CD25+ regulatory cells from HLA-DQ8 transgenic mice are capable of modulating collagen-induced arthritis.

Abstract

In the last decade, CD4+CD25+ T regulatory cells have been implicated in the protection against autoimmune diseases. The human DQ8 major histocompatibility complex (MHC) class II molecule is associated with rheumatoid arthritis (RA) and various other autoimmune diseases in humans. The human leukocyte antigen (HLA)-DQ8 transgenic mouse, containing the human DQ8 MHC class II molecule, is predisposed toward collagen-induced arthritis. However, the biologic pathways responsible for DQ8-associated autoimmunity have yet to be defined, including possible defects in the CD4+CD25+ T regulatory cell compartment. To explore this concept, we examined the suppressive capacity of CD4+CD25+ T regulatory cells from DQ8 transgenic mice in vitro and, using CD25-specific depleting antibodies, investigated their influence on collagen-induced arthritis in vivo. CD4+CD25+ T regulatory cells isolated from DQ8 transgenic mice were found to be sufficient suppressors of splenocyte proliferation and interferon (INF)-gamma production. Furthermore, depletion of these cells before immunization led to significant increases in arthritis severity, collagen-specific antibodies, and INF-gamma production. These results indicate that HLA-DQ8 mice contain naturally occurring CD25+ regulatory cells that modulate collagen-induced arthritis and imply that DQ8 expression does not hinder the development of CD25+ T regulatory cells.

Full text

Page 1

S1
Available online http://arthritis-research.com/supplements/5/S1
Autoantibodies and antigens
1
Rheumatoid arthritis — class prediction by
autoreactivity profiles
R Bergholz1, F Schumann1, S Behrens1, U Ungethüm1, G Valet2,
WA Schmidt3, GR Burmester1, JM Engel4, WJ van Venrooij5,
G Steiner6,S Bläß1
1Departmentof Rheumatology & Clinical Immunology, Charité
University Clinic, Berlin, Germany
2MPI Biochemistry, Munich, Germany
3Clinic for Rheumatology Berlin Buch, Berlin, Germany
4Rheumaklinik, Bad Liebenwerda, Germany
5Departmentof Biochemistry, University of Nijmegen,
The Netherlands
6Divison of Rhematology,Department Internal Medicine III, Vienna
General Hospital, Austria
Arthritis Res Ther 2003, 5 (suppl 1):1
Heterogeneity and multifactoriality complicate diagnostics and our
understanding of pathogenesis of rheumatoid arthritis (RA). The only
accepted serologic parameter (rheumatoid factor [RF]) is not disease
specific, nor are any of several novel RA autoantibodies. We aimed at
identifying profiles instead of individual autoreactivities allowing for
unambiguous prediction of RA.
Selected RA autoantigens were tested by ELISA (RF and anti-cyclic
citrullinated peptide [anti-CCP]) or Western blot (heavy-chain-binding
protein [BiP], heterogeneous ribonucleoprotein particle A2 [RA33/
hnRNP A2], calpastatin and calreticulin). Antibody reactivities were
assayed from serum samples of 149 RA patients and 132 patients
with other rheumatic diseases and from synovial fluids (SF) (58 RA,
65 non-RA).
No single autoreactivity was sufficient for unambiguous prediction of
RA. Frequencies of multiparameter profiles consisting of 3, 4, 5 and 6
autoreactivites were determined. Fifteen six-parameter serum profiles
were exclusively expressed in RA patients, representing a cumulative
sensitivity of 59%. Twelve SF profiles were exclusively expressed in
64% of RA patients. The self-learning classification algorithm
CLASSIF1 was capable of accurately predicting RA when these profiles
were present. Data profile analysis of RF/CCP/BiP/calpastatin/calretic-
ulin/RA33 provided specific discrimination of 64% of RA. Most impor-
tantly, RA specific profiles were observed in 64% of patients with early
disease (<12 months).
For the first time, the accurate prediction of the class RA has been
achieved by the use of multiparametric autoreactivity profiles. Because
of early expression in disease, these profiles make it possible to start a
disease-modifying therapy long before irreversible bone and joint
destruction may develop. Additional RA-specific profiles are required to
cover the entire group of RA patients.
2
Investigation of the reactivity patterns of
antifilaggrin antibodies in sera and synovial fluids
from patients with rheumatoid arthritis using
citrullinated synthetic peptides
M Brózik1, J Szakonyi2, A Magyar3, B Rojkovich1, R Tobi3,
F Hudecz3, P Gergely2, K Merétey1
1National Institute of Rheumatology, Frankel Leo u 25, Budapest,
Hungary
2Central Laboratory of Immunology, Faculty of Medicine, Semmelweiss
Medical University, Budapest, Hungary
3Peptide Chemistry Research Group, Eötvös Lóránd University,
Budapest, Hungary
Arthritis Res Ther 2003, 5 (suppl 1):2
Antifilaggrin antibodies comprise a heterogeneous population of anti-
bodies directed to citrullinated proteins.
Recent studies have shown that their production is highly specific for
rheumatoid arthritis (RA) and the initial antigenic trigger for these
autoantibodies can be localised to the inflammed synovial tissue.
The aim of our study was to compare the reactivity and specificity of anti-
bodies in sera and synovial fluids towards citrullinated epitopes. Peptide
sequence corresponding to human profilaggrin (amino acid residues
306–324) and sequences with citrulline substitution at different posi-
tions were synthetised by mutipin peptide synthesis on solid supports.
Shortened versions of the peptide were also produced by removal of
amino acid residues from its N and C terminals. Completely citrullinated
variant of the 14-mer peptide was also prepared. Peptide with no cit-
rulline replacement was used as a control antigen. We found significant
differences in the sensitivity for RA of 19 individual peptides tested (from
5% to 68%), reflecting previous results that the surrounding amino acids
play an important role in creation of an autoantigenic epitope. Further, we
tested the reactivities of paired serum and synovial fluids and found very
similar peptide recognition patterns in serum and synovial IgG from the
same individuals. Studies on larger number of samples are in progress to
evaluate the results statistically that may support further evidence of the
synovial origin of antifilaggrin autoantibodies.
Acknowledgement: This work was supported by the Hungarian grant
OTKA T037876.
3
Analysis of the peptidylarginine deiminase V gene in
rheumatoid arthritis
L Caponi1, E Petit-Teixeira3, M Sebbag2, F Bongiorni1, S Moscato1,
F Pratesi1, J Osorio1, M Guerrin-Weber3, F Cornelis3, G Serre2,
P Migliorini for European Consortium for Rheumatoid Arthritis
Families (ECRAF)
1Clinical Immunology Unit, University of Pisa, Pisa, Italy
2INSERM U563, Toulouse, France
3ECRAF and Genople EVRY, France
Arthritis Res Ther 2003, 5 (suppl 1):3
A number of rheumatoid arthritis (RA) sera contain antibodies specific
for peptides in which arginine is substituted by the deiminated form cit-
Meeting abstracts
23rd European Workshop for Rheumatology Research
Marseille, France
27 February – 2 March 2003
Received: 14 January 2003 Published: 24 February 2003
© 2003 BioMed Central Ltd (Print ISSN 1478-6354; Online ISSN 1478-6362)

Page 2

S2
rulline (AKA). These antibodies are a marker of RA, as they are absent
in other disorders. The enzyme responsible for the generation of cit-
rulline residues, peptidylarginine deiminase (PAD), has different iso-
forms, with a specific tissue distribution. PAD V, expressed in
monocytes, might be responsible for the deimination of arginine
residues of synovial proteins and thus be involved in the generation of
epitopes for RA-specific antibodies. The ECRAF genome scan showed
suggestive linkage evidence at PAD V locus on chromosome 1
(P<0.005). We decided to analyze PAD V as a candidate gene for RA,
studying a cohort of 100 RA patients (tested for AKA) and their unaf-
fected parents. Investigation (by single-strand conformation polymor-
phism [SSCP] analysis and sequencing) of the 16 exons, 5′ and 3′
regions of the PAD V gene provided polymorphisms in the 5′, exons
3, 4, and 7 and 3′ regions. Analysis used the transmission disequilib-
rium test and the haplotype relative risk for alleles and haplotypes with
Analyze and Genhunter2 programs.
We found an association between RA and one PAD V haplotype (38%
in RA versus 17% in controls) (P<0.007). The association was also
observed in the AKA+RA subgroup (41%) (P<0.03).
In conclusion, the PAD V gene may be considered one of the genetic
factors that confer susceptibility to RA. Studies are in progress to
clarify the relationship between the PAD V haplotypes, the enzyme
activity and the production of anticitrulline antibodies.
4
Subclass distribution of IgG autoantibodies to
deiminated fibrinogen in rheumatoid arthritis
S Chapuy-Regaud, L Nogueira, C Clavel, M Sebbag, C Vincent,
G Serre
Department of Epidermis Differentiation and Rheumatoid
Autoimmunity, INSERM U563, Toulouse, France
Arthritis Res Ther 2003, 5 (suppl 1):4
Background: Antifilaggrin autoantibodies, previously known as ‘antik-
eratin’ antibodies or antiperinuclear factor, are serum IgG that consti-
tute the most specific diagnostic markers of rheumatoid arthritis (RA).
We showed that they specifically recognise deiminated forms of the
α and β chains of fibrin in the rheumatoid synovium. Subsequently, we
developed a new ELISA for these autoantibodies, using in vitro deimi-
nated human fibrinogen as immunosorbent (AhFibA-ELISA). We evalu-
ated its diagnostic performance in a cohort of 617 patients with
well-characterised rheumatic diseases, including 181 patients with
established RA: at a diagnostic specificity of 98.5%, the ELISA pre-
sents a diagnostic sensitivity of 76%. It is to date the most efficient test
for the diagnosis of RA.
Objective: On the basis of this test, we undertook to determine the
subclass distribution of AhFibA.
Methods: From the AhFibA-ELISA, four ELISAs using monoclonal anti-
bodies to each IgG subclass (IgG1, IgG2, IgG3 and IgG4) were devel-
oped. The ELISAs were adjusted to allow the respective proportions of
each AhFibA subclass to be determined in each serum sample tested.
141 RA patients positive for AhFibA were analysed.
Results: For each IgG subclass, the titres (optical density [OD] values)
in the whole population of patients were found to be significantly corre-
lated with those obtained with the AhFibA-ELISA. IgG1 AhFibA
reached the highest OD values (range 0.137–3.028, median: 1.125),
followed by IgG4 AhFibA (range 0–2.846, median 0.043), IgG3
AhFibA (range 0–1.448, median 0.034) and lastly IgG2 AhFibA (range
0–0.695, median 0.040). The predominance of IgG1 AhFibA was also
observed at the individual level since, among the 141 AhFibA-positive
sera, all but one contained at least 40% of IgG1 AhFibA. In 39.7% of
the sera, one or several other subclasses accounted for more than
10% of total AhFibA. IgG4 AhFibA was the most frequently associated
subclass, 25% of the sera containing more than 10% of these antibod-
ies. Only 10.7 and 7.9% of the sera contained more than 10% of IgG2
and IgG3 AhFibA, respectively.
Conclusion: These results confirm and extend those previously
obtained by indirect immunofluorescence for ‘antikeratin’ antibodies.
The predominance of IgG1 AhFibA, and their frequent association with
IgG4 AhFibA, raises the question of the Th1/Th2 balance in RA. More-
over, the predominance of IgG1 AhFibA is compatible with effector
mechanisms involving complement activation and/or the engagement
of Fc gamma receptors.
5
Peptidylarginine deiminase isoforms expressed in
the synovial membrane of rheumatoid arthritis
patients
S Chapuy-Regaud1, M Sebbag1, R Nachat1, D Baeten2,
V Foulquier1, M Simon1, T Senshu3, M Yamada3, H Takahara4,
F De Keyser2, G Serre1
1Departmentof Epidermis Differentiation and Rheumatoid
Autoimmunity, INSERM U563, Toulouse, France
2Department of Rheumatology, Ghent University Hospital, Ghent, Belgium
3Yokohama City University, Yokohama, Japan
4Ibaraki University, Ibaraki, Japan
Arthritis Res Ther 2003, 5 (suppl 1):5
Background: Antifilaggrin autoantibodies (AFAs) are highly specific for
rheumatoid arthritis and are probably involved in its pathophysiology.
We showed that they are synthesised in the rheumatoid synovial mem-
brane and that their target antigens in the tissue correspond to variants
of the α and β chains of fibrin. The variants are generated by deimina-
tion, i.e. transformation of their arginine into citrulline residues. Deimina-
tion, mediated by a peptidylarginine deiminase (PAD) activity,
generates the epitopes recognised by AFA/antifibrin autoantibodies.
Four PAD isoforms (or types), have been identified and cloned in
humans and rodents (mouse and rat). Expression of one or several of
these isoforms has been reported in numerous tissues, but their targets
are still poorly known.
Objective: Since fibrin deimination occurs in the rheumatoid synovial
tissue, we undertook to identify which PAD types are expressed in the
tissue.
Methods: By immunising rabbits with peptides situated in the most vari-
able regions of the otherwise highly conserved PAD type sequences
(three synthetic peptides per PAD), we produced antisera specific for
each of the four PAD isoforms. The antisera were affinity-purified on the
corresponding peptides. Each set of anti-peptide antibodies was con-
firmed to be specific for one isoform by immunoblotting on recombinant
or purified PADs. Additional antisera or purified antibodies to whole
human PADs II, III and V were used to confirm the results obtained with
the anti-peptide antibodies.
The synovial tissue from seven patients with rheumatoid arthritis was
analysed. In all the tissues, the presence of deiminated proteins and
particularly of deiminated fibrin was demonstrated by immunoblotting
and/or immunohistology. Then low-salt extracts of the tissues were
analysed by immunoblotting with all the immunological tools to PADs.
Results: Expression of PAD type V was clearly detected in all seven
patients. Expression of PAD type II was observed in six patients. No reac-
tivities were observed with antibodies specific for PAD types I and III.
Conclusion: Of the four PAD isoforms, only the types II and V are sig-
nificantly expressed in the synovial tissue of patients with rheumatoid
arthritis. Their respective roles in deimination of fibrin in the tissue
remain to be determined.
6
Paratope diversity of anti-prothrombin antibodies
S Cucnik, T Kveder, B Bozic
Department of Rheumatology, University Medical Centre, Ljubljana,
Slovenia
Arthritis Res Ther 2003, 5 (suppl 1):6
To ascertain the heterogeneity of anti-prothrombin antibodies (aPT), we
compared three in-house aPT ELISAs: A) medium binding plates, phos-
phatidylserine, prothrombin, Tris-buffered saline, calcium; B) high
binding plates, prothrombin, Tris-buffered saline; and C) high binding
Arthritis Research & Therapy Vol 5 Suppl 1Abstracts of the 23rd European Workshop for Rheumatology Research

