Pristane, a Non-Antigenic Adjuvant, Induces

MHC Class II-Restricted, Arthritogenic T Cells

in the Rat
1

1.
2.
3.
4.
5.
6.

Jens Holmberg2,
Jonatan Tuncel,
Hisakata Yamada,
Shemin Lu,
Peter Olofsson3and
Rikard Holmdahl4
+Author Affiliations
1.Medical Inflammation Research, Lund University, Lund, Sweden
Next Section

Abstract
Pristane-induced arthritis (PIA) in rats, a model for rheumatoid arthritis (RA), is a T celldependent disease. However, pristane itself is a lipid and unable to form a stable complex with a
MHC class II molecule. Therefore, the specificity and function of the T cells in PIA are as
unclear as in rheumatoid arthritis. In this study, we show that activated CD4+ T cells, which
target peripheral joints, transfer PIA. The pristane-primed T cells are of oligo or polyclonal
origin as determined by their arthritogenicity after stimulation with several mitogenic antiTCRV and anti-TCRV mAbs. Arthritogenic cells secreted IFN- and TNF- (but not IL-4)
when stimulated with Con A in vitro, and pretreatments of recipient rats with either anti-IFN- or
a recombinant TNF- receptor before transfer ameliorated arthritis development. Most
importantly, we show that these T cells are MHC class II restricted, because treatment with Abs
against either DQ or DR molecules ameliorates arthritis development. The MHC class II
restriction was confirmed by transferring donor T cells to irradiated recipients that were
syngenic, semiallogenic, or allogenic to MHC class II molecules, in which only syngenic and
semiallogenic recipients developed arthritis. These data suggest that the in vivo administration of
a non-antigenic adjuvant, like pristane, activates CD4+ T cells that are MHC class II restricted
and arthritogenic.
Previous SectionNext Section

Introduction
Rheumatoid arthritis shows a strong association with the MHC region, and especially to
the MHC class II DR locus (1). This association could be explained by the sharedepitope hypothesis, in which rheumatoid arthritis (RA)5-susceptible alleles share
identical or similar motifs in the third hypervariable region of the DR -chain (2). The
RA-associated peptide binding pocket in DR molecules may favor presentation of
certain self-peptides to autoreactive T cells. Consequently, autoreactive CD4 + T cells,
which recognize peptides on MHC class II molecules, may play an important role in RA.
Moreover, the vast majority of lymphocytes infiltrating RA synovia are CD4 + T cells

(3, 4). The joint-restricted CD4+ T cell clonotypes that persist over time are preferentially
found in the CD25+ subpopulation (5). Skewing of the TCRV repertoire (e.g., V3,
V14 V17) has been shown to occur frequently in CD25 + T cells in the synovial
fluid, indicating a local activation by a common antigenic stimuli (6). However, it has
been difficult to identify clones of arthritogenic T cells or a common Ag in the joint that
could explain the development of arthritis, and the pathogenic function of infiltrating T
cells into the joint synovia is still uncertain. Several reports show a significant IFN-
secretion from synovial fluid mononuclear cells of RA patients, indicating a Th1
response in RA (7, 8).
A major problem in the investigations of the role of T cells in arthritis is the lack of
animal models that can efficiently transfer disease with specific T cells. In accordance
with this, the most commonly used animal model, collagen-induced arthritis (CIA),
cannot be as efficiently transferred compared with the corresponding model for multiple
sclerosis, the experimental autoimmune encephalomyelitis (9). Surprisingly, it is well
known that adjuvant-induced arthritides can be transferred with T cells (10, 11, 12, 13).
There are no evidences that adjuvants, such as pristane, bind MHC class II molecules,
and there is no direct link to adaptive autoimmunity. Thus, the specificity or MHC
restriction of the arthritogenic T cells is not known. We therefore set out to determine
whether pristane-activated T cells are MHC class II restricted and could transfer
arthritis. Similar to RA, pristane-induced arthritis (PIA) shows symmetry of inflammation,
chronic relapsing disease course, infiltration of T cells, and erosive destruction of
cartilagenous peripheral joints (14). Furthermore, the PIA model is characterized
genetically, and one of the identified loci contain the MHC region (PIA locus 1 (Pia1))
and another locus contains the TCR locus (PIA locus 5 (Pia5)) (15). In the current
study, we show that pristane injection in rats leads to activation of CD4 + T cells that
transfer arthritis and occurs in the arthritic synovia. The T cell-induced arthritis was
depending on IFN- and TNF-, as shown by Ab-mediated cytokine blockage in vivo.
Finally, we show that the arthritogenic T cells are MHC class II restricted, which argues
for Ag involvement in this model, although the pristane itself is non-antigenic.
Previous SectionNext Section

Materials and Methods

Rats
The DA and DA.1I (originating from Zentralinstitut Fur Versuchstierzucht, Hannover, Germany)
rat strains were kept in the animal facility of Medical Inflammation Research in a climatecontrolled environment with 12-h light/dark cycles, housed in polystyrene cages containing
wood shavings, and fed standard rodent chow and water ad libitum. The congenic DA.1I
(D20Rat42 to D20Rat49; n = 10) and the DA.1U (D20Rat47 to AA858870;n = 6) rats produced
by speed congenic technique (16) express i and u alleles, respectively, for the rat MHC class I
(RT1-A) and class II (RT1-B/D) molecules. Female DA and male DA.1I rats were intercrossed to
produce (DAxDA.1I)F1 offspring, and a DA.1Ua/uxDA.1Ua/uintercross was set up to produce
littermate offspring that were MHC syngenic (a/a), semiallogenic (a/u), or MHC allogenic (u/u)
to DA (a/a). The experiments were approved by a local (Malm/Lund, Sweden) ethical
committee (license M7-01).
PIA induction and evaluation of arthritis

