Early cytokine profiles in the joint define pathogen clearance and severity of arthritis in Chlamydia-induced arthritis in rats.

ARTHRITIS & RHEUMATISM
Vol. 54, No. 2, February 2006, pp 499–507
DOI 10.1002/art.21643
© 2006, American College of Rheumatology
Early Cytokine Profiles in the Joint
Define Pathogen Clearance and Severity of Arthritis
in Chlamydia-Induced Arthritis in Rats
Robert D. Inman and Basil Chiu
both synovial tissues and spleen compared with Lewis
rats. Local cytokine profiles demonstrated that host
resistance was characterized by enhanced synovial expression of tumor necrosis factor ␣, interferon-␥
(IFN␥), and interleukin-4.
Conclusion. These studies demonstrated that cytokines thought to be proinflammatory in nature can
play an important role in host defense in infectiontriggered arthritis and serve to highlight the dynamic
cytokine relationships that constitute effective host–
pathogen interactions.
Objective. Although Chlamydia trachomatis–
induced arthritis is among the most common rheumatic
diseases having an identified infectious trigger, the
pathogenesis of this arthritis is not well defined. We
sought to investigate the host–microbe interactions that
contribute to the severity of arthritis initiated by chlamydial infection.
Methods. We established an experimental rat
model of C trachomatis–induced arthritis that recapitulates many pathologic features of the clinical disease.
The severity of the arthritis was defined using an
established histopathologic scoring system. Host clearance of the pathogen and local cytokine production were
examined by enzyme-linked immunosorbent assays.
Results. Lewis rats were susceptible to C
trachomatis–induced arthritis, whereas BN rats were
relatively resistant to this disease. Significant differences in the histopathologic severity of arthritis were
originally observed on day 21, and this prompted an
examination of the acute phase of the arthritis. As early
as day 5 after the onset of the arthritis, pathologic
changes in Lewis rats were more severe than those in
BN rats. An evaluation of the role of complement using
cobra venom factor treatment excluded complement as
being the key to differential sensitivity, because decomplementation did not eliminate the differences in
arthritis severity between Lewis and BN rats. Host
clearance, in contrast, was significantly different between the rat strains, with BN rats showing more
prompt and effective clearance of the pathogen from
Infection as a triggering factor in chronic arthritis
continues to be an important concept in studies into the
pathogenesis of rheumatic diseases, and reactive arthritis (ReA) remains the most well-established example of
this interaction (1). In the clinical spectrum of ReA,
antecedent infection with Chlamydia trachomatis is the
most common association (2). Natural history studies
have emphasized the potential for chronicity in ReA,
and a previous study documented that chronic symptoms
persist 5 years after the onset of ReA in the majority of
patients (3). However, the host factors that dictate
postinfectious sequelae such as ReA have not been
defined and have proved difficult to resolve in the
clinical setting.
We previously developed an experimental model
for C trachomatis–induced arthritis (4). In the chronic
phase of this model, the synovial tissues are culturenegative, and the cellular infiltrate in the joint consists of
a mononuclear lymphocytic population. From the aspects of histopathology, clinical course, and immune
response, this experimental arthritis recapitulates the
events seen in clinical Chlamydia-induced arthritis. In
the present study, we examined rat strains exhibiting
differential sensitivity to C trachomatis–induced arthritis
in order to identify the early synovial cytokine profile
defining effective host clearance of the pathogen and, in
Supported by the Canadian Institutes of Health Research.
Robert D. Inman, MD, Basil Chiu, PhD: Toronto Western
Research Institute, University of Toronto, Toronto, Ontario, Canada.
Address correspondence and reprint requests to Robert D.
Inman, MD, Arthritis Center of Excellence, Toronto Western Hospital, 399 Bathurst Street, Toronto, Ontario M5T 2S8, Canada. E-mail:
Robert.inman@uhn.on.ca.
Submitted for publication July 25, 2005; accepted in revised
form November 10, 2005.
499
500
INMAN AND CHIU
turn, the chronicity of the arthritis. The choice of the
Lewis strain to establish this model initially was based on
the fact that it is susceptible to a number of experimental
arthritis models (5), including adjuvant-induced arthritis, collagen-induced arthritis (6,7), as well as Yersiniainduced arthritis (8–11). Studies on the rat model of
collagen-induced arthritis have mainly focused on the
genetic basis for the development of the disease and the
antibody response to collagen.
This genetics work was pioneered by Griffiths
and associates and has been extensively reviewed (12). It
was determined that the major histocompatibility complex (MHC) plays an important part in resistance or
susceptibility to the disease among rat strains (6,7). In
general, high antibody response to type II collagen
indicates susceptibility to the development of disease.
However, nonarthritic rats have been shown to have
high antibody titers (7,13,14). Therefore, the relationship of antibody titer and disease is not absolute. Working with uncommon WA/KIR rats, Rahman and Staines
(15) isolated and cultured antibody-producing cells from
the spleen, lymph nodes, and bone marrow of rats. They
observed no difference in anticollagen antibody production between arthritic and nonarthritic rats. Splenectomy led to more severe disease and earlier onset. They
also observed that if experimental animals were given
splenocytes obtained from classically immunized rats,
the severity of arthritis was greatly reduced. Working
mostly with DA rats, Griffiths et al (12,16) identified a
series of genetic loci that they collectively named Cia,
which modify the severity of disease development.
In order to define the pathogenesis of joint
inflammation in C trachomatis–induced arthritis, we
compared different rat strains for differential susceptibility. We included BN rats as a control, because this rat
strain is known to show relative resistance to collageninduced arthritis. Here, we report the contrasting profile
of host susceptibility, represented by Lewis rats, and
resistance, represented by BN rats, in response to experimental Chlamydia-induced arthritis.
