ARTHRITIS & RHEUMATISM
Vol. 64, No. 5, May 2012, pp 1359–1368
© 2012, American College of Rheumatology
The Loss of ?2?1 Integrin Suppresses Joint Inflammation
and Cartilage Destruction in Mouse Models of
Marvin A. Peters,1Doreen Wendholt,1Simon Strietholt,1Svetlana Frank,1
Noreen Pundt,1Adelheid Korb-Pap,1Leo A. B. Joosten,2Wim B. van den Berg,2
George Kollias,3Beate Eckes,4and Thomas Pap1
Objective. Integrin ?2?1 functions as a major
receptor for type I collagen on different cell types,
including fibroblasts and inflammatory cells. Although
in vitro data suggest a role for ?2?1 integrin in regu-
lating both cell attachment and expression of matrix-
degrading enzymes such as matrix metalloproteinases
(MMPs), mice that lack the ?2 integrin subunit
(Itga2?/?mice) develop normally and are fertile. We
undertook this study to investigate the effect of Itga2
deficiency in 2 different mouse models of destructive
arthritis: the antigen-induced arthritis (AIA) mouse
model and the human tumor necrosis factor ? (TNF?)–
transgenic mouse model.
Methods. AIA was induced in the knee joints of
Itga2?/?mice and wild-type controls. Human TNF–
transgenic mice were crossed with Itga2?/?mice and
were assessed clinically and histopathologically for
signs of arthritis, inflammation, bone erosion, and
cartilage damage. MMP expression, proliferation, fibro-
blast attachment, and ERK activation were determined.
Results. Under arthritic conditions, Itga2 defi-
ciency led to decreased severity of joint pathology.
Specifically, Itga2?/?mice showed less severe clinical
symptoms and dramatically reduced pannus formation
and cartilage erosion. Mice lacking ?2?1 integrin ex-
hibited reduced MMP-3 expression, both in their sera
and in fibroblast-like synoviocytes (FLS), due to im-
paired ERK activation. Further, both the proliferation
and attachment of FLS to cartilage were partially
dependent on ?2?1 integrin in vitro and in vivo.
Conclusion. Our findings suggest that ?2?1 in-
tegrin contributes significantly to inflammatory carti-
lage destruction by promoting fibroblast proliferation
and attachment and MMP expression.
Rheumatoid arthritis (RA) is a common inflam-
matory autoimmune disease. It is characterized by syno-
vial hyperplasia, cell activation, inflammation, and inva-
sion of the synovium, all of which ultimately lead to the
destruction of articular structures. Increasing experi-
mental evidence suggests that fibroblast-like synovio-
cytes (FLS) contribute prominently to the progression of
the disease by attaching to, invading, and degrading
cartilage (1). They show features of stable activation
resulting in increased growth, aggressive behavior, and
the production of proinflammatory cytokines and
matrix-degrading enzymes (2).
Integrins are ?/? heterodimeric transmembrane
receptors that mediate cell adhesion to the surrounding
extracellular matrix (ECM) (3). In addition, integrin
signaling is critical for different cellular functions, such
as modulation of cell growth and proliferation, as well as
inflammatory responses (3–5). Integrin-mediated signal
transduction is crucial in a variety of conditions that are
associated with altered attachment of fibroblast-like
cells to the ECM. In RA, the attachment of FLS to the
ECM is an important initiating step in the progressive
destruction of articular cartilage (6), and integrins have
been identified as critical receptors for ECM molecules
Supported by the DFG (grant SFB 829 to Dr. Eckes and grant
SFB492 to Dr. Pap).
1Marvin A. Peters, PhD, Doreen Wendholt, PhD, Simon
Strietholt, MSc, Svetlana Frank, Dipl-Ing, Noreen Pundt, PhD, Adel-
heid Korb-Pap, MD, Thomas Pap, MD: University of Muenster,
Muenster, Germany;2Leo A. B. Joosten, PhD, Wim B. van den Berg,
PhD: Radboud University Nijmegen Medical Centre, Nijmegen, The
Netherlands;3George Kollias, PhD: Alexander Fleming Biomedical
Sciences Research Center, Vari, Greece;4Beate Eckes, PhD: Univer-
sity of Cologne, Cologne, Germany.
Address correspondence to Thomas Pap, MD, Institute of
Experimental Musculoskeletal Medicine, University Hospital Muen-
ster, Domagkstrasse 3, D-48149 Muenster, Germany. E-mail:
Submitted for publication December 1, 2010; accepted in
revised form November 9, 2011.
in the disease (7). Thus, increased expression of ?1
integrins on RA FLS has been associated with their
enhanced binding to ECM (8), and Wang et al showed
that the invasion of FLS into the ECM could be blocked
by antibodies against ?1 integrins (9). Furthermore, it
was demonstrated that the proliferation of FLS is regu-
lated by various integrins (10) and that adhesion to the
ECM triggers FLS proliferation and matrix metallopro-
teinase (MMP) expression (10–13).
In this context, ?2?1 integrin is of special interest
because it functions as a major receptor for type I
collagen on a number of different cells, including fibro-
blasts (14). It has been implicated in matrix remodeling
(e.g., in the formation and turnover of focal adhesions
[15,16] or in the induction and activation of MMPs
[17–20]). Furthermore, ?2?1 integrin plays a role in
inflammatory processes and cell invasion (21–23). En-
gagement of ?2?1 integrin triggers different down-
stream signaling pathways, such as the serine/threonine
kinase Akt (24) and the MAPK/ERK or p38 pathways
(14,25,26), and regulates Fas-mediated apoptosis (25).
Consequently, ?2?1 integrin may act synergistically with
inflammatory cytokines such as tumor necrosis factor ?
(TNF?) and interleukin-1? (IL-1?) that regulate MMP
production through similar pathways. This is of interest
because under physiologic conditions, the loss of the ?2
integrin subunit has only limited effects. Thus, ?2
integrin–deficient mice are viable and fertile with mild
platelet dysfunction (27–29) and abnormal angiogenesis
(30,31), indicating a backup system of functional com-
pensation by other collagen-binding integrins.
