REVIEWS

The danger from within: alarmins

in arthritis
Meriam Nefla13, Dirk Holzinger4, Francis Berenbaum13,5 and Claire Jacques13
Abstract | Alarmins (also known as danger signals) are endogenous molecules that are released to
the extracellular milieu after infection or tissue damage. Extracellular alarmins interact with
specific receptors expressed by cells that are engaged in host defence to stimulate signalling
pathways that result in initiation of innate and adaptive immune responses, triggering
inflammation or tissue repair. Alarmins are considered to be markers of destructive processes that
occur in degenerative joint diseases (primarily osteoarthritis (OA)) and chronic inflammatory joint
diseases (such as rheumatoid arthritis, psoriatic arthritis and spondylarthropathy). In OA, high
mobility group proteinB1 (HMGB1) and S100 proteins, along with many other alarmins, are
abundantly secreted by joint cells, promoting cartilage matrix catabolism, osteophyte formation,
angiogenesis and hypertrophic differentiation. The involvement of alarmins in chronic
inflammatory arthritides is suggested by their presence in serum at high levels in these
conditions, and their expression within inflamed synovia and synovial fluid. S100 proteins,
HMGB1, IL33 and other endogenous molecules have deleterious effects on joints, and can
recruit immune cells such as dendritic cells to inflamed synovia, initiating the adaptive immune
response and perpetuating disease. Improving our understanding of the pathological
mechanisms associated with these danger signals is important to enable the targeting of new
therapeutic approaches for arthritis.

1
Sorbonne University, UPMC
University Paris 06, UMR_S
938, 75005 Paris, France.
2
INSERM UMR_S938,
UPMC University Paris 06,
75012 Paris, France.
3
Inflammation-
Immunopathology-Biotherapy
Department (DHU i2B),
184 rue du Faubourg Saint-
Antoine, 75012 Paris, France.
4
Department of Pediatric
Rheumatology and
Immunology, University
Childrens Hospital Mnster,
Domagkstr. 3, 48149
Mnster, Germany.
5
Department of Rheumatology,
Assistance Publique-Hpitaux
de Paris, Saint-Antoine
Hospital, 184 rue du Faubourg
Saint-Antoine, 75012 Paris,
France.
Correspondence to: F.B.
francis.berenbaum@sat.
aphp.fr
doi:10.1038/nrrheum.2016.162
Published online 13 Oct 2016
Arthritis is a leading cause of disability for millions of
adults. Arthritides include degenerative joint diseases,
such as osteoarthritis (OA), and different forms of
chronic inflammatory arthritis. OA is the most common
joint disease worldwide and is associated with numerous
risk factors, including ageing, joint trauma and obesity (as
reviewed elsewhere1). Although OA was initially thought
to be driven only by cartilage degradation, pathological
processes in other tissues, such as subchondral bone
and synovium, are now also considered to be involved
in its development2,3. Chronic inflammatory arthritides
are a group of conditions with primary pathologies that
involve inflammation of the synovium (synovitis). In
adults, these diseases include rheumatoid arthritis (RA),
psoriatic arthritis and spondylarthropathy, whereas the
umbrella term juvenile idiopathic arthritis (JIA) encom
passes the different chronic inflammatory arthritides of
childhood and adolescence. RA alone affects 1% of the
population globally, and is associated with considerable
morbidity and increased mortality relative to the unaf
fected population4. In chronic inflammatory arthritides,
inflamed synovia mediate progressive damage and ero
sion in joint tissues such as articular cartilage, muscle
tendons, periarticular ligaments and subchondral bone.
In arthritis, inflammatory processes are mediated
by numerous factors that are released from joint tis
sues, promoting joint destruction and pathological
progression. Endogenous proteins known as alarm
ins are released during inflammation and are involved
in joint damage5. The term alarmin was proposed by
Oppenheim and coworkers in 2005 to classify pro
teins that are rapidly released during infection or tissue
damage, activating immune cells after interaction with
their specific receptors6. Alarmins are now considered
to be markers of destructive processes in joints7,8. In this
Review, we focus on alarmins that are involved in the
pathogenesis of arthritis, and their roles in disease activity
and joint destruction.
General properties of alarmins
Alarmins, also known as danger signals or warning sig
nals, are members of the damage-associated molecular
pattern (DAMP) group of proteins. Alarmins are endog
enous molecules that are rapidly released into the extra
cellular milieu during infection or tissue damage6,9,10.
Their release involves both passive and active processes,
from cells undergoing necrosis and from activated cells
of the immune system, respectively11. Active release of

NATURE REVIEWS | RHEUMATOLOGY ADVANCE ONLINE PUBLICATION | 1

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Key points Alarmins in OA

Alarmins are endogenous molecules that are rapidly released to the extracellular
milieu during infection and tissue damage, activating receptors such as Toll-like
receptors and receptor for advanced glycosylation end products
In arthritis, including degenerative joint disease (osteoarthritis (OA)) and chronic
inflammatory arthritis (rheumatoid arthritis, psoriatic arthritis and
spondylarthropathy), extracellular levels of alarmins are associated with disease
activity and joint destruction
In OA, S100 proteins and high mobility group proteinB1 (HMGB1) are important
alarmins that induce positive-feedback loops of synovial cell reactivation,
inflammation and cartilage degradation
Alarmins involved in chronic inflammatory arthritis include S100 proteins, HMGB1,
heat-shock proteins and IL33; these alarmins represent important links between the
innate and adaptive immune systems
The level of alarmins in serum and synovial fluid could provide useful diagnostic and
prognostic biomarkers of arthritis, and inhibiting alarmin pathways could be a
therapeutically beneficial approach
alarmins is thought to be mediated by nonconventional
secretory pathways that are independent of the endo
plasmic reticulum and the Golgi apparatus10,12. Alarmins
that have been released can recruit antigen-presenting
cells (APCs), including dendritic cells (DCs). The
recruited cells then initiate innate and adaptive immune
responses, triggering inflammation or tissue regen
eration (or both)6,9,10. Alarmins bind to receptors that
are expressed on the surface of cells engaged in host
defence and tissue repair. These receptors include the
Toll-like receptors (TLRs) TLR2 and TLR4, as well as
the receptor for advanced glycosylation end products
(RAGE), and the binding of alarmins induces signal
transduction and activation of downstream signalling
pathways6,7,10,13. Alarmins are thought to crosslink var
ious receptors on the same cell to activate distinct and
synergistic signalling pathways6. In the absence of injury
or infection, alarmins have important intracellular roles
in the regulation of DNA transcription, calcium homeo
stasis and cell proliferation and differentiation. However,
high extracellular levels of alarmins are found in a wide
range of pathologies and are related to disease severity
in autoimmune conditions and inflammatory disorders,
such as sepsis, psoriasis, traumatic brain injury, acute
lung injury, inflammatory bowel disease and arthritis14.
In addition to their role in disease initiation, alarmins
also amplify and sustain inflammatory processes and have
a notable role in the pathogenesis of inflammatory condi
tions15,16. However, in addition to activation of phagocytes
and release of proinflammatory cytokines, alarmins are an
important link between the innate and adaptive immune
systems. For example, they can activate immature DCs,
which process antigens and home to secondary lym
phoid organs. In this case, they present antigenic epitopes
to naive Tcells, thereby inducing the adaptive immune
response1719. Moreover, alarmins might also be capable
of directing the commitment of Tcells toward a Thelper
2 (TH2) cell, TH17 cell or regulatory T(Treg) cell fate15.
Alarmins, therefore, contribute to a local hyperinflamma
tory environment, but might also prevent hyperactivation
of the adaptive immune system to avoid tissue damage
resulting from an overwhelming immune response (FIG.1).
Progressive cartilage loss during OA is predominantly
the result of overproduction of proteinases such as
matrix metalloproteinases (MMPs) and aggrecanases20.
In addition to remodelling of the extracellular matrix,
cartilage in OA is also an active site for chondrocyte
hypertrophy as well as proliferation and differentia
tion that promotes endochondral ossification2,20,21. The
results of experiments conducted invitro have shown
that chondrocyte damage caused by mechanical cartilage
injury stimulates the release of alarmins that activate the
migration of chondrogenic progenitor cells. These cells
can secrete large amounts of chemokines, cytokines and
destructive enzymes, leading to aberrant breakdown of
cartilage and joint inflammation22. Blocking alarmin
activity has been proposed as a strategy to prevent
post-traumatic joint inflammation23.
During OA, subchondral bone undergoes several
changes, including sclerosis, angiogenesis, microfrac
ture and the formation of cysts and osteophytes (as
reviewed elsewhere 2,24). Some of these changes are
thought to occur in the early stages of OA25. Moreover,
osteoblasts respond to mechanical loading by produc
ing inflammatory and catabolic factors, which result in
structural changes in subchondral bone, and probably
also alterations in cartilage26. Necrotic bone cells release
the alarmin high mobility group proteinB1 (HMGB1)27,
and high levels of many alarmins have been detected
in the synovial fluid of patients with OA, correlating
with disease severity2831. Several lines of evidence sug
gest that synovial inflammation aggravates joint dam
age during OA. Synovial activation occurs in 50% of
human knees affected by OA32. The synovium in OA
is composed predominantly of macrophages. However,
infiltration of Bcells and Tcells has been described33,34,
suggesting the involvement of the adaptive immune
system in OA. Activated synovial cells secrete several
degenerative enzymes and inflammatory mediators,
as well as alarmins3537. These alarmins activate pat
tern-recognition receptors (PRRs) including TLRs and
RAGE in the OAaffected cartilage and synovium13,38,
which in turn amplifies inflammation and degeneration
of cartilage.
A number of alarmins are found at high levels extra
cellularly in OA, enhancing catabolic processes and
inflammatory responses that contribute to disease pro
gression. These alarmins include S100 proteins, HMGB1
and a number of other biomolecules.
S100 proteins
S100 proteins are a family of small (1014kDa), acidic
proteins that are expressed exclusively in vertebrates and
contain two elongation factor (EF)-hand calcium-binding
motifs within their structures. S100 proteins are highly
conserved, although some variability occurs in the car
boxyl terminus and the hinge region between the two
EFhand structures, contributing to the specificity of bio
logical activity of different members39,40. S100 proteins are
found in cells as homodimers, heterodimers or multi
mers. When calcium binds to EFhand motifs the pro
teins undergo changes in their conformation, enabling

2 | ADVANCE ONLINE PUBLICATION www.nature.com/nrrheum

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

Antigen presentation
Amplication of
MHC T cell inammatory
receptor process
T-cell polarization TH1 or TH2 cells
Dendritic cell
Stress and/or
Damage
Injury T cell Tissue repair and
Age protective eects
Necrotic cell Antigen processing and
migration to Treg cells
lymphoid organs

Alarmins Alarmin
receptor

Inammatory
Infection process
Imbalance
of cytokines Other target cell Proinammatory
(such as macrophage, mediators
Activated immune cell endothelial cell
or leukocyte)

Figure 1 | A schematic, integrative overview of alarmins. Alarmins can be released from activated immune cells
during infection or from cells undergoing necrosis in response to stress and injury. Released alarmins recruit
Nature Reviews | Rheumatology
antigen-presenting cells, such as dendritic cells, by binding to specific receptors. Dendritic cells process alarmins and
home to secondary lymphoid organs, where they present antigens to naive Tcells, leading to Tcell polarization
and differentiation to Thelper 1 (TH1) cells or TH2 cells, which amplify inflammation and tissue destruction. Tcells can
also differentiate to regulatory Tcells (Treg cells) that are capable of inhibiting inflammation and participating in tissue
repair. Released alarmins can also bind to specific receptors on the surface of other target cells, such as macrophages and
endothelial cells, to stimulate the production of proinflammatory mediators or activate the endothelium, thereby leading
to inflammatory processes.