Arthritis was induced at the age of 812 wk in age-matched female rats by an intradermal
injection at the base of the tail with 500 l of pristane oil (2,6,10,14-tetramethylpentadecane;
Sigma-Aldrich). Arthritis development was monitored in all four limbs using a macroscopic
scoring system. Briefly, 1 point was given for each swollen or red toe; 1 point was given for each
swollen midfoot, digit, or knuckle; and 5 points were given for a swollen ankle (maximum score
per limb and rat was 15 and 60, respectively).
Histological analysis
Twelve days after adoptive transfer of 20 106 pristane-primed spleen cells from (DAxDA.1I)F1
donor rats, paws from DA.1I recipient rats were collected and decalcified with EDTA. Cryostat
sections were stained with hematoxylin and subjected to immunohistochemical staining with
haplotype-specific (RT1-Aa) biotinylated anti-RT1-AmAb (R3/13, clone C3) (17), followed by
streptavidin-peroxidase, and visualized using a DAB substrate kit (ChemMate;
DakoCytomation).
Isolation of synovial cells
At 14 days after pristane injection and 12 days after T cell transfer, rats were killed, and synovia
from swollen talo-crural joints were harvested. This was done by isolating the joint by cutting the
bone and removing the skin. The remaining joint capsule was excised. The tissue was
homogenized and cultured for 2 h in DMEM supplemented with heat-inactivated FCS (10%),
collagenase D (0.60 U/ml; Roche), and DNase 1 (5 U/ml; Sigma-Aldrich). A single-cell
suspension was made by passing cells through a 40-m filter, followed by a 40% Percoll
(Amersham Biosciences) density gradient separation to obtain large lymphocytes (18). The cells
were then washed and resuspended in PBS-D (Dulbeccos PBS not containing CaCl2 or MgCl2;
Invitrogen Life Technologies), supplemented with 0.5% BSA (Sigma-Aldrich) for Ab staining.
Abs and immunosuppressive reagents
The following anti-rat mAbs used for flow cytometry analysis were purchased from Pharmingen
(San Diego, CA) as FITC, PE, or biotinylated conjugates: anti-CD4 (clone OX-35), anti-CD8a
(clone OX-8), anti-CD25 (clone OX-39), anti-LCA (leukocyte common Ag) for staining of
leukocytes (clone OX-1), anti-CD45RA for staining of B cells (clone OX-33), anti-TCR
(clone R73), and anti-RT1Aa,b (clone C3). mAbs used for in vivo blocking experiments were
grown from hybridomas purchased from American Type Culture Collection: OX-6 (anti-RT1B)
(19), OX-17 (anti-RT1D) (19), DB-1 (anti-IFN-) (20), and Hy2.15 (mouse anti-rat TNP, IgG1
isotype control). Etanercept (Enbrel; Wyeth-Ayerst Pharmaceuticals), a generous gift from Dr.
H. Burkhardt (Department of Internal Medicine III and Institute of Clinical Immunology,
University of Erlangen-Nuremberg, Erlangen, Germany), was used for in vivo blocking of TNF (21), using PBS-D as control.
Flow cytometry
Cells to be stained were collected from spleen, peripheral blood, or swollen synovia at different
time points after pristane injection or T cell transfer, as indicated in the text. Single cells were
prepared and stained with FITC-, PE-, and biotin-conjugated mouse anti-rat mAbs for 20 min at
4C. Streptavidin-APC was used as a secondary reagent, and propidium iodide was added before
acquisition. Lymphocyte expansion in draining lymph nodes and spleen was determined by
comparing the number of live cells in each population with the total number of cells in the lymph
nodes and spleen, respectively. We acquired a minimum of 104 to 105 cells on a FACScan (BD
Biosciences), with gates set to include all viable cells that were further analyzed by CellQuest
(BD Biosciences) software.

Cytokine detection
Quantitative measurements of IFN-, TNF-, and IL-4 in cell supernatants were performed using
the eBioscience rat IFN- and the rat IL-4 BD OptEIA ELISA Set protocol (BD Pharmingen).
The lower limits of detection sensitivity of these assays were 1.6 pg/ml for IL-4 and 15 pg/ml for
IFN-. For the rat TNF- ELISA (22), coating anti-rat TNF- (SB/230499/GR), detecting biotinconjugated anti-rat TNF-, (S54/250499/GR), and TNF- standard (190499) were kindly
provided by Dr. S. Poole (Division of Endocrinology, National Institute for Biological Standards
and Control, Hertfordshire, U.K.). Briefly, one million spleen cells from five pristane-primed and
naive DA rats (day 14 after injection) were stimulated for 48 h at 37C with or without Con A (3
g/ml; Sigma-Aldrich) in triplicate, in 200 l of DMEM supplemented with FCS (5%), HEPES
(2.4 mg/ml), 2-ME (3.9 g/ml), and penicillin-streptamycin (104 IU/ml penicillin, 10 mg/ml
streptomycin; Invitrogen Life Technologies). Cytokines were detected by a heterogeneous timeresolved fluorometric assay (Victor/1420 Multilabel counter; Wallac Sweden AB) using Eu3+conjugated streptavidin (Delfia; PerkinElmer) and Delfia Enhancement Solution for
quantitative determination of Eu3+, reflecting quantities of IFN-, IL-4, and TNF-. All cytokine
assays were verified using recombinant proteins as positive controls.
T cell transfer
At 14 days after pristane injection, rats were killed, and inguinal lymph nodes and spleens were
removed. Cells were washed and passed through 40-m filters and reactivated in vitro with Con
A (3 g/ml; Sigma-Aldrich); IL-2 (800 pg/ml in final concentration from an IL-2-transfected
X63 cell line from our hybridoma collection); plate-bound anti-rat TCR mAbs anti-TCR
(clone R73) (23), anti-TCRV4 (clone G99) (24), anti-TCRV8.5 (clone B73) (25), antiTCRV10 (clone G101) (25), and anti-TCRV16 (clone HIS42) (26) (5 g/ml); or an anti-TNP
IgG1 isotype control (clone Hy2.15; 5 g/ml) in DMEM supplemented with FCS (5%), HEPES
(2.4 mg/ml), 2-ME (3.9 g/ml), and penicillin-streptomycin (104 IU/ml penicillin, 10 mg/ml
streptamycin; Invitrogen Life Technologies). After 48 h of incubation at 37C, cells were washed
and resuspended in PBS-D. Naive recipient rats were then injected i.v. (through the tail vein)
with different numbers of cells as indicated in the text and tables. All cell-sorting experiments
were performed according to the Dynabiotech protocol (Dynal Biotech). Anti-CD4 (clone OX35), anti-CD8a (clone OX-8), and anti-CD25 (clone OX-39) were coated on CELLection Pan
Mouse IgG Dynal beads. Naive recipient rats (DA) were in a set of experiments in which they
were given i.p. injections of 1.0 mg of the anti-rat mAbs (anti-IFN- (clone DB-1), anti-RT1B
(clone OX-6), anti-RT1D (clone OX-17), and anti-TNP (clone Hy2.15)) in 1.0 ml of PBS-D 3 h
before transfer or given s.c. injections of 0.2 mg of Etanercept (Enbrel; Wyeth Ayerst
Pharmaceuticals) in 1.0 ml of PBS-D. Additionally, 20 106 pristane-primed spleen cells from
DA rats were transferred to irradiated (600 rad) DA rats congenic for the MHC region (i.e.,
syngenic a/a, semiallogenic a/u, and allogenic u/u MHC alleles).
Statistics
Comparisons of incidences were evaluated by Fishers exact test, and the arthritis score and
cytokine measurements were evaluated with the Mann-Whitney U test or Kruskal Wallis test. In
all experiments, p < 0.05 was considered significant.
Previous SectionNext Section