Based on these data, we sought to test the
hypothesis of whether under stress conditions such as
inflammation, ?2?1 integrin becomes part of the mes-
enchymal activation machinery and contributes to dis-
ease progression. To this end, we investigated the func-
tional relevance of ?2?1 integrin in 2 different mouse
models of destructive arthritis: antigen-induced arthritis
(AIA), a well-established T cell–dependent model of
disease (32), and the human TNF–transgenic mouse
(33), in which transgenic overexpression of human TNF
leads to chronic inflammatory polyarthritis, which re-
sembles the tissue destruction seen in human RA.
MATERIALS AND METHODS
Animals. The generation of ?2 integrin–deficient mice
and heterozygous human TNF–transgenic mice has been de-
scribed elsewhere (27,33). Heterozygous human TNF–trans-
genic mice (strain Tg197; genetic background C57BL/6) and
Itga2?/?mice (genetic background C57BL/6) were interbred
to yield human TNF–transgenic/Itga2?/?mice. Animals were
backcrossed onto the C57BL/6 background for at least 8
generations. Animals of all 4 genotypes were killed by cervical
dislocation. Blood was withdrawn by heart puncture, and sera
were collected after centrifugation at 2,200 revolutions per
minute for 20 minutes. Genotyping was performed by poly-
merase chain reaction using the following primers: for the
human TNF gene, 5?-TACCCCCTCCTTCAGACACC-3?
(forward) and 5?-GCCCTTCATAATATCCCCCA-3? (re-
verse); for the Itga2 wild-type (WT) gene, 5?-AAGTTGCTC-
GCTTGCTCTA-3? (forward) and 5?-AATATCCGTAGA-
AGCTCAGC-3? (reverse); for the Itga2?/?gene, 5?-AAGTT-
GCTCGCTTGCTCTA-3? (forward) and 5?-TGGCTTT-
TCTTCCTCCTATGG-3? (reverse). WT, human TNF–trans-
genic, Itga2?/?, and human TNF–transgenic/Itga2?/?mice
were examined once a week, and weight, paw swelling, and grip
strength were scored (34). The local ethics committee ap-
proved all animal procedures.
Induction of experimental arthritis. Mice were immu-
nized with 100 ?g of methylated bovine serum albumin
(mBSA; Sigma) emulsified in 100 ?l of Freund’s complete
adjuvant (CFA; Difco). Injections were given in both flanks.
After 7 days, 2 more injections of mBSA in CFA were
administered at 2 points in the neck region to boost the first
immunization. At both time points, heat-killed Bordetella
pertussis was injected intraperitoneally as a nonspecific im-
mune activator. Twenty-one days after the initial immuniza-
tion, arthritis was induced by intraarticular (IA) injection of 60
?g mBSA in 6 ?l saline into the right knee joint. Animals were
killed 5 days after IA injection of mBSA.
Clinical assessment of arthritis. AIA severity was
scored on an arbitrary scale of 0–3, where 0 ? no proteoglycan
loss, 1 ? mild proteoglycan loss, 2 ? moderate proteoglycan
loss, and 3 ? maximum proteoglycan loss. All experiments
were scored separately and independently of each other in a
blinded manner by 2 independent observers (SS and SF).
There was agreement between scorers.
Scoring of clinical signs of arthritis and measurement
of body weight in all mice was done once weekly starting from
age 6 weeks until age 14 weeks. Paw swelling was evaluated
using a well-established semiquantitative score: 0 ? no swell-
ing, 1 ? mild swelling of toes and ankle, 2 ? moderate swelling
of toes and ankle, and 3 ? severe swelling of toes and ankle.
Grip strength was also recorded using a semiquantitative score:
0 ? normal grip strength, ?1 ? mildly reduced grip strength,
?2 ? severely reduced grip strength, and ?3 ? no grip at all.
Histologic assessment. Hind paws and total knee joints
were dissected, fixed with 4% paraformaldehyde (PFA) over-
night, and then decalcified in 20% EDTA until the bones were
pliable. Serial paraffin sections (4 ?m) of all hind paws and
knee joints were stained with toluidine blue for assessment of
proteoglycan depletion and synovial inflammation. Measure-
ments and quantifications were performed using a Zeiss
Observer.Z1 microscope and Zeiss AxioVision 4.7.1 software.
Isolation of mouse synovial fibroblasts and cell cul-
ture. Murine FLS were isolated from knee joints and tarsus of
the hind paws of WT, Itga2?/?, human TNF–transgenic, and
human TNF–transgenic/Itga2?/?mice. After careful removal
of the skin, joints were minced and digested in Dispase II
(Roche) dissolved in Dulbecco’s modified Eagle’s medium
(DMEM) (without supplements) for 2 hours with vortexing
steps every 15 minutes. After filtration through a 70-?m cell
1360 PETERS ET AL
strainer, the cell suspension was collected, centrifuged, resus-
pended, and cultured in DMEM (Invitrogen) plus 10% fetal
calf serum (FCS; PAA Laboratories) plus antibiotics/
antimycotics (PAA Laboratories). All cells were cultured in
DMEM with 10% FCS at 37°C and 5% CO2from passage 3 to
passage 5 for all in vitro assays. For enzyme-linked immuno-
sorbent assay (ELISA) and Western blot analysis, cells were
seeded on collagen-coated dishes (fibronectin-coated dishes
were used for comparison). Cells were stimulated for the
indicated time periods with 100 ng/ml human TNF? (Novitec).
MMP-3 expression in mice and mouse fibroblasts.
