Individual differences in behavior of inbred Lewis rats are associated with severity of joint destruction in adjuvant-induced arthritis.

Abstract

The aim of our study was to test the hypothesis that differences in behavioral characteristics are linked to severity of arthritis in association with neuro-endocrine and immune reactivity in an inbred strain of rats. Lewis rats were selected as high-active (HA) and low-active (LA) animals based on their exploratory activity in the open field. Subsequently, adjuvant-arthritis (AA) was induced in both groups. We observed no differences in the severity of inflammation as determined by paw swelling and redness. However, LA and HA animals differed in the severity of bone destruction as determined on radiographs taken on day 30 after induction of AA. LA rats had more osteoporosis, periostal new bone formation, and bone destruction than HA rats. There were no differences between HA and LA rats in corticosterone response after acute or chronic immune challenge. Splenocytes of LA rats had a lower mitogen-induced IL-10 and IFNgamma production during AA. Histological examination revealed more intense factor VIII staining in arthritic joints of LA animals, indicating more pronounced synovial angiogenesis. In addition, LA rats had higher plasma VEGF, an important angiogenic factor. Expression of RANKL, a crucial factor promoting bone resorption, was also higher in joints of LA animals. Our data demonstrate that activity in the open field, a behavioral trait, is associated with the severity of bone destruction in AA. Lower production of bone-protective cytokines and a higher rate of angiogenesis leading to more synovial proliferation may be responsible for the more severe joint destruction in LA animals.

Full text

Page 1

Individual differences in behavior of inbred Lewis rats are
associated with severity of joint destruction in
adjuvant-induced arthritis
Enik€ o Sajti,a,bNico van Meeteren,b,cAnnemieke Kavelaars,aJanjaap van der Net,d
Willem Hendrik Gispen,band Cobi Heijnena,*
aLaboratory for Psychoneuroimmunology, University Medical Center Utrecht, The Netherlands
bRudolf Magnus Institute for Neurosciences, University Medical Center Utrecht, The Netherlands
cDepartment of Physiotherapy, Academy of Health Sciences Utrecht, The Netherlands
dDepartment of Pediatric Physiotherapy, University Children Hospital, ?Het Wilhelmina Kinderziekenhuis?,
University Medical Center Utrecht, The Netherlands
Received 22 October 2003; received in revised form 26 November 2003; accepted 1 December 2003
Available online 28 January 2004
Abstract
The aim of our study was to test the hypothesis that differences in behavioral characteristics are linked to severity of arthritis in
association with neuro-endocrine and immune reactivity in an inbred strain of rats. Lewis rats were selected as high-active (HA) and
low-active (LA) animals based on their exploratory activity in the open field. Subsequently, adjuvant-arthritis (AA) was induced in
both groups. We observed no differences in the severity of inflammation as determined by paw swelling and redness. However, LA
and HA animals differed in the severity of bone destruction as determined on radiographs taken on day 30 after induction of AA.
LA rats had more osteoporosis, periostal new bone formation, and bone destruction than HA rats. There were no differences
between HA and LA rats in corticosterone response after acute or chronic immune challenge. Splenocytes of LA rats had a lower
mitogen-induced IL-10 and IFNc production during AA. Histological examination revealed more intense factor VIII staining in
arthritic joints of LA animals, indicating more pronounced synovial angiogenesis. In addition, LA rats had higher plasma VEGF, an
important angiogenic factor. Expression of RANKL, a crucial factor promoting bone resorption, was also higher in joints of LA
animals. Our data demonstrate that activity in the open field, a behavioral trait, is associated with the severity of bone destruction in
AA. Lower production of bone-protective cytokines and a higher rate of angiogenesis leading to more synovial proliferation may be
responsible for the more severe joint destruction in LA animals.
? 2004 Elsevier Inc. All rights reserved.
Keywords: Joint destruction; Behavior; Lewis rat; Cytokines; HPA-axis; Open field
1. Introduction
Rheumatoid arthritis (RA) is an autoimmune dis-
order characterized by synovial proliferation and in-
flammation, and subsequent destruction and deformity
of joints. Skeletal complications start with focal erosion
of cartilage followed by marginal and subchondral
bone loss. Extended joint destruction with ankylosis
and generalized bone loss are characteristic for late
complications (Feldmann et al., 1996). These long-term
skeletal complications have serious consequences as
they can lead not only to painful joint deformities but
also to progressive functional disability and increased
mortality rates (Pincus and Callahan, 1993). Joint de-
struction is characterized by pathological bone re-
sorptionduetoadisturbed
remodeling. Osteoclasts, the bone-resorbing cells, take
the forehand on bone forming osteoblasts (Goldring
and Gravallese, 2000). Moreover, the excessive ex-
pression of receptor activator of nuclear factor jB li-
balanceinbone
*Corresponding author. Fax: +31-30-2505311.
E-mail address: c.heijnen@azu.nl (C. Heijnen).
0889-1591/$ - see front matter ? 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.bbi.2003.12.004
Brain, Behavior, and Immunity 18 (2004) 505–514
BRAIN,
BEHAVIOR,
and IMMUNITY
www.elsevier.com/locate/ybrbi

