Rheumatology 1999;38:1058–1067
The elevated ratio of interferon gamma-/
interleukin-4-positive T cells found in synovial
fluid and synovial membrane of rheumatoid
arthritis patients can be changed by
interleukin-4 but not by interleukin-10 or
transforming growth factor beta
Z. Yin1,2, S. Siegert1, L. Neure1,2, M. Grolms1, L. Liu2, U. Eggens1,
A. Radbruch2, J. Braun1 and J. Sieper1,2
1Department of Medicine, Division of Nephrology and Rheumatology, Klinikum
Benjamin Franklin, Free University, Berlin and 2Deutsches Rheuma
Forschungszentrum Berlin, Berlin, Germany
Abstract
Objectives. To quantify the T-helper type ( Th) 1 cytokine interferon gamma (IFN-c)positive and the Th2 cytokine interleukin (IL)-4-positive cells in synovial fluid (SF ) and
synovial membrane (SM ) at the single-cell level in rheumatoid arthritis (RA) in comparison
to reactive arthritis (ReA), and to manipulate the cytokine pattern of RA patients in
vitro.
Methods. Eighteen patients with RA and 17 with ReA were studied. For intracellular
staining of cytokines, SF mononuclear cells (MNC ) from seven patients with RA, in
comparison to eight patients with ReA, were triple stained with anti-IFN-c, IL-4 and antiCD4 or anti-CD8 monoclonal antibodies (mAb) and analysed by flow cytometry.
Furthermore, in 13 patients with RA, immunohistology of SM was performed and compared
with seven ReA patients. In addition, in six of the RA patients, synovial T cells were grown
over 3 weeks in the presence of various cytokines and intracellular cytokine staining analysed
by flow cytometry weekly.
Results. In SF, the mean percentage of IFN-c+/CD4+ T cells in RA was almost 4-fold
higher than the number of IL-4+/CD4+ T cells (11.3 ± 5 vs 3.02±1.04; P = 0.0012), while
the ratio of IFN-c/IL-4+ CD4+ T cells was only 1.59 in ReA (P = 0.047 for the ratio
difference). A similar result was obtained for SM: the ratio of IFN-c/IL-4+ cells in RA was
4.3 (P < 0.0001 for the IFN-c/IL-4 difference), but only 1.2 for ReA (P = 0.02 for the ratio
difference). Of the CD3+ cells in SM, 2.8% were positive for IFN-c and 0.4% for IL-4 in
three RA patients. A decrease in the number of IFN-c-positive SF T cells and an increase in
the number of IL-4-positive SF T cells could be achieved in vitro through IL-4, but not by
IL-10 or transforming growth factor beta.
Conclusions. The Th1 pattern in the joint of RA patients demonstrated at the single-cell
level may be important for the pathogenesis of RA and may provide a target for future
immunotherapy. Our data suggest a therapeutic role for IL-4.
K : Th1 cytokines, Rheumatoid arthritis, Intracellular cytokines, Cytokine
manipulation.
Rheumatoid arthritis (RA) is a candidate autoimmune
disorder characterized by chronic synovitis of multiple
joints normally leading to destruction of joint cartilage
Submitted 30 December 1998; revised version accepted 30 April
1999.
Correspondence to: J. Sieper, Department of Medicine, Rheumatology, Klinikum Benjamin Franklin, Hindenburgdamm 30, 12200
Berlin, Germany.
and erosion of bone. The mechanisms underlying the
initiation and perpetuation of chronic inflammation and
tissue damage are poorly understood [1]. Undoubtedly,
macrophage-derived cytokines such as tumour necrosis
factor alpha ( TNF-a) and interleukin (IL)-1 play an
important role in this process [1]. In contrast, the role
of T cells and their cytokines in stimulation and control
of monocytes/macrophages is less clear [2].
The balance of T-helper type ( Th) 1 and Th2 subsets
1058
© 1999 British Society for Rheumatology
Th1 cytokine pattern in rheumatoid arthritis
is implicated in the regulation of many immune
responses [3]. The Th1-type cytokines interferon gamma
(IFN-c) and TNF-a are required for response against
intracellular infections, while Th2 cells (secreting IL-4,
IL-5 and IL-10) are responsible for protection against
extracellular ones. Th1 cytokines have also been implicated in the pathogenesis of autoimmune diseases and/or
their animal models, such as multiple sclerosis [4],
diabetes mellitus [5] and RA [6 ] in which a T-cell
response against an unknown self-antigen may play a
role. In contrast, the Th2-like cytokines IL-4 and IL-10
downregulate inflammation in these models [7].
Previously, we have demonstrated that in the synovium of RA patients IL-4-positive cells could be detected
in fewer patients than in reactive arthritis (ReA),
suggesting a Th1-like pattern in RA synovium and a
Th0-like pattern in ReA synovium [8]. Provided that
Th1 cells play an important role in the pathogenesis of
RA, a change of this pattern would be a therapeutic
option. However, although the Th1 pattern has been
successfully manipulated in animal models [9], no such
manipulations have been tried in RA patients.
In the present study, we have quantitatively analysed
the cytokine pattern at the single-cell level in synovial
fluid (SF ) T cells by intracellular cytokine staining and
flow cytometry, and in synovial membrane (SM ) by
immunohistology. We present further evidence that in
the affected tissue and fluid of RA patients a high ratio
of IFN-c/IL-4-positive cells is present. We went on to
investigate the possibility of manipulating the pattern in
vitro through treatment with cytokines.
Patients and methods
The characteristics of the 18 RA patients studied are
shown in Table 1. SF was investigated in patients 1–7,
1059
and SM in patients 1, 8–18. In patients 1–6, in vitro
manipulation of SF mononuclear cells (MNC ) was
performed, and in patients 9, 11 and 13 double staining
for CD3 and IFN-c or for CD3 and IL-4 was carried
out. RA was defined according to the 1987 revised
criteria of the American College of Rheumatology [10].