MMP-3 expression in sera from 12-week-old WT, Itga2?/?,
human TNF–transgenic, and human TNF–transgenic/Itga2?/?
mice or in sera collected on day 26 for AIA experiments and in
supernatants from FLS isolated from WT, Itga2?/?, human
TNF–transgenic, and human TNF–transgenic/Itga2?/?mice
was determined using a mouse MMP-3 Quantikine Colorimet-
ric Sandwich ELISA kit (R&D Systems) according to the
manufacturer’s instructions. Briefly, 1 ? 105FLS were grown
on collagen-coated 96-well microtiter plates. Supernatants
were collected, centrifuged to remove cell debris, and stored at
Proliferation assay. Cell proliferation was measured by
bromodeoxyuridine (BrdU) assay (Millipore) according to the
manufacturer’s instructions. Briefly, 5 ? 104FLS were seeded
on collagen-coated 96-well microtiter plates for 24 hours.
BrdU incorporation as a parameter for DNA synthesis was
measured by colorimetric assay at 450 nm wavelength.
wild type itga2-/-
wild type itga2-/-
wild type itga2-/-
wild type itga2-/-
(% of section area)
Figure 1. Reduced joint inflammation, proteoglycan depletion, and matrix metalloproteinase 3 (MMP-3) expression during antigen-induced
arthritis (AIA) in ?2?1 integrin–deficient (Itga2?/?) mice. a and b, Left, Shown are toluidine blue–stained sections of whole knee joints 5 days after
induction of AIA in Itga2?/?and wild-type (WT) mice. a, Middle and right, The amount of cells present in the knee joint cavity (exudate) and in
the synovium (infiltrate) was significantly less in Itga2?/?mice than in their WT controls (n ? 8 per genotype). b, Right, Proteoglycan depletion was
significantly less in Itga2?/?mice. c, Levels of secreted MMP-3 were measured in sera from 12-week-old Itga2?/?and WT mice (n ? 9 per genotype).
d, Left, Shown is immunohistochemical expression of MMP-3 (red) in representative sections of whole knee joints 5 days after induction of AIA in
Itga2?/?and WT mice. Right, Quantification was performed by measuring the MMP-3 expression intensity (n ? 8 per genotype). Values are the
mean ? SEM. ? ? P ? 0.05; ?? ? P ? 0.01 versus WT mice. Original magnification ? 100 in a, b, and d.
?2?1 INTEGRIN IN RA1361
Western blot analysis of ERK-1/2 phosphorylation.
Cell extracts were harvested in radioimmunoprecipitation as-
say buffer (phosphate buffered saline, 1% [weight/volume]
Nonidet P40, 0.5% [w/v] sodium deoxycholate, 0.1% [w/v]
sodium dodecyl sulfate [SDS]) containing 100 ?g/ml phenyl-
methylsulfonyl fluoride, 1? complete protease inhibitor cock-
tail (Roche), and 500 ?M sodium orthovanadate. Protein was
separated on an SDS 10% (w/v) polyacrylamide gel and blotted
onto a PVDF membrane (GE Healthcare). Proteins were
detected by antibodies against phosphorylated and total ERK-
1/2 (Cell Signaling Technology). Detection was performed by
the addition of horseradish peroxidase–conjugated secondary
antibodies (Dako) and using an enhanced chemiluminescence
detection kit (ECL Western Blotting Detection Reagents;
In vitro attachment assay. To study the effects of
proteoglycan loss on the attachment of synovial fibroblasts in
vitro, an attachment assay with freshly isolated murine carti-
lage explants and isolated joint fibroblasts was used as de-
scribed (35). To this end, hip joints of 5–6-week-old WT mice
were opened after the mice were killed, and cartilage from the
femoral head was obtained aseptically (36). Femoral head
cartilage was then cultivated for 24 hours in 96-well microtiter
plates in DMEM at 37°C and an atmosphere containing 5%
CO2. To induce proteoglycan loss, recombinant IL-1 (R&D
Systems) was added for 12 hours at a final concentration of 1
ng/ml (37). For attachment analysis, FLS were seeded onto the
cartilage caps for 2 hours under continuous agitation and
cultured for 12 hours in fibroblast medium as described above.
The cartilage caps and cells were then fixed with 4% PFA for
20 minutes and stained with Mayer’s hemalum for 2 minutes,
and cells were quantified using a Zeiss Stemi stereo light
Statistical analysis. Data are shown as the mean ?
SEM. Statistical analysis was performed using GraphPad Prism
Software, version 5.0c. According to data distribution and
number of groups, a parametric test (t-test) or nonparametric
test (Mann-Whitney test) was performed. For comparison of
clinical assessments, a nonparametric Mann-Whitney U test
was used. The ELISA, proliferation, and group mean values of
the histologic data were compared by Student’s paired t-test.
P values less than 0.05 were considered significant.
Integrin ?2?1 regulates joint inflammation, pro-
teoglycan depletion, and MMP-3 expression during AIA.
First, we investigated the functional significance of ?2?1
integrin during early AIA using mice constitutively
ablated for the ?2 integrin subunit, which lack the ?2?1
receptor on all cells (27). To this end, we induced AIA
in the knees of 8–10-week-old WT and Itga2?/?mice
according to an established protocol (38) and deter-
mined the degree of joint inflammation in histologic
sections of total arthritic knee joints. On day 5 after
induction of AIA, the cellular mass in the joint cavity
(exudate) and in the synovial layer (infiltrate) was
significantly lower in Itga2?/?mice than in WT controls
(57% and 40%; P ? 0.05) (Figure 1a), indicating that
?2?1 integrin promotes cell influx. To assess whether
?2?1 integrin is directly involved in proteoglycan deple-
tion, we also investigated cartilage loss in inflamed knee
joints on day 5 after AIA induction. As shown in Figure
1b, the loss of proteoglycans, as determined by toluidine
blue staining, was significantly reduced in the inflamed
joints of Itga2?/?mice and was only 54% of the loss in
WT controls (P ? 0.01).
Based on the notion that increased expression of
MMPs contributes prominently to the breakdown of
cartilage matrix in AIA, we next measured the protein
levels of different MMPs in the sera of these mice. While
the levels of MMP-9 were not different between Itga2?/?
and WT mice (Figure 2), we found that on day 5 after
AIA induction, the serum levels of MMP-3 were signif-
icantly lower in Itga2?/?mice as compared to their WT
controls (mean ? SEM 59.59 ? 4.37 ng/ml versus 96.6 ?