Page 3

S3
plates, prothrombin, PBS. One serum, exhibiting high positive aPT in all
three ELISAs, was selected as the calibrator. Sera from 47 patients
(41 with SLE, 4 with pAPS, 2 with arterial thromboses) were tested for
IgG and IgM aPT.
The results showed six different patterns: 1) similar results were
obtained with A, B and C (similar analytical sensitivity); 2) similar results
were obtained with A and B while C showed lower analitycal sensitivity;
3) B and C seemed analytically less sensitive than A; 4) A and B were
analytically less sensitive than C; 5) A showed very low analytical sensi-
tivity; 6) A and C showed lower analytical sensitivity than B.
The analysis of all the presented patterns showed noncomparable
results of the three ELISAs. In our experiments, prothrombin was rec-
ognized by relevant antibodies when the protein was bound to phos-
phatidylserine-coated microtiter plates using calcium ions or when it
was bound to high binding plates with or without calcium ions. Never-
theless, detected aPT values were not of the same fine specificity. Dif-
ferent binding conditions for prothrombin exposed different epitopes,
resulting in the detection of various subgroups of aPT.
7
High-affinity antibodies against β β2-glycoprotein I
S Cucnik, T Kveder, A Ambrozic, B Borut
Department of Rheumatology, University Medical Centre, Ljubljana,
Slovenia
Arthritis Res Ther 2003, 5 (suppl 1):7
Background: Antibodies against β2-glycoprotein I (anti-β2GPI) are
believed to be of low affinity and thus unable to bind to the free antigen
in a solution.
Objective: The aim of our study was to determine the affinity of IgG
anti-β2GPI, isolated by affinity chromatography.
Methods: A β2GPI affinity column was prepared by CNBr-activated
agarose without spacer arms and human purified unnicked β2GPI. The
IgG fraction from the protein G column was applied to the column and
bound antibodies were eluted with various solutions: A) 0.1M glycine /
0.5M NaCl / 0.1% Tween 20pH2.5; B) 0.1M glycine / 4M NaCl /
0.1% Tween 20pH2.5; C) 0.1M sodium borate pH10 and D) 25%
ethylene glycol. Eluted fractions containing anti-β2GPI antibodies were
neutralised and analysed by ELISA using various binding buffers. The
level of anti-β2GPI antibodies in each sample was derived from the
standard curve according to the defined dilutions of monoclonal anti-
bodies (AUG are arbitrary units of IgG monoclonals).
Results: Increased concentrations of sodium ions in the binding solu-
tion from 0.15, 0.25, 0.50, 1.11, 2.07 and 4.0MNaCl did not com-
pletely prevent the binding between isolated antibodies and β2-GPI
(79.8, 65.3, 36.1, 19.9, 12.0 and 8.1AUG, respectively).
Conclusion: In contrast to the common opinion that all anti-β2GPI
autoantibodies are of low affinity, we clearly showed that at least one
subset among them was of high affinity.
8
Anti-Ro/SSA antibodies in rheumatoid arthritis (RA)
F Franceschini, I Cavazzana, F Malacarne, P Airò, R Cattaneo,
N Del Papa, A Radice, RA Sinico
Piazzale Spedali Civili, Brescia, Italy
Arthritis Res Ther 2003, 5 (suppl 1):8
Background: AntiRo are found in 5–15% of RA. Significant associa-
tions were reported with sicca, vasculitis, hypergammaglobulins, ANA,
high-titer RF, toxicity to D-penicillamine and gold salts treatment. Aim of
the study: to evaluate clinical features, radiologic progression and
response to disease-modifying antirheumatic drugs (DMARDs) in
antiRo-RA patients.
Patients and methods: We studied 210 patients with RA: antiRo
were determined by CIE, with human spleen extract, and by ELISA with
recombinant Ro proteins (Pharmacia). Cutoff values for ELISA were
determined testing 177 sera from routine.
Results: AntiRo were detected in 27 patients (F:M 12.5:1). Two groups
(antiRo+and antiRo–) did not show any difference with regard to disease
duration, arthritis onset and articular erosions. AntiRo were associated
with xerophthalmia (P<0.0000001), xerostomia (P=0.0012), oral ulcers
(P=0.0067), scleritis (P=0.0067) and amyloidosis (P=0.042).
Rheumatoid factor, antiperinuclear factor and anticitrulline were recorded
in 70% in both groups; hypergammaglobulinemia, ANA, anti-dsDNA and
AMA were frequently detected in antiRo+patients. Patients were given a
mean of 3.93 DMARDs, with no statistical difference between antiRo+
and antiRo–: hydroxychloroquine, methotrexate and gold salts are the
most frequently used. Patients who were antiRo– were more frequently
treated with hydroxychloroquine and infliximab, while D-penicillamine was
used more frequently in those who were antiRo+. DMARD toxicity was
detected in 9.3% of antiRo+, with no statistically significant difference
between the two groups.
Conclusion: AntiRo, found in 12.8% of patients with RA, is associated
with extra-articular features and with an autoantibody profile unusual for
RA. No difference with respect to DMARD toxicity was found in anti-
Ro+patients.
9
Presence of anti-RNP-A and anti-RNP-C antibodies is
inversely associated with renal symptoms of systemic
lupus erythematosus
I Hoffman1, I Peene1, L Meheus2, K De Bosschere2, F Hulstaert2,
TWJ Huizinga3, L Cebecauer4, D Isenberg5, EM Veys1,
F De Keyser1
1UZG, Ghent, Belgium
2Innogenetics, Ghent, Belgium
3Leiden University Medical Center, Leiden, The Netherlands
4Research Institute for Rheumatic Diseases, Piestany, Slovakia
5University College London, London, UK
Arthritis Res Ther 2003, 5 (suppl 1):9
Background: Systemic lupus erythematosus (SLE) is an autoimmune
rheumatic disease characterised by the production of autoantibodies.
The most common serious feature of SLE is renal involvement. An
association with anti-RNP antibodies remains controversial.
Aim: To identify associations of autoantibodies and renal symptoms in
a consecutive cohort of SLE patients.
Methods: Sera and clinical data from 235 consecutive SLE patients,
fulfilling the ACR criteria for SLE, were collected in four centres. The
presence of renal disease was defined as the presence of cellular
casts in the urine, proteinuria (>0.5g/day), or glomerulonephritis during
the course of the disease. Autoantibody profiles were determined by
the INNO-LIATMANA Update (a line immunoassay with recombinant
and/or native antigens, including SmB, SmD, RNP-70, RNP-A, RNP-C,
Ro52, Ro60, SSB, and ribosomal P) and anti-dsDNA antibodies by
indirect immunofluorescence on Crithidia luciliae. Odds ratios (ORs)
and their 95% confidence intervals (CI) were computed to determine
the associations between antibodies and renal symptoms. No correc-
tion was made for multiple testing.
Results: The presence of anti-RNP-A and anti-RNP-C appeared to be
protective against renal involvement (OR=0.445, CI=0.210–0.942
and OR=0.484, CI=0.243–0.964, respectively). Concerning the indi-
vidual symptoms, anti-RNP-C was associated with a lower occurrence
of proteinuria (OR=0.470, CI=0.236–0.938), cellular casts
(OR=0.324, CI=0.150–0.696) and glomerulonephritis (OR=0.460,
CI=0.226–0.934), whereas anti-RNP-A was only significantly associ-
ated with a lower occurrence of cellular casts (OR=0.303,
CI=0.129–0.716). In contrast, antibodies to dsDNA were associated
with a higher risk for cellular casts (OR=2.014, CI=1.057–3.839).
We found no associations between renal symptoms and other specific
antinuclear reactivities. More specifically, for anti-RNP-70 a trend was
only detected for the association with the presence of cellular casts
(OR=0.411, CI=0.153–1.106).
Conclusion: Anti-RNP-A and anti-RNP-C antibodies appear to be
associated with a lower risk for renal disease.
Available online http://arthritis-research.com/supplements/5/S1

Page 4

S4
10
Cultured salivary gland epithelial cells release
exosomes that contain the Sjögren’s-syndrome-
associated autoantigenic ribonucleoproteins Ro/SSA
and La/SSB
EK Kapsogeorgou, ID Dimitriou, RF Abu-Helu,
HM Moutsopoulos, MN Manoussakis
Department of Pathophysiology, National University of Athens,
75 Mikras Asias Street, Athens, Greece
Arthritis Res Ther 2003, 5 (suppl 1):10
Sjögren’s syndrome (SS) is characterized by exocrine gland destruc-
tion associated with lymphocytic infiltrations and chronic autoimmune
antigen-driven responses against the intracellular Ro/SSA and La/SSB
ribonucleoproteins. Epithelial cells, which are the main targets of
autoimmune responses, appear to have a central role in the pathogene-
sis of SS. In recent years we have presented evidence indicating that
salivary gland epithelial cells (SGECs) of SS patients are inherently
capable of functioning as antigen-presenting cells. Recently, a novel,
cell-free mechanism of antigen presentation has been identified. This
mechanism involves exosomes, which are small (30–200nm) mem-
brane vesicles of endosomal origin secreted by a variety of cell types,
such as reticulocytes, B lymphocytes and dendritic cells. In the present
study we investigated the capacity of cultured SGEC lines established
from SS patients and disease controls to release exosomal vesicles
that contain intracellular ribonucleoproteins. Membrane vesicles were
isolated by differential centrifugation from SGEC culture supernatants
and their nature was confirmed by electron microscopy. Cultured
SGECs from patients and controls secreted significant amounts of exo-
somal vesicles, in a manner largely indistinguishable from other exosome-
secreting cells. Exosome release was not associated with apoptosis or
other cellular destruction processes. SGEC-derived exosomes were
found by Western blot analysis to contain Ro/SSA, La/SSB, and Sm
ribonucleoproteins. Our results indicate that SGECs are capable of
secreting exosomes. This mechanism may represent a pathway through
which intracellular epithelial antigens are exported and subsequently pre-
sented to the immune system. In this context, exosomes produced by
epithelial cells may have a role in the pathogenesis of SS.
11
Low frequency of phosphatidylserine/prothrombin
complex antibodies in a cohort of patients with
anticardiolipin antibodies and recent thrombosis
H Locht
Department of Autoimmunology, Statens Serum Institut, Copenhagen,
Denmark
Arthritis Res Ther 2003, 5 (suppl 1):11
Objective: Clinical data from patients who were positive for anticardi-
olipin antibodies (ACAs), tested on a routine basis in an autoimmune
laboratory, were obtained by questionnaires from the referring physi-
cians. One hundred and sixty-two individuals had experienced a recent
(within <6 months) thromboembolic event. All sera were tested for IgG
and IgM antibodies against cardiolipin, β2-glycoprotein 1 (β2-glp 1),
and the complex of phosphatidylserine/prothrombin (PPC) by in-house
ELISA methods.
Results: Among the 162 ACA-positive patients, 31 (19%) were also
positive for antibodies against PPC. In the group with ACA only, 73%
of the patients had no pre-existing rheumatic condition, compared with
32% in the PPC group (P=0.00002). Thirty-two percent in the PPC
group had SLE, vs 12% in the ACA group (P=0.016). The fractions of
patients with deep venous thrombosis (DVT), pulmonary embolism
(PE), or myocardial infarction (MI) were equal, whereas cerebrovascular
incidents (CVI) were more frequent among ACA patients; 51% vs 26%
(P=0.02). Antibodies against β2-glp 1 were also more frequent in the
PPC group 61% vs 30% (P=0.002).
Conclusion: Antibodies against both ACA and PPC seem to define a
subset of patients with autoimmune thrombophilia. More patients with
PPC antibodies had SLE and also tested positive for antibodies
against β2-glp 1. The distribution of thrombotic manifestations differed
between the two populations in that CVI was more frequent in the
ACA-only group, whereas the fractions with DVT, PE, and MI were
equal.
12
Anticardiolipin antibodies of IgG and IgM isotypes
reflect different forms of recent thromboembolic
events
H Locht, A Wiik
Department of Autoimmunology, Statens Serum Institut, Copenhagen,
Denmark
Arthritis Res Ther 2003, 5 (suppl 1):12
Objective: To correlate the distribution and levels of anticardiolipin
(ACA) and anti-β2-glycoprotein 1 (a-β2-glp 1) antibodies of IgG and
IgM isotypes to the clinical spectrum of recent (within <6 months)
thromboembolic events.
Method: During one year, all sera positive for IgG or IgM ACA submit-
ted on a routine basis from hospitals or primary-care physicians from all
parts of Denmark were recorded. Information about thromboembolic
events and any underlying rheumatic disease was obtained by ques-
tionnaires from the referring physicians. Sera were analysed for
IgG/IgM ACA and a-β2-glp 1 antibodies by in-house ELISA assays, and
the results were expressed in arbitrary units.
Results: One hundred and sixty-two patients fulfilled criteria for recent
thromboembolic disease. Cerebrovascular infarction (CVI) was present
in 82 patients, deep venous thrombosis (DVT) in 34, pulmonary
embolism (PE) in 14, myocardial infarction (MI) in 4, and other throm-
boses in 28 patients. Isolated IgG ACA was found in 31 of 48 patients
with DVT+PE (65%), but in only 21 of 82 patients with CVI (26%)
(P=0.00002). In contrast, isolated IgM ACA was found in 9 (19%) of
patients with DVT+PE, but in 46 (56%) CVI patients (P=0.00007).
IgG a-β2-glp 1 antibodies were found in 13 (16%) CVI patients and
23 (48%) DVT+PE patients (P=0.0002).
Conclusion: IgG and IgM ACA isotypes seem to define different clini-
cal subsets of patients with recent thromboembolic events, with IgG
ACA being most prevalent in the group having DVT+PE whereas IgM
ACA is found primarily among CVI patients. There was a linear correla-
tion between levels of IgG a-β2-glp 1 antibodies and IgG ACA.
13
Autoantibodies in osteoarthritis
J Menard1, J Neidel2, C Perka2, M Sparmann3, B Mueller1
1Deutsches RheumaForschungs-Zentrum, Berlin, Germany
2Charité, Humboldt-University, Berlin, Germany
3Immanuelkrankenhaus, Berlin, Germany
Arthritis Res Ther 2003, 5 (suppl 1):13
Objective: We have previously shown that inflammatory cytokines are
up-regulated in chondrocytes of osteoarthritis (OA) patients. However,
the inflammatory responses associated with OA are still ill defined. To
investigate in more detail the involvement of the immune system in the
pathogenesis of OA, we here analyzed patients for the presence of
autoantibodies.
Methods: Both sera and synovial fluids obtained from 85 OA patients
at the time of joint replacement were tested for autoreactivity. Autoanti-
bodies reacting against synovial membranes were detected performing
immunofluorescence on cryosections. Autoantibodies recognizing
lysate components of various cell lines (B-cell, T-cell, monocyte, fibrob-
last and chondrosarcoma) were shown by the use of Western blot
analyses. In preparation for the identification of the respective autoanti-
gens, we started to immunoprecipitate the proteins of interest.
Results: Sera from the OA patients reacted strongly with cells in the
synovial membranes whereas sera obtained from healthy donors did
Arthritis Research & Therapy Vol 5 Suppl 1Abstracts of the 23rd European Workshop for Rheumatology Research

