Download The danger model in deciphering autoimmunity

RHEUMATOLOGY
Rheumatology 2010;49:632–639
doi:10.1093/rheumatology/keq004
Advance Access publication 9 February 2010
Review
The danger model in deciphering autoimmunity
Anders A. Tveita1
R EV I E W
Abstract
Autoimmunity has been a topic of intensive research for several decades, yet amazingly, no uniform
hypothesis exists to explain the basis for the spectrum of autoantibody specificities seen in autoimmune
diseases. It therefore seems appropriate to consider whether our current framework for understanding
tolerance, and thus the mechanisms controlling the initiation and perpetuation of autoimmunity, may be
faulty. Adapting the paradigm of Matzinger—the ‘danger model’, a case can be made for a perspective
that appreciates the fundamental role of the tissues in controlling immune response, favouring a shift of
focus in studies on the initiation of autoimmunity. Applying the elements of this model, I set forth a number
of scenarios for how autoreactivity could emerge, with emphasis on the likely sources of the involved
autoantigens and the functional basis of their appearance. The emerging picture is one in which disruption
of tissue homeostasis takes centre stage, with the antigen-presenting cells as the key players.
Key words: Danger signalling, Tolerance, Autoimmunity, Systemic lupus erythematosus, Autoreactivity,
Apoptosis.
Introduction
The autoimmune diseases form an enigmatic class of conditions in which the immune system is engaged in
a destructive attack against varying repertoires of self
structures. Despite decades of intensive studies of
immune function and regulation of immunity in both
human and animal model systems, key elements in the
pathogenesis of autoimmunity still appear obscure. It is
tempting to speculate that this might relate to problems
with the conceptual framework of understanding immunological tolerance and its breakdown in autoimmunity.
A fresh perspective is offered by approaching the
immune system by the premises laid down by Polly
Matzinger in her ‘danger model’ [1–6]. Even today, more
than a decade after its emergence, this model offers a
fundamentally different interpretation of classic and
newly emerging concepts in immunology.
The danger model suggests that most autoimmune
states do not stem from deficiencies in the immune
response, but rather from defects in the normal physiology of tissues [4]. This review, therefore, will focus on the
1
Department of Biochemistry, Institute of Medical Biology, University
of Tromsø, Tromsø, Norway.
Submitted 18 November 2009; revised version accepted
4 January 2010.
Correspondence to: Anders A. Tveita, Department of Biochemistry,
Institute of Medical Biology, University of Tromsø, N-9037 Tromsø,
Norway. E-mail: [email protected]
fundamental principles by which the immune system interacts with the tissues making up the living human, rather
than on minute details of immunology.
A dissection of autoimmunity as a concept has to start
with definitions. Autoimmunity is a situation in which lymphocytes reactive against self structures are activated and
allowed to operate to an extent that causes sustained
self-reactivity and tissue damage. This rather pragmatic
definition excludes situations with transient activation of
autoreactive lymphocytes, such as when fighting off an
infection [7]. Indeed, such autoreactivity might be
expected to be present in any situation where there is
tissue damage, as discussed below. In this lies one of
the fundamental postulates of the danger model—that
autoreactivity per se is a normal aspect of the physiological process of tissue inflammation, whereas the defining
characteristic of pathological autoimmunity (autoimmune
disease) is the persistence of autoreactivity caused
by a continued supply of danger signals (chronic autoreactivity).
The danger model ties the emergence of autoreactivity
to the release of danger signals. Therefore, it seems
possible that the dominating targets of autoreactivity
might provide clues to the origin of the autoimmune
response and point to potential underlying causes of
danger signalling, as discussed below. For simplicity,
in the following, the term ‘site of damage’ refers not
only to the injured tissue structure, but also includes
draining lymph nodes, as this forms the point of initial
contact for lymphocyte recruitment.
! The Author 2010. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected]
Danger signalling and autoimmunity
Danger revisited
Matzinger defines danger signals as structures released
or produced by cells or tissues undergoing stress or
abnormal cell death (or even normal cell death, if the
cells are not properly scavenged and go into secondary
necrosis) that induce activation of resting antigenpresenting cells (APCs). A growing number of endogenous
substances are being identified as potential danger
signals [e.g. adenoside-50 -triphosphate (ATP), hyaluronan
breakdown products, transcription factors such as
high-mobility group box 1 (HMGB-1), certain complement
components, S100 protein family, etc.] [8], suggesting that
the term might encompass a large number of hitherto
unappreciated biochemical alterations occurring in the
micro-environment resulting from cellular damage [9].
