The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
Modulation of bone morphogenetic protein
signaling inhibits the onset and progression
of ankylosing enthesitis
Rik J.U. Lories, Inge Derese, and Frank P. Luyten
Laboratory for Skeletal Development and Joint Disorders, Department of Rheumatology, University Hospitals Leuven,
Katholieke Universiteit Leuven, Leuven, Belgium.
Joint ankylosis is a major cause of disability in the human spondyloarthropathies. Here we report that this
process partially recapitulates embryonic endochondral bone formation in a spontaneous model of arthritis
in DBA/1 mice. Bone morphogenetic protein (BMP) signaling appears to be a key molecular pathway involved
in this pathological cascade. Systemic gene transfer of noggin, a BMP antagonist, is effective both as a pre-
ventive and a therapeutic strategy in the mouse model, mechanistically interfering with enthesial progenitor
cell proliferation in early stages of the disease process. Immunohistochemical staining for phosphorylated
smad1/5 in enthesial biopsies of patients with spondyloarthropathy reveals active BMP signaling in simi-
lar target cells. Our data suggest that BMP signaling is an attractive therapeutic target for interfering with
structural changes in spondyloarthropathy either as an alternative or complementary approach to current
The human spondyloarthropathies are a group of chronic inflam-
matory joint disorders with a prevalence of about 0.3% (1), which
primarily affect individuals between 20 and 40 years of age. Differ-
ent diagnostic entities, based on clinical, genetic, and pathological
characteristics, are classified within this disease concept (2). These
consist of ankylosing spondylitis, psoriatic arthritis, reactive arthri-
tis, arthritis associated with inflammatory bowel disease, and a
juvenile and an undifferentiated form of arthritis. Clinical manifes-
tations include arthritis and ankylosis, which cause severe and per-
manent disability. Increasing evidence suggests that an anatomical
zone referred to as the enthesis, where tendons and ligaments attach
to underlying bone, is a primary target of the pathological process
(3, 4). Enthesitis, synovitis, and subchondral bone changes are asso-
ciated with enthesial cell proliferation and heterotopic cartilage and
bone formation, all of which eventually lead to joint space bridging,
a process called ankylosing enthesitis. Together, these features are
considered the hallmark of spondyloarthropathy (3–5).
Long-term treatment strategies for chronic joint diseases aim
to prevent tissue damage and loss of joint function. So-called dis-
ease modifying treatments not only target symptom control but
are also intended to influence the pathological process itself. In
spondyloarthropathy, conventional therapies, such as nonsteroidal
antiinflammatory drugs and immune modulators, are considered
symptom-controlling rather than disease-modifying treatments (6).
TNF inhibitory agents have been introduced as biologic response
modifiers (7, 8). Despite their clinical efficacy and widespread use,
and in contrast to what is seen in rheumatoid arthritis, the disease-
modifying properties of these strategies remain unclear. In chronic
joint diseases such as rheumatoid arthritis and in animal models
of these diseases (9), inflammation is associated with bone loss
and destruction. In spondyloarthropathy, new bone formation is
characteristic. Coupling of inflammation and subsequent ankylo-
sis have not been demonstrated. Progressive ankylosis may persist
despite absence of clinical disease activity (10). It is possible that
continued suppression of inflammation may result in accelerated
bone formation and ankylosis once the pathological cascade has
been triggered. Therapeutic strategies specifically targeting carti-
lage and bone formation may therefore be required, either as an
alternative or a complementary approach to immune-suppressing
drugs, for gaining full control of the disease (11).
Bone morphogenetic proteins (BMPs), discovered to be proteins
that ectopically induce a cascade of endochondral bone forma-
tion (12), play a crucial role in skeletal (13, 14) and joint morpho-
genesis (15–18). BMPs are members of the TGF-β superfamily, a
group of polypeptides capable of regulating a wide array of cellular
processes such as proliferation, differentiation, lineage determi-
nation, motility, and death (19, 20). In their canonical pathway,
BMPs induce ligand-dependent type I and type II receptor het-
erodimerization, leading to phosphorylation of smad-signaling
molecules (smad1/5) that bind smad4 and then translocate to the
nucleus. BMP signaling is controlled at many levels, including that
of extracellular antagonists such as noggin (19). Reactivation of
embryonic-signaling pathways has been suggested as an essential
part of repair and homeostasis in health and disease (21, 22). How-
ever, untimely or unwanted activation of signaling cascades fun-
damental for normal development may promote disease processes
such as spondyloarthropathies. Therefore, we hypothesized that
BMP signaling could play a direct role in joint ankylosis.