Page 5

S5
not (Fig.1). Furthermore, about 50% of the OA sera and none of the
control or RA sera reacted very strongly against lysate components of
the different cell lines tested. The corresponding synovial fluids reacted
against the same lysate components, but the signals obtained were of
reduced intensity.
Conclusion: Our results demonstrate the presence of autoantibodies
in sera and synovial fluid in about half the OA patients. The specificities
of these autoantibodies are not restricted to the joints, as reactivity was
detected towards any cell line tested. Hoping to design new diagnostic
tools and to shed light on the role of autoantibodies in the development
and progression of OA, we are in the process of identifying the corre-
sponding autoantigens.
14
The infectious origin of the antiphospholipid
syndrome: induction by passive transfer of anti-β β2GPI
antibodies induced by common bacteria
M Blank, I Krause1, M Fridkin2, N Keller1, J Kopolovic1,
I Goldberg1, A Tobar1, Y Shoenfeld1
1Center for of Autoimmune Diseases, Department of Internal Medicine
'B', The Weizmann Institute of Science, Rehovot, Israel
2Department of Organic Chemistry, The Weizmann Institute of
Science, Rehovot, Israel
Arthritis Res Ther 2003, 5 (suppl 1):14
The antiphospholipid syndrome (APS) is characterized by the presence
of pathogenic autoantibodies against β2-glycoprotein I (β2GPI). The
factors causing production of anti-β2GPI remain unidentified, but an
association with infectious agents has been reported. We recently identi-
fied a hexapeptide (TLRVYK) that is recognized specifically by a patho-
genic anti-β2GPI monoclonal antibody. In the present study we evaluated
the APS-related pathogenic potential of microbial pathogens, which
share structural homology with the this hexapeptide. Mice were immu-
nized with a panel of TLRVYK-related microbial particles and were
studied for the development of mouse anti-β2GPI autoantibodies. Mouse
IgG specific to the TLRVYK peptide were affinity purified from the immu-
nized mice and passively infused i.v. into naive mice at day 0 of preg-
nancy. APS parameters were evaluated in the infused mice on day 15 of
pregnancy. Following immunization, high titers of anti-peptide, anti-β2GPI
antibodies were observed in mice immunized with Haemophilus influen-
zae, Neisseria gonorrhoeae or tetanus toxoid. Naive mice infused with
the affinity-purified anti-peptide antibodies had a significant thrombocy-
topenia, prolonged aPTT and elevated percentage of fetal loss, similar to
the findings in a control group of mice immunized with a pathogenic anti-
β2GPI monoclonal antibody. Our study establishes a mechanism of mol-
ecular mimicry in experimental APS, demonstrating that bacteria
homologous with β2GPI structure are able to induce the generation of
pathogenic anti-β2GPI antibodies along with APS manifestations.
15
Fcγ γRI up-regulation induced by local adenoviral-
mediated IFN-γ γ production aggravates chondrocyte
death during immune-complex-mediated arthritis
K Nabbe1, PL van Lent1, AE Holthuysen1, AW Sloetjes1,
J Kolls2, JS Verbeek3, WB van den Berg1
1Department of Experimental Rheumatology and Advanced
Therapeutics, University Medical Center Nijmegen, The Netherlands
2Louisiana State University Health Science Center, New Orleans, LA, USA
3Department of Human and Clinical Genetics, University Medical
Center, Leiden, The Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):15
Using various FcγR-deficient mice, we have obtained suggestive evi-
dence that FcγRI on macrophages is responsible for severe cartilage
destruction during arthritis mediated by immune complexes (ICs) and
Th1 cells. In contrast, in arthritis mediated solely by ICs, FcγRIII seems
more important. This suggests that T-cell-mediated FcγRI up-regulation
promotes pronounced cartilage destruction. A likely Th1-cell-derived
cytokine mediating FcγRI expression is interferon-γ (IFN-γ).
In the present study we investigated whether IFN-γ is able to up-regulate
cartilage destruction during experimental immune complex-mediated arthri-
tis (ICA) and, if so, whether this mechanism is indeed regulated by FcγRI.
IFN-γ was locally overexpressed in the murine knee joint prior to ICA
induction by the use of adenoviral vectors. This had no significant effect
on joint inflammation as studied by histology. However, irreversible carti-
lage destruction as studied by the degree of chondrocyte death was
markedly enhanced. IFN-γ overexpression resulted in a fivefold increase
in chondrocyte death, in comparison with the control group, which had
received a control adenoviral vector expressing GFP (AdGFP).
To study whether this effect of IFN-γ was related to the presence of
ICs, IFN-γ was also overexpressed in a naive joint and during zymosan-
induced arthritis, which is an IC-independent arthritis model. No severe
cartilage destruction was found, implying a crucial role for ICs and their
receptors (FcγRs) in the IFN-γ effect.
When IFN-γ was overexpressed in murine knee joints, FcγRI mRNA
expression was up-regulated in synovial cells. To prove that the aggra-
vation of chondrocyte death by IFN-γ is indeed FcγRI-mediated, ICA
was raised in FcγRI–/–. IFN-γ overexpression did not result in significant
elevation of joint inflammation either in FcγRI–/–or their wild-type con-
trols. Interestingly, although IFN-γ was overexpressed, chondrocyte
death remained absent in FcγRI–/–, whereas in wild-type controls chon-
drocyte death was highly increased after IFN-γ overexpression.
These results indicate that IFN-γ can aggravate cartilage destruction in
an IC-dependent fashion, mediated by FcγRI.
16
The diagnostic significance of autoantibodies in
patients with very early rheumatoid arthritis
V Nell, K Machold, W Hueber, G Eberl, H Hiesberger, E Hoefler,
J Smolen, G Steiner
Division of Rheumatology, University Hospital of Vienna, Lainz
Hospital, Vienna, Austria
Arthritis Res Ther 2003, 5 (suppl 1):16
In the past few years, several novel autoantibodies (autoAbs) have
been described in patients with rheumatoid arthritis (RA), including an
autoAb to citrullinated antigens (anti-CCP) and anti-RA33 autoAb. It
was our aim to assess the value of these two autoAbs in relation to
rheumatoid factor (RF) in discriminating RA from non-RA in a cohort of
patients with very early arthritis.
Available online http://arthritis-research.com/supplements/5/S1
Figure 1
Immunofluorescence reveals the presence of autoantibodies in OA
sera. Cryosections of synovial membranes were incubated with
control (A) or OA (B) sera. The binding of autoantibodies was detected
via FITC-labeled anti human IgG antibodies (green). Fibroblasts were
stained with the specific antibody (5B5) and developed via SA-
rhodamine (red). Nuclei are counterstained using DAPI (blue).
The full colour version of this figure can be viewed online at
http://arthritis-research.com/supplements/5/S1/13

Page 6

S6
Ninety-four patients with arthritis of less than 3 months’ duration were
included in this prospective study. Follow-up was for at least 1 year.
Sixty-one patients had a final diagnosis of RA and 33 had other arthri-
tides. Among the RA patients, RF was present in 34 (56%), anti-CCP
in 18 (30%), and anti-RA33 in 15 (25%) at their first visit. Among
the 33 non-RA patients, 6 were RF-positive, 3 had anti-RA33, and
1 had anti-CCP. Thus, anti-CCP was very specific for RA with a posi-
tive predictive value (PPV) of 95%, while RF and anti-RA33 were
somewhat less specific, with PPVs of 85% and 83%, respectively.
However, the co-occurrence of anti-RA33 and RF was observed exclu-
sively in RA patients and thus had a PPV of 100% in this relatively
small cohort of patients. In conclusion, these data suggest that the
determination of autoantibodies such as anti-CCP and anti-A2/RA33 in
addition to RF may be quite helpful in the early diagnosis of RA.
17
A rapid ELISA based method to determine peptidyl-
arginine deiminase activity in biological samples
S Nijenhuis, AJW Zendman, JMH Raats, GJM Pruijn,
WJ van Venrooij
Department of Biochemistry, University of Nijmegen, Nijmegen,
The Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):17
Peptidylarginine deiminases (PADs; EC 3.5.3.15) are a family of
calcium-dependent enzymes that convert peptidylarginine into peptidyl-
citrulline. The recent finding that patients with rheumatoid arthritis (RA)
produce autoantibodies against citrulline-containing epitopes greatly
increased the interest in the PAD enzymes and their activities. It is not
yet known whether there is a causative relationship between the gener-
ation of antibodies targeting citrullinated epitopes and the development
of the disease. Characterisation of the structure and function of PADs
may help to understand the production process of citrullinated antigens
and possibly also the aetiology of RA.
Several assays are known to monitor PAD activity in biological
samples. However, these assays either have a low sensitivity or are
laborious. Here, we describe the development of a simple, rapid
method for the simultaneous analysis of many PAD-containing samples.
This new method is based on the binding of an antibody specifically
recognising a citrulline-containing epitope in a defined peptide. We
show that this method is very sensitive and can be applied to monitor
PAD activity in many types of biological samples, such as bacterial
lysates, mammalian cell extracts and tissue extracts.
18
Autoantibodies to deiminated fibrinogen are the
most efficient serological criterion for early
rheumatoid arthritis diagnosis
L Nogueira1, S Chapuy-Regaud1, A Constantin1, C Clavel1,
M Sebbag1, A Cantagrel2, C Vincent1, G Serre1
1Department of Epidermis Differentiation and Rheumatoid
Autoimmunity, INSERM U 563, Toulouse, France
2Department of Rheumatology, Rangueil Hospital, Toulouse, France
Arthritis Res Ther 2003, 5 (suppl 1):18
Background and objectives: Autoantibodies to deiminated proteins
are known to be the most specific serological marker for the diagnosis
of rheumatoid arthritis (RA). We recently showed that deiminated fibrin
is the major target of this family of autoantibodies in rheumatoid syn-
ovial tissue. We subsequently developed, and validated on a large
series of patients with established rheumatic diseases, an ELISA for
the detection of circulating autoantibodies to deiminated human fibrino-
gen (AhFibA). The test was shown to be the most efficient (specific
and sensitive) serological criterion for the diagnosis of RA.
Methods: We collected 352 sera from patients with arthritides of
recent onset (disease duration <1 year). The diagnosis was estab-
lished after at least 2 years’ follow-up. The patients were then classified
into two groups, 175 with RA and 177 with non-RA inflammatory
rheumatic diseases. The previously developed ELISA, using in vitro
deiminated human fibrinogen as immunosorbent, was used for detec-
tion and titration of AhFibA. Antibodies to cyclic citrullinated peptide
(CCP) were sought in accordance with the manufacturer’s procedure
(Immunoscan RA, Euro-diagnostica). Rheumatoid factor (RF) was
titrated by nephelometry (RF-reagent for Image, Beckman Coulter).
Results: The diagnostic sensitivity of AhFibA was found to be signifi-
cantly higher than those of CCP (P<0.05) and RF (P<0.001)
(Table 1). The positive predictive values (PPV) of the three tests were
all found to be very high and were not significantly different. The nega-
tive predictive values (NPV) were too low to be diagnostically useful.
Among the AhFibA-positive RA sera 83% were RF-positive, while
among the AhFibA-negative, only 13% were RF-positive.
Table 1
Diagnostic indices computed at thresholds allowing 0.985 specificity to
be reached
Antibodies toSePPV
AhFibA0.646 0.974
CCP 0.543 0.969
RF0.274 0.941
PPV, positive predictive value.
Conclusion: Unlike the case with RF, autoantibodies to deiminated
proteins are confirmed to be of diagnostic value in early arthritides. The
detection of these autoantibodies by ELISA using their synovial target
(deiminated fibrin) appears the most efficient method for the diagnosis
of early RA.
19
Anti-BIP antibodies in rheumatoid arthritis
GS Panayi, M Bodman-Smith, V Corrigall
GKT School of Medicine, Guy’s Hospital, London, UK
Arthritis Res Ther 2003, 5 (suppl 1):19
Background: We have implicated the human chaperone protein BiP in
the pathogenesis of rheumatoid arthritis (RA). Increased immunoglobu-
lin binding of RA sera to BiP is seen on Western blot analysis.
Methods: We now describe an ELISA developed to enable rapid
screening of sera for antibody reactivity to BiP.
Results: Specificity of the assay has been shown by free ligand com-
petition and extensive correlations with other immunological parame-
ters. We show no correlation between anti-BiP and rheumatoid factor
or with cyclic citrullinated peptide. We confirm the increased binding of
immunoglobulin to BiP in the sera of a group of patients with RA
(n=96) in comparison with controls (n=96). Our data show a speci-
ficity of 71% and a sensitivity of 73% for RA. Furthermore, these data
show that antibody will bind to a nonglycosylated form of BiP, since the
protein is produced in an Escherichia coli expression system.
Conclusion: We have developed a robust protocol for the detection of
antibodies to the human chaperone molecule BiP. Our data show an
elevated antibody response to BiP in RA patients and hence support a
role for this molecule in the disease.
20
Antineutrophil cytoplasmic antibodies in synovial
fluid from patients with early rheumatoid arthritis
M Puszczewicz, I Zimmermann-Górska,
G Bialkowska-Puszczewicz, E Tatarkiewicz
University of Medical Sciences, Poznañ, Poland
Arthritis Res Ther 2003, 5 (suppl 1):20
Background: Antineutrophil cytoplasmic antibodies (ANCAs) occur in
vasculitides, including Wegener’s granulomatosis and Churg-Strauss
Arthritis Research & Therapy Vol 5 Suppl 1Abstracts of the 23rd European Workshop for Rheumatology Research