Fifteen patients were investigated with ReA ( Table 2)
following infection with Chlamydia trachomatis, Yersinia
enterocolitica and Salmonella enteritidis. SF was investigated in patients 1–8, and SM in patients 9–15. A
diagnosis of ReA was made if patients had a symptomatic preceding infection of the urogenital tract or the
gut, with aetiology confirmed by positive stool culture,
the presence of C. trachomatis in a urogenital swab, or
the detection of specific antibodies [11] at the beginning
of the disease. In all patients, proliferation of SF
lymphocytes was highest to the triggering bacterium, as
expected [12].
SF was aspirated from the knee and SM was obtained
by diagnostic arthroscopy from the knee. All patients
had active disease as judged by the presence of joint
effusion, and all RA patients, except patients 5 and 6,
were treated with second-line drugs. SM samples were
placed in ‘Tissue Tek’ medium (Miles, IN, USA) and
snap-frozen in liquid nitrogen.
Cell cultures
SF MNC were separated by Ficoll-paque (Pharmacia,
Uppsala, Sweden) density centrifugation as previously
described and subsequently cultured with medium RPMI
1640 (GIBCOBRL, Life Technologies, Paisley, UK )
supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin (Biochrom KG, Berlin, Germany), 2 m
-glutamine (Biochrom KG, Berlin, Germany) and 10%
heat-inactivated fetal calf serum ( FCS) (GIBCOBRL,
Life Technologies, Paisley, UK ) [12].
T 1. Characteristics of patients with rheumatoid arthritis
T 2. Characteristics of patients with reactive arthritisa
Patient number
1
2
3
4
5
6
7
1
8
9
10
11
12
13
14
15
16
17
18
Sex
Age (yr)
f
f
m
f
f
m
m
f
f
f
f
f
f
f
m
f
f
f
f
43
70
46
70
38
60
63
43
52
43
20
28
58
44
52
51
55
61
52
RF
Positive
Positive
Negative
Positive
Positive
Positive
Negative
Positive
Positive
Positive
Negative
Positive
Positive
Negative
Negative
Negative
Positive
Positive
Negative
Disease duration
6 yr
5 yr
40 months
3 yr
6 months
6 months
4 yr
6 yr
3 yr
30 months
1 year
7 months
4 yr
19 months
3 yr
3 months
3 yr
1 year
2 yr
RF, rheumatoid factor; m, male; f, female.
Synovial fluid was investigated in patients 1–7, and synovial membrane in patients 1, 8–18.
Patient
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Age
(yr)
Sex
Disease
duration
43
49
47
41
54
8
28
14
32
23
32
27
20
33
58
f
m
m
m
m
f
m
m
m
f
f
m
m
m
m
3 weeks
2 weeks
5 yr
3 yr
3 days
5 weeks
4 yr
8 weeks
6 yr
36 months
14 months
7 weeks
2 weeks
10 yr
4 days
Triggering
bacteriuma
Chlamydia
Yersinia
Chlamydia
Chlamydia
Chlamydia
Yersinia
Chlamydia
Yersinia
Chlamydia
Chlamydia
Yersinia
Salmonella
Yersinia
Yersinia
Salmonella
HLA B27
Positive
Positive
Positive
Negative
Positive
Positive
Positive
Positive
Negative
Negative
Negative
Positive
Positive
Positive
Negative
m, male; f, female.
Synovial fluid was investigated in patients 1–8, and synovial membrane in patients 9–15.
aA diagnosis of reactive arthritis and of the triggering bacterium
was made according to the criteria described in Patients and methods.
1060
Z. Yin et al.
Intracellular cytokine staining and analysis by flow
cytometry
The staining method for intracellular cytokines has been
developed with modification from a technique described
in the literature [13]. SF MNC (1 × 106/ml ) were
stimulated with plate-bound anti-CD3 ( UCTH,
Pharmingen, San Diego, CA, USA; 10 mg/ml ) plus
soluble anti-CD28 ( Immunotech, Hamburg, Germany;
1 mg/ml ). After 4 h of activation, 2.5 m monensin
(Sigma, St Louis, MO, USA) was added and the cells
were cultured for an additional 2 h. After these 6 h of
activation, cells were then washed in phosphate-buffered
saline (PBS), fixed for 20 min with PBS containing 4%
formaldehyde (Merck, Darmstadt, Germany), washed
twice with PBS, and permeabilized by incubation in PBS
containing 0.5% saponin (Sigma, St Louis, MO, USA),
1% inactivated FCS and 0.1% sodium azide (permeabilizing buffer) for 10 min. All incubation steps were carried
out at room temperature. For triple staining, cells were
first stained with cy-chrome-labelled anti-CD4 or antiCD8 monoclonal antibodies (mAb) (Pharmingen, San
Diego, CA, USA) in PBS containing 1% inactivated
FCS and 0.1% sodium azide (staining buffer) at room
temperature for 20 min. After washing twice with staining buffer, cells were then fixed and permeabilized as
described above. Permeabilized cells were incubated with
phycoerythrin (PE )-labelled anti-human IL-4 mAb and
fluorescein isothiocyanate ( FITC )-labelled anti-human
IFN-c mAb (Pharmingen) in permeabilizing buffer. PEor FITC-labelled isotype-matched mAb (Pharmingen)
were used as a negative control. After 20 min incubation
at room temperature, cells were washed twice with the
permeabilizing buffer and further washed with staining
buffer twice, resuspended in the staining buffer and
analysed using a FACScan flow cytometer (Becton
Dickinson, San Jose, CA, USA). After gating either on
all lymphocytes, or the CD4+ or the CD8+ populations, data were analysed using CELLQuest software
and displayed as dot plots of FITC (x-axis) and PE ( yaxis) fluorescence (four decade log scales). Quadrant
markers were positioned to include >99% of control Ig
staining cells in the lower left quadrant. The percentage
of cytokine-producing cells within the gated population
was measured in the upper left for IL-4 and in the lower
right quadrants for IFN-c.