MATERIALS AND METHODS
Rats. Eight-week-old male rats were purchased from
Harlan Sprague Dawley (Indianapolis, IN). The animals were
maintained in microisolators in the animal care facility of the
Toronto Western Hospital. All animals studied were younger
than 12 weeks of age. All studies were approved by the
University Health Network Animal Care Committee.
Induction of arthritis. Arthritis was induced in the rats
by intraarticular injection of synoviocyte-packaged Chlamydia,
as previously described (4). Briefly, C trachomatis serotype L2
was inoculated onto monolayers of rat synovial fibroblasts in
tissue culture. These stable synovial fibroblast lines were
developed as previously described (4). After overnight incubation, cells were harvested and adjusted to 5 ⫻ 105 cells/ml. Rats
were anesthetized with isoflurane (Zenoca Pharma, Mississauga, Ontario, Canada), and 0.2 ml of the infected cells
containing 2 ⫻ 105 colony-forming units of Chlamydia was
injected into one knee joint of the rats of the respective strains.
Joint swelling was measured with calipers and recorded in
millimeters. Animals were killed on days 2, 7, or 21 postinjection. Mock injections on noninfected synoviocytes demonstrated only a transient inflammation in the joints (4).
Histologic analysis. At necropsy, the arthritic knee
joints were removed, fixed in formalin, and decalcified. They
were then processed and stained with hematoxylin and eosin.
The joint sections were evaluated and scored according to a
published scoring system (17). Briefly, 5 criteria were used to
grade the joints: synovitis of the synovial tissue, inflammatory
cell infiltration, joint space exudation, pannus formation, and
bone/cartilage damage. Each category was assigned a score of
0–4, where 0 ⫽ normal and 4 ⫽ extreme pathologic change
(maximum possible score ⫽ 20). The score for each joint
section was then calculated.
Chlamydia clearance in vivo. A quantitative assay for
Chlamydia in synovial tissues of the injected knee joints was
adapted from the IDEIA PCE Chlamydia enzyme-linked
immunosorbent assay (ELISA) kit (Dako, Cambridgeshire,
UK), which detects the Chlamydia genus only (antilipopolysaccharide specificity). It is a capture ELISA–based
assay using a patented polymer chain–conjugated antibody
technique and is as sensitive as current polymerase chain
reaction (PCR) kits. This kit was originally developed for
clinical use but has previously been used successfully in a
murine experimental model (18).
The joints of the test animals were carefully dissected
to harvest the synovial tissues. These tissues were weighed,
ground to a pulp, and then resuspended in saline at 40 mg of
wet weight per milliliter. The samples were then processed as
recommended by the manufacturer. A similar analysis was
applied to splenic mononuclear cells that were harvested on
day 7 postinjection (see below).
Splenocyte clearance of Chlamydia. Lewis and BN rats
were killed 7 days postinjection, and the spleens were recovered and homogenized. The amount of Chlamydia was assayed
using the Dako IDEIA PCE Chlamydia kit, adapted for use in
rats (18).
Cytokine analysis of synovial tissues. Synovial tissues
were dissected, individually weighed, and dissolved in a Tris–
saline–EDTA buffer containing 0.5 mM DTT and 1% Triton
X-100, as described by Niki et al (19). The dot enzyme
immunoassay procedure is similar in principle to a dot-blot
assay, except that the tissue lysates are dotted onto small
nitrocellulose discs in the wells of tissue culture plates instead
of dotted onto a nitrocellulose membrane in a dot-blot apparatus frame. Five milliliters of the tissue lysate was dotted onto
each disc, air-dried in the cold, and stored at ⫺70° until
assayed. The assay was performed directly in the wells.
Goat polyclonal antibodies to rat interleukin-1␤ (IL1␤), IL-4, IL-6, IL-10, interferon-␥ (IFN␥), tumor necrosis
factor ␣ (TNF␣), and transforming growth factor ␤ were
purchased from Santa Cruz Biotechnology (Santa Cruz, CA)
HOST–MICROBE INTERACTIONS IN CHLAMYDIA-INDUCED ARTHRITIS
and used at a 1:20 dilution. Antibodies to macrophage inflammatory protein 1␣ and monocyte chemoattractant protein 1
were obtained from Chemicon (Temecula, CA). Discs were
incubated for 90 minutes at 37°C. Each well containing a disc
was washed and incubated with the secondary peroxidaseconjugated anti-goat IgG antibody (1:8,000 dilution) (Sigma,
St. Louis, MO) for 90 minutes at 37°C. Thereafter, color
development utilized o-phenylenediamine and H2O2. After the
reaction was stopped, the fluid from each well was transferred
into wells of a plate for reading in an ELISA reader.
Cobra venom factor (CVF) study. In order to evaluate
the role of complement in the differential host response to the
intraarticular pathogen, we studied the impact of complement
depletion in the respective rat strains. Similar approaches have
been used in studying arthritis in the rat in different experimental models (20,21). CVF was purchased from Quidel (San
Diego, CA). It has a vender-defined stock concentration of 89
units/ml CVF and was administered intraperitoneally at 25
units/kg. The CVF was given at the same time as the intraarticular injection of Chlamydia. On day 7, the rats were killed,
and EDTA plasma was assayed for C3a desArg as a marker of
C3 activation of C3. The ELISA kit for measuring rat complement C3a desArg was from Cedarlane (Hornby, Ontario,
Canada).
Statistical analysis. All results are expressed as the
mean ⫾ 1SD. Student’s 2-tailed t-test was used for statistical
comparisons.