13.4 ng/ml; P ? 0.05) (Figure 1c). Consistent with this
observation, immunohistochemical analysis revealed sig-
nificantly reduced expression of MMP-3 in the joints of
Itga2?/?mice (Figure 1d).
Loss of ?2?1 integrin reduces joint destruction
in human TNF–transgenic mice. The observation of
reduced MMP-3 expression with decreased loss of pro-
teoglycans during AIA in Itga2?/?mice prompted us to
analyze the consequences of ?2 integrin deficiency in a
more chronic animal model of human RA that closely
resembles inflammatory joint destruction, the human
TNF?–transgenic mouse model (33). To this end, we
crossed human TNF–transgenic mice, which spontane-
ously develop an RA-like destructive arthritis, with
Figure 2. Serum MMP-9 expression level during AIA in ?2?1
integrin–deficient mice. Levels of secreted MMP-9 were measured in
sera from 12-week-old Itga2?/?and WT mice (n ? ?5 per genotype).
Circles represent individual mice. Bars show the mean ? SEM. See
Figure 1 for definitions.
1362 PETERS ET AL
Itga2?/?mice, and we assessed the clinical signs of
disease as well as the extent of joint damage weekly from
ages 6 to 14 weeks.
Clinical arthritis started at week 6 in both human
TNF–transgenic and human TNF–transgenic/Itga2?/?
mice. However, absence of ?2?1 integrin clearly ame-
liorated the clinical course of a TNF-driven arthritis
(Figure 3a), as demonstrated by enhanced grip strength
and reduced paw swelling in the human TNF–
transgenic/Itga2?/?mice as compared with the human
TNF–transgenic mice at all time points between week 8
and week 14. Body weight did not differ between these 2
genotypes. Consistent with the improved clinical signs of
arthritis, histologic analysis of the affected hind paws
revealed a significantly lower severity of arthritis in the
tarsal joints of the human TNF–transgenic/Itga2?/?mice
as compared to the human TNF–transgenic control mice
at week 12. Specifically, we observed a reduced area of
68 10 12 14
6 8 10 12 14
grip strength score
inflammation area (%)
cartilage erosion (%)
hTN tg/ itga2-/-
(% of section area)
68 10 12 14
paw swelling score
Figure 3. Deficiency of ?2?1 integrin ameliorates arthritis in the human tumor necrosis factor (TNF)–transgenic (hTNFtg) mouse model of
rheumatoid arthritis and reduces MMP expression. a, Mean ? SEM body weight, grip strength, and paw swelling were assessed once a week in WT,
Itga2?/?, human TNF–transgenic, and human TNF–transgenic/Itga2?/?(hTNFtg/Itga2?/?) mice. Paw swelling was delayed and significantly reduced
in severity and accompanied by higher grip strength in human TNF–transgenic/Itga2?/?mice when compared with human TNF–transgenic mice (???
? P ? 0.001 from week 9 to week 14). b, Shown is representative toluidine blue staining of hind paw sections from 12-week-old WT, Itga2?/?, human
TNF–transgenic, and human TNF–transgenic/Itga2?/?mice. c, Shown is quantitative analysis of synovial area of inflammation, joint destruction
(destained cartilage), and cartilage erosion in tarsal joints (n ? ?5 per genotype). Also shown is synovial fibroblast attachment, which was analyzed
by measuring the length of synovial fibroblasts attaching to cartilage (n ? 6 per genotype). d, Shown is immunohistochemical expression of MMP-3
(red) in representative sections from 12-week-old human TNF–transgenic and human TNF–transgenic/Itga2?/?mice. e, Quantification was
performed by measuring the MMP-3 expression intensity (n ? 5 per genotype). Values are the mean ? SEM. ? ? P ? 0.05; ?? ? P ? 0.01 versus
human TNF–transgenic mice. Original magnification ? 100 in b and d. See Figure 1 for other definitions.
?2?1 INTEGRIN IN RA1363
pannus tissue and cartilage destaining (P ? 0.01) in
human TNF–transgenic mice lacking ?2 integrin. The
effects were most prominent with respect to proteogly-
can depletion, in that human TNF–transgenic/Itga2?/?
mice showed a reduction in cartilage erosions (to 48.5 ?
9% of that in human TNF–transgenic mice) (P ? 0.04)
and significantly reduced attachment of pannus tissue to
the articular cartilage (to 55.5 ? 8.4% of that in human
TNF–transgenic mice) (P ? 0.04) (Figures 3b and c).
Likewise, MMP-3 expression was significantly reduced
in the joints of human TNF–transgenic/Itga2?/?mice
(Figures 3d and e).
MMP expression, proliferation, and adhesion
properties are reduced in synovial fibroblasts deficient
for ?2?1 integrin. As FLS have been demonstrated to
be involved critically in both synovial hyperplasia and
cartilage destruction, we next analyzed the consequences
of ?2?1 integrin deficiency for the disease-specific prop-
erties of arthritic FLS, such as expression of matrix-
degrading enzymes, adhesion to cartilage, and prolifer-
ation. We isolated synovial fibroblasts from the hind
paws of mice of all genotypes (WT, Itga2?/?, human
TNF–transgenic, and human TNF–transgenic/Itga2?/?)
according to an established protocol (39). MMP levels in
supernatants of FLS grown in the absence or presence of
TNF? were assessed by ELISA (40). FLS from Itga2?/?
mice produced significantly lower amounts of MMP-3
than did control fibroblasts, and this effect was particu-
larly obvious in mice of the human TNF–transgenic
background (i.e., in TNF?-stimulated [for 24 hours]
human TNF–transgenic/Itga2?/?mouse FLS compared
with human TNF–transgenic mouse FLS) (Figure 4a).