Page 2

gand (RANKL) by inflammatory T-cells and synovial
fibroblasts has a critical function in this process.
RANKL interacts with its receptor on mononuclear
phagocyte precursors and directs their differentiation
towards osteoclasts (Kong et al., 1999a,b).
Not only RANKL, but also cytokines can modulate
osteoclastogenesis. IFNc is a cytokine well known as a
protector against bone destruction suppressing potently
osteoclast development (Takayanagi et al., 2002a,b).
The synovium depends on blood supply in order to
proliferate and to invade the cartilage–bone junction.
Therefore angiogenesis, the formation of new blood
vessels, is considered to be a crucial step in pannus
formation and the subsequent destruction of cartilage
and bone. Vascular endothelial cell growth factor
(VEGF) is one of the most important angiogenic fac-
tors. It has the ability to trigger the entire sequence of
events leading to new vessel formation (Folkman, 1995;
Nagashima et al., 1995). VEGF can be produced in the
inflamed joint by several cell types: activated synovial
monocytes/macrophages and fibroblasts as well as by
synoviocytes. It is detectable in both synovial fluid and
plasma of arthritis patients and is correlated with the
course of the disease (Harada et al., 1998; Kikuchi et al.,
1998).
The occurrence and severity of late skeletal compli-
cations varies markedly among patients. We wanted to
investigate whether individual behavioral characteristics
can predict disease susceptibility or outcome. There are
a few reports available which describe a relation between
differences in behavior and the severity of inflammatory
autoimmune diseases. We have demonstrated earlier
that the behavioral reactivity of wild-type rats measured
in the resident-intruder test as the attack latency time
(ALT), was associated with EAE severity (Kavelaars
et al., 1999). Passive animals selected as rats with max-
imal ALT had less severe EAE symptoms compared to
their aggressive counterparts with a short ALT. A study
by Chover-Gonzalez investigating the severity of AA in
out-bred Wistar rats selected on the basis of their be-
havior in the learned helplessness paradigm demon-
strated that helpless rats showed significantly less paw
inflammation compared to their not helpless counter-
parts (Chover-Gonzalez et al., 2000). However, these
studies used either wild-type or outbred rats, therefore a
possible selection for genetic differences in inflammatory
properties can not be ruled out.
The purpose of the present study was to examine
whether pre-existing individual differences in behavior
of Lewis rats are associated with the severity of in-
flammation and bone destruction in AA. Moreover, we
investigated whether HA and LA rats differ with re-
spect to HPA-axis reactivity, cytokine production,
plasma level of the angiogenic factor VEGF as well as
the extent of angiogenesis and synovial proliferation in
the joint.
2. Materials and methods
The Institutional Animal Use and Care Committee of
the University Medical Center Utrecht approved all
procedures in this study involving animals.
2.1. Animals
Inbred female Lewis rats were obtained from the
University of Limburg (Maastricht, The Netherlands).
The animals were three to four weeks old at the time of
arrival. The rats were housed individually in type 3
Macrolon cages in a temperature (21?C) and humidity-
controlled room where a 12h light/dark cycle (lights on
7 AM) was maintained. Food and water was available
ad libitum.
2.2. Experimental design
Three experiments with identical behavioral selection-
procedures were executed. In these experiments rats were
pre-selected upon their behavior in the open field and
assigned to two extreme groups, HA and LA animals,
respectively. Two of the experiments were performed for
the exploration of differences in the development of ar-
thritis and its long term consequences between HA and
LA rats. The third experiment was designed in order to
characterize the cytokine production patterns and HPA-
axis reactivity of healthy HA and LA rats.
2.3. Open field test
Locomotor activity of 5–6-week-old rats was assessed
in an open field as described earlier (Stam et al., 1997;
van Meeteren et al., 1997) during the first half of the
light phase. Animals were handled for four consecutive
days before the start of the experiments. The open field
consisted of a black round polyester arena with a di-
ameter of 130cm and a 50-cm high border equally illu-
minated by a white TL tube. Prior to the start of the test,
rats were handled four times by the researcher. The
animals were tested individually on four consecutive
days during the light period of the diurnal cycle with the
radio on for constant background noise. They were
gently put in the middle of the open field and allowed to
explore the arena for 4min. Ambulation was automat-
ically registered with a computer system (Noldus In-
formation Systems, Wageningen, The Netherlands).
After the test, data were analyzed by Etho Vision
tracking software calculating covered distances for each
day. As a selection-criterion the sum of the distances
covered on day two, three, and four was chosen con-
forming to Denenberg (1969). Rats with scores in the
lower quartile were termed LA, that is they exhibited
low general activity. In contrast, animals with scores in
the upper quartile were selected as HA animals.
506
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514

