Jpn. J. Infect. Dis., 60, 220-224, 2007
Short Communication
Study of Predominant Bacterial Antigens Triggering Antibody
Response in Salmonella Reactive Arthritis: Apropos of a Case
María Gabriela Lacoste, Diego Esteban Cargnelutti, Héctor Tamashiro1 and María Silvia Di Genaro*
Laboratory of Immunology, Chemistry, Biochemistry and Pharmacy Faculty,
National University of San Luis, and 1Public Hospital of San Luis, San Luis, Argentina
(Received July 27, 2006. Accepted April 6, 2007)
SUMMARY: Reactive arthritis (ReA) is a sterile arthritis triggered by distal mucosal infection, which suggests
a contribution from bacterial products. The pathogenesis of ReA is unclear. There are no international standards
for the serological methods used to confirm ReA. In the present work, we analyzed the predominant bacterial
component that triggered an immune response in a 24-year-old woman with acute ReA. The candidate bacterial
trigger was investigated by measuring the antibacterial antibodies (all immunoglobulin classes and IgA) to
Salmonella enteritidis, Shigella flexneri and Yersinia enterocolitica. ELISA for Salmonella gave a positive
result. To identify the bacterial component triggering ReA, antibodies to crude lysate, outer membrane proteins
(OMP), cytosolic fraction, supernatant proteins and lipopolysaccharide of S. enteritidis were analyzed in sera
and synovial fluid (SF) by ELISA, dot blot and Western blot. Among the antigen preparations, the antibody
response to OMP was dominant in both serum and SF; a strong reaction to seven OMP bands (50 - 21 kDa) was
observed. We concluded that OMP were the main bacterial antigens that trigged ReA in the reported case.
Determining the triggering bacterial components in each case can help elucidate the precise causes of ReA and
will contribute to the designing of a specific serological diagnostic method for this arthritis.
pose of the present work was to report the case of a 24-yearold woman with ReA and to identify the bacterial components
triggering ReA by assessing the antibody response to different S. enteritidis antigens.
A 24-year-old woman was admitted to the Emergency
Department of the Hospital of San Luis in Argentina with
complaints of diarrhea, fever, nausea, vomiting and abdominal
cramping. She reported that the onset of the gastrointestinal
symptoms occurred 3 days after she visited a free-fork restaurant, where she ate various foods, some of them with mayonnaise dressing. Physical examination revealed that the patient
had abdominal pain and a temperature of 38.2°C, blood
pressure of 110/70 mm Hg, and a pulse of 104/min. Results
of the laboratory investigation showed increased erythrocyte
sedimentation rate (ESR) at 47 mm/h (Table 1). After this
initial diagnostic approach, empirical broad-spectrum antibiotic therapy with ciprofloxacin was started and continued
for 7 days. Four weeks later, she was admitted to the Rheumatology Section of the Hospital, and reported a history of
progressive pain and swelling in the left knee and ankle,
associated with morning stiffness. Physical examination
revealed that the patient was unable to put weight on her left
leg, and revealed florid synovitis of the left knee. She had
tenderness at the insertion of the left Achilles tendon, but not
in the lower back or buttocks. There was no recent history of
conjunctivitis. The patient reported fatigue and vague abdominal pain for 2 weeks prior to her admission. The results of the
laboratory tests are shown in Table 1. Synovial fluid (SF)
was aspirated from the left knee and revealed severe inflammation, but there was no evidence of infection, as a culture
of SF had a negative result, and no bacteria were observed
by means of Gram’s staining (Table 1). Radiographs of the
knee joint showed only soft tissue swelling, and neither
enthesopathic lesions nor erosive joint damage were visualized. Considering the clinical picture and the preceding symptomatic enteritis, a presumptive diagnosis of enteric ReA was
Reactive arthritis (ReA) is a sterile synovitis that develops
after gastrointestinal or urogenital infections, often with some
latency, at joints distant from the site of the primary infection. This genesis suggests that there is a contribution from
bacterial products in the induction of ReA (1). Because the
pathogenesis of ReA is incompletely understood, appropriate treatments are not available. Infections by Salmonella,
Shigella, Yersinia and Campylobacter spp. are implicated
in triggering enteric ReA (2). At the time of arthritis, stool
cultures are usually negative, and the background of ReA
has usually been confirmed by serological methods (2).