The principle substrate of danger need not be anything
more than unresolved cellular damage, as on the biochemical level there is no fundamental difference in
the way various kinds of tissue damage presents
themselves—there is no a priori means of telling what
caused the damage or what kind of response would be
more favourable. The logical response is therefore a broad
one: flood the area with APCs to see what is there, and
then bring in the effector cells and scavengers to start
removing it. When the job is effectively done, danger
subsides, inflammation is dampened and the process
dies out.
Approaching autoimmunity
We are now seeing the arena in which autoimmunity
unfolds. There is no reason why autoimmune diseases
should require added complexities; autoimmunity should
stem from breaches of the homeostatic apparatus
presented above. Essentially this is a matter of statistical
likelihood, with the effectiveness of tolerogenic (signal 1)
vs immunogenic (signal 1 + signal 2) antigen presentation being the crucial determinant. A shift in balance
towards the latter occurs whenever danger signals are
released and trigger the expression of signal 2, so what
we should be looking for is a source of prevailing
tissue injury.
The importance of the autoantigens vs the danger signals
Most autoimmune diseases are characteristically associated with autoreactivity against particular subsets
of antigens. A key issue is therefore to what extent
an autoimmune disease is dependent on reactivity against
such families of antigens. In other words, are the
autoreactive targets determined by the underlying
pathogenetic processes, or are they merely a reflection
of the location in which the release of danger signals
occur? Moreover, will autoreactivity persist upon the
removal of the main target structures? How important is
a particular subset of autoantigens to the maintenance of
an autoimmune disease, and what classes of antigens can
serve as sustained inducers and targets of autoreactivity?
Certain organ-specific antigens with limited expression
are presented by a small number of cells, increasing the
www.rheumatology.oxfordjournals.org
ease with which an immunogenic presentation could be
attained if danger signals are persistently released in the
containing areas. The small number of cells provides
scarce tolerogenic display in the normal state, whereas
release of danger signals will generate an inflammatory
focus, causing the emergence of autoreactivity centred
at antigens confined to the inflamed area. Conceptually,
such antigens should therefore provide good candidates
for maintenance of an autoimmune response as long as
they are presented in a dangerous context. Other, more
widely distributed antigens would be presented on
non-activated cells in other parts of the body, effectively
terminating autoreactivity against such structures. By this
line of reasoning, looking at the autoreactive target
structures could provide clues to the origin of the
danger signals, but might say nothing about the nature
of the inciting tissue damage.
The possibility that the efficacy of tolerogenic antigen
display differs for various cellular constituents provides a
model to explain the distinct panels of target antigens
seen in autoimmune diseases. It is obvious that, under
normal circumstances, inefficient tolerization of these
antigens does not pose a problem, and probabilistic
reasoning does not in itself provide an answer to why
autoimmunity can persist. It does however offer an
explanation as to why MHC loci are consistently identified
as susceptibility loci for autoimmune diseases [10, 11].
Subtle differences in MHC Class I or MHC Class II epitope
binding preferences might alter the ease with which
tolerance towards particular autoantigens is breached,
and thus increase the likelihood of autoimmunity emerging
against these structures. Indeed, data exist in support of
an association between MHC haplotype and autoantigen
profile in diseases such as SLE [12, 13]. With these
thoughts on predisposing factors in mind, we now turn
to the core of the matter—the pathogenetic basis of
sustained autoimmunity.
The pathogenesis of autoimmune diseases
Antigen display is a fundamental responsibility of most
cells, and so the emergence of autoreactivity is a
signal that this process is not properly maintained.
Consequently, autoreactivity is based on increased immunogenic presentation, diminished tolerogenic display or a
combination of these. Considering the autoantibody repertoires and the clinical manifestations of autoimmune
diseases, it is obvious that a fundamental difference
between systemic and organ-restricted disease is the
distribution of the involved autoantigens. Maintaining
autoreactivity towards ubiquitous antigens should intuitively originate in widespread deficiencies of some
aspect of APC function, whereas reactivity towards
tissue-restricted antigens could be envisioned to occur
on the basis of more ‘local’ types of aberrancy in antigen
display or tissue homeostasis.