Results
Selective accumulation of activated CD4+ T cells in PIA synovia

After an intradermal injection of pristane in DA rats, both CD4+, CD8+ T cells and B cells
significantly expand in draining lymph nodes until day 6 after injection, whereafter the
CD4+ T cells continue to expand until day 12 after injection, as shown in Table I. In the
spleen, however, the expansion of B cells ceases at day 6 after injection, followed by a
significant reduction in both T and B cells at day 12 after injection (Table I). To identify the
cells infiltrating the arthritic joints, we collected arthritic synovia and peripheral blood from rats
14 days after pristane injection. Phenotypic analysis of synovial infiltrating cells and peripheral
blood were performed by flow cytometry, with gates set to include total leukocytes (LCA + cells),
TCR+ cells, and CD4+ TCR+ cells. As shown in Fig. 1A, the arthritic snyovia is
characterized by infiltrating lymphocytes, most of which are mainly highly activated CD4 + T
cells, expressing ICAM-1, IL-2R (CD 25), RT-1D, and RT-1B (Fig. 1, Aand B). This does not
reflect the overall cell distribution after pristane injection, in which CD4+ T cells are only slightly
increased in peripheral blood. After demonstrating the selective accumulation of highly activated
CD4+ T cells in arthritic joints, we next investigated whether T cells from pristane-treated rats
could transfer arthritis.

View larger version:

In this page

In a new window

FIGURE 1.
Infiltration of highly activated CD4+T cells in the PIA synovia. Cytofluorometric analysis of
arthritic synovia and peripheral blood of DA donors (n = 4) at day 14 after injection with 500 l
of pristane. A, The T cell/B cell ratio (top panels) and the frequency (percentage) of CD4+T
cells/all TCR+ cells (bottom panels) for synovia (PIA), peripheral blood (PIA), and peripheral
blood (naive). B, The frequency (percentage) of CD25, ICAM-1, RT1B, and RT1D expression
on gated CD4+ TCR+ cells for PIA synovia (open curves), PIA peripheral blood (black curves),
and naive peripheral blood (gray curves). Significant values refer only to differences in pristaneinjected rats. Data are the mean SD. , p < 0.05 (Mann-Whitney U test).
View this table:
In this window

In a new window

Table I.
Lymphocyte expansion in draining lymph nodes after pristane injection

PIA can be transferred by activated T cells

The arthritogenic capacity of cells from peripheral lymphoid tissue was investigated by adoptive
transfer of spleen and inguinal lymph node cells collected from pristane-injected donor rats.
Initially, an optimal transfer protocol was developed in which cell number, donor organ, and
transfer day were taken into account (Table II). Both cells from the spleen and inguinal lymph
nodes transferred PIA after reactivation by Con A in vitro. At day 7 after pristane injection, both
spleen and lymph node cells could transfer a mild form of arthritis. Fourteen days after pristane
injection, however, and despite the fewer lymphocytes in the spleen, these cells were more
arthritogenic than cells from the inguinal lymph nodes. Moreover, recipient rats receiving 20
106 pristane-primed spleen cells developed more severe arthritis than rats receiving 5
106 spleen cells (Table II).
View this table:
In this window

In a new window

Table II.
Spleen cells are more arthritogenic than draining lymph node cellsa
A representative experiment from a series of four transfers using 20 106 spleen cells collected
14 days after pristane injection and challenged with Con A for 48 h is shown in Fig. 2, in which
the time course of adoptively transferred PIA is compared with that of PIA itself, both of which
reached an incidence of 100%. The adoptively transferred arthritis showed an earlier onset
compared with pristane-induced arthritis, typically 46 days after transfer, when foci of redness
and edema appeared around the paw joints. The number of arthritic joints and the severity of
inflammation in the paws reached a maximal clinical score at 810 days after transfer, followed
by a regression period of 34 wk, depending on the number of cells injected.

View larger version:

In this page

In a new window

FIGURE 2.
Arthritis developments in donors and recipients after pristane injection and adoptive PIA
transfer, respectively. PIA was induced in DA rats (n = 6) by intradermal injection of 500 l of
pristane at the base of the tail (), and PIA transfer was induced in naive DA rats (n = 6) by i.v.
injection of 20 106pristane-sensitized and Con A reactivated spleen cells ().
The need for pristane priming and in vitro reactivation was further examined by transferring
naive or pristane-primed spleen cells with or without reactivation in vitro, either with Con A or

plate-bound mitogenic anti-TCR mAb (Table III). Spleen cells that were only pristane
primed in vivo or only stimulated with Con A in vitro could not transfer arthritis, whereas cells
that were both primed with pristane in vivo and reactivated in vitro, either by Con A or with
plate-bound mitogenic anti-TCR mAb, could transfer arthritis. Arthritogenic cells also needed
to be viable to transfer arthritis, because irradiated (1500 rad) pristane-primed and Con Aactivated spleen cells could not transfer the disease. To identify the cell subset responsible for
arthritis development, we proceeded by sorting subfractions of pristane-primed spleen cells
before adoptive transfer.
View this table:
In this window