501
observed. Joint spaces were usually clear, with only light
to moderate exudate. Pannus was present, but pannus
formation in BN rats was never as aggressive as that
observed in Lewis rats. Compared with Lewis rats, BN
rats infrequently had erosions, and when erosions were
observed in BN rats, they were only superficial.
Twenty-one days postinjection, synovial tissues
from Lewis rats were hypertrophic, with extensive infiltration of leukocytes consisting of both neutrophils and
mononuclear cells (Figure 2). Bone and cartilage damage was evident, resulting in a picture of extensively
disrupted joint architecture. In contrast, synovial tissues
from BN rats showed only mild hypertrophic changes on
day 21, and the joint structure approximated that of a
normal joint.
Quantitative analysis of the histopathologic
changes confirmed the significant differences between
the 2 strains of rats. Two days after injection, the
pathology score for Lewis rats had already reached 10.0
on the Yang-Hamilton scale (17), which has a maximum
score of 20. The mean pathology scores rose quickly, to
near-maximum scores of 18.6 by day 7 postinjection, and
RESULTS
Histology. The initial impression of only transient
joint swelling in the BN rat prompted an examination of
the histopathologic characteristics of the joints in the 2
rat strains. On day 2 postinjection, both strains had
moderate infiltration of leukocytes into the synovial
tissues, with an impression of neutrophilic predominance in Lewis rats and mononuclear predominance in
BN rats. On day 7 postinjection, the contrast in pathologic changes was more dramatic (Figure 1). The injected knee joints of Lewis rats were very densely
infiltrated with neutrophils. There were focal points of
necrosis in the synovial tissue of some animals. Synovial
hyperplasia and hypertrophy were extensive. The joint
space contained a significant exudate, which was primarily neutrophilic. Pannus formation was extensive. Significant erosion of bone and articular cartilage was present.
In some cases, pannus had advanced into the subchondral bone and was seen proliferating into the bone
marrow space.
Seven days postinjection, the pathologic changes
in BN rats were much milder than those in Lewis rats.
The predominant infiltrating cells were mononuclear,
and cellular infiltration was much more limited than that
seen in Lewis rats. Hypertrophy and hyperplasia of the
synovial tissues were minimal, and necrosis was not
Figure 1. Joint histopathologic changes on day 7 postinjection in
Lewis rats (A) and BN rats (B) with Chlamydia trachomatis–induced
arthritis. Hematoxylin and eosin staining shows more aggressive synovitis in Lewis rats than in BN rats. Original magnification ⫻ 10.
502
INMAN AND CHIU
Table 1. Effect of decomplementation on differential joint swelling
in Lewis rats and BN rats*
Treatment
C trachomatis only†
Rat 1
Rat 2
Rat 3
Rat 4
Rat 5
C trachomatis ⫹ CVF‡
Rat 1
Rat 2
Rat 3
Rat 4
Rat 5
Lewis
BN
12.9
12.3
13.1
13.7
13.2
10.8
10.5
10.3
12.5
10.8
13.7
13.7
12.6
12.5
13.5
12.0
12.3
9.9
12.4
10.6
* Values are measurements (in millimeters) of the right knee on day 7.
BN rats exhibited significantly less swelling in the Chlamydia
trachomatis–injected joint than did Lewis rats, in both the absence
(P ⫽ 0.002) and the presence (P ⫽ 0.01) of cobra venom factor (CVF).
CVF had no significant impact on swelling in C trachomatis–injected
joints in either Lewis rats (P ⫽ 0.66) or BN rats (P ⫽ 0.49).
† The mean ⫾ SD values for Lewis rats and BN rats were 13.04 ⫾ 0.50
mm and 10.98 ⫾ 0.87 mm, respectively.
‡ The mean ⫾ SD values for Lewis rats and BN rats were 13.02 ⫾ 0.60
mm and 11.44 ⫾ 1.12 mm, respectively.
Figure 2. Joint histopathologic changes on day 21 postinjection in
Lewis rats (A) and BN rats (B) with Chlamydia trachomatis–induced
arthritis. Hematoxylin and eosin staining shows continued active
synovitis in Lewis rats, with BN rats showing no inflammation or joint
injury. Original magnification ⫻ 10.
remained high (16.2) on day 21. For the BN rats, the day
2 score was 6.0. The highest score reached was only 13.2;
this score was reached on day 7 and then declined
rapidly back to that of noninflamed joint architecture on
day 21. The scores of the BN rats were significantly
lower (P ⬍ 0.05) than those of the Lewis rats on days 2,
7, and 21 postinjection (n ⫽ 10 rats at each time point).
Influence of complement. We addressed the role
of complement in this process by using CVF to decomplement plasma in vivo. It was noted that on day 7
after induction of arthritis, the level of C3a desArg was
higher in BN rats (mean ⫾ SD OD units 0.87 ⫾ 0.19)
than in Lewis rats (0.44 ⫾ 0.11) (P ⬍ 0.001). After CVF
treatment, this difference was not observed (P ⫽ 0.805),
with both strains showing comparably increased levels of
C3a desArg. Thus, despite eliminating the differential
complement levels, there was no impact on the severity
of arthritis, and the differential response in swelling
between the strains was still observed after decomplementation (Table 1).