However, there are no international standards for the tests,
and the techniques vary greatly (3). Secreted proteins called
Yersina outer proteins (Yops) are unique antigens of pathogenic yersiniae, and the detection of specific IgG and IgA
by Western blot has proved to be the best diagnostic tool
for Yersinia enterocolitica-associated ReA (4). In contrast,
although antibody responses are important for achieving
protection against Salmonella infection (5), there are no
studies that show the relevant antibody response in human
Salmonella ReA. Determining this reponse could help to
define a specific diagnostic test for this arthritis. The incidence of ReA following Salmonella infection strongly fluctuates according to various factors, such as the rate of this
infection in a given area, from 1.2 to 14% (3). Salmonella
enteritidis is a major cause of human food-borne illness. It is
the most frequently detected cause of outbreaks of human
salmonellosis (6), and a high frequency of ReA has been
observed after S. enteritidis outbreaks (3,7). To date, there
have been no reported studies of ReA in Argentina. The pur*Corresponding author: Mailing address: Immunology, Laboratory of Microbiology, Chemistry, Biochemistry and Pharmacy
Faculty, National University of San Luis, Chacabuco y Pedernera,
5700, San Luis, Argentina. Tel: +54-2652-423789, Fax: +54-2652431301, E-mail: [email protected]
220
Table 1. Clinical course and laboratory tests
Time (month)
Clinical symptoms and
physical examination
Laboratory tests
ESR (mm/h)
Hemoglobin (g/dl)
WBC (cell /mm3)
PMN leukocytes (%)
MN leukocytes (%)
Platelets/mm3
CRP (mg/dl)
C3 (mg/dl)
C4 (mg/dl)
Stool culture
Blood culture
Antinuclear antibodies
Chlamydia trachomatis IgG
HLA-B27 by PCR in blood
Serum antibacterial IgA to
-Salmonella antigens
-Yersinia antigens
-Shigella antigens
Blood MN proliferation with:
-Salmonella antigens
-Yersinia antigens
-Shigella antigens
SF analysis
Cell number (cell/mm3)
PMN leukocytes (%)
MN leukocytes (%)
Protein (g/dl)
Glucose (mg/dl)
Culture
Gram staining
Crystals
C. trachomatis DNA by PCR
SF antibacterial IgA to:
-Salmonella antigens
-Yersinia antigens
-Shigella antigens
0
1
4
10 - 14
Gastrointestinal symptoms:
diarrhea, fever, nausea,
vomiting, abdominal cramping.
Abdominal pain
Temperature 38.2°C
Basal pressure 110/70 mm Hg
Pulse 104/min
Arthritis onset:
morning stiffness, pain and
swelling in the left knee and
ankle. Fatigue and vague
abdominal pain. Florid
synovitis of left knee, unable
to put weight in the left leg.
Tenderness at the insertion of
the left Achilles tendon.
Radiographic study of knee
showed soft tissue swelling.
After treatment:
less pain and swelling
in the left knee
Asymptomatic period:
without articular
inflammation
47
12.8
6,020
68
32
321,000
ND
ND
ND
ND
ND
ND
ND
ND
59
14
11,180
63
37
316,000
26
245
44.9
Negative
Negative
Negative
Positive (1/16)
ND
13
ND
ND
ND
ND
ND
ND
194
35.1
ND
ND
ND
Positive (1/16)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Positive, OMP titer: 1/200
Negative
Negative
ND
Negative
Negative
Negative
Negative
Negative
Negative
ND
ND
Positive
Negative
Negative
ND
ND
ND
590
64
36
5.37
72
Negative
No bacteria
Absent
Negative
Positive, OMP titer: 1/400
Negative
Negative
ND, not done; WBC, white blood cell count; PMN, polymorphonuclear; MN, mononuclear; ESR, erythrocyte sedimentation rate; CRP, C-reactive
protein; C3 and C4, complement 3 and 4 components; SF, synovial fluid; OMP, outer membrane proteins.