In a new window

Table III.
T cell reactivation in vitro is necessary for arthritis transfera
CD4+ T cells alone transfer PIA
Cell-dose titration showed that only 5 106 total spleen cells, delivered i.v., were enough to
provoke arthritis with 100% incidence. Spleen cells collected 14 days after injection were further
subdivided into CD4+ and CD8+ T cells or CD4, CD8, CD25, and CD134fractions before
Con A stimulation using magnetic selection. It was shown that CD4+, but not CD8+, T cells as
well as CD8, but not CD4, spleen cells could transfer arthritis with an incidence of 100%.
Furthermore, it was observed that both CD25 and CD134fractions could become arthritogenic
after Con A stimulation (Table IV). In conclusion, only CD4+T cells transfer PIA, and these
cells do not have to be CD25+ or CD134+before reactivation in vitro to transfer disease. We then
proceeded by investigating the clonality of the arthritogenic cells and whether they were
restricted to MHC class II.
View this table:
In this window

In a new window

Table IV.
CD4+ T cells alone transfer PIAa
Arthritogenic T cells that transfer PIA are not monoclonal
Pristane-primed spleen cells were reactivated in vitro with plate-bound mitogenic anti-rat TCR
mAbs with specificities for various TCR alleles (i.e., anti-TCRV4, anti-TCRV8.5, antiTCRV10, and anti-TCRV16), using Hy2.15 (anti-TNP) as an IgG1 isotype control. As shown
in Table V, anti-TCRV4, anti-TCRV10, and anti-TCRV16 mAbs, but not anti-TCRV8.5
mAb, were able to activate cells to become arthritogenic. These data show that several, but not
all, clonal specificities of T cells are arthritogenic.
View this table:
In this window

Table V.
Oligoclonal T cells transfer PIA

In a new window

MHC class II restricts arthritis development in PIA transfer

The role of MHC class II was examined in adoptively transferred arthritis by blocking MHC
class II DQ (formerly denoted RT1B in the rat) and DR (formerly denoted RT1D in the rat)
molecules with anti-RT1B and anti-RT1D mAbs, respectively (Fig. 3, A and B). Anti-RT1B
(OX-6, 1.0 mg) and anti-RT1D (OX-17, 1.0 mg) mAbs were given i.p. 3 h before adoptive
transfer of 20 106 Con A-stimulated spleen cells, using Hy2.15 (anti-TNP) as an IgG1 isotype
control. Both anti-RT1B (Fig. 3A) and anti-RT1D (Fig. 3B) significantly reduced arthritis
severity, whereas anti-RT1D treatment significantly reduced both arthritis incidence and
severity. To exclude the possibility that the anti-RT1B and anti-RT1D mAbs could result in
depletion of adoptively transferred class II-expressing T cells, we pretreated pristane-primed and
Con A-stimulated spleen cells before transfer with either anti-RT1B or anti-RT1D, using antiRT1A-stained cells as a positive control. No differences were observed between groups given
arthritogenic cells pretreated with anti-RT1B (OX-6), anti-RT1D (OX-17), or anti-RT1A (C3)
mAbs (Fig. 3C).

View larger version:

In this page

In a new window

FIGURE 3.
Arthritogenic T cells are restricted to MHC class II molecules. Rats were treated prophylactically
with anti-RT1B or anti-RT1D mAbs 3 h before adoptive PIA transfer. A, DA recipients (n = 8)
were given injections of either 1.0 mg of anti-RT1B () or isotype control () mAbs. B, DA
recipients (n = 8) were given injections of either 1.0 mg of anti-RT1D (black and white circles)
or isotype control () mAbs. C, Arthritogenic cells from DA donor rats were pretreated with
excessive amounts (50 g/ml) of anti-RT1B (), anti-RT1D (), or control anti-RT1A (black and
white circles) mAbs before transfer to DA recipients (n = 5). Error bars are the mean SD.
, p < 0.05 and , p < 0.01 (Mann-Whitney U test); , p < 0.05 (Fishers exact test).
To confirm the MHC class II restriction, spleen cells from DA donor rats (a/a alleles for the
MHC class II and classical class I region) were transferred to irradiated MHC congenic
semiallogenic (a/u) and congenic allogenic (u/u) DA rats, using irradiated syngenic DA
recipients as positive control rats. As depicted in Fig. 4A, both MHC syngenic (a/a) and
semiallogenic (a/u) recipients developed arthritis with an incidence of 100%, whereas MHC
allogenic recipients (u/u) showed no signs of arthritis. To exclude the possibility that transferred
cells were rejected by the host due to incomplete irradiation of recipients congenic for the MHC
region, peripheral blood was collected from all recipient rats shortly after onset of arthritis (i.e., 7
days after transfer) and analyzed for the absolute number of T cells. Equal volumes of
collected blood from each individual were stained with anti-TCR mAb and aquired for 60 s at
constant pressure by flow cytometry. As shown in Fig. 4B, allogenic cell transfer did not affect
the number of viable T cells indicating no host-vs-graft activity. Together, these data show
that the arthritogenic T cells are MHC class II restricted and that both DQ (RT1B) and DR
(RT1D) molecules are involved.

