen was retrieved through an enzymatic treatment.
|
We also investigated whether rickettsial infection of the endothelium can directly trigger the expression of CXCL9 and CXCL10. The mouse endothelial cell line SVEC4-10 was infected with 4 x 104 PFU/cm2 of R. conorii, and the expression of these chemokines was analyzed at the protein level by intracellular cytokine staining at various time points between 8 and 72 hours after infection. CXCL9 and CXCL10 were not detectable at any point; however, SVEC4-10 cells stimulated with IFN- or a combination of TNF- and IFN- expressed both chemokines (data not shown). The identity of the proteins detected with the antibodies used was confirmed by Western blot of lysates from SVEC4-10 cells stimulated with TNF- and IFN- for 24 hours (data not shown).
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
, and cytolytic mechanisms are all important in the effective anti-rickettsial immune response.8,9,27,31
The main target cells, endothelial cells, also play an important role in immunity. Experiments in vitro with cultured endothelial cells (mainly human umbilical vein endothelial cells) show that they become activated after rickettsial infection with concomitant production of cytokines such as interleukin (IL)-1, IL-6, and IL-811,12
and expression of adhesion molecules such as such as ICAM-1 and VCAM-1.13
Rickettsiae directly induce the activation of transcription factors of the nuclear factor-
B family, which regulate the expression of cytokines, chemokines, adhesion molecules, and other early response genes,32,33
and may initiate these responses during rickettsial infection. Activation of endothelial cells in rickettsial infection is reflected not only by the expression of markers such as the ones just discussed but more importantly by the capacity for intracellular killing of rickettsiae through nitric oxide produced by inducible nitric oxide synthase (iNOS or NOS2) after IFN- and TNF- stimulation.31
The experimental evidence indicates that a fine equilibrium between the activation of rickettsia-infected endothelial cells and the effector mechanisms of rickettsia-specific CD8+ T cells is necessary to achieve clearance of the infection. Because chemokines are known to play an important role in inducing a stop signal in lymphocytes rolling on the surface of endothelial cells through increased integrin affinity,34,35 we hypothesized that chemokines that target T cells might be important in the localization and focalization of anti-rickettsial CD8 T-cell effector functions.
As a first step we evaluated the expression of CCR2, CCR5, CX3CR1, and CXCR3 ligands in lungs of C3H/HeN mice infected with both lethal and sublethal doses of R. conorii (a well-established model of human spotted fever group rickettsioses).8,9,26,27
We studied the chemokines binding to these receptors because they are expressed mainly on mature Th1 and CD8+ T lymphocytes,36-40
which are critical in the clearance of rickettsiae from the infected vasculature. The ligands for CCR2 are CCL2 (MCP-1) and CCL12 (MCP-5); CCL3 (MIP-1) and CCL4 (MIP-1ß) are the ligands for the chemokine receptor CCR5; CX3CL1 (Fractalkine) is the ligand for CX3CR1; and CXCL9 (Mig) and CXCL10 (IP-10) are ligands for CXCR3. Real-time PCR analysis with SYBRgreen revealed that most of these chemokines had a peak of expression on day 3 after either a lethal or sublethal challenge with R. conorii (Figure 1)
; however, the CXCR3 ligands, CXCL9 and CXCL10, had much higher levels of mRNA induction with respect to control than all of the other chemokines (data for SYBRgreen-PCR not shown; Figure 3
shows results with TaqMan probes). These CXCR3 ligands have been implicated in CD8+ T-cell-mediated responses not only in infectious diseases, but also in models of allograft rejection and autoimmunity. For instance, in experimental cardiac allograft rejection, blocking of CXCL9,17
CXCL1018
or its receptor CXCR319
with neutralizing antibodies prolongs allograft survival. Furthermore, CXCR3 gene knockout mice are resistant to development of acute allograft rejection. Also, knockout of the CXCR3 gene in the RIP-GP background delays the development of insulitis and diabetes.41
More importantly, the role of these chemokines has been demonstrated in different infectious disease models. Antibody neutralization of CXCL10 in a mouse model of Toxoplasma gondii results in increased parasite burden and mortality.20
In fact, anti-CXCL10 treatment reduces the number of antigen-specific splenic CTLs. Also, CXCL9 and CXCL10 neutralization drastically reduces the number of protective infiltrating CD4+ and CD8+ T cells into the CNS of mice with acute encephalomyelitis caused by mouse hepatitis virus.21,22