interaction with various target proteins41,42. Abundant
extracellular release of some S100 proteins is associated
with several inflammatory diseases, including OA43,44.
S100A4 (also known as metastasin and calvasculin)
is highly expressed in cartilage in OA, where it binds
to RAGE, induces the production of reactive oxygen
species (ROS) and activates PYK2 (also known as pro
tein-tyrosine kinase 2), mitogen-activated protein
kinase (MAPK) and nuclear factorB (NFB) path
ways, thereby leading to production of MMP13 in
human chondrocytes45. IL7 can induce the release of
S100A4 from human chondrocytes via activation of the
Janus kinase (JAK)signal transducer and activator of
transcription (STAT) signalling pathway; S100A4 then
acts in an autocrine or paracrine manner to stimulate the
production of MMP13 via activation of RAGE43.
S100A8 (also known as migration inhibitory factor-
related protein 8 (MRP8) and calgranulin A) and S100A9
(also known as MRP14 and calgranulin B) are highly
expressed in monocytes, granulocytes and early differ
entiation stages of macrophages. These proteins have
been considered markers of a destructive process in
the joint46 and are found in the heterodimeric complex
S100A8S100A9 (calprotectin), which is abundantly
secreted by immune cells during inflammation4749. In
mice, S100A8 is the active component of this complex
(which can also form a tetramer of two heterodimers),
whereas S100A9 has regulatory functions and protects
S100A8 against degradation50,51. Large amounts of cal
protectin have been detected in synovial fluid35 and in
serum, synovium and cartilage of patients with OA, and
its presence might predict joint destruction in OA37,52.
Genetic knockout of S100A9 expression ameliorates the
pathological features of the collagenase-induced mouse
model of OA, in which synovitis is normally observed.
However, in the destabilized medial meniscus mouse
model, knockout of S100A9 expression does not have any
effect, suggesting that calprotectin is only involved when
synovial activation is present37. Production of S100A8
by macrophages can be induced by basic calcium phos
phate crystals in a manner dependent on tyrosine-protein
kinase SYK and phosphatidylinositol 3kinase (PI3K),
and can form a positive-feedback loop in immune cells,
inducing the production of interleukins and contribut
ing to the catabolic chondrocyte phenotype53. In addition,
results from an experimental model of OA suggested a
role for calprotectin in osteophyte formation54, agreeing
with results showing that the effects of calprotectin are
mediated by canonical Wnt signalling, which is known
to promote bone formation55. Calprotectin signalling is
mediated by TLR4 in chondrocytes from patients with
OA52. This finding is supported by results showing that
prophylactic treatment with paquinimod (an immuno
modulatory compound that prevents S100A9 binding to
TLRsh4) reduces the pathological effects of S100A9 in
collagenase-induced OA in mice and in the human syn
ovium in OA56. TLR4 has also been proposed to mediate
excitation of dorsal root ganglion neurons by S100A8,
contributing to the excitation of joint nociceptors and the
generation of pain during OA57.

NATURE REVIEWS | RHEUMATOLOGY ADVANCE ONLINE PUBLICATION | 3

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

S100A10 (also known as p11, cellular ligand of
annexin II and calpactin I light chain) can form a het
erotetrameric complex (AIIt) with its known ligand,
annexin II; this complex acts as an extracellular bind
ing partner for pathogens and host proteins44. S100A10
has been shown to induce the production of TNF,
IL1 and IL10 in human chondrocytes stimulated
by lipopolysaccharide via MAPK and NFB signal
ling pathways58. Production of TNF by macrophages in
response to exogenous AIIt requires TLR459, suggesting
that S100A10-mediated induction of inflammatory
cytokine production by chondrocytes also involves
TLR4 activation.
S100A11 (also known as S100C and calgizzarin) acts
as a catabolic mediator in chondrocytes60. Expression of
S100A11 is upregulated in human knee cartilage affected
by OA, and the protein is released by cultured chondro
cytes on stimulation with TNF and CXC-chemokine
ligand 8 (CXCL8, also known as IL8)21. Moreover,
recombinant S100A11 can promote hypertrophic differ
entiation of chondrocytes in a process that is depend
ent on RAGE and p38 MAPK21. S100A11 can also
induce release of the aggrecanase-generated neoepitope
NITEGE, and glycosaminoglycans, in mouse cartilage
explants60. Post-translational modification (homodimer
ization) of S100A11 by transglutaminase 2 is required
to enable signalling through RAGE and modulation
of human chondrocyte growth and cartilage matrix
catabolism60.
S100A12 (also known as calgranulin C) is abundantly
expressed by human neutrophils, monocytes and mac
rophages and is overexpressed in inflammatory compart
ments61. Its expression is markedly increased in human
cartilage affected by OA (compared with unaffected car
tilage) and its addition to cultured human osteoarthritic
chondrocytes upregulates expression and release of
MMP13 and vascular endothelial growth factor (VEGF),
leading to cartilage degradation and angiogenesis, via
RAGE and p38 MAPK and NFB pathways62. Notably,
proteomic analysis revealed high levels of S100A12
in synovial fluid from patients with OA63. Moreover,
the levels of S100A12 in synovial fluid correlate with the
symptomatic and radiographic s everity of OA29.
S100B is expressed in chondrocytes44 and stimulates
MMP13 production via extracellular signal-regulated
kinase (ERK) and NFB signalling pathways64. This
signalling is inhibited in cells pretreated with soluble
RAGE, and the results of pull-down assays have con
firmed the presence of specific binding of S100B to
RAGE in human chondrocytes64 (FIG.2).
HMGB1
HMGB1 (also known as amphoterin) is a 30kDa non
histone nuclear protein that contains two DNA-binding
boxes27. HMGB1 is a highly conserved protein that has
important intracellular and extracellular functions.
Extracellular HMGB1 acts as an alarmin binding to
multiple cell-surface receptors, cytokines and chemo
kines to stimulate the innate immune system and trigger
inflammatory responses65. HMGB1 is overexpressed in
the synovial membranes of patients with knee OA28. The
level of HMGB1 in synovial fluid is higher in patients
with knee OA than in unaffected individuals, and is pos
itively associated with disease severity66, synovitis and
pain28. HMGB1 is also overexpressed in cartilage from
patients with OA, and its expression correlates with the
Osteoarthritis Research Society International (OARSI)
histological score67,68. High levels of HMGB1 have been
detected in joint tissues in the STR/ort mouse model of
spontaneous OA69. Human osteoarthritic synoviocytes
release HMGB1 on stimulation with IL170, and chon
drocytes from patients with OA and those from healthy
mice release HMGB1 in response to several inflamma
tory mediators, including IL1 and TNF68,71. Moreover,
stimulation of human osteoarthritic chondrocytes
with HMGB1 induces the release of IL8, nitric oxide
(NO)67, IL1 and TNF68. Overexpression of MMP13 by
articular chondrocytes that have been stimulated with
HMGB1 involves binding to RAGE, which leads to ERK
and NFB activation64. However, HMGB1induced
MMP13 expression is only partially inhibited by soluble
RAGE, implying that other receptors are also involved
in mediating the effects of HMGB1 on chondrocytes. In
this context, the treatment of mouse chondrocytes and
cartilage explants with HMGB1 stimulates matrix catab
olism and chondrocyte hypertrophic differentiation that
are dependent on TLR2, TLR4 and their cytosolic adap
tor molecule myeloid differentiation primary response
protein MYD88 (REF.71).
HMGB1 that has been released from hypertrophic
human chondrocytes functions as a chemoattractant for
osteoclasts, osteoblasts and endothelial cells, and could
have a role in endochondral ossification72. A high level of
HMGB1 has been detected by immunohistochemistry at
the tidemark of cartilaginous and calcified fibrocartilage,
which again suggests a role in mineralization and endo
chondral bone formation at the cartilagebone inter
face73. Moreover, HMGB1 is known to be a bone-active
cytokine that is released from necrotic osteocytes and is
involved in bone remodelling27.
In addition to direct stimulation of cells, HMGB1 can
form immunostimulatory complexes with cytokines, as
well as with other endogenous and exogenous factors.
HMGB1 forms proinflammatory and procatabolic
extracellular complexes with IL1 in cultures of syn
oviocytes from patients with OA74, and with IL1, IL1
and microbial lipopolysaccharide in cultures of synovial
fibroblasts from patients with OA75; these complexes
promote inflammation and catabolism. These findings
suggest that some of the functions of HMGB1 in joints
affected by OA are mediated by interaction with other
proinflammatory factors (FIG.2).
HSPs
Heat shock proteins (HSPs) are highly conserved pro
teins that have considerable homology from prokaryotes
to eukaryotes. HSPs act under physiological conditions
as intracellular chaperones during synthesis of proteins
or clearance of misfolded proteins. Members of this fam
ily, in particular HSP60 and HSP70, have been identi
fied as sensors of cellular stress. Under stress (such as
high temperature, nutritional deficiency, exposure to

4 | ADVANCE ONLINE PUBLICATION www.nature.com/nrrheum

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

Healthy Osteoarthritic
joint joint
Mechanical
stress or Joint Synovium
injury/damage space Matrix products 1
Macrophage
Cytokines or
chemokines 2

3 3

Cartilage Cartilage 2 Synoviocyte

Chondrocyte
Synovium

3
Synovitis
Hypertrophic 3
7 chondrocyte
NF-B 4 NF-B
Cartilage 8
degradation MAPK MAPK MAPK
5 9 6
Subchondral Bone
bone remodelling
Vascular invasion Tidemark Calcied 10
cartilage
Osteoblast
Subchondral Osteophyte
HMGB1 ADAMTS RAGE bone
MMP 12 Chemotaxis
S100 11
TLR2 or
Cytokine or ROS TLR4 13 Bone remodelling
chemokine NO
IL-1 Dying/necrotic
HMGB1IL-1 VEGF osteocyte Osteoclast
receptor

Figure 2 | Role of S100 proteins and high mobility group proteinB1 in osteoarthritis. Cartilage breakdown during
osteoarthritis (OA) induces the release of matrix products that stimulate synovial cells andNature
amplifyReviews
synovitis| Rheumatology
(1). Activated
synovial cells produce proinflammatory and catabolic factors, S100 proteins and high mobility group protein B1 (HMGB1)
(2), leading to a positive-feedback loop of synovial-cell reactivation, inflammation and cartilage degradation (3). S100
proteins secreted by synovial cells and chondrocytes aggravate cartilage breakdown by inducing the production of a
disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS), matrix metalloproteinases (MMPs),
inflammatory mediators, reactive oxygen species (ROS) and vascular endothelial growth factor (VEGF), which stimulates
angiogenesis (4). This process also promotes hypertrophic differentiation of chondrocytes (5) and osteophyte formation
(6). These effects are mediated by receptor for advanced glycosylation end products (RAGE), Toll-like receptor 4 (TLR4),
mitogen-activated protein kinase (MAPK) and nuclear factorB (NFB) pathways. When secreted outside the cell,
HMGB1 induces the production of MMPs, inflammatory mediators and nitric oxide (NO) by chondrocytes via interaction
with RAGE and activation of MAPK and/or NFB pathways (7). It also promotes hypertrophic differentiation of
chondrocytes via TLR2 and TLR4 (8). In turn, hypertrophic chondrocytes release HMGB1 (9) to promote endochondral
ossification and osteophyte formation by acting as a chemoattractant for bone cells (10). HMGB1 can also be released by
necrotic or dying osteocytes (11), to act as a chemoattractant for osteoblasts and osteoclasts via RAGE, TLR2 and TLR4
(12) and to promote bone remodelling (13).

proinflammatory mediators or viral infections) HSP60
and HSP70 are highly expressed and are secreted into
the extracellular space 76. HSP70 is upregulated in
human osteoarthritic chondrocytes77 and in cartilage
from patients with severe OA78. HSP70 protects rabbit
chondrocytes from NOinduced apoptosis by blocking
caspase 3 activity, thereby reducing progression of OA79.
Biomechanical induction of OA in rats results in
upregulation of expression of HSP90, which might pro
mote degeneration of cartilage, remodelling of subchon
dral bone and activation of synovial macrophages80. The
use of a synthetic HSP90 inhibitor upregulates expres
sion of HSP70, suppressing these deleterious effects and
protecting cartilage80. These results suggest a mechanism
in which chondrocytes upregulate HSP70 in response
to physical activity in order to protect the extracellular
matrix against damage. However, persistent physical
stress on cartilage stimulates production of HSP90,
counteracting the protective effects of HSP70 and
promoting development of OA.
Small HSPs (sHSPs) represent a subfamily of proteins
with molecular weights <30kDa that are characterized
structurally by a conserved core domain, called the
crystallin domain81. Crystallin B chain (also known
as HSPB5) is an sHSP that is expressed at lower levels
in human osteoarthritic chondrocytes than in healthy
chondrocytes82. Differential expression of HSPB5 in rela
tion to OA is probably mainly regulated by inflamma
tory mediators, because of its downregulation following
treatment of human chondrocytes with IL1 and TNF82.
Another sHSP, heat-shock protein 1 (HSPB1, also
known as HSP27), has lower expression in chondrocytes
from patients with OA that are derived from damaged
cartilage than in those derived from intact cartilage83.