View larger version:

In this page

In a new window

FIGURE 4.
Adoptive PIA transfer depends on MHC syngenicity. A, PIA transfer of 20 106 arthritogenic
spleen cells to irradiated (600 rad) MHC-syngenic (n = 8; ), MHC-semiallogenic (n = 7; ), or
MHC-allogenic (n = 8; ) recipients.B, Total number of T cells determined by flow cytometry
through 60 s of acquisition at constant pressure and with equal volumes of peripheral blood from
all recipients 7 days after transfer. MHC-syngenic recipients are shown in black, MHCsemiallogenic recipients are shown in gray, and MHC-allogenic recipients are shown in white.
Error bars are the mean SE. , p < 0.001 (Kruskal-Wallis test).
Arthritogenic donor T cells target recipient joints
To investigate the fate of the CD4+ TCR+ cells in adoptively transferred arthritis, we used
(DAxDA.1I)F1 as donor rats, expressing both a and i alleles of the MHC class I RT1A and the
MHC class II RT1B and RT1D molecules, and irradiated DA.1I as recipient rats, only expressing
the i allele of these molecules, to follow cells in vivo. By doing so, we obtained a cell-target
system in which we avoid both graft-vs-host as well as host-vs-graft rejection, and at the same
time, we could use a mAb against the a haplotype of the RT1A molecule to detect donor cells.
As depicted in Fig. 5A, arthritic DA.1I recipient rats show a severe pannus formation in ankle
joints at day 12 after transfer, in which RT1A a-haplotype donor cells are located among
RT1A i-haplotype nucleated recipient cells. Synovial and lymphoid organ cells from the
recipient rats were isolated for flow cytometry, the gate was set to include RT1A a-haplotype
donor cells (Fig. 5B), and the tissues were analyzed for CD4 expression (Fig. 5C). Fig.
5D shows that the donor cells are located more frequently in the recipient synovia than in
blood, inguinal lymph nodes, and spleen. In all tissues examined, >96% of the donor cells were
T cells. However, in the synovia, 99% of these T cells were CD4+, in contrast to lymphoid
tissues that had a significantly lower percentage of CD4+ cells. We conclude that the transferred
CD4+ T cells reached the joints and directly caused arthritis. The next question we asked was
how these cells mediated arthritis.

View larger version:

In this page

In a new window

FIGURE 5.
CD4+ TCR+ donor cells target recipient synovia. A, Donor (DAxDA.1I)F1 cells (RT1Aa+) in
the arthritic joint of an irradiated DA.1I recipient (RT1Aa) at day 12 after transfer. ti, Tibia
bone; ta, talus bone; pa, pannus formation; arrow, RT1Aa+donor cells. B, RT1Aa+ donor cells of
all synovial cells from irradiated DA.1I recipients (n = 3) at day 12 after transfer. C, A
representative histogram of CD4+ T cells (percentage) among RT1Aa+ donor cells. D,
Distribution of donor (RT1Aa+) cells in synovia, peripheral blood, spleen, and inguinal lymph
nodes (iLN) of irradiated DA.1I recipients (n = 3) at day 12 after transfer are shown. Data are the
mean SE. , p < 0.05 (Mann-Whitney Utest).
IFN- and TNF- production has a pivotal role for arthritis development
To investigate the functional properties of the arthritogenic T cells, we first determined their
cytokine profile. Spleen cells primed in vivo with pristane and restimulated in vitro with or
without Con A were examined for the production of IFN-, TNF-, and IL-4 (Table VI).
Splenocytes from both normal and pristane-primed rats produced high amounts of IFN- and
TNF-, but not IL-4, when restimulated in vitro with Con A. No cytokine production was
detected in medium from non-restimulated cells.
View this table:
In this window

In a new window

Table VI.
Con A restimulation in vitro induces IFN- and TNF- cytokinesa
To examine the role of IFN- and TNF- in adoptively transferred arthritis, we pretreated naive
recipient DA rats with blocking amounts of anti-IFN- (DB-1) or with Etanercept (recombinant
TNF- receptor) before cell transfer. Recipients were either given i.p. injections of 1.0 mg of
anti-IFN-, using 1.0 mg of Hy2.15 (anti-TNP) as an IgG1 isotype control, or s.c. injections of
0.2 mg of Etanercept (in 1.0 ml of PBS-D), using 1.0 ml of PBS-D as a control, 3 h before
transfer. As shown in Table VII, both the anti-IFN- and the Etanercept treatments significantly
ameliorated arthritis development both regarding incidence and arthritis severity compared with
control rats.

View this table:

In this window

In a new window

Table VII.
Pretreatment with anti-IFN- or Etanercept ameliorate PIA transfera
Previous SectionNext Section

Discussion
A comprehensive understanding of the role of T cells in arthritis, both in human RA and in the
various animal models, has been difficult, both regarding their specificity, clonality, and
functional role. Detailed studies have been hampered by the lack of animal models in which
arthritis can be easily transferred with T cells. We describe in this study that T cells transferring
PIA are MHC class II restricted, oligo/polyclonal, and mediate their arthritogenic function in the
joints by secretion of cytokines such as IFN- and TNF-. This makes the PIA model useful for
investigating mechanisms of the T cell-mediated arthritis.
It is well known that T cells are able to transfer various adjuvant-induced arthritides. It was, in
fact, first shown using the mycobacteria adjuvant arthritis model with thoracic duct lymphocytes
(27), and it has later been reproduced in models such as avridine-induced arthritis (10) and also
in PIA (28). Interestingly, as in RA, all of these models also show a genetic association with the
MHC region (14, 29, 30), implicating a role for MHC-restricted T cells. Additional studies to
reveal the MHC specificity/restriction, T cell specificity, and effector function have been
hampered with a multitude of difficulties. Nevertheless, it has been reported that T cells specific
for heat shock protein can transfer arthritis; the mechanism was believed to be due to crossreactivity of the transferred T cells with various joint Ags (31). However, these studies have not
been reproduced, and histological evidence showed no or possibly only very mild synovitis (32).
In arthritis models induced by an immunization with an Ag, such as CIA, the development of
arthritis is dependent on a T cell response to the immunogen, often involving only one MHC
class II molecule and one or a few Ag-derived peptides (33). In PIA, there is, however, no reason
to assume that only one Ag or one immunodominant peptide causes the disease, and it is, in fact,
not known whether MHC class II is involved at all. Therefore, the present observation that the
arthritogenic T cells in PIA are MHC class II restricted makes the model more useful for studies
of a MHC class II-associated but complex disease as RA. It shows that classical MHC class IIrestricted T cells are of pathogenic importance and that their specificity is related to the disease
process in PIA. This indicates a role for MHC class II in a model that is not induced with an
exogenous immunogen. The peptides bound by the MHC class II molecule are presumably
derived from endogenous Ags, although we have no clue of the nature of this Ag(s). The
arthritogenic T cells were initially activated in vivo through pristane, which does not bind MHC
class II but could possibly bind other cell-surface receptors (e.g., CD1) (34). However, because
the arthritogenicity was dependent on MHC class II, a direct pathogenic role of CD1-restricted T
cells is unlikely. The finding that arthritogenic T cells use several different TCR V-genes and, in
addition, both interact with both of the major MHC class II isotypes argues that it is likely that
there is not a single self-Ag or antigenic peptide recognized. Still, at some stage, the