To address this, we studied BMP signaling in a spontaneous
model of arthritis. Grouped caging of male DBA/1 mice led to the
development of arthritis affecting hind paws and characterized by
ankylosing enthesitis (23–25). This feature clearly distinguishes
the model from other animal models of different types of joint
Nonstandard abbreviations used: BMP, bone morphogenetic protein; PCNA,
proliferating cell nuclear antigen; Sox9, sex-determining region Y–box 9; TBS,
Conflict of interest: The authors have declared that no conflict of interest exists.
However, Rik Lories and Frank P. Luyten have submitted a patent application regard-
ing the use of BMP inhibition in spondyloarthropathy.
Citation for this article: J. Clin. Invest. 115:1571–1579 (2005).
1572 The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
diseases, such as rheumatoid arthritis and osteoarthritis, that are
characterized by either synovial inflammation or progressive car-
tilage loss. Other features of the model include dactylitis and nail
changes, symptoms characteristic of psoriatic arthritis (25). This
model provides a tool for studying the mechanisms of joint anky-
losis with potential relevance to the human spondyloarthropa-
thies. In this report, we evaluate the effects of BMP signaling on
initiation and progression of arthritis by gene transfer of noggin,
a BMP antagonist.
Distinct BMPs are expressed in ankylosing enthesitis. We used
immunohistochemistry to identify specific BMPs, known to be
involved in embryonic endochondral bone formation, in ankylos-
ing enthesitis (Figure 1). The disease process was characterized
by inflammation and proliferation at the enthesis, followed by
cell differentiation into chondroblasts, prehypertrophic chon-
drocytes, and hypertrophic chondrocytes (24–27) (Supplemental
Figure 1; supplemental material available online with this arti-
cle; doi:10.1172/JCI23738DS1). The cartilage was progressively
replaced by bone, which eventually led to joint ankylosis. Differ-
ent BMPs were detected in distinct stages of ankylosing enthesi-
tis as indicated by cell morphology (Figure 1). BMP2 was seen
in spindle-shaped fibroblast-like cells in the proliferative zones
as well as in more rounded chondroblasts in early disease stages
(Figure 1, A and B). BMP2 was absent in prehypertrophic and
hypertrophic chondrocytes (data not shown). In contrast, in later
stages, BMP7 was largely restricted to prehypertrophic chondro-
cytes (Figure 1, D and E). BMP6 was found in hypertrophic chon-
drocytes (Figure 1, G and H).
Noggin gene transfer inhibits the onset and progression of spontaneous
arthritis. We used intramuscular injections of plasmids that encode
noggin cDNA under the control of a cytomegalovirus promoter
to evaluate ubiquitous suppression of BMP signaling. Injection
of 300 μg and 30 μg pcDNA3.1+noggin resulted in a dose-depen-
dent increase in local mRNA expression (Figure 2, A and B). In
the RT-negative control samples, only very discrete amounts of
residual noggin cDNA were detected after injection of 300 μg
pcDNA3.1+noggin. Protein was detectable in muscle and serum
(Figure 2C). Injection of 300-μg pcDNA3.1+ empty vector into
the contralateral muscle did not result in locally or systemically
detectable noggin. We chose to repeat injections every 3 weeks to
ensure gene expression during the course of the experiment (28,
29). To ascertain that repeated injections of noggin cDNA did not
result in the formation of antibodies that could block its effect,
mice were injected 4 times with 300 μg of pcDNA3.1+noggin at
3-week intervals. After 15 weeks, no specific antibodies were found
in their serum (data not shown).
We studied the effect of BMP inhibition on clinical incidence
and severity of spontaneous arthritis. Male mice were caged
together at the age of 12 weeks and monitored for clinical signs
of arthritis until the age of 25 weeks. Animals were treated every
3 weeks (weeks 12, 15, 18, and 21) with plasmid injections (either
300 μg pcDNA3.1+noggin, 30 μg pcDNA3.1+noggin, or similar
amounts of empty pcDNA3.1+). Injections of both 300 μg and 30
μg of noggin cDNA reduced the incidence of arthritis as compared
with injections of empty vector controls (Gehan-Wilcoxon test,
P < 0.05) (Figure 2D). Similarly, pcDNA3.1+noggin reduced sever-
ity of arthritis as compared with empty vector (Mann-Whitney
U test; P < 0.05) (Figure 2E). The effect of noggin treatment on
clinical severity during the course of the disease was evaluated by
calculating the area under the curve of the clinical severity score
(Figure 2F). Both noggin treatment groups (300 μg and 30 μg
plasmids) showed a reduction in time-integrated clinical sever-
ity scores (Mann-Whitney U test, P < 0.05). The effect was dose
dependent. We noticed a slower disease onset and slightly reduced
severity in the group that was treated with 300 μg empty vector as
compared with the group treated with 30 μg empty vector. Yet this
was not statistically significant. However, to exclude that the effect
of pcDNA3.1+noggin treatment was due to nonspecific exogenous
protein expression by gene transfer, additional mice were treated
with pcDNA3.1+ coding for an irrelevant secreted protein (mouse
Wnt5a with a mutated hydrophilic domain). No effect on arthritis
was found (data not shown).