NATURE REVIEWS | RHEUMATOLOGY ADVANCE ONLINE PUBLICATION | 5

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Transfection of human chondrocytes with small inter
fering RNA (siRNA) directed against mRNA encoding
HSPB1 results in induction of secretion of IL6 by IL1,
suggesting that the response of osteoarthritic chondro
cytes to IL1 is regulated in part by underexpression of
HSPB1 (REF.83).
To improve our understanding of the alarmin activ
ity of HSPs in OA, further studies are needed to detect
whether alarmins are present in joint spaces, and to
identify cell signalling pathways that are involved in
damage and repair of tissues.
Other alarmins
Many other molecules have putative alarmin activities
and have been shown to have important roles in the
pathogenesis of OA (TABLE1). For example, annexinII,
annexinV and annexinVI are overexpressed in human
osteoarthritic cartilage, and are major components of
matrix vesicles that are released by hypertrophic human
chondrocytes into the extracellular matrix, initiating
the mineralization process in growth-plate cartilage84,85.
These annexins can induce apoptosis in calf chondro
cytes by increasing production of basic calcium phos
phate crystals and TNF86. To activate macrophages,
annexin II, in a heterotetrameric complex with S100A10
(AIIt), requires TLR4 (REF.59). The action of annexins in
chondrocytes might also be mediated by TLR4.
ATP is secreted by porcine chondrocytes in response
to mechanical compression87. ATP has been detected in
synovial fluid from patients with OA, and its presence
is associated with knee pain, which might be explained
by the capacity of extracellular ATP to act on synovial C
nerve fibres, which act as nociceptors88. Proinflammatory
extracellular ATP stimulates release of prostaglandin E2
(PGE2) from rabbit articular chondrocytes, acting on the
P2Y2 purine receptor and activating p38, ERK1 and ERK2
MAPKs89. Moreover, high extracellular levels of ATP
induce mineralization of porcine cartilage by inducing
the production of calcium crystals90.
The soluble protein galectin 3 is overexpressed in
cartilage from patients with OA91 and in synovial tissue
during inflammatory phases of OA, in which infiltra
tion of leukocytes and macrophages occurs92, indicat
ing that galectin 3 could be released into joint spaces93.
Intra-articular injection of galectin 3 into the knees of
mice induces joint swelling and generates lesions in both
subchondral bone and cartilage tissue93. Additionally,
invitro treatment with galectin 3 inhibits osteocalcin
production by human osteoarthritic osteoblasts, and
increases expression of MMP3 and the protein a disinteg
rin and metalloproteinase with thrombospondin motifs5
(ADAMTS5) in human osteoarthritic chondrocytes93.
IL1 is a classic interleukin that is known as a danger
signal94 and is thought to be involved in OA95. IL1 is
secreted mostly by macrophages, and at lower levels by
chondrocytes and synovial fibroblasts96. It has autocrine
and paracrine actions, and promotes early degenerative
changes in patients with OA97 by binding to the mem
brane-associated IL1 receptor (IL1R)94. Higher con
centrations of IL1 have been reported in synovial fluid
in porcine knee joints affected by moderate OA than in
those affected by mild OA98. Additionally, treatment
of porcine cartilage explants with IL1 increases NO
production and total MMP activity98.
Levels of thymosin 4 are upregulated after mechan
ical loading of bovine cartilage, enhancing activation
of MMP2 and MMP9 in chondrocytes99. High levels of
thymosin 4 in the serum and synovial fluid of patients
with OA correlate with the radiographic severity of the
disease30.
Levels of uric acid in the synovial fluid of patients
with knee OA correlate with levels of IL18 and IL1,
and are positively associated with the severity of OA, as
evaluated by radiography and scintigraphy31. As pro
duction of IL18 and IL1 requires inflammasome
activation, stimulation of levels of these cytokines in
osteoarthritic joints by uric acid is thought to occur in
resident macrophages via inflammasome activation and
through TLR2 and TLR4, in the same way as in gout31.
Several extracellular matrix components are also
classified as alarmins, and have important roles in the
pathogenesis of OA. In patients with OA, the synovial
fluid and cartilage matrix contain high levels of fibronec
tin fragments100. These fragments enhance production of
NO in bovine cartilage explants101 and catabolic cytokines
(such as TNF, IL1, IL1 and IL6) in healthy human
cartilage102, as well as MMP expression via TLR2, NFB
and p38 MAPK activation in human osteoarthritic chon
drocytes103. Fibronectin fragments can also bind CD44
and several integrins, which could activate a variety of
signalling pathways leading to cartilage catabolism104.
Low-molecular-weight hyaluronan, generated by degra
dation of hyaluronan at sites of inflammation and tissue
injury105, induces production of NO and MMPs by mech
anisms dependent on CD44 in bovine articular chondro
cytes106 and MYD88 (through TLR2 and TLR4) in mouse
chondrocytes71. Low-molecular-weight hyaluronan also
regulates cartilage hypertrophy by inducing expression of
type X collagen and Runt-related transcription factor2 in
mouse chondrocytes71.
Biglycan is another extracellular matrix component
that is found in cartilage and synovial fluid from patients
with advanced OA, and it triggers catabolic processes107.
Biglycan activates human osteoarthritic chondrocytes
mainly through the TLR4NFB pathway, which upreg
ulates expression of ADAMTS4, ADAMTS5, MMPs, NO,
cathepsin K, IL6 and IL8, and downregulates expres
sion of anabolic markers (such as aggrecan and collagen
typeII), resulting in a net loss of cartilage107. In addition,
soluble biglycan induces the release of proteoglycans,
collagen, NO, IL6 and IL8 from human osteoarthritic
chondrocytes, which might result in the recruitment
of inflammatory cells and amplification of cartilage
destruction107.
Laminins which are heterotrimers consisting of one
of five -chains, one of four -chains and one of three
chains are also involved in the pathogenesis of
OA. Laminin subunit4 is overexpressed in human
osteoarthritic cartilage with high-grade lesions, and its
expression colocalizes with that of syndecan 4 in hyper
trophic areas108. Additionally, the stimulation of MMP3
expression by IL1 in primary human chondrocytes is

6 | ADVANCE ONLINE PUBLICATION www.nature.com/nrrheum

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

Table 1 | Alarmins that are involved in osteoarthritis

Alarmins Localization of Roles in OA Receptors in Refs
secreted forms joint cells
S100 proteins Serum and S100A4 is highly expressed in the cartilage of patients with OA and that RAGE and 21, 29, 35,
(A4, A8, A9, A10, synovial fluid from an experimental model of OA. It stimulates MMP13 production in TLR4 37, 43, 45,
A11, A12, B) chondrocytes via activation of PYK2, MAPK and NFB 52, 53, 54,
Calprotectin (S100A8S100A9) is overexpressed in the synovium and 57, 58, 59,
cartilage of patients with OA. It induces production of interleukins by 60, 62, 64
immune cells and aggravates osteophyte formation in an experimental
model of OA
S100A8 contributes to pain generation in a mouse model of OA
S100A10 induces production of inflammatory cytokines in human
chondrocytes via MAPK and NFB
S100A11 is upregulated in cartilage from patients with OA, where it
promotes hypertrophic differentiation of chondrocytes via p38 MAPK
and induces cartilage degradation
S100A12 expression is increased in human osteoarthritic articular
cartilage compared with normal cartilage. Synovial fluid levels of
S100A12 correlate with symptomatic and radiographic severity of OA. It
upregulates the expression and production of MMP13 and VEGF via p38
MAPK and NFB
S100B stimulates MMP13 production in human chondrocytes via
activation of ERK and NFB
HMGB1 Synovial fluid Highly expressed in the cartilage and synovium in a spontaneous mouse RAGE, TLR2 28, 64, 67,
model of OA and in patients with OA and TLR4 66,68, 69,
Expression in human osteoarthritic cartilage correlates with OARSI 71, 72, 73,
histological score 74, 75
Levels in synovial fluid of patients with OA are correlated with disease
severity, synovitis and pain
Induces the release of chemokines, cytokines and NO
Induces MMP13 expression via activation of ERK and NFB
Promotes chondrocyte hypertrophic differentiation and endochondral
ossification in mice
Forms proinflammatory and procatabolic extracellular complexes
with cytokines and lipopolysaccharide in cultures of synovial cells from
patients with OA
HSPs Not known HSP70 is upregulated in cartilage from patients with severe OA and Not known 78, 79, 80,
(HSP70, HSP90, protects rabbit chondrocytes from NOinduced apoptosis 82, 83
HSPB5, HSPB1) HSP90 inhibits HSP70 expression and has an important role in the onset
of biomechanically-induced OA in rats
Decreased expression of HSPB5 and HSPB1 might contribute to an
altered chondrocyte metabolism in patients with OA
Annexins (II, V, VI) Extracellular Overexpressed in cartilage from patients with OA TLR4* 59, 84, 85, 86
matrix of Released by hypertrophic human chondrocytes
cartilage Annexin V induces apoptosis in calf chondrocytes
ATP Synovial fluid Released by procine chondrocytes in response to mechanical load P2X and P2Y2 87, 88, 89, 90
Stimulates PGE2 release from rabbit articular chondrocytes via purino-
activation of MAPKs receptors
High levels induce mineralization of porcine cartilage
Levels in synovial fluid in patients with OA are correlated with knee pain
Galectin 3 Extracellularly Increased level of expression in human osteoarthritic articular cartilage Integrin 1 91, 92, 93
in the joint compared with normal cartilage
(synovial fluid)* Markedly present in synovial tissue of patients with OA during
inflammatory phases
Induces joint swelling and OAlike lesions in the knee joints of mice
IL1 Synovial fluid Secreted mostly by macrophages, with lesser amounts secreted by IL1R 94, 95,96,
chondrocytes and synovial fibroblasts 97, 98
Produced in the early stages of OA in humans, and acts in autocrine and
paracrine manner
Increases porcine cartilage and meniscus catabolism by upregulating
NO production and total MMP activity
Thymosin 4 Serum and Increased expression in bovine cartilage after mechanical loading Not known 30, 99
synovial fluid Enhances MMP expression and activity in bovine chondrocytes
Positive correlation between serum and synovial fluid levels and
radiographic severity in patients with knee OA
Uric acid Serum and Synovial fluid level is correlated with IL18 and IL1 production and OA TLR2* and 31
synovial fluid severity in patients with knee OA TLR4*

NATURE REVIEWS | RHEUMATOLOGY ADVANCE ONLINE PUBLICATION | 7

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Table 1 (cont.) | Alarmins that are involved in osteoarthritis

Alarmins Localization of Roles in OA Receptors in Refs
secreted forms joint cells
Fibronectin Synovial fluid Increased expression in cartilage and synovial fluid from patients with OA TLR2, integrin 100, 101,
fragments Activate procatabolic responses in human chondrocytes 51 and CD44 102, 103, 104
Low-molecular- Synovial fluid* Stimulates matrix catabolism in bovine and human cartilage TLR2, TLR4 and 71,106
weight hyaluronan Regulates hypertrophic differentiation in mouse chondrocytes CD44
Biglycan Synovial fluid Increases catabolic marker expression in human cartilage TLR4 107
Decreases anabolic marker expression in human cartilage
Induces release of NO, interleukins, glycosaminoglycans and collagen by
human chondrocytes
Laminins Chondrocyte Increased levels found in human osteoarthritic cartilage matrix Syndecan 4 108, 109
pericellular Laminin subunit 4 aggravates human cartilage degeneration
matrix Laminin subunit 1 and laminin subunit 5 might have regenerative
activities in human osteoarthritic cartilage
Promote chondrogenesis, upregulate collagen typeII and downregulate
collagen typeI in human chondrocytes
Tenascin Synovial fluid Upregulated expression in human osteoarthritic cartilage TLR4 110, 111,
Promotes release of inflammatory mediators, upregulation of ADAMTS4 112, 113, 114
by human and bovine chondrocytes and release of proteoglycan by
bovine cartilage explants
Correlation between synovial fluid tenascin levels and cartilage
degradation and disease severity in patients with OA
Promotes cartilage repair in a mouse model
Aggrecan-derived Synovial fluid Exacerbate inflammation and catabolic activities in mouse and human TLR2 115, 116
fragments joint cells via activation of NFB
ERK, extracellular signal-regulated kinase; HSP, heat-shock protein; IL1R, IL1 receptor; MAPK, mitogen-activated protein kinase; MMP, matrix metalloproteinase;
NO, nitric oxide; NFB, nuclear factorB; OA, osteoarthritis; OARSI, Osteoarthritis Research Society International; PGE2, prostaglandin E2; RAGE, receptor for
advanced glycosylation end products; TLR, Toll-like receptor; VEGF, vascular endothelial growth factor. *Speculative (not formally demonstrated).