Page 7

S7
syndrome. Their presence in sera of patients with some other
rheumatic diseases has been well documented. ANCAs have also
been detected in synovial fluid (SF) of patients with RA; however, their
presence and potential role in early stages of the disease is not known.
Objective: The aim of the study was to evaluate the prevalence and
specificity of ANCAs in SF in early RA.
Methods: SF samples were obtained from 114 patients with early RA.
Control SF were taken from 76 patients with knee osteoarthritis (OA).
ANCAs were detected by indirect immunofluorescence and by ELISA.
Proteinase 3, myeloperoxidase, elastase, lactoferrin and lysozyme, as
well as cathepsin G, were used as antigens in ELISA method. At the
same time antinuclear antibodies (ANAs) and rheumatoid factor (RA)
were detected in SF samples under study.
Results: ANCAs were found by indirect immunofluorescence in SF of
26/114 (22.8%) patients with early RA and 2/76 (2.6%) patients with
OA. A perinuclear pattern (p-ANCA) was detected in 22/26 cases
(84.6%), and atypical pattern (a-ANCA) in 4/26 (15.3%). In patients
with OA, p-ANCAs only were observed. We did not observe a cyto-
plasmic pattern (c-ANCA) in indirect immunofluorescence or in the
reactivity against proteinase 3 in ELISA. p-ANCAs yielded reactivity
against lactoferrin in SF from 15/26 (57%) and against myeloperoxi-
dase in 5/26 (19.2%) SF samples from RA patients. a-ANCAs indi-
cated reactivity against cathepsin G in 3/26 cases (11.5%) and
against lysozyme in 1/26 (3.8%). RF was present in 18/26 (69.2%)
and ANA in 14/114 (12.2%) SF samples from patients with RA.
Conclusion: ANCAs are present in SF of over 20% of patients with
early RA We think that these antibodies in SF could be one of the early
RA markers. Their role in this stage of synovitis should be clarified.
21
Biphasic decline/increase for anticitrulline and
monophasic decline in anti-type II collagen antibody
levels in recently diagnosed RA patients
J Rönnelid1,2, S Rogberg1, B Nordmark1, J Lampa1, I Dahlbom3,
T Hansson1,3, L Klareskog1
1Unit of Rheumatology, Karolinska Hospital, Stockholm
2Unit of Clinical Immunology, Uppsala University, Uppsala, Sweden
3Pharmacia Diagnostics, Uppsala
Arthritis Res Ther 2003, 5 (suppl 1):21
Objective: Antibodies against both type II collagen and citrulline
residue containing peptides can be found in rheumatoid arthritis (RA)
patients. We have noted that serum levels of both anticollagen and
anticitrulline antibodies show significant decline after inclusion in an
early arthritis clinic (EAC) cohort, and wanted to compare their kinetic
patterns of disappearance.
Methods: Two hundred and fifty-five EAC patients and 80 RA patients
were followed for 1 and 5 years, respectively, from the time of initial
referral. Anticollagen and anticitrulline antibodies were determined with
ELISA at inclusion and after 3 months and 1, 2, 3 and 5 years.
Results: At the patients’ inclusion, anticollagen and anticitrulline anti-
bodies were found in 6.3% and 48.2%, respectively, of EAC patients
and in 11.5% and 61.0% of the patients with definite RA. Serum levels
of antibodies against type II collagen declined continously for the entire
study period in both groups, being significant already at 3 months after
inclusion. Serum levels of anticitrulline antibodies, on the other hand,
showed a biphasic pattern. A significant decline from inclusion (for EAC
patients at 1 year, and for RA patients at 3 months and at 1 and 2 years)
was followed by increased serum levels (significant for RA patients
between 2 and 5 years). No correlation was found between changes in
serum levels of anticollagen and anticitrulline antibodies at any time.
Conclusion: These findings suggest that different immunological
events predispose to immunity against collagen type II and against cit-
rulline-containing peptides in RA.
22
Zinc finger domain of Ro60kD autoantigen is essential
for binding of Ro52kD and autoantibodies
JG Routsias, A Makri, C Sakarellos, M Sakarellos-Daitsiotis,
A Kosmopoulou, HM Moutsopoulos, AG Tzioufas
Department of Pathophysiology, School of Medicine, University of
Athens, Athens, Greece
Arthritis Res Ther 2003, 5 (suppl 1):22
The Ro60kD polypeptide is associated with both RNA and the Ro52kD
protein. The specific protein–RNA and protein–protein interactions are
thought to occur through the RNP and zinc-finger secondary structure
elements located on the 92–161 and 301–327 regions of Ro60kD
protein, respectively. The zinc finger domain of Ro60kD is a good can-
didate to hold a conformational epitope, because both the binding of
zinc and the redox conditions can induce specific conformational
changes. In this study, we investigated the presence of antibodies
against synthetic peptides corresponding to the zinc finger domain of
Ro60kD protein (Zif-1b), to a truncated form possessing zinc-binding
regions but lacking the intermediate loop (Zif-2b) and to the intermedi-
ate loop (310–319) of the zinc finger domain (Zif-3b). We found that
the peptide Zif-1b, corresponding to the native sequence (301–327
aa) of Ro60kD, is recognized by antibodies from the majority (83%) of
anti-Ro/La-positive patients with primary Sjögren’s syndrome (pSS), in
the absence of zinc ions. The same sera failed to react with Zif-1b in
the presence of Zn2+(2.5%). Its truncated form (Zif-2b) did not react
against the same sera, while the peptide corresponding to loop
310–319 (Zif-3b) exhibited high reactivity (85%). The presence of zinc
ions was necessary for binding of Zif-1b to recombinant Ro52kD, indi-
cating that discrete conformational states of the Ro60kD zinc finger
domain are employed in interaction with Ro52kD protein and autoanti-
bodies. The two different states of the zinc finger domain of Ro60kD
may reflect the existence of the Ro60kD autoantigen in different redox
environments (e.g. in the interior of the cell and cell membrane), where
the Cys residues have different capacities to coordinate zinc ions.
23
Differential expression of IgVH mRNAs in human RA
synovium detected by single-cell RT-PCR
S Ruzickova1,3, Z Cimburek1, J Niederlova1, O Horvath2,
O Krystufkova1, T Dörner2, J Vencovsky1
1Institute of Rheumatology and Laboratory of Gene Expression, Czech
Academy of Sciences, Prague, Czech Republic
2Institute of Microbiology, Czech Academy of Sciences, Prague,
Czech Republic
3Department of Rheumatology/Immunology,Medical Faculty, Charité,
Humboldt University, Berlin, Germany
Arthritis Res Ther 2003, 5 (suppl 1):23
Introduction: In rheumatoid arthritis (RA), the synovial membranes
contain lymphocytic infiltrates sometimes resembling germinal centres.
The production of clonally related immunoglobulin (Ig) transcripts, the
presence of plasma cells within RA synovial tissue, somatic mutations
and isotype switching in RF-specific synovial B lymphocytes have been
observed. In addition, the expression of recombination-activating
genes 1 and 2 (Rag1 and 2) has been detected in RA synovial B cells.
Objective: Analysis of the presence of plasma cells, mutational fre-
quencies of their Ig heavy-chain transcripts, the signs of isotype switch-
ing and expression of Rag1 and 2 genes in inflamed RA synovial tissue.
Methods: Synovial tissue sections were labeled with anti-CD19, anti-
CD138 and anti-CD38 antibodies and visualized using the alkaline
phosphatase technique. Individual CD19+CD38+plasma cells were
isolated from digested synovium of two caucasian RA patients using
single-cell deposition. The cDNA from each single B cell was gener-
ated, and nested polymerase chain reaction specific for VH genes and
Rag1 and Rag2 genes was performed. After sequencing, the VBASE
database was used to assign VH, DH and JH gene segments and
somatic mutations.
Available online http://arthritis-research.com/supplements/5/S1

Page 8

S8
Results: Three different subsets of CD19+CD38+plasma cells were
detected. The first subset represents cells expressing only IgM tran-
scripts (IgM+, 13.5%), the second expressed only IgG transcripts
(IgG+, 48.7%) and the third produced both IgM and IgG mRNAs (IgM+
IgG+, 37.8%). All of these detected IgVH mRNAs contained mutated
sequences, indicating their memory cell origin. However, the differ-
ences of mutational frequencies between these subsets were statisti-
cally significant (IgM+plasma cells 3.8%, IgG+plasma cells 11.2% and
IgM+IgG+cells 6.3%). Interestingly, either Rag1 or Rag2 mRNA was
observed in 83.3% of all analyzed CD19+CD38+plasma cells, with
the highest frequency in the IgM+IgG+subset of plasma cells (71.4%).
Conclusion: The population of CD19+CD38+plasma cells differentially
expressing mutated IgVH mRNAs and reinducing Rag 1 and 2 genes
was observed in RA synovium. The IgM+IgG+cells might represent cells
switching from IgM to IgG isotype and IgG+plasma cells might corre-
spond to post-switched cells producing high-affinity (auto)antibodies.
24
Ability of second-generation anti-cyclic citrullinated
peptide (CCP) to predict rheumatoid arthritis in
patients with early arthritis
A Saraux, JM Berthelot, P Le Goff, P Youinou
Department of Rheumatology, Centre Hospital Universitaire, Brest, France
Arthritis Res Ther 2003, 5 (suppl 1):24
Objective: We previously studied the diagnostic value of biological tests in
RA and found that IgG-AKA, first generation of anti-CCP, IgM-RF by
ELISA, and the latex test is the best combination in diagnosing RA. The
goal of the present study was 1)to study the diagnostic value of the
second generation of anti-CCP and 2)to determine the diagnostic value of
second-generation anti-CCP used in combination in the same population.
Methods: A cohort of 270 patients with early arthritis underwent a stan-
dardized examination, laboratory tests and radiographs in 1995–1997.
The final diagnosis was evaluated by a panel of five rheumatologists
between June and December, 1999. The respective diagnostic values
of antiperinuclear factor, antikeratin antibody, and anti-CCP (commercial
kits: Eurodiagnostica first generation [EFG], Eurodiagnostica second
generation [ESG], and Axis-Shield [AS]) carried out on sera taken at the
patients’ first visit in discriminating between patients who did (98/270;
36%) and did not have RA at the last visit was evaluated using receiv-
ing-operator characteristic curves. To evaluate the combination of anti-
CCP with other laboratory tests in discriminating between patients with
and without RA, a multiple logistic regression with backward selection
using the likelihood ratio test was applied.
Results: 1) Anti-CCP, APF and IgG AKA were not perfectly correlated
with each other. For anti-CCP EFG (cutoff 53UI), ESG (cutoff
0.120OD), and AS (cutoff 0.320OD), sensitivity and specificity were
47%–93%, 57%–94% and 58%–94%, respectively. 2) By the use of
a multiple logistic regression, second generation anti-CCP, IgG-AKA,
the latex test and IgM-RF ELISA were selected.
Conclusion: ESG and AS are the best tests for predicting RA. Com-
bining one of these tests with IgG-AKA, IgM-RF, and the latex test
slightly increases the diagnostic value.
25
Autoantibodies to hnRNP-A2 in SLE: identification of
disease-specific linear epitopes and correlation with
disease activity and clinical features
G Schett1, F Monneaux2, E Hoefler3, R Fritsch1, M Tohidast-Akrad3,
J Smolen1,3, S Muller2, G Steiner1,3
1Division of Rheumatology, University of Vienna, Austria
2Institut de Biologie Moléculaire et Cellulaire, CNRS, Strasbourg, France
3Ludwig-Boltzmann-Institute for Rheumatology, Vienna, Austria
Arthritis Res Ther 2003, 5 (suppl 1):25
The autoantigen hnRNP-A2 (RA33) is targeted by autoantibodies
(autoAbs) of patients with rheumatoid arthritis (RA), systemic lupus
erythematosus (SLE) and mixed connective tissue disease (MCTD).
To define the humoral autoimmune response in more detail, a series of
overlapping peptides covering the N-terminal part hnRNP-A2 known
to harbour the major epitopes was studied by ELISA in sera from
patients with SLE (n=40), RA (n=50), MCTD (n=11) and other
rheumatic diseases (n=86) and from healthy subjects (n=29). Anti-
peptide reactivities were detected in 25% of SLE sera but only rarely
in patients with RA or MCTD. Since most of the SLE patients investi-
gated had inactive disease, we next studied sequential sera of 15 indi-
vidual patients. Of those, only three patients were completely negative
for hnRNP-A2 antibodies, while all the other patients showed at least
one positive reaction during the observation period. These were
directed to the full-length protein and/or to 4 of the 13 peptides used:
p35–55, p50–70, p90–116 and p155–175. Interestingly, autoreactiv-
ities to the first three peptides were significantly associated with each
other but not with reactivity to p155–175 or to the complete protein.
This cluster of reactivity was also not linked to any clinical marker of
disease. In contrast, p155–175 was strongly correlated with reactivity
to the complete protein (P<0.001) and both reactivities were associ-
ated with autoAb to dsDNA and correlated significantly with clinical
disease activity, skin involvement and proteinuria (P<0.01). Remark-
ably, immunohistochemical analysis revealed overexpression of
hnRNP-A2 in affected skin of SLE patients. These data, together with
previously published findings in murine SLE models, are suggestive of
an involvement of hnRNP-A2 autoimmunity in the pathogenesis of
SLE.
26
Fcγ γ receptors complement interaction: new aspects in
pathogenesis and treatment of vasculititis
RE Schmidt
Clinical Immunology, Hannover Medical School, Hannover, Germany
Arthritis Res Ther 2003, 5 (suppl 1):26
For a long time, detailed mechanisms of immune-complex-mediated
vasculitis have been unknown. The advent of gene knockout technol-
ogy and the characterization of the various Fc and complement recep-
tors as well as cytokines now allows the various pathogenetic elements
in vasculitic inflammation to be distinguished.
Using various murine knockout models, in particular Fcγ-receptor-defi-
cient animals, mast-cell-defective animals, and complement-deficient
animals in recent years, we have shown that FcγRIII and C5a receptor
are critical for induction of immune-complex-mediated vasculitis. In
several studies, it became clear that mast cells have a critical role in the
initiation of the inflammatory process. On these mast cells, again the
FcγRIII is a critical activating receptor used by immune complexes.
When examining the effects of immune complexes in glomerular
mesangial cells and GBM nephritis, we showed that IgG immune com-
plexes had opposing regulatory effects on FcγRII and FcγRIII receptors
in glomerular mesangial cells. Whereas activation by TNF-α/IL-1β
induces substantial FcγRII expression, IFN-γ showed a complete down-
regulation of FcγRII. At the same time, IFN-γ induced the Fc receptor
γ-chain as well as the low-affinity IgG receptor FcγRIII. Triggering of
FcγRIII again induced chemoattractant protein 1, MCP-1, MCP-5 and
RANTES.
Examining the regulatory role in the cooperation of Fcγ receptors and
complement, we showed that C5a is critical in amplifying the inflam-
matory response to IgG. C5a is important on the one hand in down-
regulating FcγRII, and on the other hand in inducing the activating
FcγRIII.
Altogether, distinguishing the various components of vasculitis patho-
genesis allows for new strategies to intervene in this inflammatory
process.
Arthritis Research & Therapy Vol 5 Suppl 1Abstracts of the 23rd European Workshop for Rheumatology Research