Our mRNA data, together with the findings from other models in which CD8+ T cells are essential, prompted us to focus our analysis on the chemokines CXCL9 and CXCL10; however, before proceeding any further with our animal model, we wanted to validate the relevance of the mouse mRNA expression data by investigating the expression of these chemokines in human cases of RMSF. A collection of brainstem samples (where rickettsial lesions are particularly numerous)42
from nine pediatric cases of fatal RMSF was analyzed by immunohistochemistry using specific monoclonal antibodies against CXCL9, CXCL10, CD8, and CXCR3 (Table 2)
. Unfortunately no other tissues were available for analysis; however, we believe that despite the heterogeneous nature of the endothelium of different organs the endothelial chemokine response may be similar in different vascular beds as reflected by the immunohistochemical detection of these chemokines in different organs of rickettsia-infected mice (Figure 4)
. Although it is not possible to make firm conclusions based on the retrospective analysis of relatively few human samples that were subjectively analyzed by two pathologists, we observed that 1) the amount of rickettsial antigen detected in human cases was variable (presumably owing to anti-microbial treatment and/or the host immune response) and did not correlate with the expression of any of the other markers studied; 2) CD8+ and CXCR3+ cells were few or not detectable in cases with a short duration of illness (< 9 days) and were more abundant at later times (>12 days). This observation is consistent with the time required for maturation of the adaptive immune response; and 3) fewer CXCR3+ cells were detected than CD8+ cells. Down-regulation of CXCR3 after CXCL9/10/11 binding to this receptor could explain this finding.43
Perhaps more importantly, the data obtained from human samples validated the usefulness of our mouse model of spotted fever group rickettsioses for pathophysiological and immunological studies that may be relevant to the human spotted fever group rickettsioses.
The expression of CXCL9 and CXCL10 in C3H/HeN mice infected with R. conorii was analyzed at both the mRNA and protein levels. We used real-time PCR with TaqMan probes to analyze cDNA for CXCL9 and CXCL10 in lungs, liver, and brain throughout the course of the disease in mice undergoing both lethal and sublethal infections with R. conorii. The same tissues were also analyzed by immunohistochemistry with the objectives of validating the mRNA data and localizing the protein in the tissues analyzed. Indeed the protein data mirrored the mRNA results. We felt that the immunohistochemical approach was necessary (in contrast with other methods such as enzyme-linked immunosorbent assay for serum samples) because many chemokines, including CXCL9 and CXCL10, bind to glycosaminoglycans on the surface of endothelial cells,44 and this is the fraction that is most likely to be physiologically relevant.
In general, the peak of expression of both CXCL9 and CXCL10 occurred 3 days after intravenous injection of either lethal or sublethal inocula of R. conorii (Figure 3)
. The major exception was the expression of CXCL9 in the brain peaking at day 7 of infection in the sublethal model. On day 3 of rickettsial infection the protein signal was multifocal and extensive in both lungs and brain and essentially generalized in the liver at the level of the sinusoids (Figure 4)
. In all cases the signal localized to endothelial cells, although it was also found in infiltrating leukocytes. At earlier and later times, the signal was more focal. Unfortunately CXCL9 and CXCL10 can only be detected in formalin-fixed tissues and not in tissues fixed with precipitant fixatives such as zinc-based preparations, which allow the more sensitive detection of rickettsial antigen. Our method for detection of rickettsial antigen in formalin-fixed tissues involves enzymatic digestion of the tissue and does not identify the signal before day 4 in the sublethal rickettsial infection model (whereas zinc fixation allows the detection of rickettsial antigen on day 1). In fact, the peak of the rickettsial load in lungs and liver occurs at approximately day 5 as assessed by real-time PCR for the rickettsial gene rompB (Figure 5)
. We hypothesize that the focal expression of CXCL9 and CXCL10 on day 3 in the lungs and brain and day 1 in the liver corresponds to vascular beds infected with rickettsiae.