Manipulation of synovial fluid T-cell cytokine pattern in
vitro
SF MNC from RA patients were stimulated with coated
anti-CD3 and soluble anti-CD28 as described above for
short-term stimulation, in the absence or presence of
the following recombinant human cytokines: IL-4,
IL-10, IL-4+ IL-10, IL-12, transforming growth factor
beta ( TGF-b). IL-2 (100 U/ml ) was added to the culture
at day 3. The stimulation was repeated weekly under
the same condition at least for 3 weeks consecutively.
To measure intracellular cytokine production, cells taken
from these cell lines were stimulated at the end of each
week with PMA (5 ng/ml ) and ionomycin (1 ng/ml ) for
4 h at 37°C, and monensin was added for the last 2 h
of culture. The cells were fixed and stained for CD4 and
CD8 surface markers and for intracellular cytokines
using FITC- or PE-conjugated mAb specific for IFN-c,
IL-4, IL-10 and TNF-a as described above. The concentration was 5 ng/ml for IL-12 and 50 ng/ml for the other
cytokines in all experiments.
Immunohistology: single staining for CD3, IFN-c and
IL-4
Immunohistology was performed on sections from SM
of 12 patients with RA and six patients with ReA
as described [14]. The following primary mAb were
used: murine anti-human CD3 ( UCHT1, IgG1, Dako,
Glostrup, Denmark), murine anti-human IFN-c (B-B1,
IgG1, Biosource-International, Camarillo, CA, USA)
and murine anti-human IL-4 (IgG1, Genzyme
Diagnostics, Cambridge, MA, USA). For the staining,
the primary antibody was detected with affinity-purified
goat anti-mouse IgG antibodies (Dako) followed by
APAAP complex (Dako), and visualized with the chromogen New Fuchsin (Merck, Darmstadt, Germany).
Immunohistology: double staining for CD3/IFN-c and
CD3/IL-4
Double staining for CD3/IFN-c and CD3/IL-4 was
performed in three RA patients (patients 9, 11 and 13)
using immunofluorescence. Fixed sections were incubated with mouse anti-human IFN-c (dilution of 1:50;
IgG1, Pharmingen) or with rat anti-human IL-4 (dilution 1:50; IgG1, Pharmingen), subsequently incubated
with digoxigenin-N-hydroxysuccinimidester-conjugated
sheep IgG [F(ab) fragment] against mouse Ig for
2
IFN-c detection or with a biotin-SP-conjugated F(ab)
2
fragment donkey anti-rat IgG (H + L) (Dianova,
Hamburg, Germany) for IL-4 detection. Sections were
then incubated with an anti-digoxigenin-fluorochromeantibody (Cy@3-conjugated IgG fraction of mouse antidigoxin) (Dianova) for IFN-c detection and with
Cy@2-conjugated streptavidin (Dianova) for IL-4. For
CD3/IFN-c double labelling, cells were then incubated
with rabbit anti-human CD3 (dilution 1:100; Dako,
Glostrup, Denmark) followed by incubation with a
biotin-SP-conjugated donkey anti-rabbit IgG (H + L)
(Dianova), followed by incubation with Cy2-conjugated
streptavidin (dilution 1:200) (Dianova). For CD3/IL-4
double labelling, sections were incubated with mouse
anti-human CD3 (dilution 1:100; Dako), followed by
incubation with sheep IgG anti-mouse Ig conjugated
to digoxigenin-N-hydroxysuccinimidester (Boehringer
Mannheim, Germany), followed by incubation with
Cy@3-conjugated IgG mouse anti-digoxin (Dianova,
Denmark). All incubations were performed for 60 min
at room temperature and for the fluochrome-conjugated
antibodies in the dark. Between and after incubations,
slides were washed twice with PBS for 5 min. All dilutions were made in 3% bovine serum albumin
(BSA)/PBS.
Finally, cell nuclei were counterstained with DAPI
(4∞,6-diamidino-2-phenylindole) (Sigma, St Louis, MO,
Th1 cytokine pattern in rheumatoid arthritis
USA) for 10 min at room temperature in the dark
and the sections embedded in UV-inert medium
(Fuoromount-G; Bioproducts, Boehringer Ingelheim,
Germany) for microscopy.
1061
(A)
Counting of positively stained cells in synovial
membrane
Cytokine-positive cells were counted in the lining and
the sublining layers of synovium as described [15]. All
sections were examined under ×400 magnification with
a 1 mm graticule. For each analysis, 1–3 tissue samples
from each patient were examined. The mean number of
positive cells in the lining and sublining layer per highpower field over the entire area of three sections in
each patient was recorded. The mean count per square
millimetre was calculated using a conversion factor
(× 0.0625−1), and the number of IFN-c- and IL-4positive cells for each patient and, in the case of double
labelling, the numbers of CD3+/cytokine+ doublepositive and CD3−/cytokine+ single-positive cells are
presented. CD3- and cytokine-positive cells were
counted first using different filters. Double-positive cells
were identified subsequently by using a ‘combined’ filter.
Statistics
Student’s t-test was used to compare the mean (± ..)
between related numbers. Differences between means
were considered significant if the two-tailed P value was
<0.05. All data were analysed by the program Instat.2.1
for Macintosh.
(B)
Results
No IFN-c or IL-4 secretion by unstimulated cells
In preliminary experiments, SF MNC without stimulation with anti-CD3 plus anti-CD28 were fixed and
stained as described. No significant number of cytokinepositive cells was observed (<0.1% compared to isotype
control; data not shown).