Page 9

S9
27
Linear epitopes of two different autoantigens (La/SSB
and myelin basic protein) with a high degree of
molecular similarity cause different humoral responses
A Terzoglou, JG Routsias, C Sakarellos, M Sakarellos-
Daitsiotis, HM Moutsopoulos, AG Tzioufas
Department of Pathophysiology, School of Medicine, University of
Athens, Athens, Greece
Arthritis Res Ther 2003, 5 (suppl 1):27
Backgkround: Sequences 147–154aa of La/SSB and 139–146aa of
human myelin basic protein (MBP) present 83% sequence similarity.
Objective: We investigated the immune response of both epitopes in
rabbits and in sera from patients with autoimmune diseases.
Methods: Peptides 147–154aa of La/SSB and 139–146aa of MBP
were used for immunizations of New Zealand White rabbits. Spreading
to the other epitopes of La/SSB (289–308aa, 349–364aa) as well as
the recombinant human MBP (hMBP) and La/SSB (recLa) was identi-
fied using ELISA assays. Sera from 49 patients with systemic lupus
erythematosus (SLE), 44 patients with Sjögren’s syndrome (pSS),
and 18 with rheumatoid arthritis (RA) with anti-Ro\La reactivity were
tested against the two peptides and the hMBP.
Results: Rabbits immunized with the La epitope developed early anti-
bodies against all three La/SSB peptides, hMBP, and recLa. In con-
trast, rabbits immunized with the MBP peptide developed a late
immune response to other La epitopes, hMBP, and recLa. Inhibition
experiments using the MBP peptide as inhibitor against the hMBP
showed that the 79% of reactivity was abolished, indicating that this
peptide is the major antibody target in MBP. Twenty percent of pSS,
27% of SLE, and none from RA patients reacted with the 147–154aa
La epitope; 28% of pSS, 22% of SLE, and 17% of RA sera reacted
with the MBP peptide. Finally, 17% of pSS, 37% of SLE, and 30% of
RA sera reacted with the hMBP.
Conclusion: La 147–154aa peptide when used for animal immuniza-
tions can induce immediate epitope spreading while the mimicking
epitope MBP 139–146aa induces a delayed response against the other
La epitopes. A significant proportion of human sera reacted with both
peptides and hMBP. Thus, despite the fact that these two peptides
present molecular similarity, they induce different immune responses.
28
Autoantibodies predict progression to rheumatoid
arthritis in undifferentiated arthritis: a prospective
cohort study
FA van Gaalen, SP Linn-Rasker, WJ van Venrooij, BA de Jong,
FC Breedveld, CL Verweij, REM Toes, TWJ Huizinga
Department of Rheumatology, Leiden University Medical Center,
Leiden, The Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):28
Background: Early intervention in rheumatoid arthritis (RA) reduces
long-term disability. However, early diagnosis of RA can be difficult, as
the disease may initially be indistinguishable from other forms of arthri-
tis. Recent studies indicate that autoantibodies can be detected years
before clinical symptoms develop. In a cohort with recent-onset arthri-
tis, we aimed to assess the value of autoantibodies in predicting the
development of RA in patients with undifferentiated arthritis (UA).
Methods: IgM rheumatoid factors (IgM-RF) and anti-cyclic citrullinated
peptide (CCP) antibody tests were performed at baseline in 936 con-
secutive, newly referred patients with recent-onset arthritis. Two weeks
after inclusion arthritis, patients who could not be properly classified
were categorized as UA. Patients with UA were followed for 3 years
and evaluated for progression to RA.
Results: At 2 weeks, 346 of 936 patients (37%) with recent-onset
arthritis were classified as having UA. After 3 years of follow-up, 127
UA patients (40%) had progressed to RA. However, RA had devel-
oped in 64 of 69 UA patients with a positive anti-CCP test, giving a
positive predictive value (PPV) of 93% and a negative predictive value
(NPV) of 74%. Progression to RA was observed in 51 of 68 UA
patients with IgM-RF antibodies at baseline (PPV 75%, NPV 70%).
Conclusion: Up to 3 years before a diagnosis of RA was made,
autoantibodies were detected in patients with UA. Detection of anti-
CCP antibodies had a high predictive value for the progression to RA.
Thus, screening for anti-CCP antibodies in UA allows physicians to
predict progression to RA.
29
Autoantibodies to GPI are predominantly present in
extra-articular complications of human rheumatoid
arthritis
FA van Gaalen, REM Toes, HJ Ditzel, M Schaller, CL Verweij,
TWJ Huizinga
Department of Rheumatology, Leiden University Medical Center,
Leiden, The Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):29
Several years ago, Benoist et al. described in Cell a T-cell transgenic
mouse that spontaneously developed severe arthritis (KRN model).
Subsequent reports demonstrated that in this model, disease is caused
by antibodies against glucose-6-phosphate isomerase (GPI).
A link between this mouse model and human disease was made when
Schaller et al. reported that antibodies against GPI are present in 64%
of human rheumatoid arthritis (RA) (Nat Immunol 2001). However, this
finding remains controversial, since several other groups could not
reproduce these results. Given these apparently conflicting findings,
we hypothesized that GPI antibodies are present in a specific subset of
RA patients and set out to determine at what point autoantibodies to
GPI occur in RA. GPI antibodies were detected in only few sera of of
uncomplicated RA (2%)and healthy controls (3%). However, in RA
patients with disease manifestation outside their joints, GPI antibodies
were more common. 18% of RA patients with skin inflammation
(rheumatoid nodules) and 45% of RA patients with vascular inflamma-
tion had anti-GPI antibodies. Yet, in RA patients with arthritis compli-
cated by decreased numbers of circulating granulocytes (Felty’s
syndrome) 92% had GPI antibodies. But, since the GPI antigen was
not detected on RA granulocytes, we propose that GPI antibodies are
not likely to be directly involved in granulocyte destruction.
In summary, we conclude that anti-GPI antibodies that cause disease
in the KRN mouse model are common in human RA complicated by
inflammation outside the joint, demonstrating the relevance of the
mouse model to human disease.
30
Citrullination of synovial proteins in murine models
of rheumatoid arthritis
E Vossenaar, S Nijenhuis, MM van Helsen, A van der Heijden,
WB van den Berg, WJ van Venrooij, L Joosten
Department of Biochemistry, University of Nijmegen, Nijmegen, The
Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):30
Antibodies directed to citrulline-containing proteins are highly specific
for rheumatoid arthritis (RA) and can be detected in up to 80% of RA
patients. Citrulline is an unnatural amino acid that can be incorporated
into proteins only by post-translational modification of arginine by pep-
tidylarginine deiminase (PAD) enzymes. We investigated the presence
of anticitrulline antibodies, PAD enzymes and citrullinated antigens in
an acute and a chronic destructive mouse model for arthritis: strepto-
coccal-cell-wall arthritis and collagen-induced arthritis. In both mouse
models, PAD2 mRNA is present in the synovium but not translated into
PAD2 protein. In contrast, PAD4 mRNA, although absent from healthy
synovia, is readily transcribed and translated by polymorphonuclear
neutrophils infiltrating the synovial tissue during inflammation. As a con-
sequence, several synovial proteins are subjected to citrullination. One
of these proteins was identified as fibrin, which has been reported to
Available online http://arthritis-research.com/supplements/5/S1