proposed. The patient was treated initially with nonsteroidal
anti-inflammatory drugs (NSAIDs): diclofenac 50 mg every
8 h for 2 weeks, and then 50 mg twice daily for another 2
weeks. She initially responded well to the treatment. However,
articular inflammation returned 1 month later, and she was
then treated with sulfasalazine 2 g daily for 4 months. During
this time she gradually recovered, and the ESR returned
to normal. Follow-up after 6 months showed that she was
asymptomatic. The clinical course of the patient including
the results of the laboratory tests is shown in Table 1. The
puncturing of a vein or joint was done only when necessary
for diagnostic or therapeutic reasons. Table 1 shows the time
when clinical specimens were taken. Blood and SF were taken
on arthritis onset and a new sample of blood was taken after
3 months of evolution.
With the written consent of the patient (Declaration of
Helsinki, 2000), her serum and SF samples were remitted to
the Microbiology Laboratory of the National University of
San Luis, where the candidate bacterial trigger was investigated. Bacterial antigens were prepared from S. enteritidis,
Shigella flexneri (both isolated from clinical samples) and Y.
enterocolitica O:8 WAP (kindly provided by Dr. Kapperud,
Department of Bacteriology, National Institute of Public
Health, Oslo, Norway). Lipopolysaccharide (LPS) was obtained
by extraction with hot phenol-water as previously described
(8). Crude lysate (CL) was prepared from whole bacteria
disrupted by sonication from washed bacterium pellets (9).
Cytoplasmic fraction (CF) and outer membrane proteins
221
(OMP) were isolated from the CL (9). Culture supernatant
proteins (SN) were obtained by precipitation with ammonium
sulphate (9). The protein concentrations of the antigens were
determined by Lowry’s procedure. LPS had less than 2%
protein, and 3 - 4% 2-keto-3-deoxyoctulosonic acid (8). Total
immunoglobulins and IgA responses were studied by enzymelinked immunosorbent assay (ELISA) (10). Single 1:50 or
serial 1:50 - 12,800 dilutions of sera and SF were tested in a
plate coated with 10 μg/ml of each antigen in 0.15 M phosphate buffered saline, pH 7.2. Bound antibodies were demonstrated by reaction with horseradish peroxidase-conjugated
goat anti-human IgM, IgG and IgA (Sigma, St. Louis, Mo.,
USA) or goat anti-human IgA and peroxidase-conjugated
rabbit anti-goat IgG (Sigma). The antibody response was
considered to be positive if the serum or SF had an optical
density (OD) exceeding the mean + 2 SD of the healthy control group values. Control sera were obtained from healthy
subjects with no symptoms of infection, who required routine
laboratory tests. For SF analysis, since osteoarthritis (OA) is
not associated with infectious disease, we used SF from OA
patients as a control for the SF cut-off value calculation. To
compare the reactivity to the bacterial antigens, the antibody
titer to each antigen was determined as the last dilution with
an OD over the cut-off value.
Yops were prepared from culture supernatants as described
previously (11), and released proteins were precipitated with
trichloroacetic acid (11). The Yops were dissolved in SDS
sample buffer supplemented with 5 mM PMSF, electrophoresed in 12% polyacrylamide gels-SDS and transferred to
nitrocellulose membranes by electro-blotting. Western blot for
Yops was performed in sera and SF. In addition, since ReA
could also occur subsequent to an asymptomatic genitourinary tract infection, especially an infection with Chlamydia
trachomatis, we also investigated serum IgG specific for
this bacteria using a microimmuno-fluorescence test (MIF)
(Focus Diagnostics, Cypress, Calif., USA) (12) and searched
for C. trachomatis DNA in SF using in-house polymerase
chain reaction (PCR) (13).