High ratio of IFN-c/IL-4-producing synovial fluid
CD4+ T cells in patients with RA compared to ReA
The clearest difference was found when only CD4+
T cells were analysed (Fig. 1). In RA, the mean percentage of IFN-c+ cells out of the CD4+ T-cell population
(11.3 ± 5) was almost 4-fold higher than for IL-4positive cells (3.02 ± 1.04; P=0.0012). In contrast, in
ReA, the frequency of IFN-c-and IL-4-positive CD4+
T cells was not different (5.35 ± 4.8 and 4.34 ± 1.72,
respectively; P = 0.79), resulting in a ratio of IFN-c/
IL-4+ T cells of 1.6. This ratio was 2.5-fold higher in
RA than in ReA (P = 0.047). Examples of the cytometry
results after short-term stimulation with anti-CD3/antiCD28 from two patients with RA (nos 2 and 5) and
two patients with ReA (nos 6 and 8) are shown in
Fig. 1A. The numbers of cytokine-positive cells for all
patients are shown in Fig. 1B. The cytokine staining
could be completely inhibited by pre-incubation of the
anti-cytokine antibody with saturated concentrations of
recombinant cytokines, indicating the specificity of the
F. 1. Two-colour immunofluorescence of gated CD4+
lymphocytes for IFN-c and IL-4 expression in rheumatoid
arthritis (RA) and reactive arthritis (ReA). Synovial fluid
mononuclear cells (MNC ) were stimulated with anti-CD3 and
anti-CD28 in the presence of monensin, and stained as
described in Patients and methods. CD4+ T were gated and
quadrant markers were positioned to include >99 of isotypematched control Ig staining cells in the lower left quadrant
(not shown). (A) Dot plots are shown from two patients with
RA (nos 2 and 5) and two patients with ReA (nos 6 and 8).
(B) Results from all seven patients with RA and all eight
patients with ReA are shown.
cytokine staining. The mean percentage of CD4+ T cells
in RA patients was 52 ± 9 and of CD8+ T cells 33 ± 9;
the corresponding data in ReA patients were 51 ± 17
and 35 ± 14.
Comparing cytokine patterns in these two arthritides
directly, the mean percentage of IFN-c+ cells was 2-fold
1062
Z. Yin et al.
higher in RA than in ReA (11.3 ± 5 vs 5.35 ± 4.8;
P < 0.05). In contrast, more IL-4-positive cells were
detected in ReA than in RA (4.34 ± 1.72 vs 3.02 ± 1.04;
P = 0.23), although this difference was not significant.
When CD8+ T cells were gated, five out of seven
patients with RA and five out of eight patients with
ReA showed a T cytotoxic 1- pattern. No significant
differences between the ratios of IFN-c-positive and
IL-4-positive cells were found in the two arthritides
(9.26 ± 4.7 for IFN-c vs 6.15 ± 3.48 for IL-4 in RA,
P = 0.18, and 3.46 ± 2.29 for IFN-c vs 2.22 ± 1.09 for
IL-4 in ReA, P = 0.64). For total lymphocytes, the
mean percentage of IFN-c-positive cells (10.23 ± 4.9)
was >2-fold higher than that of IL-4-positive cells in
RA (4.8 ± 0.9; P = 0.014 for the difference), whereas
the percentage of IFN-c- (4.3 ± 3.4) and IL-4-positive
cells (3.8 ± 0.9) was similar in ReA.
To quantify further the level of cytokine secretion per
cell, the mean fluorescence intensity (MFI ) of cytokinepositive cells was compared between ReA and RA
patients. The MFI of cytokine-positive cells was divided
by the MFI of cytokine-negative cells (the lower left of
the dot plot) in order to normalize the level of MFI for
cytokine secretion. No clear difference in the intensity
of fluorescence staining was found between these two
groups of patients (data not shown).
High ratio of IFN-c/IL-4-positive cells in synovial
membrane of patients with RA compared to ReA
The mean ratio of IFN-c/IL-4+ cells in RA was 4.3
(18.5 ± 6.1/mm2 for IFN-c and 4.2 ± 2.9/mm2 for IL-4;
P < 0.0001), but only 1.2 in ReA [12.3 ± 9.7/mm2
(× 3.5) for IFN-c and 9.9 ± 8.6/mm2 (× 3.5) for IL-4;
difference not significant] (P = 0.02 for the ratio difference). The number of CD3-positive cells in SM was
~3.5-fold higher in ReA (1607 ± 1049/mm2) than in
RA (458 ± 163/mm2). Therefore, to correct the number
of cytokine-positive cells for the same number of T cells,
the number of cytokine-positive cells was divided by 3.5
for ReA patients (in Fig. 2). The linear regression shown
for RA and ReA patients in Fig. 2 indicates that the
individual IFN-c/IL-4 ratio remains about the same for
each patient group, while the number of cytokinepositive cells differs.
In three RA patients, double labelling was performed
for CD3/IFN-c and CD3/IL-4 to determine how many
of the T cells are positive for IFN-c or IL-4, and how
much of IFN-c and IL-4 is produced by CD3+ T cells
and how much by CD3-negative cells. Out of the CD3+
cells, only 2.8 ± 6.5% were positive for IFN-c and
0.4 ± 1% for IL-4. The vast majority of IFN-c was
produced by T cells. Out of the IFN-c-positive cells,
99 ± 1% were CD3+, while only 1% were CD3 negative.
In contrast, only 59 ± 17% of the IL-4-positive cells
were CD3+. Immunohistological double staining for
one of these three RA patients (patient no. 9) is shown
in Fig. 3.
Only IL-4 but not IL-10 or TGF-b changed the Th1
pattern of RA synovial T cells
By stimulating synovial MNC with anti-CD3 plus antiCD28 in the presence of various cytokines, the following
F. 2. A higher number of IFN-c-positive than IL-4-positive
cells are found in the synovial membrane of rheumatoid
arthritis (RA) patients in comparison to reactive arthritis
(ReA). Synovial tissue from 12 patients with RA and seven
patients with ReA was stained and cytokine-positive cells were
counted as described in Patients and methods. The line shows
the linear regression for RA patients and the dotted line for
ReA patients. The number of cytokine-positive cells was
divided by 3.5 in ReA patients (for details, see Results).
results were obtained for cytokine secretion by T cells
determined by flow cytometry after in vitro stimulation
with PMA/ionomycin in RA patients 1–6. Only the
results after 1 and 3 weeks are shown because the results
after 2 weeks of stimulation do not give additional
information ( Fig. 4). In general, the effect on CD4- and
CD8-positive cells was similar.