Page 10

S10
be citrullinated also in the synovia of RA patients. Although the genera-
tion of citrullinated antigens during synovial inflammation in the mice
was eminent, no anticitrullinated protein antibodies could be detected.
In conclusion, the citrullination of synovial antigens is an active process
during joint inflammation both in mouse and man, but the induction of
autoantibodies directed against these proteins is a more specific phe-
nomenon detectable only in RA patients.
31
Fibrinogen-specific T cells in rheumatoid arthritis
E Vossenaar1, R Bergholz2, F Schumann2, GR Burmester2,
JM Engel3, WJ van Venrooij1, S Bläß2
1Department of Biochemistry,University of Nijmegen, The Netherlands
2Department of Rheumatology & Clinical Immunology, Charité
University Clinic, Berlin, Germany
3Rheumaklinik, Bad Liebenwerda, Germany
Arthritis Res Ther 2003, 5 (suppl 1):31
Rheumatoid arthritis (RA) is characterized by the occurence of autoreactive
antibodies and Tcells. The family of antibodies directed to citrulline-contain-
ing antigens (anti-filaggrin, anti-CCP and anti-Sa) have the highest speci-
ficity (>98%) for RA. To investigate the presence of citrulline-specific
Tcells in RA patients, we analyzed T-cell reactivity to unmodified and citrul-
linated filaggrin and fibrinogen in modified T-cell proliferation assays.
No T-cell responses were observed when either unmodified or citrulli-
nated filaggrin was used as the stimulating antigen in serum-free T-cell
proliferation assays. With unmodified fibrinogen, however, T-cell prolifer-
ation was observed in 8/15 (53%) of RA patients, while only 1/14
control patients (SLE, SSc, PsoA, Sjö) displayed clearly positive T-cell
proliferation. The difference was highly significant (P<0.0001). These
findings were even more pronounced when using citrullinated fibrinogen
as stimulating antigen: 10/15 (67%) RA patients were positive for T-cell
proliferation and again only 1/14 control patients was clearly positive
(P<0.0001). T-cell reactivity in RA patients was significantly (P<0.05)
higher against citrullinated as compared with unmodified fibrinogen.
We conclude that fibrinogen (citrullinated or not) can induce proliferation
of RA T cells, thereby further substantiating its pathogenic relevance.
32
Elevated anti-serum amyloid P component (SAP)
antibodies in SLE patients
G Zandman-Goddard, M Blank, P Langevitz, M Pras, Y Levy,
T Witte, A Doria, J Rovensky, Y Shoenfeld
Center for Autoimmune Diseases, Sheba Medical Center,
Tel Hashomer, Israel
Arthritis Res Ther 2003, 5 (suppl 1):32
Background: Serum amyloid P component (SAP) binds to DNA and
chromatin and plays a role in the clearance of apoptotic debris. One
postulated mechanism in the dysregulation of the immune system in
systemic lupus erythematosus (SLE) is the aberrant clearance of apop-
totic cells. We hypothesize that binding of SAP to anti-SAP antibody
may alter SAP function.
Objective: The aim of the study was to determine the presence of anti-
SAP antibodies in SLE patients and investigate the correlation with
clinical disease.
Methods: Samples from 481 subjects (357 SLE patients, 124 normal
controls) were screened for the presence of elevated anti-SAP anti-
body titers by the ELISA method (optical density 405nm above 3SD
were considered elevated). Clinical parameters and SLEDAI scores
were assessed from the review of files.
Results: Elevated anti-SAP antibody titers were detected in 47% of
SLE samples, versus 2% of the control group. In a representative
group (n=112), 62% of patients had elevated anti-SAP antibodies and
anti-dsDNA antibody titers. SLEDAI scores were assessed in 83
patients, of whom 54% had elevated anti-SAP antibody titers: 59%
had a SLEDAI score of at least 8, an indication of severe disease. The
distribution of 35 clinical manifestations in 34/135 patients with ele-
vated anti-SAP antibody titers did not reveal a specific pattern. Serial
sampling of two representative SLE patients revealed a decrease in
anti-SAP antibody titers after treatment with IVIg that correlated with a
decrease in anti-dsDNA antibody titers and with clinical improvement.
Conclusion: Elevated anti-SAP antibody titers were detected in SLE
patients and correlated with disease activity. In SLE patients, elevated
anti-SAP antibody titers may serve as an additional diagnostic and
prognostic marker.
Cytokines
33
Clinical and immunological effects of anti-TNF
therapy in systemic lupus erythematosus (SLE)
M Aringer, G Steiner, WB Graninger, E Höfler, H Hiesberger,
C-W Steiner, J Smolen
Rheumatology, Internal Medicine III, University of Vienna, Austria
Arthritis Res Ther 2003, 5 (suppl 1):33
Background: Tumor necrosis factor (TNF) is increased in the sera of
patients with SLE and in lupus glomerulonephritis and is associated
with disease activity. We investigated clinical and immunological out-
comes in a pilot trial of TNF blockade in SLE.
Methods: Within an open safety study, SLE patients with nephritis or
arthritis receive the humanized chimeric anti-TNF antibody infliximab plus
azathioprine or methotrexate. Serum TNF (sTNF)was measured by
ELISA, the percentages of TNF-positive lymphocytes by fluorocytometry.
Results: In the first two lupus nephritis patients treated with infliximab,
proteinuria fell from 1.2 to 0.3g/24h and from 5.7 to 1.1g/24h,
respectively, within 3 months after the start of therapy; the (normal) cre-
atinine serum levels remained stable. Arthritis in a third SLE patient
remitted under therapy but relapsed 8 weeks after the last infusion.
Anti-dsDNA IgG antibodies increased transiently in all three patients at
around week 10 of therapy. An increase in anti-histone antibodies (as
well as anti-chromatin antibodies in two of the three patients) predated
the increase in anti-dsDNA, which was not associated with increased
disease activity. Interestingly, even the increase in anti-histone antibod-
ies was predated by an increase in sTNF (mean±SD of peak value
168±114pg/ml), while the percentage of lymphocytes carrying TNF
decreased at the same time.
Conclusion: Anti-TNF therapy improves SLE glomerulonephritis and
arthritis but leads to a transient increase in autoantibodies, which was
not associated with flares in our patients. The observation that anti-
dsDNA antibodies are predated by anti-histone antibodies and that this
increase follows the changes in TNF suggests that TNF-blocking
therapy is directly associated with an increase in anti-histone and a
subsequent transient increase in anti-dsDNA antibodies.
34
Therapy with soluble TNF receptor (etanercept) induces
apoptosis in rheumatoid arthritis (RA) synovium
AI Catrina, C Trollmo, J Lampa, E af Klint, Y Hermansson,
L Klareskog, A-K Ulfgren
Rheumatology Unit, Department of Medicine, Karolinska
Hospital/Institutet, Stockholm, Sweden
Arthritis Res Ther 2003, 5 (suppl 1):34
Objectives: This study evaluates modulation of rheumatoid arthritis
(RA) synovial apoptosis by therapy with the soluble tumor necrosis
factor (TNF) alpha receptor (etanercept).
Methods: Apoptosis (TUNEL [transferase-mediated UTP end labeling]
combined with morphology), cell-surface markers (CD3, CD68,
CD163, Fas), FLIP and granzyme B were evaluated by immunohisto-
chemistry in synovial biopsies from 12 RA patients before and after
8 weeks of treatment with etanercept. Moreover, the in vitro effect of
etanercept on FLIP expression and on the death of mononuclear cells
(MNCs) derived from synovial fluid (SF) was determined in five RA
patients by flow cytometry.
Arthritis Research & Therapy Vol 5 Suppl 1Abstracts of the 23rd European Workshop for Rheumatology Research

Page 11

S11
Results: Etanercept treatment increased synovial apoptosis and
decreased the number of CD68-positive and CD163-positive mono-
cyte/macrophages and FLIP expression (P<0.05). No significant
changes were observed for the expression of CD3, Fas and
granzyme B. In vitro, low concentrations of etanercept (1 and 10µg/ml)
increased cell death (P<0.05) in the SF CD14-positive monocyte/
macrophage population after 24 hours’ incubation, while higher con-
centrations (100µg/ml) had no effect. FLIP expression in this popula-
tion did not change after in vitro culture with either low or high doses.
The in vitro culture with etanercept did not induce any changes in FLIP
expression or apoptosis level in SF CD3-positive cells.
Conclusion: Therapy with etanercept at clinically relevant concentra-
tions increased RA synovial monocyte/macrophage apoptosis, sug-
gesting an alternate pathway to explain the decrease in synovial
cellularity observed after anti-TNF therapy.
35
The effect of methotrexate and mycophenolic acid on
monokine production in vitro
S de Lathouder1, AH Gerards2, ER de Groot1, LA Aarden1
1Immunopathology, Sanquin Research at CLB, Amsterdam,
The Netherlands
2Rheumatology, VU medical center, Amsterdam, The Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):35
Methotrexate (MTX) and mycophenolic acid (MPA) are used clinically
for their immunosuppressive properties. MTX is widely used for the
treatment of RA. MPA is used to prevent graft rejection and is now
experimentally used in SLE and RA. The precise mechanism of action
is still debated. Both drugs, though in different ways, inhibit the
de novo synthesis of DNA and RNA. We have analysed cytokine pro-
duction in short cell cultures in whole blood and isolated cells by
ELISA. We have shown before that both drugs inhibit the production of
several cytokines after T-cell stimulation, and we concluded that MTX
leads to irreversible elimination of activated T cells by apoptosis,
whereas MPA reversibly prevents activation of resting T cells.
We now show that when monocytes are stimulated by SAC or
lipopolysaccharide, both MTX and MPA decrease TNF-α, IL-6 and IL-8
production. However the inhibition is not as profound as after T-cell
stimulation. An exception is the effect on the production of the proin-
flammatory cytokine IL-1β. The production of IL-1β is not influenced by
MTX after SAC or LPS stimulation, whereas the production is increased
by MPA when cells are stimulated with LPS, but not with SAC. We have
investigated the cause for this increase. The expression of IL-1β mRNA
is not influenced by MPA. Immunoprecipitation and ELISA show that
MPA leads to a decrease in the intracellular pro- form of IL-1β. We con-
clude that MPA leads to enhanced cleavage of pro-IL-1β.
36
Changes in rheumatoid factor reflect the inflammatory
response (CRP and ESR) to infliximab treatment
L De Rycke, E Kruithof, N Van Damme, IEA Hoffman, F Van den
Bosch, EM Veys, F De Keyser
Department of Rheumatology, Ghent University Hospital, Ghent, Belgium
Arthritis Res Ther 2003, 5 (suppl 1):36
Background: We showed previously that there is a reduction in both
rheumatoid factor (RF) Waaler Rose (WR) and RF latex fixation (LF)
titres after infliximab treatment, but that no differences are seen for anti-
cyclic citrullinated peptide antibodies.
Objectives: To analyze changes in C-reactive protein (CRP) and ery-
throcyte sedimentation rate (ESR) in relation to changes in RF.
Patients and methods: Sixty-two patients with refractory rheumatoid
artritis (RA) were treated with infliximab in an early-access program.
They received 3mg/kg infliximab IV at weeks 0, 2 and 6 and every
8 weeks thereafter in combination with methotrexate. Serum samples
were obtained at baseline and at week 30 and tested for RF WR, RF
LF, CRP and ESR. For statistical analysis, Mann–Whitney tests were
used. P-values ≤0.05 were considered significant.
Results: Patients with an increase in RF WR titre (n=9) were likely to
have an increase in CRP (median CRP increase of 0.59mg/dl) and
ESR (median ESR increase of 9mm/h), whereas patients with no
increase in RF WR titre (n=53) were more likely to have a decrease in
CRP (median CRP decrease of 0.89mg/dl) and ESR (median ESR
decrease of 5.5mm/h) during infliximab treatment. The changes in CRP
and ESR were significantly different when comparing patients with an
increase in RF WR titre and patients with no increase in RF WR titre
(P=0.046 for CRP and P=0.032 for ESR). When additionally also the
RF LF titre increased in the group of patients with an increase in RF
WR titre, all patients (n=3) had an increase in CRP (median CRP
increase of 2.3mg/dl) and ESR (median ESR increase of 16mm/h).
Conclusion: After infliximab treatment, RA patients with an increase in
RF WR titre are more likely to have increased inflammatory parameters,
in comparison with patients who had no increase in RF WR. Especially
when both RF titres increased, an increase of CRP and ESR was
present in all patients.
37
Promoter polymorphisms in the IL-18 gene are
associated with rheumatoid arthritis in two
independent clinical cohorts
JA Gracie1, N Koyama2, M Field1, F McGarry1, A Schobel2,
IB McInnes1, B Moller2
1Centre for Rheumatic Diseases, University of Glasgow, Glasgow, UK
2Rheumatology, University of Frankfurt, Frankfurt, Germany
Arthritis Res Ther 2003, 5 (suppl 1):37
Introduction: IL-18 in synovial tissues of patients with rheumatoid
arthritis (RA) promotes local inflammation via effects on innate and
adaptive immune responses. Promoter polymorphisms may modulate
IL-18 expression. Using two independent clinical cohorts, we deter-
mined the frequency of single nucleotide polymorphisms (SNPs) in the
IL-18 promoter for RA patients.
Methods: DNA was extracted from peripheral blood mononuclear cells
of 102 RA patients and 114 healthy donors in Frankfurt and from 153
RA patients and 187 healthy donors in Glasgow. In Frankfurt, DNA was
amplified by polymerase chain reaction (PCR) and the presence of
IL-18 SNP at positions -607 and -137 was determined by the restric-
tion fragment length technique. Independently, in Glasgow, allele-spe-
cific PCR was performed for the same SNP sites.
Results: Frankfurt cohort: The -607 C and -137 G alleles were signifi-
cantly more frequent in the RA population. Highly significant associa-
tions (P<0.001) for RA were found for the homozygous C genotype in
positions -607 and -137. Glasgow cohort: The -137 C allele was more
frequent in RA patients (P<0.01), but no effect was seen at position -
607. Moreover, the -137CC genotype was more frequent in RA
(P<0.01). However in contrast to the Frankfurt dataset, no effect was
seen at the -607 site.
Conclusion: SNPs at the -137 position in the IL-18 promoter appear to
contribute to the genetic background in RA pathogenesis. Importantly,
this has been independently identified in two clinical cohorts, by the
use of distinct methodologies, in Germany and the UK. IL-18 is a
promising therapeutic target and, as such, defining factors that modu-
late its regulation in rheumatoid tissues is important.
38
Combination of the proinflammatory cytokines IL-1,
TNF-α α and IL-17 leads to enhanced expression and
additional recruitment of AP-1 family members, Egr-1
and NF-κ κB in osteoblast-like cells
C Granet, P Miossec
INSERM U403, Department of Immunology and Rheumatology,
Hôpital E Hérriot, Lyon, France
Arthritis Res Ther 2003, 5 (suppl 1):38
Objectives: To determine the contribution of IL-1, TNF-α and IL-17 on
AP-1, NF-κB and Egr-1 activation in cytokine-induced bone destruction
Available online http://arthritis-research.com/supplements/5/S1