Dot blot was performed by seeding Salmonella antigens
on nitrocellulose strips, incubating them with the serum or
SF (dilution 1:50) and then with horseradish peroxidaseconjugated goat anti-human IgM, IgG and IgA (Sigma). The
reaction was developed using 4-chloro-1-naphtol (BioRad,
Hercules, Md., USA) as chromogen. The membrane strip with
the seeded antigens was stained with Ponceau S for protein
detection. Moreover, Western blot of IgM, IgG and IgA for
all Salmonella antigens was carried out (6).
Four months after the gastrointestinal symptoms, blood
was obtained and proliferation assays of peripheral blood
mononuclear cells (PBMC) were performed. Cells were
cultured at 5 × 105 viable cells per well in flat-bottom 96well tissue plates, and duplicate wells were stimulated with
Salmonella, Yersinia and Shigella CL (2.5 μg/ml). After
incubation at 37°C in 5% CO2 for 3 days, cell proliferation
was measured by MTT reduction with the colorimetric assay
described by Mosmann (14). Results were expressed in terms
of the stimulation index (SI), defined as the ratio of proliferation (OD) induced by the antigen to that in medium alone,
and an SI ≥ 2 was considered to be a positive response.
Because HLA-B27 positivity is linked to ReA, the presence
of HLA-B27 antigen was determined using PCR (12). Briefly,
blood-extracted DNA was amplified with the primers B27/
E136 as 5´-CGGCGGTCCAGGAGCT-3´ and B37/E91 s+ 5´GGGTCTCACACCCTCCAGAAT-3´. HLA-B27 product
(136 bp) was detected on agarose electrophoresis and visualized with ethidium bromide.
ELISA with Salmonella antigens had a positive result in
SF upon arthritis onset (Table 1, Fig. 1A). After this result, as
in other previous studies (10), Salmonella was considered to
be the triggering bacterium that was the probable cause of the
ReA. In addition, after 3 months of arthritis evolution, PBMC
showed a positive proliferative response with Salmonella
CL (Fig. 1A), which indicated again that this bacterium
was a possible trigger of the ReA. In addition, Yops Western
blot had a negative result, which excluded Yersinia as the
cause of ReA. In addition, C. trachomatis IgG titer was 1/16.
However, this bacterium was also excluded as the trigger of
the ReA since this titer was too low to be associated with the
arthritis. Thus, another study (12) concluded that titers of ≥512
in the IgG fraction indicate ongoing infection, and IgG titers
of 1/32 - 1/256 indicate previous infection. The unchanging
1/16 titer of our patient in both serum samples (at arthritis
onset and in month 4) could indicate anti-Chlamydia antibody persistence. A fourfold increase in paired sera, or a single
endpoint of ≥1/512 is considered to indicate a possible C.
trachomatis acute infection in the MIF assay that was used;
however, this assay has not been established for diagnosing
chronic infections. Moreover, cross-reactivity may also occur
due to exposure to more than one Chlamydia spp. as described
in the performance characteristics of the Chlamydia MIF IgG
assay that was used. Furthermore, since several reports have
demonstrated that C. trachomatis can be detected in the SF
obtained from ReA patients (15), and the PCR test may be
superior to MIF with respect to its sensitivity and specificity for the diagnosis of C. trachomatis, we investigated
Chlamydia DNA in the SF of our patient by PCR. This PCR
had a negative result. In addition, our patient had no pertinent
sexual history; however, gastrointestinal symptoms preceded
the arthritic symptoms by 4 weeks. Moreover, the strong
immune response to Salmonella detected in serum and SF
upon the arthritis onset became negative when the arthritic
symptoms reduced. Based on the history of enteric symptoms
and the results of laboratory tests, we considered that the ReA
of our patient was associated with a Salmonella infection.