In vitro stimulation in the presence of IL-4 (Fig. 4).
IL-4 application resulted in a significant reduction of
IFN-c secretion compared to stimulation in the presence
of no cytokines. IL-4 plus IL-10 led to a similar but not
higher reduction of IFN-c secretion. This reduction
already started in the first week. Furthermore, the
number of IL-4-positive T cells could be significantly
enhanced by stimulation in the presence of IL-4, and
again IL-4 plus IL-10 had a similar but no greater effect.
These effects were visible as early as at week 1 and were
similar for CD4- and CD8-positive T cells. IL-4 also led
to an increase of IL-4/IFN-c-double-positive cells,
although significance was only reached for CD8+ T cells
at week 3.
In vitro stimulation in the presence of IL-10. In general,
the number of IL-10-secreting T cells was low. IL-10
alone had no significant effect on the number of
IFN-c-, IL-4-or IL-10-positive T cells over 3 weeks
( Fig. 4).
In vitro stimulation in the presence of TGF-b. TGF-b
induced a small but partly significant reduction of
IL-4-producing cells (IL-4 single positive: P = 0.08,
P = 0.04, P = 0.08 at week 1, 2 and 3, respectively, for
CD4+ T cells, no significant reduction for CD8+
T cells; IL-4/IFN-c-double-positive cells: P = 0.005,
P = 0.04, P = 0.02 at week 1, 2 and 3, respectively, for
CD4+ T cells, no significant reduction for CD8+
Th1 cytokine pattern in rheumatoid arthritis
1063
F. 3. Double staining of synovial membrane sections (magnification ×400) for CD3+/IFN-c+ and CD3+/IL-4+ cells in
one patient with rheumatoid arthritis (patient no. 9). Colour changes with the intensity of antibody binding (green colour:
strong binding; blue colour with green membrane staining: less strong binding of anti-CD3 antibody). Arrows indicate cells
single positive for cytokine staining. Quantified numbers are given in Results.
T cells). However, the number of IFN-c-positive cells
remained unchanged.
In vitro stimulation in the presence of IL-12. IL-12 had
a small enhancing effect on IFN-c secretion, which was
not significant, while the number of IL-4-secreting cells
was significantly and most effectively reduced, an effect
which was stronger than that induced by TGF-b (data
not shown).
The number of TNF-a-positive T cells. The number of
TNF-a-positive T cells was not significantly influenced
by any of the cytokines tested (data not shown).
Discussion
Our results demonstrate a Th1-like pattern in the joint
of RA patients, particularly in comparison with ReA.
We could show (i) by flow cytometry analysis and
immunohistology that the number of IFN-c-positive
T cells in RA SF and SM is ~4-fold higher than those
of IL-4-positive ones, while, for comparison, the ratio
of IFN-c/IL-4-positive synovial T cells is only ~1.5 in
ReA, (ii) that in RA SM only a small number of T cells
produce cytokines and that IFN-c is produced nearly
exclusively by T cells while a substantial proportion of
IL-4 is also produced by non-T cells and (iii) that a
change of the Th1 pattern of RA SF T cells could be
achieved in vitro only by treatment with IL-4, but not
with IL-10 or TGF-b.
The role of T cells in the pathogenesis of RA is still
a matter of controversy, especially in long-standing
disease [2, 16, 17]. The abundance of macrophagederived cytokines such as TNF-a and IL-1 in SF and
SM [1], and the therapeutic benefit of antagonizing
these cytokines [18], point to an important role for
macrophages in the pathogenesis of RA. However, such
treatment has not yet led to long-lasting remission. The
evidence that a Th1-like response might be deleterious
and a Th2-like response beneficial in autoimmune disease came first from animal models [4–7]. Early studies
in man failed to detect T-cell cytokines in the rheumatoid
SM [16 ]. However, there are now a growing number of
reports on the presence of T-cell cytokines in the joint
of RA patients [8, 19–26 ]. Most of these studies report
a Th1-type pattern with more IFN-c than IL-4, using
the techniques of polymerase chain reaction (PCR),
semi-quantitative PCR, in situ hybridization, immunohistology, measurement of soluble or intracellular cytokines in the joint or peripheral blood [27]. Despite these
studies, there is still a controversy about the role of Th1
cytokines in RA [1, 2, 17, 23].
This study is the first to quantify the number of
IFN-c- and IL-4-positive cells at the single-cell level in
SF and SM, and to compare the number of cytokinepositive cells with those found in another form of
inflammatory arthritis. Most interestingly, a similar pattern was found in both SF and SM ( Figs 1 and 2): the
number of IFN-c-positive cells was about four times
higher than that of IL-4-positive cells in RA. In a recent
investigation analysing SM-derived CD4+ T cells by
flow cytometry, an even higher ratio was reported which
was clearly higher than in PB [26 ]. For comparison, in
ReA the number of IFN-c-positive cells was about the
same as that of IL-4-positive cells. The presence of both
T-cell cytokines in two different forms of arthritis
stresses the importance of always analysing a control
disease for comparison, to quantify the cytokine-positive
1064
Z. Yin et al.
F. 4. Percentage of IFN-c- and IL-4-single-positive and
double-positive CD4+ and CD8+ T cells, and percentage of
IL-10-positive CD4+ and CD8+ T cells after stimulation in
the presence of different cytokines or normal medium after
1 week or 3 weeks as described in Patients and methods. The
mean ± .. calculated from six patients is shown. *P < 0.05
compared to stimulation only in the presence of medium.
cells and to compare ratios instead of absolute numbers.