Page 3

2.4. Induction and clinical evaluation of the arthritis
model
Four days after open field selection HA and LA
rats were pre-immunized according to the protocol of
Zhang et al. (1999) with the following modifications:
rats were pre-immunized with 100ll (instead of 50ll)
of squalane (Sigma–Aldrich, Steinheim, Germany)
intradermally into the base of the tail instead of the
footpad.
After two weeks, arthritis was induced by an intra-
dermal injection of 100ll of Freund?s complete adjuvant
containing 5mg of heat-killed Mycobacterium tubercu-
losis (strain H37Ra; Difco, Detroit, MI) in 1ml Freund?s
incomplete adjuvant (Difco) in the base of the tail
(Pearson, 1956). Rats were observed and examined daily
in a blinded fashion for clinical signs of arthritis using
standard methods. Each paw was scored from 0 to 4 for
the degree of erythema, swelling, and joint deformity.
The highest achievable score is 16 per animal (Trentham
et al., 1977).
All time points were considered in relation to the AA
induction day, designated as day 1. Arthritis in one of
the joints was developed in 100% of the treated animals
by day 11.
2.5. Radiological studies
Onday30afterimmunizationwithFreund?sComplete
Adjuvant (CFA), rats were sacrificed by decapitation,
and the hind paws were examined radiographically. Joint
radiographs were taken at 10mA, 50kV, and 0,25s
of exposure to Kodak Insight occlusal film. The paws
were radiographed at a 25cm focus to film distance. Ra-
diographs were evaluated in a blinded manner, using the
grading scale described by Ackerman et al. (1979).
2.6. Histology
Whole ankle joints of the hind paws were fixed in
formaline for at least 2 days before decalcification with
EDTA for a period of three weeks. After decalcification,
the joints were imbedded in paraffin, serially sectioned
onto microscope slides, and then deparaffinized. Tissue
sections (7lm) were stained with haematoxylin–eosine.
Immunohistochemicalstaining
RANKL and factor VIII. For RANKL, tissue sections
were treated with 3% H2O2 for 20min at room tem-
perature. Sections were incubated for 20min with
10mM citrate (pH 6.0) at 100?C and incubated again
with the primary antibody directed against RANKL
(rabbit polyclonal antibody raised against the epitope
corresponding to amino acids 46–317 of RANKL of
human origin [FL-317], Santa Cruz Biotechnology,
Santa Cruz, CA). Rabbit IgG antibody (Dako X0936;
Dako, Carpinteria, CA) was used as an isotype control.
wasperformed for
After rinsing, sections were incubated for 30min with
biotinylated goat anti-rabbit horseradish peroxidase
(Dako), developed with 3;30-diaminobenzidine (Sigma,
St. Louis, MO) and counterstained with haematoxylin.
For vascular endothelial specific staining, sections were
stained with anti-human factor VIII antibody.
2.7. HPA axis activation
Nine-week-old HA and LA rats were i.p. injected at 9
AM with 1lg/kg, 2.5lg/ml LPS (E. coli, Sigma, St.
Louis, MO). The control groups were injected with
equal volumes of saline. Rats were decapitated 120min
after the injection.
2.8. Corticosterone assay
After decapitation trunk blood was collected in
EDTA-containing tubes on ice, immediately centrifuged
(3000rpm, 10min, 4?C) and stored ()20?C) until assay.
Plasma corticosterone levels were determined by radio-
immuno assay as described previously (Nijsen et al.,
2000). The sensitivity was 0.4ng/ml; the intra- and in-
terassay variations were 4 and 7%, respectively. Plasma
corticosterone concentrations are expressed as lg/dl.
2.9. Peritoneal macrophages isolation and stimulation
After decapitation, 20ml of ice-cold RPMI-1640
(Gibco, Grand Island, NY) was injected into the peri-
toneum, and peritoneal macrophages were harvested
after 5min of gentle massage of the abdomen. Cells were
cultured in 24-well plates (5?105cell/well) in the pres-
ence of varying concentrations of LPS. Culture super-
natants were collected after 24h of incubation at 37?C
and stored at )80?C until assay.
2.10. Splenocyte stimulation
After decapitation, spleens were removed, mashed
gently through filter chambers (NPBI, Emmer-Compa-
scum, The Netherlands). Cells were resuspended in
culture medium (RPMI-1640,) supplemented with 5%
fetal calf serum (Gibco, Grand Island, NY), 2mM L L-
glutamine, 100U/ml penicillin, 100lg/ml streptomycin,
100lg/ml gentamycin, and 50lM 2-mercaptoethanol.
Splenocytes were cultured in quadruplicate in 200ll
round-bottom microtiter wells (Costar, Cambridge,
MA) at 105cells/well in the presence of 1lg/ml Conca-
navalin A (Calbiochem, La Jolla, CA) and different
concentrations of dexamethasone. Cultures were incu-
bated for 96h at 37?C in a humidified atmosphere, 5%
CO2. For the final 18h cultures were pulsed with 1lCi/
well [3H]thymidine (Amersham, Bucks, UK). Thymidine
uptake was measured using a liquid scintillation b
counter.
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514
507

Page 4

2.11. TNFa, IL-10, and IFNc production
Supernatants of Concanavalin A stimulated spleno-
cytes were collected after 48h of culture. IL-10 and
IFNc content was determined by ELISA (U-Cytech,
Utrecht, The Netherlands) according to the manufac-
turer?s procedure.
Supernatants of LPS stimulated peritoneal macro-
phages were collected after 24h of culture. IL-10 and
TNFa was determined by ELISA (U-Cytech, Utrecht,
The Netherlands) according to the manufacturer?s
procedure.
2.12. Statistics
Clinical course of AA, cytokine production by peri-
toneal macrophages and dexamethasone inhibition of
splenocyte proliferation were evaluated with GLM re-
peated measures. Differences in X-ray scores, plasma
VEGF, and splenocyte cytokines were tested with Stu-
dent?s T test.
Plasma corticosterone levels after LPS injection were
analyzed with two-way ANOVA. Differences in plasma
corticosterone levels in arthritic rats were analyzed with
Student?s T test.
p < :05 was considered statistically significant. Sta-
tistical analysis was performed using SPSS, version 10.
3. Results
3.1. Selection of HA and LA animals
At six weeks of age rats were subjected to an open
field test to study general locomotor activity as de-
scribed in Section 2. The total distance covered by each
rat was calculated for each test day. As a selection cri-
terion we have chosen for the sum of the distances
covered on day two, three, and four. The activity scores
from test day two onward are stable for each individual
animal (Denenberg, 1969). Rats with scores in the upper
quartile were assigned to the HA group whereas rats
with scores in the lower quartile to the LA group. Fig. 1
shows the scores of a representative selection procedure.
The 30 highest and lowest active animals of the tested
120 rats were used for further experiments. The rest of
the animals, the middle group, were used in other ex-
periments.
3.2. Clinical course of adjuvant arthritis in HA and LA
rats
As shown in Fig. 2A we did not observe any differ-
ence in the severity of AA, as assessed by the clinical
score based on inflammatory signs like joint swelling,
redness, and limitation in joint motion. In particular,
there was no difference in kinetics or intensity in onset,
peak or the late phase of the disease. Furthermore, the
groups had a similar weight loss during the entire course
of arthritis (Fig. 2B).
Fig. 1. Selection of HA and LA rats. HA rats with scores in the upper
quartile and LA rats with scores in the lower quartile. The Y-axis rep-
resentsthe sumofthe distancescoveredontestdaytwo,three,andfour.
Fig. 2. (A) Clinical course of AA. After preimmunization with squa-
lane, AA was induced by intradermal injection of Complete Freund?s
Adjuvant (CFA) containing 5mg/ml Mycobacterium tuberculosis into
the base of the tail. Rats were examined daily in a blinded protocol for
clinical signs of disease. HA rats (m), LA rats (j). (B) Mean weight
gain as a percentage of the body weight on day 11. HA rats (m), LA
rats (j). Data represent means and SEM. N ¼ 16 rats per group.
GLM, repeated measures, p ¼ ns.
508
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514