markedly decreased following inactivation of laminin
subunit4 with a blocking antibody108. Laminins might,
therefore, have a regulatory role in MMP expression,
leading to deleterious effects in osteoarthritic cartilage108.
Other chains of laminin monomers (laminin1 and
laminin5) have also been found to promote chondro
genesis and to be present at high levels in the pericellu
lar matrix in osteoarthritic cartilage109. Indeed, laminins
upregulate expression of collagen typeII, cartilage oli
gomeric matrix protein and aggrecan, and downregulate
expression of collagen typeI in human chondrocytes109.
Laminin expression might be activated as part of the
response to repair damaged tissue, eventually resulting
in destruction of cartilage108.
Tenascin is an extracellular matrix glycoprotein
that has a dual role. A high level of tenascin immuno
staining has been observed in areas of damaged human
osteoarthritic cartilage relative to normal cartilage110,111.
Moreover, tenascin levels in synovial fluid correlate
with the degree of cartilage degradation112 and the radi
ographic severity of OA113. Addition of tenascin to pri
mary human and bovine chondrocytes invitro induces
expression of ADAMTS4 and production of IL6, IL8,
PGE2 and nitrate through TLR4 signalling111. However,
genetic deficiency of tenascin in knockout mice is also
associated with accelerated development of OA and
delayed repair of cartilage damage114. A possible expla
nation for these results is that tenascin fragments gen
erated by proteases can induce a positive-feedback loop,
leading to cartilage catabolism. However, intact tenascin
promotes cartilage matrix repair112. Aggrecan-derived
fragments have also been found in synovial fluid from
patients with OA115. A 32residue fragment of aggrecan
has been identified as an OAassociated danger sig
nal that accelerates cartilage destruction by inducing
production of MMPs and cytokines via activation of
TLR2NFB in chondrocytes, synovial fibroblasts and
macrophages from both humans and mice116.
Several other molecules have roles in arthritis, but
confusion exists regarding their classification as alarm
ins. One example is neopterine117, which is found at
high levels in synovial fluid of patients with OA118. In
addition, some isoforms of 1433 proteins are released
extracellularly in the joint119,120, and have been iden
tified in exosomes released by activated immune
cells, which suggests that they are actively secreted
during inflammation120,121. The 1433 isoform, which
has been detected in synovial fluid of patients with OA,
is released from osteoblasts in response to mechanical
stress, and is similar to other alarmins (such as ATP
and thymosin4) in inducing degradation of cartilage
matrix119. Extracellular isoforms of 1433 proteins
could, therefore, be classified as alarmins that are derived
from activated or damaged joint cells. However, further
studies are required to identify the cell receptors and sig
nalling pathways associated with these proteins in joints.
Alarmins in inflammatory arthritis
During chronic inflammatory arthritis, synovial inflam
mation and joint-tissue damage are the main pathologi
cal processes that lead to the clinical symptoms of pain,
swelling, stiffness and loss of function122,123. Hallmarks
of chronic synovitis are hyperplasia of synovial lining
layer cells, infiltration of leukocytes and angiogenesis.

8 | ADVANCE ONLINE PUBLICATION www.nature.com/nrrheum

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

The synovial membrane becomes hyperplastic, expands
and forms villi. In humans, pannus the osteoclast-rich
portion of the synovial membrane destroys bone,
whereas enzymes secreted by neutrophils, synoviocytes
and chondrocytes degrade cartilage124.
Cells of the innate and adaptive immune systems
that are expanded or recruited to the inflamed syn
ovium include neutrophils, monocytes/macrophages,
Tcells, Bcells and DCs, as well as mast cells and res
ident fibroblast-like synoviocytes and macrophage-like
synoviocytes. APCs present arthritis-associated antigens
to Tcells. Concurrently, CD4+ Tcells that secrete IL2
and IFN infiltrate the synovial membrane. Bcells con
tribute via antigen presentation, and also through the
production of autoantibodies and cytokines. Activation
of Tcells and Bcells results in production of cytokines
and chemokines, creating a feedback loop for additional
interactions among Tcells, macrophages and Bcells124.
Innate immune cells are responsible for major
inflammatory effector mechanisms because they pro
duce activated proteinases (such as MMPs), ROS,
antimicrobial peptides, complement and a number of
chemokines and cytokines125. The release of cytokines
such as TNF, IL1 and IL6 has numerous proinflam
matory effects on recruited leukocytes, endothelial cells
and mesenchymal cells including fibroblast-like synovi
ocytes, chondrocytes and osteoclasts126. A network of
these signalling molecules links effector cells of innate
and adaptive immune responses, such as DCs and auto
reactive lymphocytes.
S100 proteins
Evidence of the importance of S100 proteins in inflam
matory arthritis includes their high extracellular expres
sion in inflamed joints, their role in cellular recruitment
and their cytotoxic effects127,128. The clinical relevance of
the proteins is suggested by their potential as markers of
inflammatory disease activity129 and radiographic pro
gression130. Elevated serum levels and expression within
the inflamed synovium or synovial fluid (or both) of
S100A8, S100A9, calprotectin and S100A12 has been
reported in several chronic inflammatory arthritides,
such as RA131,132, JIA133,134, psoriatic arthritis131,135 and
spondylarthropathy136.
Results from experimental models of arthritis indi
cate a direct role for S100 proteins in synovitis137. The
expression of S100A8 and S100A9 is closely correlated
with the intensity of inflammation in both antigen-
induced arthritis (AIA) and immune-complex-induced
arthritis in mice, and their expression reflects the activa
tion of synovial macrophages138,139. The role of these pro
teins is demonstrated in AIA in S100a9/ mice, which
have reduced joint swelling and proteoglycan depletion
relative to wild-type mice, and little MMP-mediated
cartilage destruction. Accordingly, intra-articular injec
tion of S100A8 into wild-type mice induces synovitis,
depletes cartilage proteoglycan levels and upregulates
synovial levels of mRNAs encoding S100A8, S100A9,
IL1 and MMPs44. Additionally, S100A8 mediates joint
inflammation and cartilage destruction in AIA140,141.
Calprotectin complexes can be used as biomarkers for
molecular imaging of local inflammatory activity142.
S100A12 can also trigger synovial inflammation in
collagen-induced arthritis in mice143.
In RA, activated phagocytes in the synovium express
S100A8 and S100A9144, with strong expression at the
cartilagepannus junction132. However, in psoriatic
arthritis, S100A8, S100A9 and S100A12 are expressed
in a perivascular pattern in the subsynovial layer131,135.
S100 levels are about tenfold higher in synovial fluid
than in serum in patients with active inflammatory
arthritis127,135, which might reflect an intra-articular
origin from activated myeloid cells145.
Because of the strong correlation between S100 pro
tein concentrations in serum and synovitis activity, these
proteins are useful diagnostic biomarkers for monitor
ing arthritis146. In JIA, calprotectin and S100A12 can
be used to monitor the response to treatment, to detect
subclinical inflammation133,147,148 and to support a diag
nosis of systemic JIA in a fever of unknown origin128.
Calprotectin has been established as a valuable diagnostic
and prognostic biomarker in RA129,130,149 (FIG.3).
HMGB1
The role of HMGB1 in chronic inflammatory arthritis
has been extensively studied. HMGB1 is released locally
at the site of joint inflammation150152. Intra-articular
HMGB1 injection induces arthritis in mice153, and
therapeutic blockade of HMGB1 prevents progres
sion in mouse models of arthritis150,154156. In mouse
collagen-induced arthritis, HMGB1 induces synovitis,
which can be blocked by HMGB1specific monoclonal
antibodies154. Levels of HMGB1 are upregulated in the
extracellular space and in the cytoplasm of macrophages,
synoviocytes, fibroblasts and endothelial cells in humans
in response to injury, infection or other inflammatory
stimuli157. HMGB1, TNF and IL1 are particularly
abundant in regions with tissue destruction, where pro
liferating synovial tissue invades cartilage and bone158.
Hypoxia might cause the release of HMGB1145. HMGB1
enhances the activity of tissue plasminogen activator and
MMPs, and activates osteoclastogenesis, thereby leading
to destruction of cartilage and bone159,160.
In human arthritis, extranuclear expression of HMGB1
can be detected locally in synovial fluid and serum of
patients with RA150,151,161 and JIA, and its expression is high
est in patients with early-onset disease162. HMGB1 levels
are higher in synovial fluid from patients with RA than in
synovial fluid from those with OA, and its expression is
prominent in endothelial cells and macrophages152. In RA,
HMGB1 might induce extension of the inamed synovium
by accelerating angiogenesis via activation of hypoxia-
inducible factor1163. HMGB1 also stimulates mac
rophages derived from synovial fluid to release proinflam
matory cytokines such as TNF, IL1 and IL6152. IL1,
along with bacterial DNA, viral RNA, endotoxin and other
microbial molecules, can form bifactorial complexes with
HMGB1 that enhance the innate immune response to both
factors synergistically164. HMGB1lipopolysaccharide
complexes might contribute to the development of RA,
because these complexes promote transformation of syn
ovial fibroblasts to a fibroblast-like phenotype165 (FIG.3).

NATURE REVIEWS | RHEUMATOLOGY ADVANCE ONLINE PUBLICATION | 9

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Joint Self-HSP-specic Treg cell

Healthy Chronically space
joint inamed
joint IL-10
Synovium
B cell 3 TGF
Angiogenesis 12
T cell Eector
IL-10
DC TNF T cell
Neutrophil 5
Synovium Monocyte
Hypertrophic Endothelium
Synoviocyte synovial activation
1 Synovitis Synoviocyte
lining
Cartilage
Synovium

Macrophage
Fibroblast
MAC1

Macrophage DC 10
Cartilage maturation
degradation 4 2
Monocyte Cartilage * 9 11
Subchondral Bone
bone erosion TNF
Neutrophil
IL-6
IL-1
Chondrocyte
Osteoclasts Calcied
cartilage 6
Necrotic
Monocyte cell
5
Activated
phagocyte Hypoxia
Complement
activation
Proinammatory
HMGB1 MMP HSP60 or HSP70 cytokines
8
S100 ROS ICAM Osteoclast 7
Cytokine or TLR2 or VCAM Chemotaxis
chemokine TLR4
RAGE Subchondral Osteoclast
HMGB1IL-1 HSP
bone activity
receptor * Prostaglandin

Figure 3 | Role of S100 proteins, high mobility group proteinB1 and heat-shock proteins in chronic
inflammatory arthritis. Synovial inflammation and joint-tissue damage are the main pathological Nature Reviews | Rheumatology
processes that occur
during chronic inflammatory arthritis. Hallmarks of chronic synovitis are hyperplasia of the synovial lining layer cells (1),
infiltration of leukocytes (2) and angiogenesis (3). Pannus (the osteoclast-rich portion of the synovial membrane) destroys
bone, whereas enzymes secreted by neutrophils, synoviocytes and chondrocytes degrade cartilage (4). S100 proteins and
high mobility group protein B1 (HMGB1) secreted from activated phagocytes and released from necrotic cells (5) activate
other phagocytes through Toll-like receptor (TLR) and receptor for advanced glycosylation end products (RAGE) signalling
to produce proinflammatory and catabolic factors (6), enhance osteoclastic activity (7) and lead to accumulation of other
phagocytes through chemotaxis (8). HMGB1 amplifies the effect of other cytokines (for example, through formation of
HMGB1IL1 complexes) (9) and induces maturation of dendritic cells (DCs) (10). S100 proteins activate endothelial cells
and induce extravasation of leukocytes by intercellular adhesion molecule (ICAM) and vascular cell adhesion molecule
(VCAM) upregulation on endothelial cells and macrophage 1 antigen (MAC1; also known as integrin M2) on leukocytes
(11). By contrast, regulatory Tcells (Treg cells) specific to self-heat-shock proteins (self-HSPs) inhibit effector Tcells and
dampen phagocyte activation (12). ROS, reactive oxygen species.