Indeed, the balance of these two T-cell cytokines seems
to be crucial for the outcome of an immune response
[3]. Previous studies trying to quantify T-cell cytokines
at the single-cell level in RA did this only in SF [25, 28]
or in SM [26, 29], and did not calculate a ratio of
IFN-c/IL-4-positive cells because IL-4 was not detected
or not looked for [25, 29].
Synovial fluid T cells were analysed by using the
technique of intracellular cytokine staining and the
number of cytokine-positive T cells was quantified by
flow cytometry [13, 30]. We and others [20, 25, 26, 30]
have shown that the frequency of cytokine-positive
T cells, especially IL-4-positive ones, is very low and
therefore normally not detectable by flow cytometry or
ELISA without in vitro stimulation. Therefore, we used
short-term mitogenic in vitro stimulation.
However, it was possible to quantify the number of
cytokine-positive cells in SM (without stimulation) probably because of a higher frequency compared to SF. By
double staining RA SM for CD3 and either IFN-c or
IL-4, we could show that only a minority of T cells
(<3%) secrete cytokines ( Fig. 3). A similar result has
been reported before for IFN-c, but not for IL-4 [29].
Cytokine-positive regulatory T cells in the range of
1–2% seem indeed to be sufficient to orchestrate a whole
immune response, as suggested by animal models [31].
We could also show for the first time in RA synovium
that the vast majority of IFN-c-positive cells were T cells
(99%), while as many as 49% of IL-4-positive cells were
non-T cells. The most likely non-T-cell source for
IFN-c would be NK cells and, for IL-4 mast cells,
basophils or NK1+ T cells.
The mean disease duration was clearly longer in
patients with RA than with ReA. However, two recent
studies demonstrated that the cell infiltrates in SM [32]
and the amount of IFN-c in RA SM [21] are similar in
early and late RA, indicating that ‘early’ RA may
already have a chronic phenotype. In contrast, a distinct
Th1/Th2 cytokine pattern was found in new-onset
synovitis vs chronic RA by other investigators [33].
Nonetheless, the Th1-type cytokine pattern in the joints
of patients with chronic RA demonstrated in our study
could be a target of immunointervention independently
of whether such a pattern is also present in early RA.
There was no difference in the cytokine pattern in
patients treated with second-line drugs compared to
those not treated, indicating that activity might be more
important than treatment.
ReA was chosen in this study for comparison because
of the Th0/Th2-like pattern present in the joint [8, 34],
which makes it a good control disease for RA. The
finding of a Th1 pattern in the joints of some patients
with ReA in this study suggests a more heterogeneous
spectrum which needs further clarification. As can be
seen from Table 2, there is no simple correlation of the
cytokine pattern with disease duration, HLA-B27 status
or the triggering bacterium. Using a nested PCR technique for the detection of cytokine mRNA, one recent
study could not find a difference between RA and ReA
membrane in IFN-c expression [23], while another
report found less IFN-c in ReA than in RA [24].
According to our results, the IFN-c/IL-4 ratio is more
relevant than the absolute numbers for IFN-c alone.
Investigating the cytokine pattern in SF and SM by
PCR, a recent study detected a higher IL-4/IFN-c ratio
in patients with pauciarticular juvenile rheumatoid arthritis and with juvenile spondyloarthropathy compared
to RA [35]. This would be in line with our results
Th1 cytokine pattern in rheumatoid arthritis
because of some similarity between these forms of
juvenile arthritis and ReA [36 ].
Therefore, based on the evidence that both IFN-cand IL-4-positive cells are present in the joint of RA
patients, but that the IFN-c-positive cells clearly outnumber the IL-4-positive ones, an obvious question is
whether such an established Th1 pattern could be
changed. Current approaches to the modulation of a
Th1 pattern are discussed elsewhere in more detail [37].
Among them, IL-4 and IL-10 are so far the best candidates as agents of intervention. IL-4 has been reported
to downregulate secretion of pro-inflammatory cytokines of cells derived from RA joints [38], to suppress
IFN-c production of T cells in vitro [39] and to improve
animal models of arthritis [7, 40]. In this study, we
show that IL-4 is able to change an established Th1
pattern in vitro of T cells derived from RA joints by
suppressing IFN-c-positive cells and by expanding
IL-4-positive ones, both IL-4 single-positive and IL-4/
IFN-c-double-positive cells. It is possible that these
effects are more pronounced if T cells (or patients) are
treated for longer than 3 weeks. From our experiments,
we cannot answer the question whether this was due to
a true reversion from Th1 to Th2 cells. However, because
there is good experimental evidence that a polarized
Th1 or Th2 pattern is irreversible [39], it is more likely
that IL-4 worked as an inhibitor of IFN-c secretion and
that the new IL-4-positive cells were derived from expansion of IL-4+ T cells and/or from uncommitted precursor cells of the Th0 type by differentiation and clonal
outgrowth [39, 40].
In contrast, the presence of IL-10 in our T-cell lines
over 3 weeks did not change the number of T
cells secreting IFN-c, IL-10 or IL-4 significantly.
Furthermore, IL-10 added to IL-4 proved no more
effective than IL-4 alone. Thus, IL-10 neither suppressed
IFN-c secretion nor did it induce an expansion of
IL-10+ T cells. These results are of special interest
because recent evidence suggests that IL-10 may be a
good candidate for the treatment of RA: IL-10 suppresses inflammatory cytokines of RA MNC in vitro
[19, 41] and is an effective treatment of arthritis in
animal models of the disease [7, 42]. Furthermore, a
new T-cell type has been described, characterized by
high IL-10 and low IL-4 secretion, which has the capacity to suppress ongoing Th1 responses [43]. However,
our data suggest that manipulation of T cells, potentially
capable of mediating long-lasting suppression [43],
seems to be more difficult after disease onset.