inflammatory response, and possibly also immune recognition, is likely to be joint specific
because only cartilagenous peripheral joints, with a similar distribution as in RA, are attacked. A
joint-specific inflammatory attack may not, however, necessarily need a joint-specific T cell
recognition. This has been shown from the observations that glucose-6-isomerase, a systemically
occurring protein, is involved in arthritis in mice (35). In fact, also in CIA, the arthritis
distribution is more limited than the distribution of the target self-Ag, type II collagen (36).
However, in both the G6PI-induced arthritis and in the CIA model, immune recognition has been
shown to occur in the lymph nodes draining the joints (37,38), suggesting that the arthritogenic
immune recognition is related to joints. Thus, for the pristane-injected donor rat, it is possible
that the initial Ag/MHC recognition occurs in draining lymph nodes shortly after injection where
we can observe an expansion of lymphocytes. However, the situation might be more complicated
because T cells derived from the spleens are more arthritogenic than those from the lymph nodes.
The reason for this remains to be investigated, but it is possible that the first pristane-mediated
activation of the T cells is not MHC class II restricted/Ag dependent, implying a second step of
Ag activation. An alternative explanation could be that both aggressive and regulatory T cells are
activated and that regulatory T cells are more dominating in lymph nodes draining the joints.
The arthritis in both PIA and the arthritis transferred by T cells from pristine-injected rats occurs
only in peripheral cartilagenous joints; in fact, they are more specific than CIA (39). This is not
related to where pristane or the T cells are injected but could be related to the source of the
endogenous peptides recognized. We know from studies in CIA that the nature of such
endogenous proteins could be quite complex because the T cell recognition in this case is
dependent on various forms of glycosylation of type II collagen (40). In addition, this
glycosylation is dependent on the status of chondrocytes in the joints that could vary depending
on inflammatory conditions (41). In the present experiments, we could show that the T cellinduced inflammatory cascade in the joints is dependent on cytokines such as IFN- and TNF-,
which are believed to be secreted by T cells and macrophages. Infiltration of mononuclear cells
into the joint synovium characterizes the chronic inflammatory process in RA and results from
the migration of lymphocytes and monocytes through the endothelium of postcapillary venules
(42). Both IFN- and TNF- are able to activate endothelial cells allowing traffic of arthritogenic
cells into the joint synovia. Blocking of TNF- in RA patients is clearly therapeutic (43),
whereas blocking of IFN- gave some initial effects but had a less promising long-term outcome
(44). There are certainly many different mechanisms of the arthritogenic actions of IFN- and
TNF- in both RA and PIA that need to be investigated. Therefore, the present T cell-mediated
arthritis model could be a useful tool to study in detail downstream effector mechanisms of TNF and other cytokines in a T cell/self-peptide/MHC class II-mediated disease.
Previous SectionNext Section

Acknowledgments
We thank Carlos Palestro, Isabell Bohlin, Sandy Liedholm, and Rebecka Ljungqvist for taking
care of the animals and Margareta Svejme for preparing the histology slides.
Previous SectionNext Section

Disclosures

The authors have no financial conflict of interest.

Previous SectionNext Section

Footnotes

The costs of publication of this article were defrayed in part by the payment of page charges.
This article must therefore be hereby marked advertisement in accordance with 18 U.S.C.
Section 1734 solely to indicate this fact.

1 This work was supported by the Crafoord, Lundberg, Kock, and sterlund Foundations, the
Swedish Association against Rheumatism, the Swedish Medical Research Council, the Swedish
Foundation for Strategic Research, Arexis AB, and the European Union framework 5 program
(C5RD-CT-200-00267) Van ae B.T. och aer det nodvaendigt.

2 Current address: Pharmexa A/S, Kogle Alle 6, DK-2970 Hrsholm, Denmark.

3 Current address: Arexis AB, Arvid Wallgrensbacke 20, SE-413 46 Gteborg, Sweden.

4 Address correspondence and reprint requests to Dr. Rikard Holmdahl, Section for Medical
Inflammation Research, Slvegatan 19, I11 Biomedical Center, Lund University, S-221 84 Lund,
Sweden. E-mail address: rikard.holmdahl@inflam.lu.se

5 Abbreviations used in this paper: RA, rheumatoid arthritis; CIA, collagen-induced arthritis;
PIA, pristane-induced arthritis.
Received October 13, 2004.
Accepted October 24, 2005.
Copyright 2006 by The American Association of Immunologists
Previous Section

References
1.

Stastny, P.. 1978. Association of the B-cell alloantigen DRw4 with rheumatoid arthritis. N.
Engl. J. Med. 298: 869-871.
CrossRefMedline

2.

Gregersen, P. K., J. Silver, R. J. Winchester. 1987. The shared epitope hypothesis. An

approach to understanding the molecular genetics of susceptibility to rheumatoid
arthritis. Arthritis Rheum. 30: 1205-1213.
CrossRefMedline

3.

Janossy, G., G. Panayi, O. Duke, M. Bofill, L. W. Poulter, G. Goldstein. 1981.

Rheumatoid arthritis: a disease of T-lymphocyte/macrophage
immunoregulation.Lancet 2: 839-842.
Medline

4.

Klareskog, L., U. Forsum, A. Scheynius, D. Kabelitz, H. Wigzell. 1982. Evidence in

support of a self-perpetuating HLA-DR dependent delayed-type hypersensitivity reaction
in rheumatoid arthritis. Proc. Natl. Acad. Sci. USA 79: 3632-3636.
Abstract/FREE Full Text

5.

Rittner, H. L., A. Zettl, M. C. Jendro, P. Bartz-Bazzanella, J. J. Goronzy, C. M.