HSPs protein 1), TLRs, CD36 (also known as platelet glyco

HSPs have been identified as immune targets in RA and
JIA. Notably, in contrast to other alarmins, HSPs seem
to be associated with resolution rather than induction
of inflammation. No specific receptor has yet been iden
tified for the effects of secreted HSPs that are not anti
gen-specific, such as the release of cytokines (including
TNF, IL1 and IL12) and induced maturation of DCs;
potential candidates for such a receptor are CD91 (also
known as prolow-density lipoprotein receptor-related
protein 4), CD40 (also known as TNF receptor super
family member 5) and CD14 (REF.166).
The regulatory role of immune responses to HSPs
in inflammatory diseases was first shown for adjuvant
arthritis induced by mycobacteria167. The protective
effects of microbial HSPs have also now been seen in
many other experimental models of disease76.
In humans, HSP60 induces production of a subtype
of CD4+CD25+FOXP3+ Tregcells, which is relevant for

10 | ADVANCE ONLINE PUBLICATION www.nature.com/nrrheum

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

the resolution of inflammation168. In JIA, the presence of
these Treg cells has a favourable effect on prognosis168,169.
In RA, HSP60reactive Tcells suppress TNF production
by autologous peripheral blood mononuclear cells170,
and disease severity in RA might depend in part on the
ability of human HSP60 to induce the production of
anti-inflammatory IL10 by these Tcells171. Moreover,
HSP-specific Tregcells can suppress the activity of proin
flammatory effector Tcells by secreting IL4, IL10 and
transforming growth factor-168,172. Finally, transfer of
these HSP-specific Treg cells could inhibit inflammation
in animal models of arthritis76,172 (FIG.3).
IL33
IL33 is a member of the IL1 family. It is primarily
expressed as a nuclear protein but is released into the
extracellular milieu upon cell damage or stress to act
as an alarmin. IL33 has proinflammatory properties
through binding to IL1 receptor-like1 (also known
as ST2), which is expressed in cells of the innate and
adaptive immune systems. IL33 signalling is involved in
pathogenesis in several experimental models of arthritis,
contributes to erosion of bone and cartilage and induces
neutrophil migration173. However, induction of AIA and
immune-complex-induced arthritis is not affected by
IL33 deficiency in mice, and IL33 might have a mod
ifying rather than a pivotal role in the development of
immune-mediated arthritis174. Levels of IL33 are higher
in serum and synovial fluid from patients with RA than
in such samples from healthy donors or from patients
with OA, and these levels decrease with successful treat
ment, although this correlation with disease activity has
not been observed in all studies174.
Mitochondrial DAMPs
Mitochondrial DAMPs (mtDAMPs), such as oxidized
mitochondrial DNA (mtDNA) and Nformyl peptides,
can be released by damaged mitochondria, triggering
inflammation. In mice, human and mouse mtDNAs
induce arthritis mediated by monocytes and mac
rophages, along with the release of TNF from mouse sple
nocytes175. Additionally, extracellular mtDNA has been
detected in synovial fluid from patients with RA but not
in controls without RA175,176. Moreover, a higher mtDNA
mutation frequency has been observed in synovial tissue
from patients with RA or psoriatic arthritis than in con
trols177. The frequency of mtDNA mutations correlates
positively with synovial inflammation, and reduces after
anti-TNF therapy, in line with disease activity177. In vitro
treatment with TNF significantly induces mtDNA muta
tion177. This result suggests that mtDNA mutation and
oxidative damage driven by TNF are implicated in the
pathogenesis of inflammatory arthritis177.
The role of mtDAMPs in the pathogenesis of sys
temic inflammatory response syndrome (SIRS) has been
investigated. Extracellular mitochondrial Nformyl pep
tides and mtDNA activate human polymorphonuclear
leukocytes through fMet-Leu-Phe receptor (also known
as Nformyl peptide receptor) and TLR9, respectively,
leading to migration of polymorphonuclear neutrophils
and degranulation invitro and invivo178. Furthermore,
mtDAMPs are potent activators of human polymor
phonuclear neutrophils, causing shedding of Lselectin,
production of ROS, degranulation and release of IL8
(REF.179). A role for mtDNA in a mouse model of lupus-
like disease has also now been described180. mtDNA in
neutrophil extracellular traps (NETs) is proinflammatory
and induces the production of IFN and other cytokines
in a process that is dependent on the DNA sensor stimu
lator of interferon genes (STING) and TLR signalling180.
Other alarmins
The list of DAMPs that are released by activated or
necrotic cells, and of extracellular matrix molecules
that are upregulated by injury (or degraded after tissue
damage), continues to grow. The list includes mole
cules that induce inflammatory arthritis invivo in mice,
such as fibronectin extra domain A, which induces
TLR4dependent activation of mast cells161, and tenascin,
which induces TLR4dependent cytokine release181. In an
AIA model, tenascin-deficient mice were also protected
against joint erosion and tissue destruction181. Inhibition
of neutrophil elastase with ONO5046 reduces the inci
dence and the severity of collagen-induced arthritis in
both rats and mice182. Studies of synovial fluid or cartilage
extracts from patients with chronic inflammatory arthritis
suggest roles for molecules such as low-molecular-weight
hyaluronan183 (which is involved in activating APCs and
Tcells184) and biglycan185. Finally, ATP induces chemot
axis, release of cytokines and generation of ROS via P2X
and P2Y receptors186.
Clinical utility of alarmins
The level of alarmins in serum and synovial fluid can be
a useful diagnostic and prognostic biomarker of arthritis.
Similar to inflammatory biomarkers such as Creactive
protein and erythrocyte sedimentation rate, serum levels
of S100 proteins correlate with disease activity in sev
eral inflammatory conditions, such as sepsis, Kawasaki
disease, psoriasis and cystic fibrosis14. Their concentra
tions also correlate with synovitis activity and are useful
diagnostic markers for monitoring arthritis146. In various
inflammatory arthritides, serum levels of calprotectin
show better correlation with disease activity and joint
damage than do classical markers of inflammation135,187.
Additionally, serum concentrations of S100A12 in chil
dren with systemic-onset JIA can differentiate this condi
tion from infections with high sensitivity and specificity,
in contrast to Creactive protein and erythrocyte sedi
mentation rate188. Alarmins also provide more-specific
indications than traditional markers, because of their
local expression and direct release in response to tissue
injuries14.
Alarmins could be promising biomarkers for the
prediction of responses to biological therapies in arthri
tis. Measurement of calprotectin could add value to the
assessment of patients with RA who undergo biologic
treatment129,149. A decreasing serum level of S100A12
could reflect attenuated synovial neutrophil activa
tion during successful anti-inflammatory therapy (via
intra-articular corticosteroids) in RA145. Similarly, serum
and synovial fluid levels of IL33 are high in patients

NATURE REVIEWS | RHEUMATOLOGY ADVANCE ONLINE PUBLICATION | 11

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Table 2 | Main roles of key alarmins in osteoarthritis and inflammatory arthritis

Alarmins Main effects
Osteoarthritis Inflammatory arthritis
High mobility group Synovial inflammation Synovial inflammation
proteinB1 Cartilage degradation Cartilage and bone destruction
Osteophyte formation Angiogenesis
Endochondral ossification Activation of innate immune response
Bone-cell recruitment Proinflammatory cytokine production
Hypertrophic differentiation
Production of proinflammatory cytokines
and chemokines
Heat-shock proteins Synovial activation Resolution of inflammation
Matrix catabolism
Bone remodelling
Protective effects against chondrocyte
apoptosis and cartilage degeneration
IL33 No known effect Cartilage and bone erosion
Neutrophil migration
S100 proteins Synovial inflammation Synovial inflammation
Cartilage degradation Cartilage destruction
Angiogenesis Activation and recruitment of
Osteophyte formation inflammatory cells
Inflammatory cytokine production
Hypertrophic differentiation

with RA, but decrease with treatment174. Because of its Conclusions

association with disease severity, HMGB1 is also thought
be a biomarker in arthritis that could predict response
to treatment189.
Alarmins could serve as targets for future therapeu
tic interventions in arthritis. Inhibition of alarmin path
ways could be therapeutically beneficial by preventing
destruction of joints. Blockade of the proinflammatory
effects of S100A8 and S100A9 has been effective in
mouse models190, and evidence suggests that S100A12-
dependent cell activation and recruitment can be inhib
ited by either antiS100A12 antibodies or soluble RAGE
constructs191. In addition to the use of agents to block
alarmin pathways, immune intervention by epitope-
specific therapy might be a promising approach,
potentially leading to immune tolerance to alarmins.
Preliminary clinical trials with epitope-specific immu
notherapy have shown promising results in patients with
RA192,193. However, further confirmation of this benefi
cial effect needs to be demonstrated in larger cohorts,
with long-term followup.
Alarmins have important roles in the pathological pro
cess during arthritis. The excessive and chronic release
of alarmins results in aberrant inflammation of joints
and destruction of cartilage. However, some of these
molecules can promote tissue repair and the resolution
of inflammation after joint damage. A variety of alarm
ins, originating from distinct joint cells, are released
extracellularly to create networks that are able to form
feedback loops, amplifying damage and inflammation of
joints. Among the different forms of arthritis, common
alarmins include (but are not limited to) HMGB1, HSPs,
IL33 and S100 proteins; these proteins have notable sim
ilarities and differences in their effects (TABLE2). When
released into the extracellular milieu, these mediators
might crosslink different receptors on the same cell and
thus simultaneously activate many signal transduction
pathways within joint cells, thereby leading to destructive
inflammation. The development of preventive and thera
peutic strategies for arthritis might, therefore, depend on
the early detection and inhibition of these danger signals.

1. Palazzo,C. etal. Risk factors and burden of
osteoarthritis. Ann. Phys. Rehabil. Med.59, 134138
(2016).
2. Abramson,S.B. & Attur,M. Developments in the
scientific understanding of osteoarthritis. Arthritis.
Res. Ther.11, 227 (2009).
3. Bijlsma,J.W.etal. Osteoarthritis: an update with
relevance for clinical practice. Lancet.377,
21152126 (2011).
4. Firestein,G.S. Evolving concepts of rheumatoid
arthritis. Nature 423, 356361 (2003).
5. Foell,D. Wittkowski,H. & Roth,J. Mechanisms of
disease: a DAMP view of inflammatory arthritis. Nat.
Clin. Pract. Rheumatol.7, 382390 (2007).
6. Oppenheim,J.J. & Yang,D. Alarmins: chemotactic
activators of immune responses. Curr. Opin. Immunol.
17, 359365 (2005).
7. Liu-Bryan,R. & Terkeltaub,R. Emerging regulators of
the inflammatory process in osteoarthritis. Nat. Rev.
Rheumatol. 11, 3544 (2015).
8. Houard,X., Goldring,M.B. & Berenbaum,F. Homeostatic
mechanisms in articular cartilage and role of inflammation
in osteoarthritis. Curr. Rheumatol. Rep. 15, 375 (2013).
9. Bianchi,M.E. DAMPs, PAMPs and alarmins: all we need
to know about danger. J. Leukoc. Biol. 81, 15 (2007).
10. Harris,H.E. & Raucci,A. Alarmin(g) news about
danger: workshop on innate danger signals and
HMGB1. EMBO Rep. 7, 774778 (2006).
11. Pugin, J. Dear SIRS, the concept of alarmins makes a
lot of sense! Intensive Care Med. 34, 218221 (2008).
12. Pisetsky,D.S. etal. HMGB1 and microparticles as
mediators of the immune response to cell death.
Antioxid. Redox. Signal. 15, 22092219 (2011).
13. Loeser,R.F. etal. Osteoarthritis: a disease of the joint
as an organ. Arthritis Rheum. 64, 16971707 (2012).
14. Chan,J.K. etal. Alarmins: awaiting a clinical
response. J.Clin. Invest. 122, 27112719 (2012).
15. Manfredi,A.A. etal. Regulation of dendritic- and
Tcell fate by injury-associated endogenous signals.
Crit. Rev. Immunol. 29, 6986 (2009).
16. Tamaki,Y. etal. Expression of Toll-like receptors and
their signaling pathways in rheumatoid synovitis.
J.Rheumatol. 38, 810820 (2011).
17. Dumitriu,I.E. etal. Release of high mobility group
BOX1 by dendritic cells controls Tcell activation via
the receptor for advanced glycation end products.
J.Immunol. 174, 75067515 (2005).
18. Yang,D. etal. High mobility group box1 protein
induces the migration and activation of human
dendritic cells and acts as an alarmin. J.Leukoc. Biol.
81, 5966 (2007).
19. Yang,D. etal. The alarmin functions of high-mobility
group proteins. Biochim. Biophys. Acta 1799,
157163 (2010).
20. Goldring,M.B. & Goldring,S.R. Osteoarthritis.
J.Cell. Physiol. 213, 626634 (2007).
21. Denise,L.C. etal. Inflammation-induced chondrocyte
hypertrophy is driven by receptor for advanced
glycation end products. J.Immunol. 175, 82968302
(2005).