It has also been reported that TGF-b might be a
potent suppressor of Th1 responses [4], but the only
effect in our study was seen in a reduction of
IL-4-positive T cells. IL-12 did not expand the population of IFN-c-positive T cells significantly during the
3 weeks, probably reflecting an already maximally polarized Th1 pattern. Anti-IL-12 could also be a therapeutic
option for the downregulation of Th1 responses [44],
although it was not tested in this study.
In summary, by quantifying IFN-c- and IL-4-positive
single cells in SF and SM, we demonstrated a shift
1065
towards a high IFN-c/IL-4 ratio ( Th1 like) in the joints
of RA patients in comparison to patients with ReA.
Only IL-4 induced a suppression of Th1 cytokines
(IFN-c) and an expansion of Th2 cytokines (IL-4)
in vitro, while IL-10 was not effective. These results are
of special interest with regard to the first clinical studies
treating RA patients with IL-10 [45] or IL-4 (P. Miossec,
personal communication). Although IL-10 might be
effective if monocytes [41] and not T cells are the targets,
new approaches in the treatment of RA should rather
aim for immunomodulation of the T cell response [46 ].
Acknowledgements
We are grateful to N. A. Mitchison for discussion and
critical review of the manuscript. This study was supported by a grant from the Bundesministerium für
Bildung, Wissenschaft, Forschung und Technologie in
Germany.
References
1. Feldmann M, Brennan FM, Maini RN. Rheumatoid
arthritis. Cell 1996;85:307–10.
2. Fox DA. The role of T cells in the immunopathogenesis
of rheumatoid arthritis. Arthritis Rheum 1997;40:598–609.
3. Abbas AK, Murphy KM, Sher A. Functional diversity of
helper T lymphocytes. Nature 1996;383:787–93.
4. Khoury SJ, Hancock WW, Weiner HL. Oral tolerance to
myelin basic protein and natural recovery from experimental autoimmune encephalomyelitis are associated with
downregulation of inflammatory cytokines and differential
upregulation of transforming growth factor beta, interleukin 4, and prostaglandin E expression in the brain.
J Exp Med 1992;176:1355–64.
5. Rapoport MJ, Jaramillo A, Zipris D et al. Interleukin 4
reverses T cell proliferative unresponsiveness and prevents
the onset of diabetes in nonobese diabetic mice. J Exp
Med 1993;178:87–99.
6. Mauri C, Williams RO, Walmsley M, Feldmann M.
Relationship between Th1/Th2 cytokine patterns and the
arthritogenic response in collagen-induced arthritis. Eur
J Immunol 1996;26:1511–8.
7. Joosten LA, Lubberts E, Durez P et al. Role of
interleukin-4 and interleukin-10 in murine collageninduced arthritis. Protective effect of interleukin-4 and
interleukin-10 treatment on cartilage destruction. Arthritis
Rheum 1997;40:249–60.
8. Simon AK, Seipelt E, Sieper J. Divergent T-cell cytokine
patterns in inflammatory arthritis. Proc Natl Acad Sci
USA 1994;91:8562–6.
9. Hesse M, Bayrak S, Mitchison A. Protective major histocompatibility complex genes and the role of interleukin-4
in collagen-induced arthritis. Eur J Immunol 1996;
26:3234–48.
10. Arnett FC, Edworthy SM, Bloch DA et al. The American
Rheumatism Association 1987 revised criteria for the
classification of rheumatoid arthritis. Arthritis Rheum
1988;31:315–24.
11. Kingsley G, Sieper J. Third International Workshop on
Reactive Arthritis. Ann Rheum Dis 1996;55:564–84.
12. Sieper J, Braun J, Brandt J et al. Pathogenetic role of
Chlamydia, Yersinia and Borrelia in undifferentiated
oligoarthritis. J Rheumatol 1992;19:1236–42.
1066
Z. Yin et al.
13. Assenmacher M, Schmitz J, Radbruch A. Flow cytometric
determination of cytokines in activated murine T helper
lymphocytes: expression of interleukin-10 in interferongamma and in interleukin-4-expressing cells. Eur J
Immunol 1994;24:1097–101.
14. Braun J, Bollow M, Neure L et al. Use of immunohistologic and in situ hybridization techniques in the examination of sacroiliac joint biopsy specimens from patients
with ankylosing spondylitis. Arthritis Rheum 1995;
38:499–505.
15. Mulherin D, Fitzgerald O, Bresnihan B. Synovial tissue
macrophage populations and articular damage in rheumatoid arthritis. Arthritis Rheum 1996;39:115–24.
16. Firestein GS, Alvaro Gracia JM, Maki R, Alvaro Garcia
JM. Quantitative analysis of cytokine gene expression in
rheumatoid arthritis. J Immunol 1990;144:3347–53.
17. Miossec P, van den Berg W. Th1/Th2 cytokine balance in
arthritis. Arthritis Rheum 1997;40:2105–15.
18. Butler DM, Maini RN, Feldmann M, Brennan FM.
Modulation of proinflammatory cytokine release in
rheumatoid synovial membrane cell cultures. Comparison
of monoclonal anti-TNF-alpha antibody with the
interleukin-1 receptor antagonist. Eur Cytokine Netw
1995;6:225–30.
19. Cohen SB, Katsikis PD, Chu CQ et al. High level of
interleukin-10 production by the activated T cell population within the rheumatoid synovial membrane. Arthritis
Rheum 1995;38:946–52.
20. Dolhain RJ, van der Heiden AN, ter Haar NT, Breedveld
FC, Miltenburg AM. Shift toward T lymphocytes with a
T helper 1 cytokine-secretion profile in the joints of
patients with rheumatoid arthritis. Arthritis Rheum
1996;39:1961–9.
21. Ulfgren AK, Lindblad S, Klareskog L, Andersson J,
Andersson U. Detection of cytokine producing cells in the
synovial membrane from patients with rheumatoid arthritis. Ann Rheum Dis 1995;54:654–61.