When all class antibodies to different Salmonella antigens
were studied by ELISA, we detected positive responses to
CL, OMP and LPS in serum and to all Salmonella antigen
preparations in SF with titers of 1/100 - 1/200 (Fig. 1B). This
result could suggest different kinetics between the systemic
and local clearance of the arthritogenic bacteria, and differences between the systemic and local immune responses.
The highest antibody responses were responses against OMP
in serum and responses to both OMP and LPS in SF (Fig.
1B). In the joint, OMP and LPS could trigger the synthesis of
pro-inflammatory cytokines, thus contributing to the recruitment of leukocytes into the synovium and articular inflammation. In addition, a higher Salmonella-specific IgA response
was detected in SF with titers of 1/100 - 1/400 (Fig. 1B), which
could indicate a pathogenic link between gut inflammation
and ReA. By ELISA, we observed only a slight IgA antibody
response to CF in SF, indicating that Salmonella cytosolic
antigens could not play a central role as an ReA trigger.
In contrast, we found that the IgA antibody response was
stronger to OMP, indicating that they were the major antigenic target in the reported case of Salmonella-induced ReA.
The serum OMP-specific IgA response decreased with the
reduction of the arthritis symptoms (from 1/200 at onset to
negative in month 4) indicating that this response may reflect
222
Fig. 1. Immune response to Salmonella antigens. (A) ELISA of antibacterial antibodies in synovial fluid (SF) on arthritis onset. The
response is represented as the optical density (OD) obtained for each bacterial antigen. Dotted line represents the cut-off value
(OD mean + 2 SD of the control group). CL, crude lysate; OMP, outer membrane proteins; CF, cytosolic fraction; LPS,
lipopolysaccharide; SN, supernatant proteins. Proliferative response of peripheral blood mononuclear cells to CL of Salmonella,
Yersinia and Shigella after 3 months of arthritis onset. SI, stimulation index. (B) ELISA and Dot blot of antibody response to
Salmonella antigens. (C) Western blot of IgM, IgG and IgA to Salmonella antigens. S, serum; SF, synovial fluid. (D) HLA-B27
PCR. MW, molecular weight marker; C(–): negative control without DNA; P, patient DNA; C(+), positive control DNA.
the disease activity (Table 1). Using dot blot, we analyzed the
antigen preparations without dilution to avoid any contamination. Using Ponceau S-staining, we observed a higher protein
concentration in CF. This result accords with the protein
determination of each preparation by Lowry’s method (Fig.
1B). Among the antigen preparations, the strongest reaction
in serum was again against OMP (Fig. 1B). Similar results
were obtained in SF (data not shown). In addition, a reaction
to Salmonella antigens (stronger for OMP) was also detected
in the serum 3 months after the initial ReA symptoms (Fig.
1B). Because of its high sensitivity, Western blot showed IgM,
IgG and IgA reactions with protein bands of all Salmonella
antigens (Fig. 1C). However, these reactions were stronger
to OMP. Reactions against flagellin (50 kDa), porins (38 - 36
kDa), OmpA (34 kDa) and unidentified OMP (32, 22, 21, 16
kDa) were observed and defined by their molecular weights
as previously reported (6). Chronic Salmonella ReA (prolonged arthritis of >1 year extension) has been reported in
only 20% of cases (3). In the present work, the patient showed
higher anti-OMP antibody response, and she did not develop
chronic ReA. In addition, HLA-B27 positiveness is associated with chronic and relapsing arthritis. In our patient, HLAB27 was negative (Fig. 1D), indicating the absence of a prognostic marker for chronicity (2). The frequency of HLA-B27
is not high (less than 50%) in Salmonella-induced ReA (3).
On the other hand, porins play a role in Salmonella pathogenesis (6), and anti-porin antibodies, in contrast with antiLPS anti-bodies, have bactericidal activity (16). Therefore, fur223
antibody memory response. Immunology, 117, 59-70.
6. Ochoa-Repáraz, J., Sesma, B., Alvarez, M., et al. (2004): Humoral
immune response in hens naturally infected with Salmonella enteritidis
against outer membrane proteins and other surface structural antigens.
Vet. Res., 35, 291-298.