Page 5

3.3. Long-term bone destruction in HA and LA rats
Despite a similar clinical course of arthritis, LA rats
suffered from more pronounced bone destruction than
HA animals as seen on radiographs taken on day 30 of
the disease when almost no active inflammatory signs
were visible anymore. The differences in the total X-ray
scores, assessing osteoporosis, periostal new bone for-
mation, erosions, and swelling are depicted in Fig. 3A.
LA animals had a more pronounced decrease in bone-
density, destruction of bony structures, and ossifications
not contiguous with the normal bone-line. The differ-
ences in bone destruction between HA and LA animals
are also evident on the two radiographs representing the
ankle of a rat with a representative score from each
group (Figs. 3C and D). The presence of marked bone
resorption in combination with adjacent new bone for-
mation suggested strongly the presence of a disordered
pattern of bone remodeling in the LA animals.
3.4. Histology
In haematoxylin–eosin stained sections of ankle joints
of LA rats the normal tissue architecture was completely
lost. A dramatic joint-space loss was visible with the
presence of a massive proliferative and erosive pannus.
The cartilage lining the bones was severely eroded. In
contrast, sections of ankle joints of HA animals revealed
significantly less severe lesions that were mainly re-
stricted to the lateral parts of the joints. The integrity of
the cartilage was preserved in the HA animals, especially
in the centers of the joint surfaces. Furthermore, the HA
rats exhibited markedly less cortical and trabecular bone
loss (Fig. 4A vs. D).
3.5. RANKL and vascularization of the arthritic joint
RANKL has a critical role in bone remodeling pro-
moting bone resorption. Specific immuno-histochemical
staining for RANKL showed an abundant presence of
RANKL in ankle joints of LA animals in contrast to the
joints of HA animals (Fig. 4B vs. E).
Angiogenesis is known to play an important role in
the proliferation of the bone-destroying pannus. To in-
vestigate the level of vascularization in the joint, we
stained slides for the endothelial marker factor VIII. As
is shown in Fig. 4F, a more intensive staining for factor
VIII was observed in the ankles of LA rats, suggesting a
more pronounced vascularization of the invasive pannus
compared to the ankles of HA rats (Fig. 4C).
One of the crucial angiogenic factors produced in
arthritis is VEGF. We measured VEGF in the plasma of
HA and LA animals on day 30 after induction of AA. In
Fig. 5 we demonstrate that HA animals (with a better
bone integrity) had significantly lower VEGF levels as
compared to the LA rats.
3.6. Cytokine profiles of HA and LA rats
To investigate whether cytokine patterns were differ-
ent in HA and LA animals we measured the capacity of
splenocytes to produce IL-10 and IFNc and the capacity
of peritoneal macrophages to produce TNF-a and IL-
10. At baseline conditions there was no difference in
mitogen-induced splenocyte in vitro IL-10 production
between HA and LA animals. However, after AA, IL-10
production by splenocytes from HA rats was signifi-
cantly higher than from LA rats (Fig. 6A). The capacity
to produce IFNc was higher in HA animals both before
the onset and after AA (Fig. 6B).
To investigate whether peritoneal macrophages of the
two groups of animals differed in their capacity to pro-
duce cytokines, peritoneal macrophages of HA and LA
animals were in vitro stimulated with different concen-
trations of LPS. In Figs. 7A and B it is shown that LPS
induces TNF-a and IL-10 production dose dependently,
but no differences between the HA and LA animals were
observed.
Fig. 3. Bone destruction in the ankles of HA and LA rats. Radiographs taken on day 30 after induction of AA were evaluated using the grading scale
described by Ackerman et al. assessing osteoporosis, periostal new bone formation, erosions and swelling (A). Data represent the mean and SEM.
N ¼ 16 rats per group. tð24Þ ¼ 3:2; p < :01, **p < :01. Radiographs of the hindlimb of a control rat (B), HA rat (C), and LA rat (D). Radiographs
were taken on day 30 after induction of AA.
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514
509