Page 12

S12
as in rheumatoid arthritis, we investigated the effect of these proinflam-
matory cytokines on transcription factor activation in osteoblasts.
Methods: Osteoblast-like ROS 17/2.8 cells were cultured with IL-1,
TNF-α and IL-17 alone and in combination. Effects of each cytokine
were explored by RT-PCR and immunocytochemistry.
Results: IL-1 and TNF-α induced most of these transcription factors
while IL-17 had a weak effect. IL-1 induced egr-1 and all AP-1 member
expression, except fosB and junD. TNF-α also induced AP-1 member
expression, with a longer expression for fra-1 and fra-2. These two
cytokines individually induced nuclear translocation at 30 min, except
for JunB. IL-17 enhanced fra-2 and egr-1 mRNA between T30-T120
with a peak at T90, while nuclear localisation was noticed at T30 for
Fra-1, JunD and NF-κB. More importantly, when these cytokines were
used at low concentrations with no effect when used alone, their com-
binations showed a synergistic effect on transcription and nuclear
translocation of AP-1 members, Egr-1 and NF-κB. Moreover, cytokine
combinations allowed an enhanced recruitment of factors not express
by cytokines used alone.
Conclusion: AP-1, Egr-1 and NF-κB pathways in osteoblast cells are
very sensitive to the combined effect of proinflammatory cytokines
through additive or synergistic mechanisms.
39
Expression of chemokines and their receptors in
synovial tissue of patients with rheumatoid arthritis,
osteoarthritis and reactive arthritis
JJ Haringman, TJM Smeets, MC Kraan, PP Tak
Clinical Immunology and Rheumatology, Academic Medical
Center/University of Amsterdam, Amsterdam, The Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):39
Background: Chemokine receptors and their ligands play a crucial role
in the recruitment of leukocyte subsets into inflamed tissue. Their exact
expression in synovial tissue (ST) of patients with various forms of
arthritis has yet to be determined.
Objective: The objective of this study was to determine the expression
of an extensive number of chemokines and their receptors in patients
with rheumatoid arthritis (RA) and other forms of arthritis.
Methods: Synovial biopsies were obtained with a Parker–Pearson
needle from patients with RA (n=23), osteoarthritis (OA) (n=16), and
reactive arthritis (ReA) (n=8). ST was studied in patients with both
early and late stages of disease. Sections were analyzed by immuno-
histochemistry using monoclonal antibodies against CD3 (T-cells),
CD68 (macrophages), CD13 (aminopeptidase N), CCR1, CCR2b,
CCR5, CXCR4, CCL2 (MCP-1), CCL8 (MCP-2), CCL7 (MCP-3),
CCL14 (hCC-1), CCL15 (hCC-2), CCL16 (hCC-4), and CCL5
(RANTES). Digital image analysis was used to quantify the staining and
for statistical analysis the Kruskal–Wallis H test and the Mann–Whitney
U test were used.
Results: All chemokines and chemokine receptors were detected in
inflamed synovium. There was abundant expression of especially
CD13, CCR1, CXCR4, CCR5, CCL7, and CCL8 in all forms of arthri-
tis. The ligands CCL7, CCL8, CCL14, CCL15, and CCL16, which
have not previously been described in ST, were all expressed in
inflamed ST. The expression of CCL5 (RA, 11182±2239 (mean inte-
grated optical density±SEM); OA, 8759±2930; ReA, 2081±1181;
P=0.010) and CCL15 (RA, 4993±1601 (mean integrated optical
density±SEM); OA, 1738±573; ReA, 1252±1919; P=0.043) was
significantly higher in rheumatoid ST compared with disease controls.
Conclusion: These results suggest a potentially important role for a
variety of chemokines and their receptors in the migration of inflamma-
tory cells towards the synovial compartment. Disruption of the
chemokine network might represent a novel therapeutic approach in
RA, but also in other arthritides.
40
Osteoprotegerin protects from generalized bone loss
in TNF-transgenic mice
S Hayer1, K Redlich1, J Zwerina1, B Bolon2, C Dunstan2, B Görtz1,
H Bergmeister3, G Kollias4, G Steiner1, J Smolen1, G Schett1
1Division of Rheumatology, Department of Internal Medicine III,
University of Vienna, Vienna, Austria
2Department of Pathology, Amgen, Inc., Thousand Oaks, CA, USA
3Center for Biomedical Research, University of Vienna, Vienna, Austria
4Institute of Immunology, Alexander Fleming Biomedical Sciences
Research Center, Vari, Greece
Arthritis Res Ther 2003, 5 (suppl 1):40
Background: Chronic inflammatory conditions, such as rheumatoid
arthritis, are characterized by generalized loss of bone mass. Proinflam-
matory cytokines, such as TNF, are believed to play a central role in this
process by increasing bone resorption.
Objective: We have investigated systemic bone changes in human
TNF-transgenic (hTNFtg) mice, which spontaneously develop severe
inflammatory arthritis.
Results: Osteodensitometry revealed a significant decrease of trabec-
ular bone mineral density (BMD) (–37%) in hTNFtg mice, and histomor-
phometry revealed a dramatic loss of bone volume (–85%) in
comparison with wild-type controls. Osteoclast-covered bone surface
and serum levels of deoxypyridinolin crosslinks were significantly ele-
vated, suggesting increased osteoclast-mediated bone resorption in
hTNFtg mice. Osteoprotegerin (OPG) completely blocked TNF-medi-
ated bone loss by increasing BMD (+89%) and bone volume (+647%).
Most strikingly, formation of primary spongiosa was dramatically
increased (+563%) in hTNFtg mice after OPG treatment. Osteoclast-
covered bone surface and serum levels of deoxypyridinolin crosslinks
were significantly decreased by OPG, suggesting effective blockade of
osteoclast-mediated bone resorption. OPG did not influence levels of
hTNF, TNF-receptor-1, IL-1β and IL-6. However, OPG decreased bone
formation parameters, which were elevated in hTNFtg mice. In contrast
to OPG, bisphosphonates and anti-TNF treatment did not affect gener-
alized bone loss in hTNFtg mice.
Conclusion: These data indicate that TNF-mediated generalized bone
loss is primarily dependent on RANKL/RANK signaling and can be
blocked by OPG. Thus, OPG may represent a potent tool to prevent
generalized loss of bone mass in chronic inflammatory disorders, espe-
cially rheumatoid arthritis.
41
Influence of anti-TNF therapy on monocyte gene
expression in rheumatoid arthritis
M Hernandez, L Martinez, O Kiesslich, N Tandon, M Janitz,
H Lehrach, F Wagner, T Häupl, GR Burmester, B Stuhlmüller
Department of Rheumatology, Charité, Berlin, Germany
Arthritis Res Ther 2003, 5 (suppl 1):41
Background and objective: The joint in rheumatoid arthritis (RA) is
characterized by pannus formation and cartilage/bone destruction. In
RA, not only tissue macrophages but also blood monocytes (MO) are
known to be activated and spontaneously release inflammatory media-
tors. However, genes and functional pathways involved in RA mono-
cyte activation are recognized and characterized only in part.
Of all techniques for differential mRNA expression analysis, DNA arrays
raise most expectations and are applied for drug discovery and compound
testing in preclinical or clinical studies. This technique may also enable the
identification of genes and expression patterns associated with the type of
disease, stage of activation, disease progression and molecular therapeu-
tic mechanisms. For functional interpretation, the experimental concept is
important. Considering drawbacks and benefits, analysis of defined cell
populations is advantageous for functional interpretation.
Methods: Therefore, we analyzed the expression profile of untouched
negatively separated MO using the Affymetrix U133A/B array system.
MO were obtained from normal donors, patients with active RA before
Arthritis Research & Therapy Vol 5 Suppl 1Abstracts of the 23rd European Workshop for Rheumatology Research

Page 13

S13
and after several months of anti-TNF antibody therapy. Bioinformatic
analysis of gene expression included background correction, normaliza-
tion, comparative statistics and cluster analysis.
Results: So far, six characteristic clusters have been obtained, includ-
ing genes relevant to RA and induced or decreased upon therapy. Fur-
thermore, RA regulated genes were identified, which returned to
‘normal’ upon anti-TNF treatment, indicating selective molecular effects
of the drug.
Conclusion: These promising results could open new avenues for the
identification of novel genes for a general definition of MO/macrophage
activation patterns and for pathways activated or blocked during differ-
ent stages of the disease or therapy-induced remission in RA.
42
Endogenous IL-12p40 is crucial for antigen-induced
arthritis and is partly involved in chronic relapsing
streptococcal-cell-wall arthritis
M Jacobs, L Joosten, MMA Helsen, E Lubberts, WB van den
Berg
Rheumatology Research Laboratory, University Medical Center
Nijmegen, Nijmegen, The Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):42
Background: Interleukin 12 (IL-12) is a proinflammatory cytokine with
important immunoregulatory activities and is critical in determining the
differentiation and generation of Th1 cells.
Objectives: For the present study, we investigated the role of endoge-
nous IL-12p40, component of both IL-12 and IL-23, in the pathogenesis
of antigen- and chronic relapsing streptococcal-cell-wall-(SCW-)
induced arthritis. To this end we used C57Bl6 (wild-type; WT) and
IL-12p40-deficient (IL-12 ko) mice. Chronic SCW was induced by intra-
articular (i.a.) injection of 25µg SCW fragments at days 0, 7, 14 and 21.
Joint swelling and histology were examined at day 28 after the first injec-
tion. Antigen-induced arthritis (AIA) was initiated by i.a. injection of
60µg mBSA in preimmunized mice. Joint swelling was measured at
days 3, 7 and 14, whereas histology was examined at days 7 and 14.
Results: At day 28, joint swelling of the mainly macrophage-driven
chronic SCW arthritis was significantly reduced in IL-12p40 ko mice in
comparison with WT mice (P=0.0008). Although cell influx was
decreased (P=0.0194), no reduction in cartilage proteoglycan deple-
tion was found. In a predominantly T-cell-mediated process such as
AIA, joint swelling was significantly suppressed at days 3 and 7 in
IL-12p40 ko mice in comparison with WT (P=0.0021 and P=0.0260
respectively). Moreover, histopathology was drastically reduced in
IL-12p40 ko mice; the number of inflammatory cells was strongly sup-
pressed in IL-12p40 ko mice in comparison with WT animals (1.6±0.4
vs 0.2±0.2 and P<0.0001). In line with these findings matrix proteo-
glycan depletion was completely absent in IL-12p40 ko mice
(1.9±0.35 vs 0.05±0.11 and P<0.0001).
Conclusion: These results indicate that IL-12p40 plays a pivotal role in
antigen-induced arthritis and a minor role in chronic relapsing strepto-
coccal-cell-wall-induced arthritis.
43
IL-18 directly promotes joint inflammation and
induces cartilage destruction through IL-1
L Joosten, E Lubberts, MJM Wagenmans, MMA Helsen,
FAJ van de Loo, WB van den Berg
Rheumatology Research Laboratory, University Medical Center
Nijmegen, Nijmegen, Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):43
Background: Interleukin-18 is a member of the IL-1 family of proteins
that exerts proinflammatory effects and induces cartilage destruction
in vitro.
Objective: The goal of the present study was to investigate whether
IL-18 mediates joint destruction in vivo, directly or via induction of other
cytokines.
Results: To this end, we performed both in vitro and in vivo kinetic
studies. 35S-labeled explants of C57Bl/6 mice were cultured for 24 to
72 hours in either IGF-1, IGF-1/IL-18 or IGF-1/IL-1β in combination
with or without IL-1 receptor antagonist (IL-1Ra) or an ICE inhibitor.
Wild-type, TNFα-deficient and IL-1α,β-deficient mice were used for
in vivo IL-18 exposure studies. Mice were injected intra-articularly with
107pfu mIL-18 adenovirus at day 0, and histopathology was examined
at days 4, 7 and 14. In vitro IL-18 exposure for 24 or 48 hours did not
induce cartilage degradation, determined as release of prelabeled pro-
teoglycans. Cartilage degradation by IL-18 was only found after a
72-hour culture period. Blocking of IL-1 with IL-1Ra or ICE-inhibitor
resulted in almost complete protection against IL-18 mediated cartilage
degradation in vitro. Application of 107pfu AdmIL-18 resulted in pro-
longed elevated levels of IL-18, for up to 14 days. Histology at
days 4, 7, and 14 revealed that local overexpression of IL-18 resulted in
increasing joint inflammation and cartilage destruction in the wild-type
mice. Of high interest, IL-18 gene transfer in IL-1α,β–/–mice did not
show cartilage damage at the various time points, although joint inflam-
mation was similar to that in the wild-type animals. Overexpression of
IL-18 in TNFα-deficient mice showed that TNFα was partly involved in
IL-18-induced joint inflammation.
Conclusion: Here we showed that IL-18 induces joint inflammation
independently of IL-1. In addition, we showed that IL-1 generation, due
to IL-18 exposure, was essential for marked cartilage degradation both
in vitro and in vivo. These findings imply that IL-18 contributes, through
separate pathways, to joint inflammation and cartilage destruction.
44
TNF polymorphisms are associated with the
development of joint erosions in psoriatic arthritis
D Kane1, J Balding2, W Livingstone2, L Mynett-Johnson2,
B Bresnihan1, O Smith2, O Fitzgerald1
1Department of Rheumatology, St Vincent’s University Hospital, Dublin,
Irish Republic
2Department of Genetics, Trinity College, Dublin, Irish Republic
Arthritis Res Ther 2003, 5 (suppl 1):44
Objective: To determine whether functional cytokine gene polymor-
phisms influence disease susceptibility and phenotype in patients with
psoriatic arthritis (PsA).
Methods: DNA was obtained from 147 PsA patients (mean age of
arthritis onset=35.2±13.4 years; oligoarticular=43 [29%], polyarticu-
lar=104 [71%]) and 389 healthy Irish blood donors. Seven functional
proinflammatory (IL-1β +3953, IL-6 -174, TNF-α -308, TNF-β +252)
and anti-inflammatory (IL-10 -1082, IL-10 -592, IL-1Ra [intron 2, 86bp
VNTR]) gene polymorphisms were detected by PCR and RFLP assays.
Results: No significant difference in genotype frequencies was
observed between the control population and the PsA population, and
no association with ACR functional class, disease classification (pol-
yarticular or oligoarticular), the presence of spinal involvement, or age
of PsA onset was observed. Age of onset of psoriasis was significantly
associated with the TNF-β (P=0.0011) and TNF-α (P=0.01) polymor-
phisms. The TNF-β B2/B2 and TNF-α -308 AA genotypes were associ-
ated with the earliest mean age of psoriasis onset. Plain radiographs of
the hands and feet were obtained in 114 patients (64 with joint ero-
sions, 48 with periostitis). The presence of joint erosions was signifi-
cantly associated with the TNF-α -308 A (P<0.0001) and TNFB1
(P=0.0009) alleles. Sequential radiographs were obtained in 52/61
patients who presented with early PsA (<2 years’ duration) with 19/52
(37%) increasing the number of joint erosions (progressors) over a
median interval of 24 months. The TNF-α -308 A allele and the
TNFB B1 allele were significantly increased in the progressor group
(P=0.014 and P=0.048 respectively).
Conclusion: The TNF-α -308 and TNFB polymorphisms are signifi-
cantly associated with the presence of joint erosions in PsA and with
the progression of joint erosions in early PsA.
Available online http://arthritis-research.com/supplements/5/S1