7. Locht, H., Molbak, K. and Krogfelt, K. (2002): High frequency of
reactive joint symptoms after an outbreak of Salmonella enteritidis. J.
Rheumatol., 29, 767-771.
8. Di Genaro, M., Muñoz, E., Aguilera, C., et al. (2000): Yersinia
enterocolitica O:8 and O:5 lipopolysaccharide arthritogenicity in hamsters. Rheumatology, 39, 73-78.
9. Michiels, T., Wattieau, P., Brasseur, R., et al. (1990): Secretion of Yop
proteins by Yersiniae. Infect. Immun., 58, 2840-2849.
10. Fendler, C., Leiko, S., Soerensen, H., et al. (2001): Frequency of triggering bacteria in patients with reactive arthritis and undifferentiated
oligoarthritis and the relative importance of the tests used for diagnosis.
Ann. Rheum. Dis., 60, 337-346.
11. Trček, J., Wilharm, G., Jacovi, C., et al. (2002): Yersinia enterocolitica
Yop Q: strain-dependent cytosolic accumulation and post-translational
secretion. Microbiology, 148, 1457-1465.
12. Huhtinen, M., Laasila, K., Granfors, K., et al. (2002): Infectious background of patients with a history of acute anterior uveitis. Ann. Rheum.
Dis., 61, 1012-1016.
13. Golijow, C.D., Abba, M.C., Mourón, S.A., et al. (2005): Chlamydia
trachomatis and human papillomavirus infections in cervical disease in
Argentina women. Gynecol. Oncol. 96:181-186.
14. Mosmann, T. (1983): Rapid colorimetric assay for cellular growth and
survival: application to proliferation and cytotoxicity assays. J. Immunol.
Methods, 65, 55-63.
15. Taylor-Robinson, D., Gilroy, C.B., Thomas, B.J., et al. (1992). Detection of Chlamydia trachomatis DNA on joints of reactive arthritis
patients by polymerase chain reaction. Lancet, 340, 81-82.
16. Rodriguez-Morales, O., Fernández-Mora, M., Hernández-Lucas, I., et
al. (2006): Salmonella enterica serovar Typhimurium ompS1 and ompS2
mutans are attenuated for virulence in mice. Infect. Immun., 74, 13981402.
ther studies could elucidate whether anti-OMP antibody response could play a role in the protection against chronic ReA.
Our data stress that OMP are the immunodominant antigens
to antibody response in the reported Salmonella ReA case.
The results of our work could contribute to elucidating the precise causes of ReA and to the designing of a specific serological diagnostic method for this arthritis.
ACKNOWLEDGMENTS
This study was supported by Consejo Nacional de Investigaciones
Científicas y Técnicas (CONICET, PIP 6224) and Universidad Nacional de
San Luis (Proj. 0401).
Silvia Di Genaro is member career and Gabriela Lacoste fellow of
CONICET.
The technical assistance of Lic María Amelia Rodriguez in Chlamydia
PCR is gratefully acknowledged.
REFERENCES
1. Granfors, K., Jalkanen, S., Lindberg, A., et al. (1990): Salmonella
lipopolysaccharide in synovial cells from patients with reactive arthritis.
Lancet, 335, 685-688.
2. Leirisalo-Repo, M. (2005): Reactive arthritis. Scand. J. Rheumatol., 34,
251-259.
3. Colmegna, I., Cuchacovich, R. and Espinoza, L. (2004): HLA-B27associated reactive arthritis: pathogenetic and clinical consideration.
Clin. Microbiol. Rev., 17, 348-369.
4. Heesemann, J., Egger, C. and Shroeder, J. (1987): Serological diagnosis
of yersiniosis by immunoblot technique using virulence-associated
antigens of enteropathogenic yersiniae. Contrib. Microbiol. Immunol.,
9, 285-289.
5. Secundino, I., López-Macías, C., Cervantes-Barragán, L., et al. (2005):
Salmonella porins induce a sustained, lifelong specific bactericidal
224