The effect of BMP inhibition as a therapeutic strategy was evalu-
ated by pcDNA3.1+noggin or empty vector administration (300
μg) after symptoms of arthritis had developed. In a first set of
experiments, individual mice were injected immediately after the
BMP immunohistochemistry in ankylosing enthesitis. (A–C) The
first phase of ankylosing enthesitis is characterized by enthesial
cell proliferation and cartilage differentiation; immunoreactivity
for BMP2 is detected in spindle-shaped enthesial fibroblast-like
cells (arrowheads, inset in B) and in more rounded early chon-
droblast-like cells (arrows) (A; B, detail of boxed area in A). At
this stage of the disease process, no cartilage hypertrophy is
seen. (C) Negative control staining using IgG. (D–I) Cartilage
formation in ankylosing enthesitis shows BMP7-positive prolif-
erating and prehypertrophic chondrocyte-like cells (D; E, detail
of boxed area in D; arrows indicate zone of positive cells) and
BMP6-positive hypertrophic chondrocyte-like cells (G; H, detail
of boxed area in G; arrows indicate positive cells). No BMP7
expression is seen in hypertrophic chondrocyte-like cells (E,
arrowheads). (F and I) Negative control staining using IgG. Scale
bars: 200 μm in A and D; 100 μm in G; 80 μm in E and F; 50 μm
in B and C; 25 μm in H and I.
The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
36. Francois, R.J., Gardner, D.L., Degrave, E.J., and
Bywaters, E.G. 2000. Histopathologic evidence that
sacroiliitis in ankylosing spondylitis is not merely
enthesitis. Arthritis Rheum. 43:2011–2024.
37. Nordling, C., Kleinau, S., and Klareskog, L. 1992.
Down-regulation of a spontaneous arthritis in
male DBA/1 mice after administration of mono-
clonal anti-idiotypic antibodies to a cross-reactive
idiotope on anti-collagen antibodies. Immunology.
38. Maksymowych, W.P. 2000. Ankylosing spondylitis–
at the interface of bone and cartilage. J. Rheumatol.
39. Tani, Y., Sato, H., and Hukuda, S. 1997. Autoanti-
bodies to collagens in Japanese patients with anky-
losing spondylitis. Clin. Exp. Rheumatol. 15:295–297.
40. Hohler, T., and Marker-Hermann, E. 2001. Psori-
atic arthritis: clinical aspects, genetics, and the role
of T cells. Curr. Opin. Rheumatol. 13:273–279.
41. Pizette, S., and Niswander, L. 2000. BMPs are
required at two steps of limb chondrogenesis: For-
mation of prechondrogenic condensations and
their differentiation into chondrocytes. Dev. Biol.
42. Balemans, W., and Van Hul, W. 2002. Extracellular
regulation of BMP signaling in vertebrates: a cock-
tail of modulators. Dev. Biol. 250:231–250.
43. Fowler, M.J., Jr., et al. 1998. Induction of bone
morphogenetic protein-2 by interleukin-1 in
human fibroblasts. Biochem. Biophys. Res. Commun.
44. Lories, R.J.U., Derese, I., Ceuppens, J.L., and Luy-
ten, F.P. 2003. Bone morphogenetic proteins 2 and
6, expressed in arthritic synovium, are regulated
by proinflammatory cytokines and differentially
modulate fibroblast-like synoviocyte apoptosis.
Arthritis Rheum. 48:2807–2818.
45. Fukui, N., Zhu, Y., Maloney, W.J., Clohisy, J., and
Sandell, L.J. 2003. Stimulation of BMP-2 expres-
sion by pro-inflammatory cytokines IL-1 and TNF-
alpha in normal and osteoarthritic chondrocytes.
J. Bone Joint Surg. Am. 85(Suppl. 3):59–66.
46. Lin, J., et al. 2005. Kielin/chordin-like protein, a
novel enhancer of BMP signaling, attenuates renal
fibrotic disease. Nat. Med. 11:387–393.
47. van Beuningen, H.M., Glansbeek, H.L., van der
Kraan, P.M., and van Den Berg, W.B. 1998. Dif-
ferential effects of local application of BMP-2 or
TGF-beta 1 on both articular cartilage composi-
tion and osteophyte formation. Osteoarthr. Cartil.
48. Uusitalo, H., et al. 2001. Induction of periosteal
callus formation by bone morphogenetic protein-2
employing adenovirus-mediated gene delivery.
Matrix Biol. 20:123–127.
49. Persson, U., et al. 1998. The L45 loop in type I
receptors for TGF-beta family members is a critical
determinant in specifying Smad isoform activa-
tion. FEBS Lett. 434:83–87.