Page 6

3.7. Plasma corticosterone levels in HA and LA rats
The HPA-axis is known to play an important role
in interfering with immune processes by secreting
corticosterone. We tested the reactivity of the HPA-
axis in HA and LA animals upon an acute immune
challenge as well as after the induction of the chronic
inflammatory stress of AA. As an acute immune
stimulus we used a low dose of LPS. Healthy animals
were injected i.p. with 1lg/kg LPS and decapitated
after 120min, at the peak of the corticosterone re-
sponse. As shown in Fig. 8, LPS induced an increase
in plasma corticosterone concentration in both groups
compared to the saline injections (two-way ANOVA,
significant treatment effect, p < :001) but there were no
significant differences in the magnitude of the response
between the HA and LA rats (two-way ANOVA, in-
teraction effect not significant). Moreover, the plasma
levels of corticosterone of HA and LA animals 30 days
after induction of the disease were similar as well
(Fig. 8).
3.8. Glucocorticoid receptor sensitivity in HA and LA rats
To test if an altered communication between corti-
costerone and immune cells could contribute to the
observed differences in bone destruction in HA and LA
animals, we investigated the glucocorticoid receptor
sensitivity of splenocytes by co-incubating the cells with
different concentrations of dexamethasone in a mitogen-
induced proliferation assay. Dexamethasone is known
to inhibit the proliferation of cells dose-dependently.
However, no differences were found in the slopes of the
dexamethasone inhibition curves of proliferating cells of
the two groups of rats (Fig. 9).
Fig. 4. Paraffin imbedded sections of ankle joints of HA and LA rats on day 30 after induction of AA. (A) and (D) Haematoxylin–eosin staining
(magnification 5?). Note the complete joint-space loss with the presence of a massive proliferative and erosive pannus in the LA animals (B). In HA
animals lesions were restricted to the lateral parts of the joints. The integrity of the cartilage was preserved, especially in the centers of the joint
surfaces (A). (B) and (E) Immunohistochemistry for RANKL (magnification 20?) LA animals (E) show high expression of RANKL in the invasive
pannus (arrow heads). In the HA (B) animals RANKL expression in the cartilage (arrow heads). (C) and (F) immunohistochemistry for factor VIII
(magnification 20?). A more intensive staining in LA animals (F), indicating the presence of more blood vessels (arrow heads) in the pannus as
compared to the HA rats (C).
Fig. 5. Plasma VEGF measured on day 30 after the induction of AA.
Data represent means and SEM. N ¼ 6 rats per group. tð10Þ ¼ 2:3;
p < :05, *p < :05.
510
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514

Page 7

4. Discussion
One important obstacle in the treatment of RA is the
limited ability to select and to recognize those patients
who will develop severe long-term complications. In the
current study we show that a pre-existing behavioral
characteristic is associated with the severity of bone
destruction in an animal model of arthritis.
Individual differences in behavioral profiles are driven
by genetic and environmental factors. In contrast to
most behavioral studies, we used an inbred strain of
rats. This approach provided the possibility to reduce
the contribution of genetic background to the relation
between behavior and disease outcome. It should be
noted, however, that we cannot exclude that genetic
differences between the animals contribute to our find-
ings. We demonstrated that rats selected upon a be-
havioral characteristic in the open field, prior to the
induction of arthritis, show marked differences in bone
destruction in the late stage of the disease.
The behavior of a rat in the open field embraces
several aspects. Denenberg clearly demonstrated that
the activity measured on the first test day differs from
the activity measured on the second, third, and fourth
days. Subsequent factor analysis led him to suggest that
a different behavioral dimension is measured by the
activity on day one as compared to the activity of the
animal measured on consecutive test days (Denenberg,
1969). On the first day the novelty aspect predominates,
therefore the activity mirrors a fear response that
Fig. 7. Cytokine production by peritoneal macrophages of HA rats (m)
and LA rats (j). Isolated peritoneal macrophages were stimulated
with different concentrations of LPS in vitro. After 24h of culture Il-10
(A) and TNFa (B) were measured in the supernatants by ELISA.
GLM, repeated measures, p ¼ ns.
Fig. 6. (A) IL-10 production of Concanavalin A stimulated splenocytes
before the induction of AA and on day 30 of AA. tð8Þ ¼ 4:8; p < :01.
(B) IFNc production of Concanavalin A stimulated splenocytes before
the induction of AA and on day 30 after induction of AA.
tð20Þ ¼ 2:2; p < :05 and tð8Þ ¼ 3:2; p < :05 for basal conditions and
for arthritic groups, respectively. Data represent means and SEM.
N ¼ 11 rats per group for the basal conditions and N ¼ 5 for the ar-
thritis groups. *p < :05; **p < :01.
Fig. 8. HPA-axis reactivity. Healthy HA and LA rats were injected i.p.
with 1lg/kg LPS or saline and decapitated after 120min. Two-way
ANOVA, interaction effect, p ¼ ns, treatment effect, p < :001. Basal
corticosterone in the plasma of HA and LA animals on day 30 after
induction of AA was measured as well. Student?s T test, p ¼ ns.
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514
511