To test whether the loss of ?2?1 integrin alters
the ability of murine FLS to adhere to predamaged
cartilage matrix, we performed an established attach-
ment assay that uses IL-1–treated mouse cartilage ex-
100 ng TNFalpha-+-+
100 ng TNFalpha-+-+
Figure 4. Regulation of attachment, proliferation, matrix metalloproteinase (MMP) expression, and ERK-1/2 phosphorylation by ?2?1 integrin. a,
Basal and TNF?-induced expression levels of MMP-3 in supernatant from wild-type (WT) and ?2?1 integrin–deficient (Itga2?/?) mouse fibroblasts
(left) and from human TNF–transgenic and human TNF–transgenic/Itga2?/?mouse fibroblasts (right) were determined by enzyme-linked
immunosorbent assay (n ? 10 per genotype). b, Shown is quantification of adhesion of fibroblasts from all 4 groups in an assay using
interleukin-1–treated murine cartilage explants (n ? ?3 per genotype). c, Effects of TNF? stimulation on proliferation of fibroblasts from WT and
Itga2?/?mice was measured by bromodeoxyuridine (BrdU) assay. d, Shown is basal and TNF?-induced activation of phospho–ERK-1/2 in WT and
Itga2?/?mouse fibroblasts seeded on collagen. e, Shown is basal and TNF?-induced activation of phospho–ERK-1/2 in WT and Itga2?/?mouse
fibroblasts seeded on fibronectin. Values are the mean ? SEM. ? ? P ? 0.05; ?? ? P ? 0.01; ??? ? P ? 0.001. See Figure 3 for other definitions.
1364PETERS ET AL
plants as adhesive substrate to mimic the inflammatory
proteoglycan loss found in early RA (35). Deficiency in
?2?1 integrin considerably decreased the attachment of
murine FLS to the cartilage, and again, this effect was
particularly pronounced in mice of the human TNF–
transgenic background (i.e., in human TNF–transgenic/
Itga2?/?mouse FLS compared with human TNF–
transgenic mouse FLS) (Figure 4b). Of interest, BrdU
incorporation assays clearly demonstrated strongly re-
duced proliferation of Itga2?/?mouse FLS compared to
WT mouse FLS (P ? 0.001) (Figure 4c). This difference
was particularly pronounced in TNF?-stimulated FLS,
and it may explain the reduced synovial hyperplasia
found in both mouse models of RA (Figures 1 and 3).
To study the signaling pathways involved and to
better understand the mechanisms by which loss of ?2
integrin decreases TNF?-induced fibroblast prolifera-
tion and MMP expression, we investigated 2 key medi-
ators of the MAPK cascade, ERK-1 and ERK-2, in FLS
from mice of all genotypes. Exposure to TNF? for 20
minutes strongly activated the MAPK cascade in WT
mouse FLS but not Itga2?/?mouse FLS when seeded
onto collagen (Figure 4d). Furthermore, we observed a
clearly reduced basal phosphorylation of ERK-1/2 in
mouse FLS lacking ?2 integrin as compared with WT
mouse FLS in these experiments (Figure 4d). Consistent
with the notion that ?2?1 integrin binds collagen but not
fibronectin, these effects on the phosphorylation of
ERK were not seen when FLS were seeded on fibronec-
tin instead of collagen as the adhesion substrate (Figure
4e). Likewise, we found similar ERK-1/2 phosphoryla-
tion results when comparing human TNF–transgenic
mouse FLS and human TNF–transgenic/Itga2?/?mouse
FLS (Figure 5).
One hallmark in the disease process in RA is the
progressive breakdown of articular cartilage. Besides
chondrocytes and neutrophils, RA FLS play a dominant
role in the degradation of cartilage. The critical role of
FLS in the process of cartilage destruction and bone
erosion in RA is most likely due to the synthesis of
mediators of inflammation, cytokines, and MMPs. In
RA, FLS are constantly stimulated by inflammatory
cytokines such as IL-1 and TNF?, which ultimately
results in a stable, tumor-like activation state, and in the
continuous production of large amounts of matrix-
degrading enzymes, particularly MMPs. These eventu-
ally lead to the irreversible destruction of the articular
cartilage. Of note, it has been suggested that the attach-
ment of FLS to components of the ECM is an important
step in the activation of FLS and sensitizes these cells to
inflammatory stimuli and growth factors, thus promoting
their transformation. Among the different adhesion
molecules, ?1 integrins appear to be of particular im-
portance in this process (7,8,41,42).
The ?2?1 integrin was identified as a collagen-
binding receptor early on and is expressed in various
tissues and cells, including fibroblasts (43). It is abun-
dantly expressed on many epithelial cell types and
platelets, but it is also present at increased levels on
inflammatory cells and FLS of RA patients (8,43). First
reports demonstrated expression during late stages of T
cell activation (44), and the fact that ?2?1 integrin is
able to mediate leukocyte adhesion suggested that it may
play an important role in inflammation. This concept is
supported by the finding that treating mice with
100 ng TNFalpha-+-+
100 ng TNFalpha-+-+
Figure 5. Regulation of ERK-1/2 phosphorylation by ?2?1 integrin.
a, Basal and TNF?-induced activation of phospho–ERK-1/2 in human
TNF–transgenic and human TNF–transgenic/?2?1 integrin–deficient
(Itga2?/?) mouse fibroblasts seeded on collagen. b, Basal and TNF?-
induced activation of phospho–ERK-1/2 in human TNF–transgenic
and human TNF–transgenic/Itga2?/?mouse fibroblasts seeded on
fibronectin. See Figure 3 for other definitions.
?2?1 INTEGRIN IN RA1365
collagen-induced arthritis (CIA) with ?2 integrin–
specific monoclonal antibodies inhibits the effector
phase of inflammatory responses in the CIA and has a
beneficial effect on the arthritis score (22). Therefore,
we were interested in analyzing the role of ?2?1 integrin
in animal models of human RA, and we challenged ?2
integrin–deficient mice in order to understand the im-
plications of a loss of ?2?1 integrin for inflammatory
disorders, especially for inflammatory cartilage degrada-
tion. Two groups of investigators had independently
shown that mice lacking ?2?1 integrin were viable,
fertile, and apparently normal (27,28), but exhibited
dysfunction of specific cell types upon challenge in vivo
(21,29–31,45) and in vitro (15,16).