There appears to be a tendency to try to pinpoint a
particular subset of antigens that constitute the pathogenic basis of the disease being studied. The search for
direct causality between autoreactivity and clinical
633
Anders A. Tveita
manifestations might be a considerable source of
confusion in designing and interpreting studies on autoimmunity, especially the systemic diseases, in which the
source of the triggering autoantigens is not obvious, as
discussed below. An additional and important cause for
caution lies in the heterogenicity of laboratory and clinical
features of the individual autoimmune syndromes, with the
possibility that some current disease entities represent
artificial clustering of aetiologically unrelated illnesses.
Such precautions might be particularly relevant to the
systemic diseases, with conditions overlapping clinically
and serologically, further adding to the uncertainty.
Systemic autoimmunity: SLE
SLE is considered a prototypic systemic autoimmune disease, and has been extensively studied for decades.
Particular interest has been paid to the nephritis accompanying the disease. To some extent, the dominating
autoantibody specificities in SLE appear to correlate
with specific end-organ manifestations. Autoantibodies
against chromatin particles [double-stranded DNA
(dsDNA) and nucleosomes] are by many regarded as pivotal in the development of lupus nephritis [14, 15].
Nucleosomes have been found to be deposited within
the glomerular matrix and capillary membranes in patients
with lupus nephritis, and these deposits co-localize with
glomerular autoantibody deposits [16]. A key problem in
this respect is the origin of such extracellular chromatin
deposits. Are they the result of an intrinsic renal pathology, or is deposition caused by passive accumulation of
pre-formed circulating immune complexes?
The example of lupus nephritis effectively demonstrates
major problems associated with studies of systemic autoimmunity. The ubiquitous nature of the major autoantigens
makes it hard to infer what the origin of autoimmunity is,
and the resulting inflammatory response against exposed
antigenic structures makes it difficult to separate primary
and secondary events.
Targets of autoreactivity in systemic autoimmunity
Autoantibody production in systemic autoimmune diseases is remarkable in being largely directed at antigens
with more or less universal tissue distribution, including
DNA/chromatin and RNA-containing complexes. There
are few situations that should allow for persistent
autoreactivity against such very abundant cellular constituents, and appreciating this considerably limits the
possibilities of how systemic autoimmunity could conceivably be initiated.
Autoantibody production requires antigen presentation
on activated APCs. The most obvious source of autoantigens for display on APCs is phagocytosed dead cell
remnants in the form of apoptotic bodies or necrotic
debris, with the former being the predominant one in the
healthy individual. A large number of in vitro studies have
indicated that phagocytic ingestion of normal apoptotic
cells is immunologically silent or even triggers
anti-inflammatory cytokine production by the phagocytosing cell [17, 18]. The topographic origin of an autoimmune
634
response should be a site where APCs are effectively
activated by danger signals and where autoantigens are
displayed in a manner that is not sufficiently countered by
tolerogenic display on non-activated APCs. Within a
danger-signalling framework, the mechanistic explanation
for an autoimmune disease, therefore, must encompass
two basic elements: (i) a sustained source of danger signals and (ii) a means to explain the selection of target
autoantigens.
Listed below are four hypothetical scenarios that could
fulfil these criteria, and thus form the pathogenic basis of
an autoimmune response. In various forms, elements
of each of these scenarios have all been suggested as
aetiological factors in SLE.
Scenario 1: waste management and apoptotic cell antigens. MHC Class II antigen presentation is largely
restricted to APCs, and accordingly, T-helper and B-cell
activation is controlled by these cell types. As a consequence, tolerance is closely reliant on presentation
by non-activated APCs. Through the process of
cross-presentation, phagocytosed material is also made
accessible for presentation on MHC Class I, which gives
APCs a role in the regulation of cytotoxic T cells as well.
From these considerations, it follows that normal APC
function is of key importance to the maintenance of
tolerance.
It is possible that some antigens, by virtue of their relative inaccessibility to MHC loading compartments, would
only rarely be presented on the cell surface. When cells
die, however, such antigens would be sequestered
in apoptotic blebs and taken up by phagocytosing cells.