Page 8

describes the emotional reactivity of the animal and can
be regarded as a state parameter. The activity of the first
day is positively correlated with defecation, another
frequently used parameter for emotional reactivity in the
open field. In contrast, activity on consecutive days is
negatively correlated with both activity on day one and
defecation. It represents an independent and stable be-
havioral dimension that could be interpreted as a trait
characteristic.
A study by Chover-Gonzalez et al. (1999) examining
the associations between behavioral dimensions de-
scribing open field emotional reactivity and AA disease
activity reports no correlation between either activity or
number of deposited fecal pellets during the first expo-
sure to the open field and AA susceptibility or severity.
In our study as well, trait was not associated with the
inflammatory component of arthritis in a way that HA
and LA animals did not differ with respect to the in-
flammatory signs of AA. However, our data show that
bone destruction—as reflected by osteoporosis, erosions
of bone, and cartilage loss—was profound in LA ani-
mals, whereas HA animals had only mild bone damage.
Furthermore, in contrast to HA animals, LA rats had
severe pathological bone proliferation and ankylosis
leading to substantial joint deformities. Moreover, we
recently described that using the same experimental set-
up, this trait characteristic was also correlated to tumor
growth, tumor metastasis and tumor angiogenesis (Sajti
et al., 2004). We propose therefore that trait character-
istics may represent stable behavioral dimensions that
may have an organizing influence on physiological
processes and thereby tune disease outcome.
The observed uncoupling of inflammation and joint
destruction in our data is in line with earlier reports on
disease characteristics in humans. Rheumatoid arthritis
patients having the same inflammatory score can present
with a destructive or non-destructive phenotype. In a
clinical study, Bresnihan et al. (1998) showed that
treatment with IL-1 receptor antagonist had beneficial
effects on bone destruction. The inflammatory process,
however, remained unaffected. Similar observations
have been made in rodents. IL-4 gene therapy in
experimental arthritis in mice has cartilage and bone-
protective effects despite ongoing inflammation (Lub-
berts et al., 1999, 2000). Another important finding
demonstrating complete abolishment of bone loss with
essentially no effect on inflammation is osteoproteregin
(OPG) treatment. Systemic administration of OPG, a
decoy receptor for RANKL, in arthritic rats was shown
to result in a marked reduction in cortical and trabec-
ular bone loss and a dramatic decrease in osteoclast
numbers without significant effects on joint inflamma-
tion (Kong et al., 1999b). OPG treatment did not reduce
joint inflammation quantitatively, implying that the in-
flammatory process associated with arthritis does not
depend on the RANK/RANKL pathway.
In our histological preparations, we observed a
marked difference in the expression of RANKL, an es-
sential factor in osteoclatogenesis and bone erosion, in
the two groups of rats. LA rats showed an abundant
expression of RANKL in the erosive pannus suggesting
that a difference in expression of RANKL could be a
potential mechanism explaining the differences in bone
damage between the two groups. The expression of
RANKL is under control of different cytokines. IL-10
and IFNc have been suggested to be bone-protective
regulatory mediators (Moore et al., 2001). IFNc sup-
pressesosteoclastogenesis
RANKL/RANK pathway thereby protecting bone for-
mation (Lubberts et al., 1999; Takayanagi et al., 2000).
However, IFNc is known to have a dual role in the in-
flammation accompanying arthritis, enhancing in early
stages whereas suppressing in later stages (Boissier et al.,
1995). We measured the production of these two cyto-
kines in splenocytes of our HA and LA rats. LA ani-
mals, with a pronounced joint damage, had lower levels
of IFNc both before and after the induction of arthritis.
Regarding IL-10, only the HA animals showed an in-
crease in the production of this protective cytokine upon
the induction of the disease. Although measured in a
distant immune compartment, these results suggest that
HA rats have a more favorable cytokine milieu for re-
tention of bone integrity thereby preventing long-term
skeletal complications. Furthermore, there is evidence
that cytokine production by splenocytes is closely re-
lated to disease activity in AA (Cobelens et al., 2002).
Another important mechanism, which could explain
the differences in bone erosion between the two groups,
is angiogenesis. The significantly higher plasma VEGF
concentration in LA animals strongly suggests that a
more pronounced angiogenesis contributes to the
byinterferingwith the
Fig. 9. Dexamethasone sensitivity of splenocytes of HA rats (m) and
LA rats (j). Splenocytes were incubated with different concentrations
of dexamethasone. After three days of culture proliferation was mea-
sured by [3H]thymidine incorporation. GLM, repeated measures,
p ¼ ns.
512
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514