We first used the AIA model, a T cell–mediated
mouse model of human RA. We found that Itga2?/?
mice displayed significantly less inflammatory pannus
formation and also significantly decreased proteoglycan
depletion. Consistent with earlier reports that ?2?1
integrin is able to modulate MMP expression (18–
20,45), we found reduced levels of secreted MMP-3 in
the serum of mice lacking ?2?1 integrin. This was also
reflected by diminished synovial MMP-3 concentrations.
This is of importance because MMP-3 has been reported
to play a key role in cartilage destruction, as it not only
degrades collagen but can also activate other metallo-
proteinases (46,47). Interestingly, the number of cells
present in the synovium (infiltrate) and in the knee joint
cavity (exudate) was significantly lower in ?2 integrin–
deficient mice, which potentially explains the reduced
MMP-3 expression in the AIA model system. On the
other hand, low MMP-3 levels in specific cell popula-
tions may also be due to the intracellular signaling
initiated by ?2?1 integrin. In RA, the synovium trans-
forms into an inflammatory pannus, which invades the
cartilage. This inflammatory pannus consists of several
cell types, such as macrophages, neutrophils, B and T
cells, and FLS (48), most of which express ?2?1 integrin.
Because AIA largely represents a T cell–
mediated autoimmune arthritis, we aimed to verify the
relevance of ?2?1 integrin to pathologic processes of
inflammatory joint destruction in the cytokine-
dependent human TNF–transgenic mouse model of
arthritis. According to the AIA model, human TNF–
transgenic/Itga2?/?mice showed reduced severity of
clinical signs of arthritis. Histologic analysis revealed less
severe joint inflammation and, in particular, significantly
less cartilage damage as compared with human TNF–
transgenic control animals. The absence of ?2?1 integ-
rin resulted in a decrease in proteoglycan loss. Similar
to the decreased MMP-3 expression that we found in
?2 integrin–deficient mice after induction of AIA, we
showed significantly reduced MMP-3 expression in the
inflamed pannus of human TNF–transgenic/Itga2?/?mice.
Two ways in which RA FLS differ from normal
FLS are altered attachment and high invasiveness. FLS
have the ability to remodel mesenchymal structures
through the production of cytokines or chemokines and
various MMPs. Presumably due to the chronic exposure
of these cells to the inflammatory environment of RA,
the transformed human RA FLS or human TNF–
transgenic murine FLS show a high invasiveness that
correlates with the elevated production of inflammatory
cytokines and MMPs (49,50). In addition, the expression
of certain adhesion molecules such as integrins is in-
creased in RA tissues and FLS, and it has been hypoth-
esized that they are crucial for the transformation of RA
FLS and for their invasion into cartilage. We showed
that Itga2?/?mouse FLS were nearly unable to attach to
murine cartilage that had been pretreated with IL-1 to
induce proteoglycan loss, whereas human TNF–
transgenic mouse FLS adhered well to this damaged
cartilage. Our in vitro data agree well with our in vivo
observations, in which histomorphometric analysis dem-
onstrated significantly less FLS attachment in human
TNF–transgenic/Itga2?/?mice than in human TNF–
Depending on the specific cell type, ?2?1 integ-
rin is able to regulate proliferation (10) and inhibit Fas
ligand expression or activation-induced cell death
(25,51). In the present study, we showed that FLS
lacking ?2?1 integrin had decreased proliferation in
vitro, supporting our findings of a reduction in inflam-
matory pannus formation in human TNF–transgenic/
Itga2?/?mice and a reduction in cell density in the
synovium after AIA induction. Taken together, these
findings indicate that reduced joint damage in mice
lacking ?2?1 integrin is caused by both reduced invasion
and reduced proliferation of mesenchymal-derived fi-
broblasts in the synovium. Consistent with our observa-
tions of lower serum concentrations of MMP-3, we
detected lower MMP-3 production by Itga2?/?mouse
fibroblasts. When searching for alterations in signaling
pathways known to be elicited by ?2?1 integrin activa-
tion (24,26), we found a less pronounced activation of
the ERK-1/2 cascade in FLS lacking ?2?1 integrin. This
effect was caused specifically by the loss of ?2?1 integ-
rin, since FLS grown on fibronectin did not show altered
From the findings of the present study, we con-
clude that the ?2?1 integrin receptor contributes to the
development of inflammatory RA by at least 2 mecha-
1366 PETERS ET AL
nisms. First, ?2?1 integrin is crucial for the adherence of
fibroblasts to damaged cartilage, as exemplified by de-
creased cell influx and reduced adhesion in vitro. Sec-
ond, ?2 integrin is involved in the expression of matrix-
degrading enzymes and in the regulation of proliferation
in the inflamed joint. Thus, our studies clearly show that
blockade of ?2?1 integrin protects against cartilage
destruction and thereby identify ?2 integrin as a poten-
tial therapeutic target for RA.
We would like to thank Borna Truckenbrod and Vera
Eckervogt for excellent technical assistance.
All authors were involved in drafting the article or revising it
critically for important intellectual content, and all authors approved
the final version to be published. Dr. Pap had full access to all of the
data in the study and takes responsibility for the integrity of the data
and the accuracy of the data analysis.
Study conception and design. Peters, Wendholt, Kollias, Eckes, Pap.
Acquisition of data. Peters, Wendholt, Strietholt, Frank, van den Berg.
Analysis and interpretation of data. Peters, Wendholt, Strietholt,
Frank, Pundt, Korb-Pap, Joosten, Pap.