22. Bucht A, Larsson P, Weisbrot L et al. Expression of
interferon-gamma (IFN-gamma), IL-10, IL-12 and transforming growth factor-beta (TGF-beta) mRNA in synovial fluid cells from patients in the early and late phases
of rheumatoid arthritis (RA). Clin Exp Immunol
1996;103:357–67.
23. Kotake S, Schumacher HR Jr, Yarboro CH et al. In vivo
gene expression of type 1 and type 2 cytokines in synovial tissues from patients in early stages of rheumatoid,
reactive, and undifferentiated arthritis. Proc Assoc Am
Physicians 1997;109:286–301.
24. Smeets TJ, Dolhain RJ, Breedveld FC, Tak PP. Analysis
of the cellular infiltrates and expression of cytokines in
synovial tissue from patients with rheumatoid arthritis
and reactive arthritis. J Pathol 1998;186:75–81.
25. Ronnelid J, Berg L, Rogberg S, Nilsson A, Albertsson K,
Klareskog L. Production of T-cell cytokines at the singlecell level in patients with inflammatory arthritides:
enhanced activity in synovial fluid compared to blood. Br
J Rheumatol 1998;37:7–14.
26. Morita Y, Yamamura M, Kawashima M et al. Flow
cytometric single-cell analysis of cytokine production by
CD4+ T cells in synovial tissue and peripheral blood
from patients with rheumatoid arthritis. Arthritis Rheum
1998;41:1669–76.
27. Schulze Koops H, Lipsky PE, Kavanaugh AF, Davis LS.
Elevated Th1- or Th0-like cytokine mRNA in peripheral
circulation of patients with rheumatoid arthritis.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
Modulation by treatment with anti-ICAM-1 correlates
with clinical benefit. J Immunol 1995;155:5029–37.
Kusaba M, Honda J, Fukuda T, Oizuma K. Analysis of
type 1 and type 2 T cells in synovial fluid and peripheral
blood of patients with rheumatoid arthritis. J Rheumatol
1998;25:1466–71.
Dolhain RJ, ter Haar NT, Hoekfakker S et al. Increased
expression of interferon (IFN )-gamma together with IFNgamma receptor in the rheumatoid synovial membrane
compared with synovium of patients with osteoarthritis.
Br J Rheumatol 1996;35:24–32.
Picker LJ, Singh MK, Zdraveski Z et al. Direct demonstration of cytokine synthesis heterogeneity among human
memory/effector T cells by flow cytometry. Blood
1995;86:1408–19.
Brocke S, Gijbels K, Allegretta M et al. Treatment of
experimental encephalomyelitis with a peptide analogue
of myelin basic protein. Nature 1996;379:343–6.
Tak PP, Smeets TJ, Daha MR et al. Analysis of the
synovial cell infiltrate in early rheumatoid synovial tissue
in relation to local disease activity. Arthritis Rheum
1997;40:217–25.
Kanik KS, Hagiwari E, Yarboro CH, Schumacher HR,
Wilder RL, Klinman DM. Distinct patterns of cytokine
secretion characterize new onset synovitis versus chronic
rheumatoid arthritis. J Rheumatol 1998;25:16–22.
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.
Murray KJ, Grom AA, Thompson SD, Lieuwen D, Passo
MH, Glass DN. Contrasting cytokine profiles in the
synovium of different forms of juvenile rheumatoid
arthritis and juvenile spondyloarthropathy: prominence
of interleukin 4 in restricted disease. J Rheumatol
1988;25:1388–98.
Sieper J, Braun J, Döring E et al. The aetiological role of
reactive arthritis-associated bacteria in pauciarticular
juvenile chronic arthritis. Ann Rheum Dis 1993;
51:1208–14.
Mueller B, Gimsa U, Mitchison NA, Radbruch A, Sieper
J, Yin Z. Modulating the Th1/Th2 balance in inflammatory arthritis. Semin Immunopathol 1998;20:181–96.
Miossec P, Briolay J, Dechanet J, Wijdenes J, MartinezValdez H, Banchereau J. Inhibition of the production of
proinflammatory cytokines and immunoglobulins by
interleukin-4 in an ex vivo model of rheumatoid arthritis.
Arthritis Rheum 1992;35:874–83.
Murphy E, Shibuya K, Hosken N et al. Reversibility of
T helper 1 and 2 populations is lost after long-term
stimulation. J Exp Med 1996;183:901–13.
Horsfall AC, Butler DM, Marinova L et al. Suppression
of collagen-induced arthritis by continuous administration
of IL-4. J Immunol 1997;159:5687–96.
Katsikis PD, Chu CQ, Brennan FM, Maini RN, Feldmann
M. Immunoregulatory role of interleukin 10 in rheumatoid
arthritis. J Exp Med 1994;179:1517–27.
Persson S, Mikulowska A, Narula S, O’Garra A,
Holmdahl R. Interleukin-10 suppresses the development
of collagen type II-induced arthritis and ameliorates sustained arthritis in rats. Scand J Immunol 1996;44:607–14.
Groux H, O’Garra A, Bigler M et al. A CD4+ T-cell
subset inhibits antigen-specific T-cell responses and prevents colitis. Nature 1997;389:737–42.
Saito K, Yagita H, Hashimoto H, Okumara K, Azuma
M. Effect of CD80 and CD86 blockade and anti-
Th1 cytokine pattern in rheumatoid arthritis
interleukin-12 treatment on mouse acute graft-versus-host
disease. Eur J Immunol 1996;26:3098–106.
45. Maini RN, Paulus H, Breedveld FC et al. rHu IL-10 in
subjects with active rheumatoid arthritis: a phase I and
cytokine response study. Arthritis Rheum 1998;40:S224.
1067
46. Sieper J, Kary S, Sorensen H et al. Oral type II collagen
treatment in early rheumatoid arthritis. A double-blind,
placebo-controlled, randomized trial. Arthritis Rheum
1996;39:41–51.