Weyand. 1997. Multiple mechanisms support oligoclonal T cell expansion in rheumatoid
synovitis. Mol. Med. 3: 452-465.
Medline

6.

Howell, M. D., J. P. Diveley, K. A. Lundeen, A. Esty, S. T. Winters, D. J. Carlo, S. W.

Brostoff. 1991. Limited T-cell receptor -chain heterogeneity among interleukin 2
receptor-positive synovial T cells suggests a role for superantigen in rheumatoid
arthritis. Proc. Natl. Acad. Sci. USA 88: 10921-10925.
Abstract/FREE Full Text

7.

Steiner, G., M. Tohidast-Akrad, G. Witzmann, M. Vesely, A. Studnicka-Benke, A. Gal, M.

Kunaver, P. Zenz, J. S. Smolen. 1999. Cytokine production by synovial T cells in
rheumatoid arthritis. Rheumatology (Oxford) 38: 202-213.
Abstract/FREE Full Text

8.

Morita, Y., M. Yamamura, M. Kawashima, S. Harada, K. Tsuji, K. Shibuya, K.

Maruyama, H. Makino. 1998. Flow cytometric single-cell analysis of cytokine production
by CD4+ T cells in synovial tissue and peripheral blood from patients with rheumatoid
arthritis. Arthritis Rheum. 41: 1669-1676.
CrossRefMedline

9.

Ben-Nun, A., H. Wekerle, I. R. Cohen. 1981. The rapid isolation of clonable antigenspecific T lymphocyte lines capable of mediating autoimmune encephalomyelitis. Eur. J.
Immunol. 11: 195-199.
CrossRefMedline

10.
Taurog, J. D., G. P. Sandberg, M. L. Mahowald. 1983. The cellular basis of adjuvant
arthritis. II. Characterization of the cells mediating passive transfer. Cell
Immunol. 80: 198-204.
CrossRefMedline

11.
Taurog, J. D., G. P. Sandberg, M. L. Mahowald. 1983. The cellular basis of adjuvant
arthritis. I. Enhancement of cell mediated passive transfer by concanavalin A and by
immunosuppressive pretreatment of the recipient. Cell Immunol. 75: 271-282.
CrossRefMedline

12.
Svelander, L., A. Mussener, H. Erlandsson-Harris, S. Kleinau. 1997. Polyclonal Th1 cells
transfer oil-induced arthritis. Immunology 91: 260-265.
CrossRefMedline

13.
Cannon, G. W., D. S. Harper, F. Clayton, M. M. Griffiths. 1993. Passive transfer of
adjuvant-induced arthritis into irradiated DA recipient rats. Autoimmunity 15: 267-274.
CrossRefMedline

14.
Vingsbo, C., P. Sahlstrand, J. G. Brun, R. Jonsson, T. Saxne, R. Holmdahl.1996. Pristaneinduced arthritis in rats: a new model for rheumatoid arthritis with a chronic disease course
influenced by both major histocompatibility complex and non-major histocompatibility
complex genes. Am. J. Pathol. 149: 1675-1683.
Medline

15.
Vingsbo-Lundberg, C., N. Nordquist, P. Olofsson, M. Sundvall, T. Saxne, U. Pettersson,
R. Holmdahl. 1998. Genetic control of arthritis onset, severity and chronicity in a model
for rheumatoid arthritis in rats. Nat. Genet. 20: 401-404.
CrossRefMedline

16.
Wakeland, E., L. Morel, K. Achey, M. Yui, J. Longmate. 1997. Speed congenics: a classic
technique in the fast line (relatively speeking). Immunol. Today 18: 472-477.
CrossRefMedline

17.
Yousaf, N., J. C. Howard, B. D. Williams. 1986. Studies in cobra venom factor treated rats
of antibody coated erythrocyte clearance by the spleen: differential influence of red blood
cell antigen number on the inhibitory effects of immune complexes on Fc dependent
clearance. Clin. Exp. Immunol. 66: 654-660.
Medline

18.
Kozak, R. W., C. E. Moody, L. Staiano-Coico, M. E. Weksler. 1982. Lymphocyte
transformation induced by autologous cells. XII. Quantitative and qualitative differences
between human autologous and allogeneic reactive T lymphocytes. J. Immunol. 128: 17231727.
Abstract

19.
Offner, H., S. W. Brostoff, A. A. Vandenbark. 1986. Antibodies against I-A and I-E
determinants inhibit the activation and function of encephalitogenic T-lymphocyte
lines. Cell Immunol. 100: 364-373.
CrossRefMedline

20.

Van der Meide, P. H., M. Dubbeld, K. Vijverberg, T. Kos, H. Schellekens. 1986. The
purification and characterization of rat gamma interferon by use of two monoclonal
antibodies. J. Gen. Virol. 67: 1059-1071.
Abstract/FREE Full Text

21.
Bendele, A. M., E. S. Chlipala, J. Scherrer, J. Frazier, G. Sennello, W. J. Rich, C. K.
Edwards, III. 2000. Combination benefit of treatment with the cytokine inhibitors
interleukin-1 receptor antagonist and PEGylated soluble tumor necrosis factor receptor
type I in animal models of rheumatoid arthritis. Arthritis Rheum. 43:2648-2659.
CrossRefMedline

22.
Rees, G. S., C. K. Gee, H. L. Ward, C. Ball, G. M. Tarrant, S. Poole, A. F. Bristow. 1999.
Rat tumour necrosis factor-: expression in recombinant Pichia pastoris, purification,
characterization and development of a novel ELISA. Eur. Cytokine Netw. 10: 383-392.
Medline

23.
Hnig, T., H. J. Wallny, J. K. Hartley, A. Lawetzky, G. Tiefenthaler. 1989. A monoclonal
antibody to a constant determinant of the rat T cell antigen receptor that induces T cell
activation. Differential reactivity with subsets of immature and mature T lymphocytes. J.
Exp. Med. 169: 73-86.
Abstract/FREE Full Text

24.
Torres-Nagel, N. E., G. Giegerich, D. P. Gold, T. Hunig. 1994. Identification of rat Tcra-V
4 and 8 gene products by monoclonal antibodies and cDNA
sequence.Immunogenetics 39: 367-370.
Medline

25.
Torres-Nagel, N., D. Gold, T. Hnig. 1993. Identification of rat Tcrb-V8.2, 8.5, and 10
gene products by monoclonal antibodies. Immunogenetics 37: 305-308.
Medline

26.
Kampinga, J., F. G. Kroese, G. H. Pol, P. Nieuwenhuis, F. Haag, P. B. Singh, B. Roser, R.
Aspinall. 1989. A monoclonal antibody to a determinant of the rat T cell antigen receptor
expressed by a minor subset of T cells. Int. Immunol. 1: 289-295.
Abstract/FREE Full Text

27.
Whitehouse, M. W., K. J. Orr, F. W. J. Beck, C. M. Pearson. 1974. Freunds adjuvants:
relationsip to arthritogenicity and adjuvanticity in rats to vehicle
composition. Immunology 27: 311-330.
Medline

28.