Chlamydia clearance. The clearance of Chlamydia in synovial tissues on day 7 postinjection was
quantitated by ELISA (Figure 3). Lewis rats had a
significantly greater Chlamydia load than did BN rats
(mean OD values 0.791 and 0.141, respectively; P ⫽
0.0018). This difference was also reflected in systemic
clearance of Chlamydia, using spleen as the target tissue
Figure 3. Clearance of Chlamydia in synovial tissues obtained from
Lewis (Lew) and BN rats. On day 7, quantitative analysis of Chlamydia
showed more effective host clearance of the pathogen from the joints
of BN rats (n ⫽ 3) than from the joints of Lewis rats (n ⫽ 3). Values
are the mean and SD. OD ⫽ optical density.
HOST–MICROBE INTERACTIONS IN CHLAMYDIA-INDUCED ARTHRITIS
503
correlation of synovial tissue weight (reflecting inflammatory infiltration, hypertrophy, and edema) and synovial IFN␥ expression (r ⫽ ⫺0.86). This inverse correlation was also observed for TNF␣ and synovial tissue
mass (r ⫽ ⫺0.70).
DISCUSSION
Figure 4. Effect of cobra venom factor (CVF) on recovery of Chlamydia trachomatis (Ct) from splenocytes. On day 7, clearance of C
trachomatis from splenocytes was enhanced in BN rats (n ⫽ 3)
compared with Lewis rats (n ⫽ 3), both in the absence (P ⬍ 0.005) and
the presence (P ⬍ 0.05) of CVF treatment. Values are the mean and
SD. OD ⫽ optical density.
to assess dissemination and persistence. Splenocytes
from BN rats showed less recovery of the organism than
did splenocytes from Lewis rats, and this was true in the
presence and in the absence of systemic decomplementation with CVF (Figure 4).
Synovial cytokine profiles in Lewis rats and BN
rats. An analysis of synovial cytokine expression in Lewis
rats and BN rats revealed no differences between the 2
strains for the following cytokines: IL-1, IL-6, IL-10,
MIP, and MCP-1 (Figure 5). However, BN rats showed
significantly higher synovial expression of IL-4, IFN␥,
and TNF␣ (P ⬍ 0.05 for each comparison). When the
tissues were analyzed collectively, there was an inverse
Figure 5. Cytokine levels in synovial tissue on day 5. Local production
of interleukin-4 (IL-4), interferon-␥ (IFN␥), and tumor necrosis factor
␣ (TNF␣) was enhanced in BN rats (n ⫽ 6) compared with Lewis rats
(n ⫽ 6) (P ⬍ 0.05 for all 3 cytokines). OD ⫽ optical density; TGF ⫽
transforming growth factor; MIP ⫽ macrophage inflammatory protein;
MCP ⫽ monocyte chemoattractant protein. Values are the mean and
SD.
ReA refers to an aseptic, seronegative arthritic
condition in which genetically susceptible individuals
experience joint inflammation as a sequela of certain
bacterial infections. Patients in whom ReA develops
commonly, but not invariably, express HLA–B27 or one
of the related cross-reactive group (CREG) antigens of
the class I MHC. Arthritis develops following either a
gastrointestinal infection with gram-negative organisms
such as Salmonella, Yersinia, Shigella, or Campylobacter
species or a genitourinary tract infection with C trachomatis. Despite the strong circumstantial clinical evidence
linking chlamydial infection to subsequent arthritis, the
mechanism underlying this process has not been defined.
Recently, there has been mounting evidence that
bacteria may indeed be found in the joint in this setting.
The demarcation line between septic arthritis and ReA
has since been blurred as a result of such evidence.
Several groups of investigators, using immunofluorescence techniques, identified Chlamydia-like inclusion
bodies in synovial biopsy specimens from patients with
ReA (22–24). Schumacher et al (22) also observed
characteristic electron-dense bodies of Chlamydia in
synovial specimens. Subsequently, it was reported that
chromosomal DNA material encoding Chlamydia ribosomal RNA (rRNA) could be demonstrated in synovial
samples (25). Besides the gene sequence for 16S rRNA,
Bas et al (26) detected a major outer membrane protein
(MOMP) gene sequence and 2 chlamydial plasmid
sequences. Several investigators then reported obtaining
positive PCR results for 16S rRNA (27,28). Gerard et al
(29) took synovial specimens that tested positive for 16S
rRNA by PCR and then further identified chlamydial
gene products (glyQS, r-protein S5 and L5, and Hsp60)
within a majority of these biopsy specimens.
Nonetheless, as molecular detection technologies
began to show the existence and the importance of
chlamydial genetic material within the inflamed synovial
tissue of ReA joints, several studies raised some question about the specificity of PCR results in synovial
tissue. Using PCR, Schumacher et al observed C trachomatis 16S rRNA as well as MOMP gene materials in
synovial biopsy specimens from normal volunteers and
patients with rheumatoid arthritis (30) and subsequently
504
reported that the chlamydial 16S and MOMP, as well as
plasmid gene materials, could be found by PCR in some
biopsy specimens from osteoarthritic joints (31).
In view of the difficulties defining precise mechanisms of C trachomatis–induced arthritis in the clinical
setting, we developed an animal model that recapitulates
many of the features seen in clinical ReA (4). What
begins as a septic process in the acute phase develops
into a chronic arthritis that is resistant to antibiotic
treatment and from which viable organisms cannot be
recovered. However, the host factors that mediate this
process were not defined, and this may shed light on the
earliest events in the pathogenesis of C trachomatis–
induced arthritis.
In the course of our studies, we discovered that
BN rats demonstrated a relative resistance to intraarticular challenge with C trachomatis, while Lewis rats
demonstrated an aggressive arthritis following this challenge. In BN rats, clinical events following the challenge
were much milder than those observed in Lewis rats,
with less joint swelling and a transient clinical course.