This is, to the best of our knowledge, the first study addressing the kinetics and localization of expression of T-cell-targeting chemokines throughout the natural history of an infectious disease. One of the most interesting observations was that the peak of expression of these chemokines, which have been regarded as specific for memory/effector Th1 and CD8 T cells, occurred at a time when no infiltrating CD8+ T cells were found in the tissues of rickettsia-infected mice (Figure 3
and Figure 4, h and i
). Because CXCR3, the receptor for CXCL9 and CXCL10, is also expressed by natural killer (NK) cells,45
it is possible that this early peak of chemokine expression might be important for the effector functions of the innate immune system. On the other hand, our data from a mouse endothelial cell line suggest that rickettsial infection does not directly trigger the expression of CXCL9 and CXCL10 and that IFN- or a combination of IFN- and TNF- are necessary to elicit their expression. We have previously found that both cytokines are detectable in the sera of rickettsia-infected mice on the second day of infection.46
A potential early source of these chemokines might be NK cells, and, if this is the case, they could be part of an early positive feedback loop. Unfortunately it was not possible to evaluate the presence of NK cells in our animal model because NK cells of C3H/HeN mice do not have exclusive markers that could allow for their identification in tissues; however, we have shown in two other models of rickettsioses that NK cells become activated.46
Another yet more intriguing possibility isthat rickettsia-infected endothelial cells directly prime antigen-specific CD8+ T cells very early after infection,which can produce an early rise in IFN- before they proliferate and become detectable as infiltrating cells in infected tissues. This hypothesis is plausible because there is evidence that endothelial cells can prime T-cell responses.47-52
This is, therefore, a subject that should be addressed in the rickettsial model because this system offers the unique opportunity to study lymphocyte-endothelial interactions in the context of an intracellular bacterial infection that targets these cells specifically.
It is also possible that the early rise in the expression of CXCL9 and CXCL10 may participate in the process of Th1 polarization and effector cell generation that are so important in the clearance of rickettsiae from infected hosts. There is evidence that these chemokines, particularly CXCL10, play that role in models of arthritis,53 T. gondii infection,20 and in CXCL10 gene knockout mice.54 These knockout mice exhibit decreased Th1 responses including decreased contact hypersensitivity and alloresponses, defective clearance of a virus (mouse hepatitis virus), and decreased numbers of effector CD8+ T lymphocytes.
Although we do not know the importance of endothelial proliferation and angiogenesis in the repair process during a rickettsial infection, given the fact that CXCR3 ligands are angiostatic chemokines,55,56 it is possible that high levels of CXCL9 and CXCL10, such as the ones found in our study, might have a detrimental effect in the repair process of the vasculature. This is also an issue that merits consideration for future studies.
The generalized expression of CXCL9 and CXCL10 in the sinusoids of the liver and the almost generalized expression of these chemokines in the lungs 3 days after rickettsial infection were surprising findings as well because this pattern of expression does not help explain the precise targeting and localization of T cells around infected endothelial cells. It is possible, however, that most of these chemokines on the surface of endothelial cells are partially or totally inactivated by enzymatic modification. Recent evidence indicates that this is a plausible explanation because dipeptidyl peptidase IV (CD26) modifies CXCR3 ligands to become antagonists or low-potency agonists.57
In conclusion, we have shown that among several different T-cell-targeting chemokines the CXCR3 ligands CXCL9 and CXCL10 have the highest levels of expression during endothelial-target infections caused by spotted fever group rickettsiae. Our findings in a mouse model were validated in fatal human cases of RMSF. Also, the morphological evidence showed that endothelial cells are an important source of these chemokines. The fact that the timing and distribution of expression of CXCL9 and CXCL10 do not seem to agree with their expected role as chemoattractants for effector T cells in this rickettsial infection model opens interesting new questions in regard to the function of the early rise and wide distribution of these chemokines. It is likely that they also play a role during the induction phase of the adaptive immune response and/or in guiding effector cells of the innate immune system including macrophages and NK cells. In the future it will be important to determine whether these chemokines are intact or modified by CD26 or other peptidases and whether functional blocking with specific antibodies can affect the outcome of the infection. It will also be important to identify the source of early IFN- production during an experimental rickettsial infection because this is the most likely stimulus that triggers the production of CXCR3 ligands and other chemokines.
|
Footnotes |
|---|
Supported by a grant from the National Institutes of Health (grant RO1 AI21242 to D. H. W.), a College of American Pathologists Foundation Scholars Research Fellowship (to G. V.), and a James W. McLaughlin Predoctoral Fellowship (to G. V.).