12 | ADVANCE ONLINE PUBLICATION www.nature.com/nrrheum

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

22. Seol,D. etal. Chondrogenic progenitor cells respond
to cartilage injury. Arthritis Rheum. 64, 36263637
(2012).
23. Buckwalter,J.A. etal. The roles of mechanical
stresses in the pathogenesis of osteoarthritis:
implications for treatment of joint injuries. Cartilage 4,
286294 (2013).
24. Glyn-Jones,S. etal. Osteoarthritis. Lancet 386,
376387 (2015).
25. Weinans,H. etal. Pathophysiology of peri-articular
bone changes in osteoarthritis. Bone 51, 190196
(2012).
26. Sanchez,C. etal. Regulation of subchondral bone
osteoblast metabolism by cyclic compression. Arthritis
Rheum. 64, 11931203 (2012).
27. Bidwell,J.P. etal. Is HMGB1 an osteocyte alarmin?
J.Cell. Biochem. 103, 16711680 (2008).
28. Ke,X. etal. Synovial fluid HMGB1 levels are
associated with osteoarthritis severity. Clin. Lab.61,
809818 (2015).
29. Wang,L.C. etal. S100A12 levels in synovial fluid may
reflect clinical severity in patients with primary knee
osteoarthritis. Biomarkers 18, 216220 (2013).
30. Wei,M. etal. Increased thymosin 4 levels in the
serum and SF of knee osteoarthritis patients correlate
with disease severity. Regul. Pept.185, 3436
(2013).
31. Denoble,A.E. etal. Uric acid is a danger signal of
increasing risk for osteoarthritis through
inflammasome activation. Proc. Natl. Acad. Sci. U.S.A.
108, 20882093 (2011).
32. Ayral,X. etal. Synovitis: a potential predictive factor
of structural progression ofmedialtibiofemoralknee
osteoarthritis: results of 1 year longitudinal
arthroscopic studi in 422 patients. Osteoarthritis
Cartilage 13, 361367 (2005).
33. Benito,M.J. etal. Synovial tissue inflammation in
early and late osteoarthritis. Ann. Rheum. Dis.64,
12631267 (2005).
34. Da,R.R. etal. B cell clonal expansion and somatic
with osteoarthritis genes in the synovial membrane of
patients hypermutation of Ig variable heavy chain.
J.Immunol.178, 557565 (2007).
35. Sunahori,K. etal. The S100A8/A9 heterodimer
amplifies proinflammatory cytokine production by
macrophages via activation of nuclear factor kappa B
and p38 mitogen-activated protein kinase in
rheumatoid arthritis. Arthritis Res. Ther.8, 69
(2006).
36. Daghestani,H.N.etal. Inflammatory biomarkers in
osteoarthritis. Osteoarthritis Cartilage 23,
18901896 (2015).
37. Van Lent,P.L. etal. Active involvement of alarmins
S100A8 and S100A9 in the regulation of synovial
activation and joint destruction during mouse and
human osteoarthritis. Arthritis Rheum. 64,
14661476 (2012).
38. Liu-Bryan,R. & Terkeltaub,R.T. The growing array of
innate inflammatory ignition switches in osteoarthritis.
Arthritis Rheum. 64, 20552058 (2012).
39. Marenholz,I. etal. S100 proteins in mouse and man:
from evolution to function and pathology (including an
update of the nomenclature). Biochem. Biophys. Res.
Commun. 322, 11111122 (2004).
40. Donato,R. Intracellular and extracellular roles of
S100 proteins. Microsc. Res. Tech. 60, 540551
(2003).
41. Ruse,M. etal. S100A7, S100A10, and S100A11 are
transglutaminase substrates. Biochem. 40,
31673173 (2001).
42. Meijer, B., Gearry, R. B. & Day, A. S. The role of
S100A12 as a systemic marker of inflammation. Int.
J.Inflam.2012, 907078 (2012).
43. Yammani,R.R. etal. Interleukin7 stimulates
secretion of S100A4 by activating the JAK-STAT
signaling pathway in human articular chondrocytes.
Arthritis Rheum. 60, 792800 (2009).
44. Donato,R. etal. Functions of S100 proteins. Curr.
Mol. Med. 13, 2457 (2013).
45. Yammani,R.R. etal. Increase in production of matrix
metalloproteinase 13 by human articular
chondrocytes due to stimulation with S100A4: role of
the receptor for advanced glycation end products.
Arthritis Rheum.54, 29012911 (2006).
46. Van den Berg,W.B. Osteoarthritis year 2010 in
review: pathomechanisms. Osteoarthritis Cartilage
19, 338341 (2011).
47. Van Lent,P.L. Stimulation of chondrocyte-mediated
cartilage destruction by S100A8 in experimental
murine arthritis. Arthritis Rheum. 58, 37763787
(2008).
48. Nacken,W. etal. S100A9/S100A8: myeloid
representatives of the S100 protein family as
prominent players in innate immunity. Microsc. Res.
Tech.60, 569580 (2003).
49. Vogl,T. etal. Pro-inflammatory S100A8 and S100A9
proteins: self-assembly into multifunctional native and
amyloid complexes. Int. J.Mol. Sci. 13, 28932917
(2012).
50. Van Lent,P.L. etal. Myeloid-related proteins S100A8/
S100A9 regulate joint inflammation and cartilage
destruction during antigen-induced arthritis. Ann.
Rheum. Dis. 67, 17501758 (2008).
51. Vogl,T. etal. Mrp8 and Mrp14 are endogenous
activators of Toll-like receptor 4, promoting lethal,
endotoxin-induced shock. Nat. Med. 13, 10421049
(2008).
52. Schelbergen,R.F.P. etal. Alarmins S100A8 and
S100A9 elicit a catabolic effect in human
osteoarthritic chondrocytes that is dependent on Toll-
like Receptor 4. Arthritis Rheum. 64, 14771487
(2012).
53. Cunningham,C.C. etal. Osteoarthritis-associated
basic calcium phosphate crystals induce pro-
inflammatory cytokines and damage-associated
molecules via activation of Syk and PI3 kinase. Clin.
Immunol. 144, 228236 (2012).
54. Schelbergen,R.F.P. etal. Alarmins S100A8/S100A9
aggravate osteophyte formation in experimental
osteoarthritis and predict osteophyte progression in
early human symptomatic osteoarthritis. Ann. Rheum.
Dis.75, 218225 (2014).
55. Van Den Bosch,M.H. etal. The alarmins S100A8/A9
induce canonical Wnt signaling in murine knee joints;
implications for osteoarthritis. Arthritis
Rheumatol.68, 152163 (2016).
56. Schelbergen,R.F.P. etal. Prophylactic treatment with
S100A9 inhibitor paquinimod reduces pathology in
experimental collagenase-induced osteoarthritis. Ann.
Rheum. Dis. 74, 22542258 (2015).
57. Miller,R. etal. Damage-associated molecular patterns
generated in osteoarthritis directly excite murine
nociceptive neurons through Toll-like receptor 4.
Arthritis Rheum. 67, 29332943 (2015).
58. Song, C. etal. Regulation of inflammatory response in
human chondrocytes by lentiviral mediated RNA
interference against S100A10. Inflamm. Res. 61,
12191227 (2012).
59. Swisher,J. F., Burton, N., Bacot, S. M., Vogel, S. N. &
Feldman, G. M. Annexin A2 tetramer activates human
and murine macrophages through TLR4. Blood115,
549558 (2010).
60. Cecil,D.L.& Terkeltaub,R. Transamidation by
transglutaminase 2 transforms S100A11 calgranulin
into a procatabolic cytokine for chondrocytes.
J.Immunol. 180, 83788385 (2008).
61. Kosaki,A. etal. Increased plasma S100A12
(ENRAGE) levels in patients with type2 diabetes.
J.Clin. Endocrinal. Metab.89, 54235428 (2004).
62. Nakashima,M. etal. Role of S100A12 in the
pathogenesis of osteoarthritis. Biochem. Biophys. Res.
Commun.422, 508514 (2012).
63. Han,M. etal. Identification of osteoarthritis
biomarkers by proteomic analysis of synovial fluid.
J.Int. Med. Res. 40, 22432250 (2012).
64. Loeser,R.F. etal. Articular chondrocytes express the
receptor for advanced glycation end products:
potential role in osteoarthritis. Arthritis Rheum. 52,
23762385 (2005).
65. Magna,M.& Pisetsky,D.S. The role of HMGB1 in the
pathogenesis of inflammatory and autoimmune
diseases. Mol. Med. 20, 138146 (2014).
66. Li,Z.C. etal. Correlation of synovial fluid HMGB1
levels with radiographic severity of knee osteoarthritis.
Clin. Invest. Med. 34, 298303 (2011).
67. Amin,A.R. & Islam,A.B. Genomic analysis and
differential expression of HMG and S100A family in
human arthritis: upregulated expression of
chemokines, IL8 and nitric oxide by HMGB1. DNA
Cell Biol.33, 550565 (2014).
68. Terada,C. etal. Gene expression and localization of
high-mobility group box chromosomal protein1
(HMGB1) in human osteoarthritic cartilage. Acta
Med. Okayama 65, 369377 (2011).
69. Kyostio-Moore,S. etal. STR/ort mice, a model for
spontaneous osteoarthritis, exhibit elevated levels of
both local and systemic inflammatory markers. Comp.
Med.61, 346355 (2011).
70. Garcia-Arnandis,I. etal. Haem oxygenase1 down-
regulates high mobility group box 1 and matrix
metalloproteinases in osteoarthritic synoviocytes.
Rheumatology (Oxford) 49, 854861 (2010).
71. Liu-Bryan,R. & Terkeltaub,R. Chondrocyte innate
immune MyD88dependent signaling drives pro-
catabolic effects of the endogenous TLR2/TLR4
ligands LMWHA and HMGB1. Arthritis Rheum. 62,
20042012 (2010).
72. Taniguchi,N. etal. Stage-specific secretion of HMGB1
in cartilage regulates endochondral ossification. Mol.
Cell. Biol.27, 56505663 (2007).
73. Heinola,T. etal. High mobility group box1 (HMGB1)
in osteoarthritic cartilage. Clin. Exp. Rheumatol. 28,
511518 (2010).
74. Garca-Arnandis,I. etal. High mobility group box1
potentiates the pro-inflammatory effects of
interleukin1 in osteoarthritic synoviocytes. Arthritis
Res. Ther. 12, R165 (2010).
75. Whmaa,H. etal. High mobility group box protein 1
in complex with lipopolysaccharide or IL1 promotes
an increased inflammatory phenotype in synovial
fibroblasts. Arthritis Res. Ther. 13, R136 (2011).
76. Van Eden,W., van der Zee,R. & Prakken,B. Heat-
shock proteins induce Tcell regulation of chronic
inflammation. Nat. Rev. Immunol. 5, 318330 (2005).
77. Kubo,T. etal. Stress-inducedproteins in chondrocytes
from patients with osteoarthritis. Arthritis Rheum.28,
11401145 (1985).
78. Takahashi,K. etal. Localization of heat shock protein
in osteoarthritic cartilage. Scand. J. Rheumatol. 26,
368375 (1997).
79. Terauchi,R. etal. Hsp70preventsnitricoxide-
inducedapoptosisinarticularchondrocytes. Arthritis
Rheum.48, 15621568 (2003).
80. Siebelt,M. etal. Hsp90inhibitionprotects against
biomechanically induced osteoarthritis in rats.
Arthritis Rheum.65, 21022112 (2013).
81. Lambrecht,S. Heat-shock proteins in stromal joint
tissues: innocent bystanders or disease-initiating
proteins? Rheumatology (Oxford)53, 223232
(2014).
82. Lambrecht,S. etal. Differential expression of
BCrystallin and evidence of its role as a mediator of
matrix gene expression in osteoarthritis. Arthritis
Rheum. 60, 179188 (2009).
83. Lambrecht,S. etal. Proteome characterization of
human articular chondrocytes leads to novel insights
in the function of small heat-shock proteins in
chondrocyte homeostasis. Osteoarthritis Cartilage
18, 440446 (2010).
84. Kirsch,T. etal. Activation of annexin II and V
expression, terminal differentiation, mineralization
and apoptosis in human osteoarthritic cartilage.
Osteoarthritis Cartilage 8, 294302 (2000).
85. Pfander,D., Swoboda,B. & Kirsch,T. Expression of
early and late differentiation markers (proliferating cell
nuclear antigen, syndecan3, annexin VI, and alkaline
phosphatase) by human osteoarthritic chondrocytes.
Am. J.Pathol. 159, 17771783 (2001).
86. Ea,H.K. et al. Annexin 5 overexpression increased
articular chondrocyte apoptosis induced by basic
calcium phosphate crystals. Ann. Rheum. Dis. 67,
16171625 (2008).
87. Graff,R.D. etal. ATP release by mechanically loaded
porcine chondrons in pellet culture. Arthritis Rheum.
43, 15711579 (2000).
88. Kumahashi,N. etal. Correlation of changes in pain
intensity with synovial fluid adenosine triphosphate
levels after treatment of patients with osteoarthritis of
the knee with high-molecular-weight hyaluronic acid.
Knee 18, 160164 (2011).
89. Berenbaum,F. etal. Concomitant recruitment of
ERK1/2 and p38MAPK signalling pathway is required
for ativation of cytoplasmic phapholipase A2 via ATP
in articular chondrocytes. J.Biol. Chem.278,
1368013687 (2003).
90. Ryan,L.M. etal. ATP-induced chondrocalcinosis.
Arthritis Rheum. 35, 15201525 (1992).
91. Guvremont,M. etal. Galectin3 surface expression
on human adult chondrocytes: a potential substrate
for collagenase3. Ann. Rheum. Dis. 63, 636643
(2004).
92. Ohshima,S. etal. Galectin 3 and its binding protein in
rheumatoid arthritis. Arthritis Rheum.48,
27882795 (2003).
93. Janelle-Montcalm,A. etal. Extracellular localization of
galectin3 has a deleterious role in joint tissues.
Arthritis Res. Ther. 9, R20 (2007).
94. Hirsiger,S. etal. Danger signals activating the
immune response after trauma. Mediators
Inflamm.2012, 315941 (2012).
95. Goldring,M.B. etal. Osteoarthritis and cartilage: the
role of cytokines. Curr. Rheumatol. Rep.2, 459465
(2000).