Olofsson, P., J. Holmberg, J. Tordsson, S. Lu, B. kerstrm, R. Holmdahl. 2003.

Positional identification of Ncf1 as a gene that regulates arthritis severity in rats. Nat.
Genet. 33: 25-32.
CrossRefMedline

29.
Vingsbo, C., R. Jonsson, R. Holmdahl. 1995. Avridine-induced arthritis in rats; a T celldependent chronic disease influenced both by MHC genes and by non-MHC genes. Clin.
Exp. Immunol. 99: 359-363.
Medline

30.
Kawahito, Y., G. W. Cannon, P. S. Gulko, E. F. Remmers, R. E. Longman, V. R. Reese, J.
Wang, M. M. Griffiths, R. L. Wilder. 1998. Localization of quantitative trait loci
regulating adjuvant-induced arthritis in rats: evidence for genetic factors common to
multiple autoimmune diseases. J. Immunol. 161: 4411-4419.
Abstract/FREE Full Text

31.
Holoshitz, J., A. Mattitau, I. R. Cohen. 1984. Arthritis induced in rats by cloned T
lymphocytes responsive to Mycobacteria but not to collagen type II. J. Clin.
Invest.73: 211-215.
CrossRefMedline

32.
Stanescu, R., O. Lider, W. van Eden, J. Holoshitz, I. R. Cohen. 1987. Histopathology of
arthritis induced in rats by active immunization to mycobacterial antigens or by systemic
transfer of T lymphocyte lines. A light and electron microscopic study of the articular
surface using cationized ferritin. Arthritis Rheum.30: 779-792.
CrossRefMedline

33.
Brunsberg, U., K. Gustafsson, L. Jansson, E. Michalsson, L. hrlund-Richter, S.
Pettersson, R. Mattsson, R. Holmdahl. 1994. Expression of a transgenic class II Ab gene
confers susceptibility to collagen-induced arthritis. Eur. J. Immunol. 24: 1698-1702.
CrossRefMedline

34.
Kronenberg, M., Y. Kinjo. 2005. Infection, autoimmunity, and glycolipids: T cells detect
microbes through self-recognition. Immunity 22: 657-659.
CrossRefMedline

35.
Korganow, A. S., H. Ji, S. Mangialaio, V. Duchatelle, R. Pelanda, T. Martin, C. Degott, H.
Kikutani, K. Rajewsky, J. L. Pasquali, C. Benoist, D. Mathis. 1999. From systemic T cell
self-reactivity to organ-specific autoimmune disease via
immunoglobulins. Immunity 10: 451-461.
CrossRefMedline

36.
Holmdahl, R., V. Malmstrm, E. Vuorio. 1993. Autoimmune recognition of cartilage
collagens. Ann. Med. 25: 251-264.
CrossRefMedline

37.
Mandik-Nayak, L., B. T. Wipke, F. F. Shih, E. R. Unanue, P. M. Allen. 2002. Despite
ubiquitous autoantigen expression, arthritogenic autoantibody response initiates in the
local lymph node. Proc. Natl. Acad. Sci. USA 99: 14368-14373.
Abstract/FREE Full Text

38.
Staines, N. A., N. Harper, F. J. Ward, V. Malmstrm, R. Holmdahl, S. Bansal.1996.
Mucosal tolerance and suppression of collagen-induced arthritis (CIA) induced by nasal
inhalation of synthetic peptide 184-198 of bovine type II collagen (CII) expressing a
dominant T cell epitope. Clin. Exp. Immunol. 103: 368-375.
Medline

39.
Hansson, A. S., S. Lu, R. Holmdahl. 2002. Extra-articular cartilage affected in collageninduced, but not pristane-induced, arthritis models. Clin. Exp. Immunol.127: 37-42.
CrossRefMedline

40.
Bcklund, J., S. Carlsen, T. Hger, B. Holm, L. Fugger, J. Kihlberg, H. Burkhardt, R.
Holmdahl. 2002. Predominant selection of T cells specific for glycosylated collagen type
II peptide (263-270) in humanized transgenic mice and in rheumatoid arthritis. Proc. Natl.
Acad. Sci. USA 99: 9960-9965.
Abstract/FREE Full Text

41.
Dzhambazov, B., M. Holmdahl, H. Yamada, S. Lu, M. Vestberg, B. Holm, O. Johnell, J.
Kihlberg, R. Holmdahl. 2005. The major T cell epitope on type II collagen is glycosylated
in normal cartilage but modified by arthritis in both rats and humans.Eur. J.
Immunol. 35: 357-366.
CrossRefMedline

42.
Ziff, M.. 1989. Pathways of mononuclear cell infiltration in rheumatoid
synovitis.Rheumatol. Int. 9: 97-103.
Medline

43.
Feldmann, M., R. N. Maini. 2001. Anti-TNF alpha therapy of rheumatoid arthritis: what
have we learned?. Annu. Rev. Immunol. 19: 163-196.
CrossRefMedline

44.

Cannon, G. W., R. D. Emkey, A. Denes, S. A. Cohen, P. A. Saway, F. Wolfe, A. M. Jaffer,

A. L. Weaver, B. J. Manaster, K. A. McCarthy. 1993. Prospective 5-year followup of
recombinant interferon- in rheumatoid arthritis. J. Rheumatol. 20: 1867-1873.