Page 9

development of pronounced skeletal complications in
these rats. Furthermore, specific staining for factor VIII,
an endothelial marker, revealed more blood vessels in
the proliferating synovium of LA rats.
To dissect possible endocrine correlates we measured
the reactivity of the HPA-axis in our two groups of rats.
The results showed no significant differences between the
corticosterone responses of HA and LA animals, neither
upon a chronic nor acute immune stimulus. These
findings are in contrast with the hypothesis stating that
susceptibility to autoimmune disease occurs in the
presence of a hyporeactive HPA-axis (Sternberg et al.,
1989a,b). More recent studies, however, have shown
that in other rat strains, corticosterone is not the dom-
inating endocrine mediator. Harbuz et al. (2003) clearly
conclude from their studies that susceptibility for and
severity of autoimmune diseases are more complex
phenomena and depend on more than one parameter.
Our present results are in substantial agreement with
other behavioral studies suggesting a link between be-
havioral characteristics of the individual and subsequent
disease severity regardless HPA-axis reactivity. An ear-
lier study by our group (Kavelaars et al., 1999) dem-
onstrated that the latency of a rat to attack an intruder
was inversely correlated with the severity of experi-
mental autoimmune encephalomyelitis. A study by
Chover-Gonzalez et al. (2000) using the learned help-
lessness paradigm to divide single populations of rats
into subpopulations showed that learned helpless rats
develop paw swelling in AA sooner and with increased
severity compared to the not helpless rats. Strikingly, in
both studies a different corticosterone response to acute
stress was not predictive for disease severity.
The impact of the level of glucocorticoids on the in-
flammatory process will ultimately depend on the ca-
pacity of the specific receptors on peripheral tissues to
respond to circulating glucocorticoids. Therefore we
also tested the sensitivity of splenocytes to the synthetic
glucocorticoid analogue dexamethasone. However, no
differences were found between the groups in the sensi-
tivity of splenocytes to respond to dexamethasone,
suggesting no differences in the communication between
the HPA-axis and the immune system between the two
groups of animals.
Considering species differences, one should be cau-
tious in attempting to extrapolate these results to hu-
mans. However, studies on the course of RA have
provided a basis for an impact of psychological factors
on disease outcome. There is evidence that neuroticism,
a relatively stable personality trait, strongly influences
self-rated disease symptoms, and wellbeing in RA
(Persson and Sahlberg, 2002). Furthermore, neuroticism
at the moment of diagnosis of RA is significantly related
to psychological distress at 3- and 5-year follow-up
(Evers et al., 2002). Whether neuroticism is also related
to joint impairment has to be further investigated.
Joint impairment has a big impact on functional
ability; therefore the control of bone erosive processes is
one of the most challenging objectives in the treatment
of RA. The considerable heterogeneity (clinical, patho-
logical, and immunological) of the disease makes it im-
portant to tailor the treatment on the individual patient.
Reliable prediction of those at risk for joint destruction
will almost certainly translate into long-term prevention
of disability as patients are aggressively treated early in
their disease course.
Taken these data together, we would like to speculate
that stable behavioral traits might become of clinical use
to identify patients at high risk to develop severe joint
destruction in rheumatoid arthritis allowing early and
more effective treatment of the disease.
Acknowledgments
The authors are greatly indebted to Jan Brakee for his
help with the animal experiments and to Jitske Zijlstra
for excellent technical assistance.
References
Ackerman, N.R., Rooks, W.H., Shott, L., Genant, H., Maloney, P.,
West, E., 1979. Effects of naproxen on connective tissue changes in
the adjuvant arthritic rat. Arthritis Rheum. 22, 1365–1374.
Boissier, M.C., Chiocchia, G., Bessis, N., Hajnal, J., Garotta, G.,
Nicoletti, F., Fournier, C., 1995. Biphasic effect of interferon-
gamma in murine collagen-induced arthritis. Eur. J. Immunol. 25,
1184–1190.
Bresnihan, B., Alvaro-Gracia, J.M., Cobby, M., Doherty, M., Doml-
jan, Z., Emery, P., Nuki, G., Pavelka, K., Rau, R., Rozman, B.,
Watt, I., Williams, B., Aitchison, R., McCabe, D., Musikic, P.,
1998. Treatment of rheumatoid arthritis with recombinant human
interleukin-1 receptor antagonist. Arthritis Rheum. 41, 2196–2204.
Chover-Gonzalez, A.J., Harbuz, M.S., Tejedor-Real, P., Gibert-
Rahola, J., Larsen, P.J., Jessop, D.S., 1999. Effects of stress on
susceptibility and severity of inflammation in adjuvant-induced
arthritis. Ann. N. Y. Acad. Sci. 876, 276–286.
Chover-Gonzalez, A.J., Jessop, D.S., Tejedor-Real, P., Gibert-Rahola,
J., Harbuz, M.S., 2000. Onset and severity of inflammation in rats
exposed to the learned helplessness paradigm. Rheumatology
(Oxford) 39, 764–771.
Cobelens, P.M., Kavelaars, A., van der Zee, R., van Eden, W.,
Heijnen, C.J., 2002. Dynamics of mycobacterial HSP65-induced T-
cell cytokine expression during oral tolerance induction in adjuvant
arthritis. Rheumatology (Oxford) 41, 775–779.
Denenberg, V.H., 1969. Open-field behavior in the rat: what does it
mean? Ann. N. Y. Acad. Sci. 159, 852–859.
Evers, A.W., Kraaimaat, F.W., Geenen, R., Jacobs, J.W., Bijlsma,
J.W., 2002. Longterm predictors of anxiety and depressed mood in
early rheumatoid arthritis: a 3 and 5 year followup. J. Rheumatol.
29, 2327–2336.
Feldmann, M., Brennan, F.M., Maini, R.N., 1996. Rheumatoid
arthritis. Cell 85, 307–310.
Folkman, J., 1995. Angiogenesis in cancer, vascular, rheumatoid and
other disease. Nat. Med. 1, 27–31.
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514
513