NATURE REVIEWS | RHEUMATOLOGY ADVANCE ONLINE PUBLICATION | 13

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

96. Pelletier,J.P. etal. Are cytokines involved in
osteoarthritic pathophysiology? Semin. Arthritis
Rheum.20 (Suppl. 2), 1225 (1991).
97. Towle,C.A. etal. Detectionofinterleukin1in
thecartilageofpatientswith osteoarthritis: apossible 120. Maksymowych,W.P.& Marotta,A. 1433: a novel
autocrine/paracrine roleinpathogenesis.
Osteoarthritis Cartilage5, 293300 (1997).
98. McNulty,A.L. etal. Synovial fluid concentrations and
relative potency of interleukin1 alpha and beta in
cartilage and meniscus degradation. J.Orthop.
Res.31, 10391045 (2013).
99. Blain,E.J.etal. The effect of thymosin 4 on articular
cartilage chondrocyte matrix metalloproteinase
expression. Biochem. Soc. Trans.30, 879882 (2002).
100. Homandberg,G.A. etal. Cartilage damaging
activities of fibronectin fragmentsderived from
cartilage and synovialfluid. Osteoarthritis Cartilage6,
231244 (1998).
101. Pichika,R. & Homandberg,G.A. Fibronectin
fragments elevate nitric oxide (NO) and inducible NO
synthetase (iNOS) levels in bovine cartilage and iNOS
inhibitors block fibronectin fragment mediated
damage and promote repair. Inflamm. Res.53,
405412 (2004).
102. Homandberg,G.A. & Hui,F. Associationof
proteoglycan degradation with catabolic cytokine
andstromelysin release from cartilageculturedwith
fibronectin fragments. Arch. Biochem. Biophys.334,
325331 (1996).
103. Hwang,H.S. etal. Fibronectin ragment induced
expression of matrix metallo-proteinases is mediated
by MyD88dependent TLR2 signalingpathway in
human chondrocytes. Arthritis Res. Ther.17, 320
(2015).
104. Yasuda,T. Cartilage destruction by matrix
degradation products. Mod. Rheumatol. 16,
197205 (2006).
105. Jyang,D. etal. Hyaluronan in tissue injury and repair.
Annu. Rev. Cell Dev. Biol.23, 435461 (2007).
106. Ohno,S. etal. Hyaluronan oligosaccharides induce
matrix metalloproteinase 13 via transcriptional
activation of NFB and p38 MAP kinase in articular
chondrocytes. J.Biol. Chem.281, 1795217960
(2006).
107. Barreto,G. etal. Solublebiglycan: a potential
mediator of cartilage degradation inosteoarthritis.
Arthritis Res. Ther. 17, 379 (2015).
108. Fuerst,F.C. etal. Regulation of MMP3 by laminin
alpha 4 in human osteoarthritic cartilage. Scand.
J.Rheumatol.40, 494496 (2011).
109. Schminke,B. etal. Lamininsandnidogensin
thepericellularmatrixofchondrocytes:
theirroleinosteoarthritis
andchondrogenicdifferentiation. Am. J.Pathol. 186,
410418(2016).
110. Nakoshi,Y. etal. Regulation of tenascinC expression
by tumor necrosis factor-alphainculturedhuman
osteoarthritischondrocytes. J.Rheumatol.35,
147152 (2008).
111. Patel,L. etal. TenascinC induces inflammatory
mediators and matrix degradation in osteoarthritic
cartilage. BMC Musculoskelet. Disord.12, 164
(2011).
112. Chockalingam,P.S. etal. TenascinC levels in synovial
fluid are elevated after injury to the human and canine
joint and correlate with markers of inflammation and
matrix degradation. Osteoarthritis Cartilage 21,
339345 (2013).
113. Hasegawa,M. etal. TenascinC concentration in
synovial fluid correlates with radiographic progression
of knee osteoarthritis. J.Rheumatol.31, 20212026
(2004).
114. Okamura,N.etal. Deficiency of tenascinC delays
articular cartilage repair in mice. Osteoarthritis
Cartilage18, 839848 (2010).
115. Lohmander,L.S. etal. The structure of aggrecan
fragments in human synovial fluid. Evidence that
aggrecanase mediates cartilage degradation in
inflammatory joint disease, joint injury, and
osteoarthritis. Arthritis Rheum.36, 12141222
(1993).
116. Lees,S. etal. Bioactivity in an aggrecan 32mer
fragment is mediated via Toll-like receptor 2. Arthritis
Rheum. 67, 12401249 (2015).
117. Manson,J., Thiemermann,C. & Brohi,K. Trauma
alarmins as activators of damage-induced
inflammation. Br. J.Surg. 99, 1220 (2012).
118. Zhou,S.J., Sun,Z.X. & Liu,J. Neopterin
concentrations in synovial fluid may reflect disease
severity in patients with osteoarthritis. Scand. J.Clin.
Lab. Invest. 73, 344348 (2013).
119. Priam,S. etal. Identification of soluble 1433 as a 141. Grevers,L.C. etal. S100A8 enhances osteoclastic
novel subchondral bone mediator involved in cartilage
degradation in osteoarthritis. Arthritis Rheum. 65,
18311842 (2013).
13651375 (2011).
biomarker platform for rheumatoid arthritis. Clin. Exp.
Rheumatol.32, 3539 (2014).
121. Thry,C. etal. Membrane vesicles as conveyors of
immune responses. Nat. Rev. Immunol. 9, 581593
(2009).
122. Prakken,B., Albani,S. & Martini,A. Juvenile
idiopathic arthritis. Lancet 377, 21382149
(2011).
123. Lee,D.M. & Weinblatt,M.E. Rheumatoid arthritis.
Lancet 358, 903911 (2001).
124. Choy,E. Understanding the dynamics: pathways
involved in the pathogenesis of rheumatoid arthritis.
Rheumatology (Oxford) 51 (Suppl. 5), v3v11 (2012).
125. Muller-Ladner,U., Pap,T., Gay,R.E., Neidhart,M. & 146. Foell,D. & Roth,J. Proinflammatory S100 proteins in
Gay,S. Mechanisms of disease: the molecular and
cellular basis of joint destruction in rheumatoid
arthritis. Nat. Clin. Pract. Rheumatol.1, 102110
(2005).
126. McInnes,I.B. & Schett,G. Cytokines in the
pathogenesis of rheumatoid arthritis. Nat. Rev.
Immunol. 7, 429442 (2007).
127. Frosch,M. etal. Myeloid-related proteins 8 and 14
are specifically secreted during interaction of
phagocytes and activated endothelium and are useful
markers for monitoring disease activity in
pauciarticular-onset juvenile rheumatoid arthritis.
Arthritis Rheum. 43, 628637 (2000).
128. Kessel,C., Holzinger,D. & Foell,D. Phagocyte-derived
S100 proteins in autoinflammation: putative role in
pathogenesis and usefulness as biomarkers. Clin.
Immunol. 147, 229241 (2013).
129. Choi,I.Y. etal. MRP8/14 serum levels as a strong
predictor of response to biological treatments in
patients with rheumatoid arthritis. Ann. Rheum. Dis.
74, 499505 (2015).
130. Hammer,H.B. etal. Calprotectin (a major S100
leucocyte protein) predicts 10year radiographic
progression in patients with rheumatoid arthritis. Ann.
Rheum. Dis. 69, 150154 (2010).
131. Foell,D. etal. Expression of the pro-inflammatory
protein S100A12 (ENRAGE) in rheumatoid and
psoriatic arthritis. Rheumatology (Oxford) 42,
13831389 (2003).
132. Youssef,P. etal. Expression of myeloid related
proteins (MRP) 8 and 14 and the MRP8/14
heterodimer in rheumatoid arthritis synovial
membrane. J.Rheumatol. 26, 25232528 (1999).
133. Foell,D. etal. Monitoring neutrophil activation in
juvenile rheumatoid arthritis by S100A12 serum
concentrations. Arthritis Rheum. 50, 12861295
(2004).
134. Frosch,M. etal. Expression of myeloid-related
proteins 8 and 14 in systemic-onset juvenile
rheumatoid arthritis. Arthritis Rheum. 48,
26222626 (2003).
135. Kane,D. etal. Increased perivascular synovial
membrane expression of myeloid-related proteins in
psoriatic arthritis. Arthritis Rheum. 48, 16761685
(2003).
136. Kruithof,E. etal. Synovial histopathology of
psoriatic arthritis, both oligo- and polyarticular,
resembles spondyloarthropathy more than it does
rheumatoid arthritis. Arthritis Res. Ther. 7,
R569R580 (2005).
137. Imamichi,T. etal. Expression and cloning of migration
inhibitory factor-related protein (MRP)8 and MRP14
in arthritis-susceptible rats. Biochem. Biophys. Res.
Commun. 194, 819825 (1993).
138. van Lent,P. etal. The inhibitory receptor FcRII
reduces joint inflammation and destruction in
experimental immune complex-mediated arthritides
not only by inhibition of FcRI/III but also by efficient
clearance and endocytosis of immune complexes. Am.
J.Pathol. 163, 18391848 (2003).
139. Nabbe,K.C. etal. Coordinate expression of activating
Fc receptors I and III and inhibiting Fc receptor
typeII in the determination of joint inflammation and
cartilage destruction during immune complex-
mediated arthritis. Arthritis Rheum. 48, 255265
(2003).
140. van Lent,P.L. etal. S100A8 causes a shift toward
expression of activatory Fc receptors on
macrophages via Toll-like receptor 4 and regulates Fc
receptor expression in synovium during chronic
experimental arthritis. Arthritis Rheum. 62,
33533364 (2010).
bone resorption invitro through activation of Toll-like
receptor 4: implications for bone destruction in
murine antigen-induced arthritis. Arthritis Rheum. 63,
142. Vogl,T. etal. Alarmin S100A8/S100A9 as a
biomarker for molecular imaging of local inflammatory
activity. Nat. Commun. 5, 4593 (2014).
143. Hofmann,M.A. etal. RAGE and arthritis: the G82S
polymorphism amplifies the inflammatory response.
Genes Immun. 3, 123135 (2002).
144. Odink,K. etal. Two calcium-binding proteins in
infiltrate macrophages of rheumatoid arthritis. Nature
330, 8082 (1987).
145. Wittkowski,H. etal. Effects of intra-articular
corticosteroids and anti-TNF therapy on neutrophil
activation in rheumatoid arthritis. Ann. Rheum. Dis.
66, 10201025 (2007).
arthritis and autoimmune disease. Arthritis Rheum.
50, 37623771 (2004).
147. Holzinger,D. etal. The Toll-like receptor 4 agonist
MRP8/14 protein complex is a sensitive indicator for
disease activity and predicts relapses in systemic-
onset juvenile idiopathic arthritis. Ann. Rheum. Dis.
71, 974980 (2012).
148. Moncrieffe,H. etal. A subgroup of juvenile idiopathic
arthritis patients who respond well to methotrexate
are identified by the serum biomarker MRP8/14
protein. Rheumatology (Oxford) 52, 14671476
(2013).
149. Hammer,H.B., Fagerhol,M.K., Wien,T.N. &
Kvien,T.K. The soluble biomarker calprotectin (an
S100 protein) is associated to ultrasonographic
synovitis scores and is sensitive to change in patients
with rheumatoid arthritis treated with adalimumab.
Arthritis Res. Ther. 13, R178 (2011).
150. Hamada,T. etal. Extracellular high mobility group
box chromosomal protein 1 is a coupling factor for
hypoxia and inflammation in arthritis. Arthritis
Rheum. 58, 26752685 (2008).
151. Kokkola,R. etal. High mobility group box
chromosomal protein 1: a novel proinflammatory
mediator in synovitis. Arthritis Rheum. 46,
25982603 (2002).
152. Taniguchi,N. etal. High mobility group box
chromosomal protein 1 plays a role in the
pathogenesis of rheumatoid arthritis as a novel
cytokine. Arthritis Rheum. 48, 971981 (2003).
153. Pullerits,R. etal. High mobility group box
chromosomal protein 1, a DNA binding cytokine,
induces arthritis. Arthritis Rheum. 48, 16931700
(2003).
154. Kokkola,R. etal. Successful treatment of collagen-
induced arthritis in mice and rats by targeting
extracellular high mobility group box chromosomal
protein 1 activity. Arthritis Rheum. 48, 20522058
(2003).
155. Ostberg,T. etal. Oxaliplatin retains HMGB1
intranuclearly and ameliorates collagen
typeIIinduced arthritis. Arthritis Res. Ther. 10, R1
(2008).
156. Tian,J. etal. Toll-like receptor 9dependent activation
by DNA-containing immune complexes is mediated by
HMGB1 and RAGE. Nat. Immunol. 8, 487496
(2007).
157. Lotze,M.T. &Tracey,K.J. High-mobility group box 1
protein (HMGB1): nuclear weapon in the immune
arsenal. Nat. Rev. Immunol.5, 331342 (2005).
158. Palmblad,K. etal. Morphological characterization of
intra-articular HMGB1 expression during the course
of collagen-induced arthritis. Arthritis. Res. Ther. 9,
R35 (2007).
159. Parkkinen,J. & Rauvala,H. Interactions of
plasminogen and tissue plasminogen activator (tPA)
with amphoterin. Enhancement of tPAcatalyzed
plasminogen activation by amphoterin. J.Biol. Chem.
266, 1673016735 (1991).
160. Yamoah,K. etal. High-mobility group box proteins
modulate tumor necrosis factor-alpha expression in
osteoclastogenesis via a novel deoxyribonucleic
acid sequence. Mol. Endocrinol. 22, 11411153
(2008).
161. Gondokaryono,S.P. etal. The extra domain A of
fibronectin stimulates murine mast cells via toll-like
receptor 4. J.Leukoc. Biol. 82, 657665 (2007).
162. Schierbeck,H. etal. HMGB1 levels are increased in
patients with juvenile idiopathic arthritis, correlate
with early onset of disease, and are independent of
disease duration. J.Rheumatol. 40, 16041613
(2013).