Accepted for publication June 18, 2003.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. D. Paddock, S. Fernandez, G. A. Echenique, J. W. Sumner, W. K. Reeves, S. R. Zaki, and C. E. Remondegui Rocky Mountain Spotted Fever in Argentina Am J Trop Med Hyg, April 1, 2008; 78(4): 687 - 692. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Rydkina, A. Sahni, D. J. Silverman, and S. K. Sahni Comparative analysis of host-cell signalling mechanisms activated in response to infection with Rickettsia conorii and Rickettsia typhi J. Med. Microbiol., July 1, 2007; 56(7): 896 - 906. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Fang, N. Ismail, L. Soong, V. L. Popov, T. Whitworth, D. H. Bouyer, and D. H. Walker Differential Interaction of Dendritic Cells with Rickettsia conorii: Impact on Host Susceptibility to Murine Spotted Fever Rickettsiosis Infect. Immun., June 1, 2007; 75(6): 3112 - 3123. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Rydkina, A. Sahni, R. B. Baggs, D. J. Silverman, and S. K. Sahni Infection of Human Endothelial Cells with Spotted Fever Group Rickettsiae Stimulates Cyclooxygenase 2 Expression and Release of Vasoactive Prostaglandins Infect. Immun., September 1, 2006; 74(9): 5067 - 5074. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. H.J. Looijenga, R. Hersmus, A. J.M. Gillis, R. Pfundt, H. J. Stoop, R. J.H.L.M. van Gurp, J. Veltman, H. B. Beverloo, E. van Drunen, A. Geurts van Kessel, et al. Genomic and Expression Profiling of Human Spermatocytic Seminomas: Primary Spermatocyte as Tumorigenic Precursor and DMRT1 as Candidate Chromosome 9 Gene Cancer Res., January 1, 2006; 66(1): 290 - 302. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Zeng, T. A. Moore, M. W. Newstead, J. C. Deng, S. L. Kunkel, A. D. Luster, and T. J. Standiford Interferon-Inducible Protein 10, but Not Monokine Induced by Gamma Interferon, Promotes Protective Type 1 Immunity in Murine Klebsiella pneumoniae Pneumonia Infect. Immun., December 1, 2005; 73(12): 8226 - 8236. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. R. Clifton, E. Rydkina, R. S. Freeman, and S. K. Sahni NF-{kappa}B Activation during Rickettsia rickettsii Infection of Endothelial Cells Involves the Activation of Catalytic I{kappa}B Kinases IKK{alpha} and IKK{beta} and Phosphorylation-Proteolysis of the Inhibitor Protein I{kappa}B{alpha} Infect. Immun., January 1, 2005; 73(1): 155 - 165. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. P. Widney, Y. Hu, A. K. Foreman-Wykert, K. C. Bui, T. T. Nguyen, B. Lu, C. Gerard, J. F. Miller, and J. B. Smith CXCR3 and Its Ligands Participate in the Host Response to Bordetella bronchiseptica Infection of the Mouse Respiratory Tract but Are Not Required for Clearance of Bacteria from the Lung Infect. Immun., January 1, 2005; 73(1): 485 - 493. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. VALBUENA and D. H. WALKER EFFECT OF BLOCKING THE CXCL9/10-CXCR3 CHEMOKINE SYSTEM IN THE OUTCOME OF ENDOTHELIAL-TARGET RICKETTSIAL INFECTIONS Am J Trop Med Hyg, October 1, 2004; 71(4): 393 - 399. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Li, Z. Ma, Z.-L. Tang, T. Stevens, B. Pitt, and S. Li CpG DNA-mediated immune response in pulmonary endothelial cells Am J Physiol Lung Cell Mol Physiol, September 1, 2004; 287(3): L552 - L558. [Abstract] [Full Text] [PDF]
|
|
| ||||||