Page 10

Goldring, S.R., Gravallese, E.M., 2000. Mechanisms of bone loss in
inflammatory arthritis: diagnosis and therapeutic implications.
Arthritis Res. 2, 33–37.
Harada, M., Mitsuyama, K., Yoshida, H., Sakisaka, S., Taniguchi, E.,
Kawaguchi, T., Ariyoshi, M., Saiki, T., Sakamoto, M., Nagata, K.,
Sata, M., Matsuo, K., Tanikawa, K., 1998. Vascular endothelial
growth factor in patients with rheumatoid arthritis. Scand. J.
Rheumatol. 27, 377–380.
Harbuz, M.S., Chover-Gonzalez, A.J., Jessop, D.S., 2003. Hypothal-
amo–pituitary–adrenal axis and chronic immune activation. Ann.
N. Y. Acad. Sci. 992, 99–106.
Kavelaars, A., Heijnen, C.J., Tennekes, R., Bruggink, J.E., Koolhaas,
J.M., 1999. Individual behavioral characteristics of wild-type rats
predict susceptibility to experimental autoimmune encephalomy-
elitis. Brain Behav. Immun. 13, 279–286.
Kikuchi, K., Kubo, M., Kadono, T., Yazawa, N., Ihn, H., Tamaki,
K.,1998.Serum concentrations
growth factor in collagen diseases. Br. J. Dermatol. 139, 1049–
1051.
Kong, Y.Y., Feige, U., Sarosi, I., Bolon, B., Tafuri, A., Morony, S.,
Capparelli, C., Li, J., Elliott, R., McCabe, S., Wong, T., Campa-
gnuolo, G., Moran, E., Bogoch, E.R., Van, G., Nguyen, L.T.,
Ohashi, P.S., Lacey, D.L., Fish, E., Boyle, W.J., Penninger, J.M.,
1999a. Activated T cells regulate bone loss and joint destruction in
adjuvant arthritis through osteoprotegerin ligand. Nature 402,
304–309.
Kong, Y.Y., Yoshida, H., Sarosi, I., Tan, H.L., Timms, E., Capparelli,
C., Morony, S., Oliveira-dos-Santos, A.J., Van, G., Itie, A., Khoo,
W., Wakeham, A., Dunstan, C.R., Lacey, D.L., Mak, T.W., Boyle,
W.J., Penninger, J.M., 1999b. OPGL is a key regulator of
osteoclastogenesis, lymphocyte development and lymph-node or-
ganogenesis. Nature 397, 315–323.
Lubberts, E., Joosten, L.A., Chabaud, M., Van Den, B.L., Oppers, B.,
Coenen-De Roo, C.J., Richards, C.D., Miossec, P., Van Den Berg,
W.B., 2000. IL-4 gene therapy for collagen arthritis suppresses
synovial IL-17 and osteoprotegerin ligand and prevents bone
erosion. J. Clin. Invest. 105, 1697–1710.
Lubberts, E., Joosten, L.A., Van Den, B.L., Helsen, M.M., Bakker,
A.C., van Meurs, J.B., Graham, F.L., Richards, C.D., Van Den
Berg, W.B., 1999. Adenoviral vector-mediated overexpression of
IL-4 in the knee joint of mice with collagen-induced arthritis
prevents cartilage destruction. J. Immunol. 163, 4546–4556.
Moore, K.W., de Waal, M.R., Coffman, R.L., O?Garra, A., 2001.
Interleukin-10 and the interleukin-10 receptor. Annu. Rev. Immu-
nol. 19, 683–765.
Nagashima, M., Yoshino, S., Ishiwata, T., Asano, G., 1995. Role of
vascular endothelial growth factor in angiogenesis of rheumatoid
arthritis. J. Rheumatol. 22, 1624–1630.
Nijsen, M.J., Croiset, G., Stam, R., Bruijnzeel, A., Diamant, M.,
de Wied, D., Wiegant, V.M., 2000. The role of the CRH type 1
receptor in autonomic responses to corticotropin- releasing
hormone intherat. Neuropsychopharmacology
399.
ofvascular endothelial
22,388–
Pearson, C.M., 1956. Development of arthritis, periarthritis and
periotitis in rats given adjuvants. Proc. Soc. Exp. Biol. Med. 91, 95–
101.
Persson, L.O., Sahlberg, D., 2002. The influence of negative illness
cognitions and neuroticism on subjective symptoms and mood in
rheumatoid arthritis. Ann. Rheum. Dis. 61, 1000–1006.
Pincus, T., Callahan, L.F., 1993. The ?side effects? of rheumatoid
arthritis: joint destruction, disability and early mortality. Br. J.
Rheumatol. 32 (Suppl. 1), 28–37.
Sajti, E., Kavelaars, A., van Meeteren, N., Teunis, M., Gispen, W.H.,
Heijnen, C.J., 2004. Tumor angiogenesis and metastasis formation
are associated with individual differences in behavior of inbred
Lewis rats. Brain Behav. Immun. in press.
Stam, R., Croiset, G., Akkermans, L.M., Wiegant, V.M., 1997.
Behavioural and intestinal responses to novelty in rats selected for
diverging reactivity in the open field test. Behav. Brain Res. 88,
231–238.
Sternberg, E.M., Hill, J.M., Chrousos, G.P., Kamilaris, T., Listwak,
S.J., Gold, P.W., Wilder, R.L., 1989b. Inflammatory mediator-
induced hypothalamic–pituitary–adrenal axis activation is defective
in streptococcal cell wall arthritis-susceptible Lewis rats. Proc.
Natl. Acad. Sci. USA 86, 2374–2378.
Sternberg, E.M., Young III, W.S., Bernardini, R., Calogero, A.E.,
Chrousos, G.P., Gold, P.W., Wilder, R.L., 1989a. A central
nervous system defect in biosynthesis of corticotropin-releasing
hormone is associated with susceptibility to streptococcal cell wall-
induced arthritis in Lewis rats. Proc. Natl. Acad. Sci. USA 86,
4771–4775.
Takayanagi, H., Kim, S., Matsuo, K., Suzuki, H., Suzuki, T., Sato, K.,
Yokochi, T., Oda, H., Nakamura, K., Ida, N., Wagner, E.F.,
Taniguchi, T., 2002a. RANKL maintains bone homeostasis
through c-Fos-dependent induction of interferon-beta. Nature
416, 744–749.
Takayanagi, H., Kim, S., Taniguchi, T., 2002b. Signaling crosstalk
between RANKL and interferons in osteoclast differentiation.
Arthritis Res. 4 (Suppl. 3), S227–S232.
Takayanagi, H., Ogasawara, K., Hida, S., Chiba, T., Murata, S., Sato,
K., Takaoka, A., Yokochi, T., Oda, H., Tanaka, K., Nakamura,
K., Taniguchi, T., 2000. T-cell-mediated regulation of osteoclasto-
genesis by signalling cross-talk between RANKL and IFN-gamma.
Nature 408, 600–605.
Trentham, D.E., Townes, A.S., Kang, A.H., 1977. Autoimmunity to
type II collagen an experimental model of arthritis. J. Exp. Med.
146, 857–868.
van Meeteren, N.L., Brakkee, J.H., Helders, P.J., Croiset, G., Gispen,
W.H., Wiegant, V.M., 1997. Recovery of function after sciatic
nerve crush lesion in rats selected for diverging locomotor activity
in the open field. Neurosci. Lett. 238, 131–134.
Zhang, L., Mia, M.Y., Zheng, C.L., Hossain, M.A., Yamasaki, F.,
Tokunaga, O., Kohashi, O., 1999. The preventive effects of
incomplete Freund?s adjuvant and other vehicles on the develop-
ment of adjuvant-induced arthritis in Lewis rats. Immunology 98,
267–272.
514
E. Sajti et al. / Brain, Behavior, and Immunity 18 (2004) 505–514