These events were mirrored in significant differences in
the histopathologic findings in the affected joints. The
BN rats had a mononuclear leukocyte synovial response
to the intraarticular injection of Chlamydia, while the
acute response in Lewis rats was dominated by neutrophils. In Lewis rats, lymphoid cells were not apparent
until the chronic phase of disease. More precise immunostaining was hampered by the decalcification process,
which complicates detection of surface markers on infiltrating cells in the joint. Despite the lack of phagocytic
cells, BN rats were able to clear bacteria from the
injected knees much more effectively than were Lewis
rats. In Lewis rats, massive infiltration of phagocytic cells
was not effective in this respect and likely contributed to
the extensive joint damage seen in these animals. BN
rats are very different in this regard and responded in a
most effective way to counter the microbial challenge in
the joint. The lymphoid infiltration seemed to be part of
a self-limiting process that is associated with much less
collateral joint damage.
The differences in host response to the pathogen
occur in the acute phase of arthritis, which suggests that
innate immune differences are playing a key role here.
In this regard, we evaluated whether differences in host
complement might be playing an important part, but
decomplementation was not associated with any differential response that would support this notion. Curfs et
al (10) found that Lewis rats were susceptible to a
Yersinia-induced ReA, while BN and Fischer rats were
resistant. Hill and Yu (8) also observed that Fischer,
INMAN AND CHIU
Buffalo, and DA rats are resistant to Yersinia-induced
ReA. It is of interest that the BN rat is recognized for its
tendency to develop certain autoimmune diseases. The
best-studied model is the asthma model in which BN rats
preimmunized with ovalbumin are challenged with an
aerosol of the immunogen, resulting in a pulmonary
infiltration of eosinophils, neutrophils, and mononuclear
cells. Lewis rats are resistant in this model (32).
The host response to chlamydial infection is a
complex interrelationship of cellular and humoral immunity (2). Increased production and secretion of proinflammatory mediators in the clinical setting of C
trachomatis–induced arthritis have been examined by
several groups of investigators (33,34). Studies suggest
that host cell responses are diverse and dependent on
the temporal course. Analyses of clinical samples in the
chronic phase have suggested that a mixture of monocytes and T cell cytokines, as well as proinflammatory
mediators produced by resident synoviocytes, is present
at sites of inflammation. Mononuclear cells in synovial
fluid have been shown to express high levels of IL-4 and
IL-10. Consistent with culture data, more IL-4–
expressing cells than IFN␥-expressing cells were observed in the synovial membrane (35).
The time course of most clinical cases of C
trachomatis–induced arthritis cannot be defined precisely, but studies to date have focused on chronic
changes, in part because it is difficult to examine tissue
and serum during the induction of arthritis, because
patients are rarely seen at that time. In vitro studies of
Chlamydia-infected synoviocytes have demonstrated the
production of soluble mediators by resident synoviocytes
that could influence local immune and inflammatory
responses. Rodel et al (36) showed that chlamydial
infection up-regulates indoleamine 2,3-dioxygenase
gene expression, and the latter was enhanced by TNF␣.
Chlamydial infection induced type I IFN␣/␤ in these
cells, presumably via activation of IFN regulatory factor
1 and IFN-stimulated gene factor 3. The precise cellular
difference accounting for the differences between Lewis
and BN rats has not been fully defined. We have
conducted in vitro studies examining chlamydia invasion
into synoviocytes from the 2 rat strains but found no
difference in this primary event.
The role of cytokines in the pathogenesis of ReA
has proven to be a complex issue to resolve definitively.
It is known that IFN␥ and TNF␣ are indispensable for
an effective defense against bacterial infection. In ReA,
failure of effective bacterial elimination at the initiation
of the disease may be attributable to a relative lack of
Th1 cytokine production. Several studies have demon-
HOST–MICROBE INTERACTIONS IN CHLAMYDIA-INDUCED ARTHRITIS
strated a relative decrease in Th1 cytokines in peripheral
blood and synovium of patients with ReA (35,37). In a
study of 11 patients with ReA, it was shown that
stimulation of synovial fluid mononuclear cells resulted
in low levels of IFN␥ and TNF␣ but high amounts of
IL-10 (35). In T cell clones derived from the synovial
fluid of patients with ReA, disease-related bacterial
antigens predominantly induced Th1 cytokine secretion
(38–40).
It is interesting to note that in Chlamydia-induced
ReA, synovial fluid levels of IFN␥ were lower in HLA–
B27–positive patients than in HLA–B27–negative patients (41). This bears an interesting relationship to our
findings in the present study, in which as-yetuncharacterized genetic susceptibility conferring sensitivity to experimental C trachomatis–induced arthritis
was also associated with a relative lack of IFN␥ production locally. A comparable analogy can be seen in the
significant associations observed between low TNF␣
secretion and a more chronic course of ReA in HLA–
B27–positive patients, but that is not always the case
(47,42). This also could represent the clinical parallel to
that seen in the current study, in which relatively impaired host TNF␣ production was associated with the
chronicity and severity of the experimental C
trachomatis–induced arthritis. It is interesting to note
that the TNF alleles that have been associated with
ankylosing spondylitis are those that confer lower, not
higher, TNF production (43). This paradoxical finding
has been difficult to explain. Our findings in experimental
arthritis suggest that an impairment during the acute phase
of arthritis might set the stage for a more chronic course.