1. Firestein GS. Evolving concepts of rheumatoid arthritis. Nature
2. Muller-Ladner U, Pap T, Gay RE, Neidhart M, Gay S. Mecha-
nisms of disease: the molecular and cellular basis of joint destruc-
tion in rheumatoid arthritis. Nat Clin Pract Rheumatol 2005;1:
3. Hynes RO. Integrins: bidirectional, allosteric signaling machines.
4. Lee JW, Juliano R. Mitogenic signal transduction by integrin- and
growth factor receptor-mediated pathways. Mol Cells 2004;17:
5. Legate KR, Wickstrom SA, Fassler R. Genetic and cell biological
analysis of integrin outside-in signaling. Genes Dev 2009;23:
6. Pap T, Franz JK, Hummel KM, Jeisy E, Gay R, Gay S. Activation
of synovial fibroblasts in rheumatoid arthritis: lack of expression of
the tumour suppressor PTEN at sites of invasive growth and
destruction. Arthritis Res 2000;2:59–64.
7. Agarwal SK, Brenner MB. Role of adhesion molecules in synovial
inflammation. Curr Opin Rheumatol 2006;18:268–76.
8. Rinaldi N, Schwarz-Eywill M, Weis D, Leppelmann-Jansen P,
Lukoschek M, Keilholz U, et al. Increased expression of integrins
on fibroblast-like synoviocytes from rheumatoid arthritis in vitro
correlates with enhanced binding to extracellular matrix proteins.
Ann Rheum Dis 1997;56:45–51.
9. Wang AZ, Wang JC, Fisher GW, Diamond HS. Interleukin-
1?–stimulated invasion of articular cartilage by rheumatoid syno-
vial fibroblasts is inhibited by antibodies to specific integrin
receptors and by collagenase inhibitors. Arthritis Rheum 1997;40:
10. Sarkissian M, Lafyatis R. Integrin engagement regulates prolifer-
ation and collagenase expression of rheumatoid synovial fibro-
blasts. J Immunol 1999;162:1772–9.
11. Semaan N, Alsaleh G, Gottenberg JE, Wachsmann D, Sibilia J.
Etk/BMX, a Btk family tyrosine kinase, and Mal contribute to the
cross-talk between MyD88 and FAK pathways. J Immunol 2008;
12. Hoberg M, Rudert M, Pap T, Klein G, Gay S, Aicher WK.
Attachment to laminin-111 facilitates transforming growth factor
?-induced expression of matrix metalloproteinase-3 in synovial
fibroblasts. Ann Rheum Dis 2007;66:446–51.
13. Zeisel MB, Druet VA, Wachsmann D, Sibilia J. MMP-3 expres-
sion and release by rheumatoid arthritis fibroblast-like synovio-
cytes induced with a bacterial ligand of integrin ?5?1. Arthritis
Res Ther 2005;7:R118–26.
14. Heino J. The collagen receptor integrins have distinct ligand
recognition and signaling functions. Matrix Biol 2000;19:319–23.
15. Zhang ZG, Bothe I, Hirche F, Zweers M, Gullberg D, Pfitzer G,
et al. Interactions of primary fibroblasts and keratinocytes with
extracellular matrix proteins: contribution of ?2?1integrin. J Cell
16. Zhang ZG, Lambert CA, Servotte S, Chometon G, Eckes B, Krieg
T, et al. Effects of constitutively active GTPases on fibroblast
behavior. Cell Mol Life Sci 2006;63:82–91.
17. Dumin JA, Dickeson SK, Stricker TP, Bhattacharyya-Pakrasi M,
Roby JD, Santoro SA, et al. Pro-collagenase-1 (matrix metallo-
proteinase-1) binds the ?2?1integrin upon release from keratino-
cytes migrating on type I collagen. J Biol Chem 2001;276:
18. Zigrino P, Drescher C, Mauch C. Collagen-induced proMMP-2
activation by MT1-MMP in human dermal fibroblasts and the
possible role of ?2?1 integrins. Eur J Cell Biol 2001;80:68–77.
19. Riikonen T, Westermarck J, Koivisto L, Broberg A, Kahari VM,
Heino J. Integrin ?2?1 is a positive regulator of collagenase
(MMP-1) and collagen ?1(I) gene expression. J Biol Chem
20. Langholz O, Rockel D, Mauch C, Kozlowska E, Bank I, Krieg T,
et al. Collagen and collagenase gene expression in three-dimen-
sional collagen lattices are differentially regulated by ?1?1 and
?2?1 integrins. J Cell Biol 1995;131:1903–15.
21. Edelson BT, Li Z, Pappan LK, Zutter MM. Mast cell–mediated
inflammatory responses require the ?2?1 integrin. Blood 2004;
22. De Fougerolles AR, Sprague AG, Nickerson-Nutter CL, Chi-
Rosso G, Rennert PD, Gardner H, et al. Regulation of inflamma-
tion by collagen-binding integrins ?1?1 and ?2?1 in models of
hypersensitivity and arthritis. J Clin Invest 2000;105:721–9.
23. Lee SH, Hatakeyama S, Yu SY, Bao X, Ohyama C, Khoo KH, et
al. Core3 O-glycan synthase suppresses tumor formation and
metastasis of prostate carcinoma PC3 and LNCaP cells through
down-regulation of ?2?1 integrin complex. J Biol Chem 2009;284:
24. Ivaska J, Nissinen L, Immonen N, Eriksson JE, Kahari VM, Heino
J. Integrin ?2?1 promotes activation of protein phosphatase 2A
and dephosphorylation of Akt and glycogen synthase kinase 3?.
Mol Cell Biol 2002;22:1352–9.
25. Gendron S, Couture J, Aoudjit F. Integrin ?2?1 inhibits Fas-
mediated apoptosis in T lymphocytes by protein phosphatase
2A-dependent activation of the MAPK/ERK pathway. J Biol
26. Ivaska J, Reunanen H, Westermarck J, Koivisto L, Kahari VM,
Heino J. Integrin ?2?1 mediates isoform-specific activation of p38
and upregulation of collagen gene transcription by a mechanism
involving the ?2 cytoplasmic tail. J Cell Biol 1999;147:401–16.