Page 14

S14
45
Defect of Th1 immune response of whole blood cells
from active patients with rheumatoid arthritis (RA)
M Kawashima, P Miossec
Department of Immunology and Rheumatology and INSERM U-403,
Hôpital Edouard Hérriot, Lyon, France
Arthritis Res Ther 2003, 5 (suppl 1):45
Background: Cases of severe tuberculosis have been reported in
patients with RA during anti-TNF-alpha (TNF-α) treatment. In addition
to TNF-α, IFN gamma (IFN-γ) and the cellular immunity are known to be
important to prevent the development of tuberculosis infection.
Objective and methods: We examined mRNA expression of IFN-γ,
IL-4, T-bet, GATA-3, and TNF-α by peripheral whole blood from RA
patients (n=28) and healthy controls (n=16), using real-time RT-PCR.
Results: IFN-γ, IL-4, and TNF-α expression was significantly higher in
RA than in healthy blood (P=0.0023, 0.003, and 0.0004, respec-
tively). In RA samples, T-bet expression correlated with that of IFN-γ
(r=0.738, P<0.0001) and negatively correlated with serum CRP
levels (r=–0.516, P=0.004). No correlation between T-bet and IFN-γ
expression was observed in healthy blood. Next, we separated RA
patients into two groups according to their serum CRP levels. In com-
parison with healthy or mild RA (CRP <40mg/l, n=19), active RA
blood (CRP >40mg/l, n=9) showed significantly lower ratio of
T-bet/β-actin (P=0.014, P=0.0009) or T-bet/GATA-3 (P=0.0035,
P=0.0085, respectively). Active RA blood expressed significantly
lower IFN-γ mRNA in comparison with that of mild RA (P=0.014).
There was no difference of IFN-γ expression between active RA and
healthy blood, but active RA showed higher expression of IL-4 than that
of healthy samples (P=0.025). TNF-α expression by active RA blood
was lower than that of mild RA (P=0.013) and healthy blood, though
not significantly so.
Conclusion: These findings demonstrate that IFN-γ, T-bet, or TNF-α
expression by whole blood cells may be influenced by disease activity
of RA, and the combination of the defect of TNF-α and Th1 immune
response indicated by low IFN-γ and T-bet with high IL-4 expression in
the active RA blood, with the additional effect of blockade of TNF acti-
vation on T cells, may represent critical conditions for the development
of tuberculosis.
46
TNF-α α dependency of IL-17-induced joint pathology
differs under naive and arthritis conditions in vivo
M Koenders, E Lubberts, L Joosten, B Oppers, L van den
Bersselaar, J Kolls, WB van den Berg
University Medical Center Nijmegen, Nijmegen, The Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):46
Background: T-cell IL-17 is a proinflammatory cytokine present in the
synovium of RA patients. IL-17 is an inducer of other cytokines, such as
IL-1 and TNF. It can have both additive and synergistic effects on
cytokine induction and tissue destruction with these cytokines, but may
have direct pathological effects as well.
Objective and methods: In the present study, we examined the
dependency of TNF-α in the IL-17-induced joint inflammation and carti-
lage damage under naive and arthritis conditions using an adenoviral
vector expressing mIL-17 (AdIL-17).
Results: IL-17 overexpression in the knee joint of naive mice resulted in
joint inflammation and cartilage proteoglycan depletion, which gradually
increased with time. No effects were noted with the same dose of the
control vector. IL-17 induced elevated expression of IL-1 mRNA levels
in the synovium in comparison with the control group. However, no dif-
ference in IL-17-induced joint pathology was noted in IL-1α/β-deficient
mice. Of high interest, using TNF-α deficient mice, the IL-17-induced
joint inflammation and cartilage damage were almost completely
absent. This strongly indicates that under naive conditions in vivo the
IL-17-induced joint inflammation and cartilage destruction are mediated
by TNF and not necessarily by IL-1.
Since it is known that in streptococcal-cell-wall (SCW) arthritis, TNF
and IL-1 induction can be uncoupled, we did similar experiments in this
acute SCW arthritis model. Overexpression of T-cell IL-17 in this
macrophage-mediated model results in an elevation of joint inflammation
and cartilage proteoglycan depletion in comparison with the control
vector group. Furthermore, this T-cell cytokine turns this acute model
into a more chronic one. Although knocking out TNF-α did have some
effect on inflammation and cartilage damage, the dependency of TNF in
IL-17 induced pathology was not as strong as under naive conditions.
Conclusion: These data show a direct relation of IL-17 and TNF-α
induction under naive conditions in vivo. However, the presence of IL-1
during arthritis conditions and the collaboration of IL-17 with IL-1 mod-
ulates the TNF-α dependency during arthritis.
47
Differential effects of leflunomide and methotrexate
on cytokine production in RA
MC Kraan1, TJM Smeets1, MJ van Loon1, FC Breedveld2,
BAC Dijkmans3, PP Tak1
1Division of Clinical Immunology and Rheumatology, Academic
Medical Center/University of Amsterdam, Amsterdam, The Netherlands
2Department of Rheumatology, Leiden University Medical Center, Leiden,
The Netherlands
3Department of Rheumatology, Free University, Amsterdam, The
Netherlands
Arthritis Res Ther 2003, 5 (suppl 1):47
Background: T cells are considered to be pivotal cells in the pathogen-
esis of rheumatoid arthritis (RA) and therefore represent a potential
target for treatment. The novel disease-modifying antirheumatic drug
(DMARD) leflunomide inhibits pyrimidine biosynthesis. T cells are espe-
cially susceptible to inhibition of this enzyme due to increased demand
for pyrimidines after activation, together with the absence of a salvage
pathway.
Objective: We investigated the effects of leflunomide on cytokine pro-
files in vivo and in vitro to provide more insight into the mechanism of
action of leflunomide in RA.
Methods: Serum samples from 100 RA patients, treated with either
leflunomide (n=50) or methotrexate (n=50), were collected at base-
line, after 12 weeks and after 1 year of treatment. In these samples,
serum levels of interleukin-6 (IL-6) and interferon gamma (IFN-γ) were
determined by ELISA. The effects of the active metabolite of lefluno-
mide (A77-1726; 0–200µM) on IL-6 and IFN-γ production by periph-
eral blood mononuclear cells (PBMCs) from healthy volunteers (n=6)
and RA patients (n=3) were studied by ELISA after activation (with
phytohemagglutinin, lipopolysaccharide, and αCD3/αCD28) by ELISA.
In addition, monocytes and lymphocytes were isolated from two healthy
volunteers by density-gradient centrifugation methods, and effects of
A77-1726 on IL-6 production after activation (with phytohemagglutinin
or lipopolysaccharide) were measured by ELISA and PCR. Effects on
cell proliferation (3H-thymidine incorporation), were measured as well.
Results: Serum levels of IFN-γ were significantly reduced after lefluno-
mide treatment (baseline 43pg/ml±10 [mean±SEM]; 1 year 29±7
[P=0.015]), whereas we did not observe a change in IL-6 concentra-
tions in serum (baseline 158±41, 4 months 151±48). In contrast, both
IFN-γ and IL-6 serum levels were significantly reduced after methotrex-
ate treatment. Consistent with these data, in vitro experiments revealed
a dose-dependent inhibition of IFN-γ production by activated PBMCs in
both healthy volunteers and RA patients in the presence of A77-1726,
without a clear-cut effect on IL-6 production. IL-6 production by mono-
cytes was not inhibited (measured by ELISA and PCR). Production of
IFN-γ by lymphocytes was inhibited by A77-1726.
Conclusion: The results presented here show inhibition of IFN-γ pro-
duction by leflunomide without a clear-cut effect on IL-6 production.
This differential effect supports the hypothesis that leflunomide prefer-
entially affects activated T cells. The effects on T cells could be
explained by both DHODH inhibition and effects on signal transduction
pathways.
Arthritis Research & Therapy Vol 5 Suppl 1Abstracts of the 23rd European Workshop for Rheumatology Research

Page 15

S15
48
Alveolar macrophages in scleroderma lung disease
are activated by TGF-β β and IL-4 but not TNF-α α
I Luzina, S Atamas, B White
University of Maryland, Baltimore, MD, USA
Arthritis Res Ther 2003, 5 (suppl 1):48
Background: We reported previously that alveolar macrophages from
scleroderma patients with lung inflammation are characterized by an
activated gene expression profile.
Objective: The purpose of this study was to identify the molecular
mediators that have activated alveolar macrophages in scleroderma
lung disease, through an analysis of transcription factor activation.
Methods: Freshly isolated alveolar macrophages were negatively
selected by depleting CD3+cells from BAL cells of six patients with
lung inflammation. Nuclear translocation of 55 transcription factors was
evaluated using a protein/DNA array system (Panomics). Data were
analyzed using hierarchical clustering.
Results and conclusion: Three distinct clusters of translocated tran-
scription factors were notable. One cluster included Smad3/4 as well
as other factors involved in or activated by TGF-β signaling, including
CREB/p300, Fast-1, the glucocorticoid receptor element, and
GAS/ISRE. A second cluster included Stat5/6 and other factors acti-
vated by IL-4 signaling, including GATA, Sp1, and early-response
element. The third cluster included factors that were translocated to
the nucleus in low levels in only a few samples. This cluster included
NF-κB, which is critical in signaling by TNF-α and IL-1, and several
other transcription factors activated by TNF-α signaling. In animal
models, macrophages that are activated by TGF-β or IL-4 have been
associated with the subsequent development of fibrosis. Thus, we
directly demonstrate cellular activation by the profibrotic factors TGF-β
and IL-4 in involved tissues from scleroderma patients.
49
Circulating bioactive TNF in rheumatoid arthritis
patients treated with infliximab: link to clinical response
H Marotte, P Miossec
Clinical Immunology Unit, Departments of Immunology and
Rheumatology, Hôpital Edouard Hérriot, Lyon, France
Arthritis Res Ther 2003, 5 (suppl 1):49
Background: Infliximab is an anti-tumor-necrosis-factor (anti-TNF) anti-
body used for the treatment of rheumatoid arthritis.
Objective: To clarify heterogeneity in patient response, a bioassay was
established to measure functional circulating tumor necrosis factor (TNF).
Results: Culture of TNF activated synoviocytes with plasma samples
before and after infliximab treatment showed an increased production
of IL-6 that was strongly reduced by treatment. This pattern was asso-
ciated with a good clinical response, and bioactive TNF levels detected
with this bioassay were correlated with TNF levels measured by ELISA.
This was not observed in patients with a poor response. The functional
circulating bioactive TNF was significantly (P<0.001) more elevated in
subset with a good biological response than bad responders. No cor-
relation was observed with levels of soluble receptors.
Conclusion: This result suggests the importance of systemic bioactive
TNF in such response.
50
Shared epitope and rheumatoid arthritis severity:
association with infliximab treatment in a
postmarketing study
H Marotte, P Gaudin, C Alexandre, P Miossec
Clinical Immunology Unit, Departments of Immunology and
Rheumatology, Hôpital Edouard Hérriot, Lyon, France
Arthritis Res Ther 2003, 5 (suppl 1):50
Background: Modalities of infliximab treatment for active rheumatoid
arthritis (RA) have been established following randomised clinical trials.
Objective: In a postmarketing study, we compared the clinical presen-
tation of RA patients from the same geographic area treated by inflix-
imab or not. We checked for markers linked with selection for infliximab
treatment focusing on the shared epitope (SE).
Methods: RA patients (897 ) from the Rhone-Alpes area, France, were
enrolled. Clinical indices of disease activity and joint destruction,
including age, sex, disease duration, Ritchie articular index, and right
Larsen wrist index, were collected. Biological data included erythrocyte
sedimentation rate, rheumatoid factor (RF), and SE status. SE determi-
nation was performed by enzyme-linked oligosorbent assay.
Results and conclusion: All patients had the typical clinical and biologi-
cal features of RA. The patients were predominantly white women and
64.6% were RF-positive. The distribution of the SE was 44.74% for 0
copy, 43.33% for 1 copy, and 11.93% for 2 copies. The risk of develop-
ing joint destruction was associated with the presence of the SE, with a
dose effect. SE heterozygote patients were almost twice as likely to
develop erosions as SE-negative patients (OR=2.18; 95% CI
1.48–3.22, P<0.001). Patients with two copies of the SE were more
likely to develop erosions (OR=4.73, CI 2.67–8.36, P<0.001). When
patients treated with infliximab were isolated, RA severity parameters,
such as the Ritchie and Larsen indices, RF, and the presence of the SE,
were associated with infliximab treatment. The frequency of the SE in RA
patients treated with infliximab was approximately twice as high for SE
heterozygote carriers (OR=2.18; 95% CI 1.43–3.31, P<0.001), and 4
times as high for SE homozygotes (OR=3.88; 95% CI 2.25–6.68,
P<0.001). Thus RA patients with the SE were more often selected for
infliximab treatment extending the link between SE and severity.
51
High-molecular-weight PEG precipitates from
synovial fluid induce more TNF-α α than those from
serum of RA patients, which is in contrast to patients
with other inflammatory arthritides
L Mathsson1, N Mohammad Noor1, G ElGhazali1, O Sjöberg1,
L Bengtsson1, K Nilsson Ekdahl1, B Nilsson1, J Rönnelid1,2
1Unit of Clinical Immunology, Uppsala University, Stockholm, Sweden
2Unit of Rheumatology, Karolinska Institute, Stockholm, Sweden
Arthritis Res Ther 2003, 5 (suppl 1):51
Background: We have earlier shown that polyethylene glycol (PEG)
precipitates from sera of active SLE patients can induce monokine pro-
duction, possibly through immune complexes.
Objective: We have now investigated monokine production induced by
PEG precipitates from arthritis patients.
Methods: Paired synovial fluid (SF) and serum samples from 26 RA
patients and 20 patients with other inflammatory arthritides were sub-
jected to combined PEG precipitation and purification. The IgG
content was determined by ELISA. Precipitates were added to periph-
eral blood mononuclear cell (PBMC) cultures, and supernatant levels of
TNF-α and IL-10 determined by ELISA after 20 hours.
Results: In both the investigated PBMC donors, the highest median
values of TNF-α were found in SF-PEG cultures. In one of the PBMC
donors, this difference was significant in paired analysis for all patients
and for the RA group only, but not for non-RA patients only. The corre-
lation between the TNF-α and IL-10 production was higher for RA than
for non-RA patients, and higher for SF than for serum precipitates. No
correlation was noted between the IgG content of the precipitates and
cytokine levels.
Conclusion: In this preliminary study, PEG precipitates from RA SF
induced higher production of TNF-α than did corresponding serum pre-
cipitates, possibly through stimulation by immune complexes in SF. Since
no correlation was observed between the IgG content and cytokine
induction, it is likely that immune-complex composition rather than the
IgG content of the precipitates determines the cytokine response.
Available online http://arthritis-research.com/supplements/5/S1