In comparison with synovial fluid levels of TNF␣
in rheumatoid arthritis, synovial fluid levels of TNF␣ in
ReA are lower, despite comparable levels of IL-2 receptor. A lower expression of TNF␣ has been observed in
peripheral blood lymphocytes from HLA–B27–positive
individuals; therefore, impairment of such microbicidal
cytokines may set the stage for an altered host response
to infection. This possibility is supported by our experimental studies in which deficiency of TNF␣ was associated with more severe infection and more severe arthritis (44). It is difficult in the clinical setting to resolve
whether the relative shift to Th2-type cytokine production in ReA (in contrast to rheumatoid arthritis) reflects
a genetic predilection for a blunted host response to
such infections, or whether this is a profile that reflects
the host attempt at gradual, but eventual, clearance of
the pathogen. The role of innate immunity in the host
response to Chlamydia has just recently come under
scrutiny (45). Netea et al reported that Chlamydia
505
pneumoniae stimulates IFN␥ synthesis through MyD88dependent, Toll-like receptor 2 (TLR-2)– and TLR-4–
independent induction of IL-18 release (46). Polymorphisms in CD14 may confer differential host
responsiveness to chlamydia (47). Eng et al recently
reported a CD14 promoter polymorphism associated
with CD14 expression and Chlamydia-stimulated TNF␣
production (48). Recent studies have demonstrated that
neutrophil Rac and TLR-4 may play important roles in
this context (49).
The role of cytokines in the joint has become an
important issue, not only for understanding the pathogenesis of arthritis but also because of the advent of
anticytokine therapies. Such treatment in patients with
spondylarthropathies may have a tendency to normalize
impaired Th1 cytokine production, but this issue has not
been finally resolved (50–52). Our findings highlight the
fact that cytokine expression in the joint is dynamic, with
beneficial and deleterious effects varying according to
the time of observation. Furthermore, the complex
microenvironment of the joint reflects a finely balanced
system, and interrupting this cytokine circuitry may have
unexpected effects. This is not yet resolved in clinical
ReA.
REFERENCES
1. Inman RD, Perl A, Phillips PE. Infectious agents in chronic
rheumatic diseases. In: Koopman WJ, Moreland LW, editors.
Arthritis and allied conditions: a textbook of rheumatology. 15th
ed. Philadelphia: Lippincott Williams & Wilkins; 2005. p. 647–8.
2. Inman RD, Whittum-Hudson JA, Schumacher HR, Hudson AP.
Chlamydia and associated arthritis. Curr Opin Rheumatol 2000;
12:254–60.
3. Thomson GT, DeRubeis DA, Hodge MA, Rajanayagam C, Inman
RD. Post-Salmonella reactive arthritis: late clinical sequelae in a
point source cohort. Am J Med 1995;98:13–21.
4. Inman RD, Chiu B. Synoviocyte-packaged Chlamydia trachomatis
induces a chronic aseptic arthritis. J Clin Invest 1998;102:1776–82.
5. Joe B, Griffiths M, Remmers EF, Wilder RL. Animal models of
rheumatoid arthritis and related inflammation. Curr Rheumatol
Rep 1999;1:139–48.
6. Griffiths MM, Eichwald EJ, Martin JH, Smith CB, DeWitt CW.
Immunogenetic control of experimental type II collagen–induced
arthritis. I. Susceptibility and resistance among inbred strains of
rats. Arthritis Rheum 1981;24:781—94.
7. Griffiths MM, DeWitt CW. Genetic control of collagen-induced
arthritis in rats: the immune response to type II collagen among
susceptible and resistant strains and evidence of multiple gene
control. J Immunol 1984;132:2830–6.
8. Hill JL, Yu DT. Development of an experimental animal model
for reactive arthritis induced by Yersinia enterocolitica infection.
Infect Immun 1987;55:721–6.
9. Toivanen A, Merillahti-Palo R, Gripenberg C, Lahesman-Rantala
R, Soderstrom KO, Jaakkola UM. Yersinia-associated arthritis in
the rat: experimental model for human reactive arthritis. Acta
Path Microb Immun Scand 1986;94:261–9.
10. Curfs JH, Meis JF, van der Lee HA, Mulder JA, Kraak WA,
506
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
INMAN AND CHIU
Hoogkamp-Korstanje JA. Persistent Yersinia enterocolitica infection in three rat strains. Microb Pathog 1995;19:57–63.
Odinot RT, Curfs JH, Meis JF, Melchers WJ, HoogkampKorstanje JA. Local expression of cytokine mRNA in spleen and
Peyer’s patches of rats is involved in resistance against infection
with Yersinia enterocolitica. Cytokine 1998;10:206–12.
Joe B, Griffiths MM, Remmers EF, Wilder RL. Animal models of
rheumatoid arthritis and related inflammation. Curr Rheumatol
Rep 1999;1:139–48.
Clague RB, Morgan K, Shaw MJ, Holt PJ. Native type II collagen
induced arthritis in rat. J Rheumatol 1980;7:775–82.
Larsson P, Kleinan S, Holmdahl R, Klareskog L. Homologous
type II collagen–induced arthritis in rats: characterization of the
disease and demonstration of clinically distinct forms of arthritis in
two strains of rats after immunization with the same collagen
preparation. Arthritis Rheum 1990;33:693–701.
Rahman J, Staines NA. Contribution of the spleen, lymph nodes
and bone marrow to the antibody response in collagen induced
arthritis in the rat. Clin Exp Immunol 1991;85:48–54.
Griffiths MM, Wang J, Joe B, Dracheva S, Kawahito Y, Cannon
GW, et al. Identification of four new quantitative trait loci
regulating arthritis severity and one new quantitative trait locus
regulating autoantibody productions in rats with collagen-induced
arthritis. Arthritis Rheum 2000;43:1278–89.
Yang YH, Hamilton JA. Dependence of interleukin-1–induced
arthritis on granulocyte–macrophage colony-stimulating factor.