27. Holtkotter O, Nieswandt B, Smyth N, Muller W, Hafner M,
Schulte V, et al. Integrin ?2-deficient mice develop normally, are
fertile, but display partially defective platelet interaction with
collagen. J Biol Chem 2002;277:10789–94.
28. Chen J, Diacovo TG, Grenache DG, Santoro SA, Zutter MM. The
?2integrin subunit-deficient mouse: a multifaceted phenotype
?2?1 INTEGRIN IN RA1367
including defects of branching morphogenesis and hemostasis. Download full-text
Am J Pathol 2002;161:337–44.
29. Gruner S, Prostredna M, Aktas B, Moers A, Schulte V, Krieg T, et
al. Anti–glycoprotein VI treatment severely compromises hemo-
stasis in mice with reduced ?2?1 levels or concomitant aspirin
therapy. Circulation 2004;110:2946–51.
30. Zweers MC, Davidson JM, Pozzi A, Hallinger R, Janz K, Quon-
damatteo F, et al. Integrin ?2?1 is required for regulation of
murine wound angiogenesis but is dispensable for reepithelializa-
tion. J Invest Dermatol 2007;127:467–78.
31. Woodall BP, Nystrom A, Iozzo RA, Eble JA, Niland S, Krieg T, et
al. Integrin ?2?1 is the required receptor for endorepellin angio-
static activity. J Biol Chem 2008;283:2335–43.
32. Van den Berg WB, Joosten LA, van Lent PL. Murine antigen-
induced arthritis. Methods Mol Med 2007;136:243–53.
33. Keffer J, Probert L, Cazlaris H, Georgopoulos S, Kaslaris E,
Kioussis D, et al. Transgenic mice expressing human tumour
necrosis factor: a predictive genetic model of arthritis. EMBO J
34. Zwerina J, Redlich K, Polzer K, Joosten L, Kronke G, Distler J, et
al. TNF-induced structural joint damage is mediated by IL-1. Proc
Natl Acad Sci U S A 2007;104:11742–7.
35. Lefevre S, Knedla A, Tennie C, Kampmann A, Wunrau C, Dinser
R, et al. Synovial fibroblasts spread rheumatoid arthritis to unaf-
fected joints. Nat Med 2009;15:1414–20.
36. Little CB, Mittaz L, Belluoccio D, Rogerson FM, Campbell IK,
Meeker CT, et al. ADAMTS-1–knockout mice do not exhibit
abnormalities in aggrecan turnover in vitro or in vivo. Arthritis
37. Hajiioannou JK, Nikolidakis A, Naumidi I, Helidonis E,
Tzanakakis G, Velegrakis GA. In vitro enzymatic treatment and
carbon dioxide laser beam irradiation of morphologic cartilage
specimens. Arch Otolaryngol Head Neck Surg 2006;132:1363–70.
38. Van Lent PL, Grevers L, Lubberts E, de Vries TJ, Nabbe KC,
Verbeek S, et al. Fc? receptors directly mediate cartilage, but not
bone, destruction in murine antigen-induced arthritis: uncoupling
of cartilage damage from bone erosion and joint inflammation.
Arthritis Rheum 2006;54:3868–77.
39. Pap T, Shigeyama Y, Kuchen S, Fernihough JK, Simmen B, Gay
RE, et al. Differential expression pattern of membrane-type
matrix metalloproteinases in rheumatoid arthritis. Arthritis
40. Hayer S, Pundt N, Peters MA, Wunrau C, Kuhnel I, Neugebauer
K, et al. PI3K? regulates cartilage damage in chronic inflammatory
arthritis. FASEB J 2009;23:4288–98.
41. Ishikawa H, Hirata S, Andoh Y, Kubo H, Nakagawa N, Nishi-
bayashi Y, et al. An immunohistochemical and immunoelectron
microscopic study of adhesion molecules in synovial pannus for-
mation in rheumatoid arthritis. Rheumatol Int 1996;16:53–60.
42. Nakayamada S, Saito K, Fujii K, Yasuda M, Tamura M, Tanaka Y.
?1 integrin–mediated signaling induces intercellular adhesion
molecule 1 and Fas on rheumatoid synovial cells and Fas-mediated
apoptosis. Arthritis Rheum 2003;48:1239–48.
43. Zutter MM, Santoro SA. Widespread histologic distribution of the
a2?1integrin cell-surface collagen receptor. Am J Pathol 1990;137:
44. Hemler ME. VLA proteins in the integrin family: structures,
functions, and their role on leukocytes. Annu Rev Immunol
45. Grenache DG, Zhang Z, Wells LE, Santoro SA, Davidson JM,
Zutter MM. Wound healing in the ?2?1 integrin-deficient mouse:
altered keratinocyte biology and dysregulated matrix metallopro-
teinase expression. J Invest Dermatol 2007;127:455–66.
46. Uitto VJ, Schwartz D, Veis A. Degradation of basement-mem-
brane collagen by neutral proteases from human leukocytes. Eur
J Biochem 1980;105:409–17.
47. Burrage PS, Mix KS, Brinckerhoff CE. Matrix metalloproteinases:
role in arthritis. Front Biosci 2006;11:529–43.
48. Sweeney SE, Firestein GS. Rheumatoid arthritis: regulation of
synovial inflammation. Int J Biochem Cell Biol 2004;36:372–8.
49. Konttinen YT, Ainola M, Valleala H, Ma J, Ida H, Mandelin J, et
al. Analysis of 16 different matrix metalloproteinases (MMP-1 to
MMP-20) in the synovial membrane: different profiles in trauma
and rheumatoid arthritis. Ann Rheum Dis 1999;58:691–7.
50. Noss EH, Brenner MB. The role and therapeutic implications of
fibroblast-like synoviocytes in inflammation and cartilage erosion
in rheumatoid arthritis. Immunol Rev 2008;223:252–70.
51. Aoudjit F, Vuori K. Engagement of the ?2?1 integrin inhibits Fas
ligand expression and activation-induced cell death in T cells in a
focal adhesion kinase-dependent manner. Blood 2000;95:2044–51.
1368 PETERS ET AL