14 | ADVANCE ONLINE PUBLICATION www.nature.com/nrrheum

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

163. Park,S.Y. etal. HMGB1 induces angiogenesis in
rheumatoid arthritis via HIF1 activation. Eur.
J.Immunol. 45, 12161227 (2015).
164. Hreggvidsdottir,H.S. etal. The alarmin HMGB1 acts
in synergy with endogenous and exogenous danger
signals to promote inflammation. J.Leukoc. Biol. 86,
655662 (2009).
165. Qin,Y. etal. HMGB1LPS complex promotes
transformation of osteoarthritis synovial fibroblasts to
a rheumatoid arthritis synovial fibroblast-like
phenotype. Cell Death Dis. 5, e1077 (2014).
166. Srivastava,P. Roles of heat-shock proteins in innate
and adaptive immunity. Nat. Rev. Immunol. 2,
185194 (2002).
167. van Eden,W. etal. Cloning of the mycobacterial
epitope recognized by T lymphocytes in adjuvant
arthritis. Nature 331, 171173 (1988).
168. Kamphuis,S. etal. Tolerogenic immune responses to
novel Tcell epitopes from heat-shock protein 60 in
juvenile idiopathic arthritis. Lancet 366, 5056
(2005).
169. de Kleer,I.M. etal. CD4+CD25bright regulatory
Tcells actively regulate inflammation in the joints of
patients with the remitting form of juvenile idiopathic
arthritis. J.Immunol. 172, 64356443 (2004).
170. van Roon,J.A.etal. Stimulation of suppressive Tcell
responses by human but not bacterial 60kD heat-
shock protein in synovial fluid of patients with
rheumatoid arthritis. J.Clin. Invest. 100, 459463
(1997).
171. Macht,L.M. etal. Relationship between disease
severity and responses by blood mononuclear cells
from patients with rheumatoid arthritis to human
heat-shock protein 60. Immunology 99, 208214
(2000).
172. Prakken,B.J. etal. Inhibition of adjuvant-induced
arthritis by interleukin-10driven regulatory cells
induced via nasal administration of a peptide analog
of an arthritis-related heat-shock protein 60Tcell
epitope. Arthritis Rheum. 46, 19371946 (2002).
173. Palmer,G. & Gabay,C. Interleukin33 biology with
potential insights into human diseases. Nat. Rev.
Rheumatol. 7, 321329 (2011).
174. Talabot-Ayer,D. etal. Immune-mediated experimental
arthritis in IL33 deficient mice. Cytokine 69, 6874
(2014).
175. Collins,L.V. etal Endogenously oxidized
mitochondrial DNA induces in vivo and in vitro
inflammatory responses. J. Leukoc. Biol.75,
9951000 (2004).
176. Hajizadeh,S. etal. Extracellular mitochondrial DNA
and oxidatively damaged DNA in synovial fluid of
patients with rheumatoid arthritis. Arthritis Res. Ther.
5, 234240 (2003).
177. Harty,L.C. etal. Mitochondrial mutagenesis
correlates with the local inflammatory environment in
arthritis. Ann. Rheum. Dis. 71, 582588 (2012).
178. Zhang,Q. etal. Circulatingmitochondrial
DAMPscause inflammatoryresponsestoinjury.
Nature464, 104107 (2010).
179. Hazeldine,J. etal. NFormylpeptides
drivemitochondrialdamage associated molecular
pattern induced neutrophil activation throughERK1/2
and P38 MAP kinase signalling pathways. Injury46,
975984 (2015).
180. Lood,C. etal. Neutrophilextracellular traps enriched
in oxidized mitochondrial DNAare interferogenic and
contribute to lupus-likedisease. Nat. Med.22,
146153 (2016).
181. Midwood,K. etal. TenascinC is an endogenous
activator of Toll-like receptor 4 that is essential for
maintaining inflammation in arthritic joint disease.
Nat. Med. 15, 774780 (2009).
182. Kakimoto,K.etal. Suppressive effect of a neutrophil
elastase inhibitor on the development of collagen-
induced arthritis. Cell. Immunol. 165, 2632 (1995).
183. Witter,J. etal. The immunologic detection and
characterization of cartilage proteoglycan degradation
products in synovial fluids of patients with arthritis.
Arthritis Rheum. 30, 519529 (1987).
184. Powell,J.D. & Horton,M.R. Threat matrix: low-
molecular-weight hyaluronan (HA) as a danger signal.
Immunol. Res. 31, 207218 (2005).
185. CsSzabo,G., Roughley,P.J., Plaas,A.H. & Glant,T.T. article, to substantial discussions of its content, to writing the
Large and small proteoglycans of osteoarthritic and
rheumatoid articular cartilage. Arthritis Rheum. 38,
660668 (1995).
186. Di Virgilio,F.etal. Leukocyte P2 receptors: a novel
target for anti-inflammatory and anti-tumor therapy.
Curr. Drug Targets Cardiovasc. Haematol. Disord. 5,
8599 (2005).
187. Hammer,H.B.etal. Calprotectin (a major leucocyte
protein) is strongly and independently correlated with
joint inflammation anddamage in rheumatoid arthritis.
Ann. Rheum. Dis. 66, 10931097(2007).
188. Wittkowski,H.etal. S100A12 is a novel molecular
marker differentiating systemic-onset juvenile
idiopathic arthritis from other causes of fever of
unknown origin. Arthritis Rheum. 58, 39243931
(2008).
189. OReilly,S. Poundthealarm:
dangersignalsinrheumaticdiseases. Clin. Sci.
(Lond.)128, 297305 (2015).
190. Srikrishna,G. etal. Carboxylated glycans mediate
colitis through activation of NFB. J.Immunol. 175,
54125422 (2005).
191. Hofmann,M.A. etal. RAGE mediates a novel
proinflammatory axis: a central cell surface receptor
for S100/calgranulin polypeptides. Cell 97, 889901
(1999).
192. Prakken,B.J. etal. Epitope-specific immunotherapy
induces immune deviation of proinflammatory Tcells
in rheumatoid arthritis. Proc. Natl. Acad. Sci. U.S.A.
101, 42284233 (2004).
193. Koffeman,E.C. etal. Epitope-specific immunotherapy
of rheumatoid arthritis: clinical responsiveness occurs
with immune deviation and relies on the expression of
a cluster of molecules associated with Tcell tolerance
in a double-blind, placebo-controlled, pilot phaseII
trial. Arthritis Rheum. 60, 32073216 (2009).
Acknowledgements
The authors were supported by INSERM, the University of
Paris 6 (Pierre and Marie Curie), Arthritis R&D (ROAD:
Research on Osteoarthritic Diseases) and the Agence
Nationale de la Recherche. D.H. acknowledges support from
the Interdisciplinary Centre of Clinical Research, Mnster,
Germany (Clinical Research Award CRA0) and the German
Ministry for Education and Science (BMBF 01GM08100,
AID-Net).
Author contributions
All authors contributed equally to researching data for the
article and to reviewing and editing of the manuscript before
submission.
Competing interests statement
D.H. declares that he has received speakers fees from
Novartis. M.N., F.B. and C.J. declare no competing interests.
Review criteria
We searched MEDLINE via PubMed with the following MeSH
terms: alarmin, DAMP, osteoarthritis, rheumatoid
arthritis, psoriatic arthritis, spondylarthropathy, inflam-
mation, HMGB1, S100, HSP, IL33, annexins,
ATP, galectins, IL1, tenascinc, thymosins, uric
acid, fibronectin, laminin, biglycan, and low molecular
hyaluronan, with no restriction on the year of publication (up
to February 2016).

NATURE REVIEWS | RHEUMATOLOGY ADVANCE ONLINE PUBLICATION | 15

2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.