Arthritis Rheum 2001;44:111–9.
Haranaga S, Yamaguchi H, Friedman H, Izumi SI, Yamamoto Y.
Chlamydia pneumoniae infects and multiplies in lymphocytes in
vitro. Infect Immun 2001;69:7753–9.
Niki Y, Yamada H, Seki S, Kikuchi T, Takaishi H, Toyama Y, et
al. Macrophage- and neutrophil-dominant arthritis in human IL-1
transgenic mice. J Clin Invest 2001;107:1127–113.
Gaede K, Baumeister E, Heeseman J. Decomplementation by
cobra venom factor suppresses Yersinia-induced arthritis in rats.
Infect Immun 1995;63:3697–701.
Goodfellow RM, Williams AS, Levin JL, Williams BD, Morgan
BP. Soluble complement receptor one (sCR1) inhibits the development and progression of rat collagen induced arthritis. Clin Exp
Immunol 2000;19:210–6.
Schumacher HR Jr, Magge S, Cherian PV, Sleckman J, Rothfuss
S, Clayburne G, et al. Light and electron microscopic studies on
the synovial membrane in Reiter’s syndrome: immunocytochemical identification of chlamydial antigen in patients with early
disease. Arthritis Rheum 1988;31:937–46.
Taylor-Robinson D, Thomas BJ, Dixey J, Osborn MF, Furr PM,
Keat AC. Evidence that Chlamydia trachomatis causes seronegative arthritis in women. Ann Rheum Dis 1988;47:295–9.
Inman RD, Chiu B, Katz A. Immunological features of “reactive
arthritis” due to Chlamydia trachomatis. Arthritis Rheum 1989;32
Suppl 1:S113.
Rahman MU, Cheema MA, Schumacher HR, Hudson AP. Molecular evidence for the presence of chlamydia in the synovium of
patients with Reiter’s syndrome. Arthritis Rheum 1992;35:521–9.
Bas S, Griffais R, Kvien T, Glennas A, Melby K, Vischer TL.
Amplification of plasmid and chromosome Chlamydia DNA in
synovial fluid of patients with reactive arthritis and undifferentiated seronegative oligoarthropathies. Arthritis Rheum 1995;38:
1005–13.
Nanagara R, Li F, Beutler A, Hudson A, Schumacher HR Jr.
Alteration of Chlamydia trachomatis biologic behavior in synovial
membranes: suppression of surface antigen production in reactive
arthritis and Reiter’s syndrome. Arthritis Rheum 1995;38:1410–7.
Branigan PJ, Gerard HC, Hudson AP, Schumacher HR Jr, Pando
J. Comparison of synovial tissue and synovial fluid as the source of
nucleic acids for detection of Chlamydia trachomatis by polymerase chain reaction. Arthritis Rheum 1996;39:1740–6.
29. Gerard HC, Branigan PJ, Schumacher HR Jr, Hudson AP.
Synovial Chlamydia trachomatis in patients with reactive arthritis/
Reiter’s syndrome are viable but show aberrant gene expression.
J Rheumatol 1998;25:734–42.
30. Schumacher HR, Arayssi T, Crane M, Lee J, Gerard H, Hudson
AP, et al. Chlamydia trachomatis nucleic acid can be found in the
synovium of some asymptomatic subjects. Arthritis Rheum 1999;
42:1281–4.
31. Olmez N, Wang GF, Li Y, Zhang H, Schumacher HR. Chlamydial
nucleic acid in synovium in osteoarthritis: what are the implications. J Rheumatol 2001;28:1874–80.
32. Hylkema MN, Hoekstra MO, Luinge M, Timens W. The strength
of the OVA-induced airway inflammation in rats is strain dependent. Clin Exp Immunol 2002;29:390–6.
33. Kotake S, Schumacher HR, Arayssi TK, Gerard HC, Branigan PJ,
Hudson AP, et al. Gamma interferon and interleukin-10 gene
expression in synovial tissues from patients with early stages of
Chlamydia-associated arthritis and undifferentiated oligoarthritis
and from healthy volunteers. Infect Immun 1999;67:2682–6.
34. Muller B, Gimsa U, Mitchison NA, Radbruch A, Sieper J, Yin Z.
Modulating the Th1/Th2 balance in inflammatory arthritis.
Springer Semin Immunopathol 1998;20:181–96.
35. Yin Z, Braun J, Neure L, Wu P, Liu L, Eggens U, et al. Crucial
role of interleukin-10/interleukin-12 balance in the regulation of
the type 2 T helper cytokine response in reactive arthritis. Arthritis
Rheum 1997;40:1788–97.
36. Rodel J, Groh A, Hartmann M, Schmidt KH, Lehmann M,
Lungerhausen W, et al. Expression of interferon regulatory factors
and indoleamine 2,3-dioxygenase in Chlamydia trachomatis-infected synovial fibroblasts. Med Microbiol Immunol (Berl) 1999;
187:205–12.
37. Braun J, Yin Z, Spiller I, Siegert S, Rudwaleit M, Liu L, et al. Low
secretion of tumor necrosis factor ␣, but no other Th1 or Th2
cytokines, by peripheral blood mononuclear cells correlates with
chronicity in reactive arthritis. Arthritis Rheum 1999;42:2039–44.
38. Lahesmaa R, Yssel H, Batsford S, Luukkainen R, Mottonen T,
Steinman L, et al. Yersinia enterocolitica activates a T helper type
1-like T cell subset in reactive arthritis. J Immunol 1992;148:
3079–85.
39. Schlaak J, Hermann E, Ringhoffer M, Probst P, Gallati H, Meyer
zum Buschenfelde KH, et al. Predominance of Th1-type T cells in
synovial fluid of patients with Yersinia-induced reactive arthritis.
Eur J Immunol 1992;22:2771–6.
40. Simon AK, Seipelt E, Wu P, Wenzel B, Braun J, Sieper J. Analysis
of cytokine profiles in synovial T cell clones from chlamydial
reactive arthritis patients: predominance of the Th1 subset. Clin
Exp Immunol 1993;94:122–6.
41. Bas S, Kvien TK, Buchs N, Fulpius T, Gabay C. Lower level of
synovial fluid interferon-␥ in HLA-B27-positive than in HLA-B27negative patients with Chlamydia trachomatis reactive arthritis.
J Rheumatol 2003;42:461–7.
42. Butrimiene I, Jarmalaite S, Ranceva J, Venalis A, Jasiuleviciute L,
Zvirbliene S. Different cytokine profiles in patients with chronic and
acute reactive arthritis. Rheumatology (Oxford) 2004;43:1300–4.
43. Rudwaleit M, Siegert S, Yin Z, Eick J, Thiel A, Radbruch A, et al.
Low T cell production of TNF␣ and IFN␥ in ankylosing spondylitis: its relation to HLA-B27 and influence of the TNF-308 gene
polymorphism. Ann Rheum Dis 2001;60:36–42.
44. Zhao YX, Zhang H, Chiu B, Payne U, Inman RD. Tumor necrosis
factor receptor p55 controls the severity of arthritis in experimental Yersinia enterocolitica infection. Arthritis Rheum 1999;42:
1662–72.
45. Prebeck S, Brade H, Kirschning CJ, da Costa CP, Durr S, Wagner
H, et al. The Gram-negative bacterium Chlamydia trachomatis L2
stimulates tumor necrosis factor secretion by innate immune cells
independently of its endotoxin. Microbes Infect 2003;5:463–70.
46. Netea MG, Kullberg BJ, Jacobs LE, Verver-Jansen TJ, van der
HOST–MICROBE INTERACTIONS IN CHLAMYDIA-INDUCED ARTHRITIS
Ven-Jongekrijg J, Galama JM, et al. Chlamydia pneumoniae
stimulates IFN␥ synthesis through MyD88-dependent, TLR2- and
TLR4-independent induction of IL-18 release. J Immunol 2004;
173:1477–82.
47. Rupp J, Goepel W, Kramme E, Jahn J, Solbach W, Maass M.
CD14 promoter polymorphism –159C⬎T is associated with susceptibility to chronic Chlamydia pneumoniae infection in peripheral blood monocytes. Genes Immun 2004;5:435–8.
48. Eng HL, Wang CH, Chen CH, Chou MH, Cheng CT, Lin TM. A
CD14 promoter polymorphism is associated with CD14 expression
and Chlamydia-stimulated TNF␣ production. Genes Immun 2004;
5:426–30.
49. Zhang X, Glogauer M, Zhu F, Kim TH, Chiu B, Inman RD.
Innate immunity and arthritis: neutrophil Rac and Toll-like recep-
507
tor 4 expression define outcomes in infection-triggered arthritis.
Arthritis Rheum 2005;52:1297–304.
50. Baeten D, Van Damme N, Van den Bosch F, Kruithof E, De Vos
M, Mielants H, et al. Impaired Th1 cytokine production in
spondyloarthropathy is restored by anti-TNF␣. Ann Rheum Dis
2001;60:750–5.
51. Zou J, Rudwaleit M, Brandt J, Thiel A, Braun J, Sieper J.
Down-regulation of the nonspecific and antigen-specific T cell
cytokine response in ankylosing spondylitis during treatment with
infliximab. Arthritis Rheum 2003;48:780–90.
52. Baeten D, Vandooren B, De Rycke L, Veys EM, De Keyser F.
Effect of infliximab treatment on T cells cytokine responses in
spondylarthropathy: comment on the article by Zou et al [letter].
Arthritis Rheum 2004;50:1015–6.
DOI 10.1002/art.21515
Clinical Images: Dysphagia is sometimes easier to swallow than spinal surgery
The patient, a 71-year-old man, was referred with a long history of dysphagia. Lateral radiographs of the cervical spine had
unexpectedly revealed a pair of huge, beak-shaped osteophytes arising from the anterior vertebral bodies of C2 and C3 (left). The
patient described a 10-year-history of intermittent dysphagia to solids, exacerbated in certain neck positions. Videofluoroscopic
examination (right) revealed that the epiglottis (E) did not move past the osteophytes during swallowing. This interrupted the flow
of the bolus of barium (B) into the esophagus (O) and resulted in silent aspiration (A) of barium into the trachea. The patient
declined neurosurgical referral, explaining that he was usually able to swallow normally. While anterior osteophytes cause dysphagia
due to compression of the esophagus in rare instances (Benhabyles M, Brattstrom H, Sunden G. Dysphagia and dyspnea as
complications in spondylarthritis anklyopoetica with cervical osteophytes. Acta Orthop Scand 1970;41:396–401), this symptom is
extremely common after surgery to the anterior cervical spine (Smith-Hammond CA, New KC, Pietrobon R, Curtis DJ, Scharver
CH, Turner DA. Prospective analysis of incidence and risk factors of dysphagia in spine surgery patients: comparison of anterior
cervical, posterior cervical, and lumbar procedures. Spine 2004;29:1441–6).
R. Rajendram, MRCP, MBBS, BSc, AKC
J. Ehtisham, MRCP, MBBChir, PhD
R. W. Smith, FRCP
Milton Keynes General Hospital
Milton Keynes, UK