Upon phagocytosis, these antigens are made accessible
for MHC Class I and MHC Class II presentation, forming a
population of dead cell ‘ghost antigens’ effectively only
presented on APCs. The appearance of such antigens
on APCs would not constitute a problem, provided
phagocytosis of apoptotic cells is maintained as an
immunologically silent process, which is the case for
normal apoptotic cell clearance.
However, if the phagocytic process triggers danger signalling, and thus activates the phagocytosing APC, a
fertile ground is created for autoreactivity against the
ghost antigens. Such a situation could emerge if phagocytosis was inefficient, allowing apoptotic cells to undergo
secondary necrotic transformation. Indeed, there have
been reports of impaired phagocytic function in SLE
[19]. Deficiencies of several soluble factors, notably C1q
and serum amyloid P, as well as a growing number of cell
surface receptors have been implicated as possible
sources of aberrant apoptotic cell clearance [20, 21].
C1q deficiency is a rare hereditary disease known to
elicit a lupus-like syndrome, and it would seem plausible
that deficiencies of similar factors could serve as aetiological factors in SLE [22].
Another important aspect of apoptotic cell death is its
immunomodulatory effects on surrounding cells, including
APCs. It is tempting to speculate that alterations in these
interactions could serve as a permissive factor for the
development of autoimmune diseases. Recent reports
www.rheumatology.oxfordjournals.org
Danger signalling and autoimmunity
indicate the existence of several subtypes of apoptotic
cell death, characterized by the presence or absence
of certain secreted or membrane-bound antigens that
bring about different immunological responses [23]. By
means of one or more ill-defined ligand(s), apoptotic
cells under some circumstances appear to elicit an
anti-inflammatory response in contacting cells [24]. In
contrast, surface expression of calreticulin on apoptotic
cells has been shown to cause rapid activation of APCs,
and preventing its binding on the surface blocks this effect
altogether [25]. Subsequent release of potential danger
signals including heat-shock proteins and HMGB1, such
as by secondary necrotic transformation, may further
influence the APC response and mode of antigen display.
In these ways, factors in the local micro-environment of
the dying cell might serve as sources of immunological
priming against phagocytosed self antigens, either
through prolonged persistence caused by inefficient
uptake or by their inappropriate extracellular exposure
by secretion or surface display. Indeed, there are data
suggestive of this type of mechanism influencing the
effects of apoptotic cells on macrophages from mouse
models of both SLE and diabetes mellitus [26].
Scenario 2: digestion and fate of antigens ingested by
APCs. Looking beyond the uptake of apoptotic debris, it
is obvious that the further fate of ingested material is
important in avoiding danger signalling. Several important
danger-sensing structures, including Toll-like receptors
(TLR) 7 and 9, are active within endocytic compartments
and could be triggered by contact with internalized structures like chromatin and certain RNA structures. It is
known that genomic DNA contains CpG dinucleotide
motifs, and under certain conditions has the ability to activate TLR9 signalling [27]. Therefore, if phagocytosed
chromatin is not promptly and completely degraded,
one could risk the activation of the APC by retained
DNA. At the same time, accumulation of undegraded
chromatin would provide an abundant source of antigens
accessible for MHC presentation. Effectively, incomplete
digestion of DNA could therefore provide both a danger
signal and a source of autoreactivity focused against
nuclear antigens. Similarly ill-defined cellular defects
causing accumulation of endogenous DNA elements
within the cytosol appear to trigger the so-called
INF-stimulated DNA response. This mechanism serves
as a source of pro-inflammatory signalling, causing
autoinflammatory conditions with lupus-like manifestations (e.g. chilblain lupus caused by mutations in the
trex1 gene) [28].
Kawane et al. [29] reported that mice deficient in the
endonuclease DNase II, thought to play a key role in the
degradation of nucleosomal DNA in phagocytosed material in macrophages, are incapable of degrading ingested
DNA, possibly triggering ‘activation of the innate immune
system’. Several other DNase deficiencies have been proposed as potential sources of anti-DNA autoimmunity,
and data to support this include reports of DNase I deficiency in two SLE patients [30] and findings of lupus-like
disease manifestations in murine knockouts of DNase I
www.rheumatology.oxfordjournals.org
[31], as well as reports of DNase-g (DNase1L3) mutations
in spontaneous models of murine SLE [32]. Of note, other
ANAs are also reported to be present in these mice,
although at lower levels.
Scenario 3: cryptic antigens—‘altered self’. Antigens processed in an unusual manner can lead to the formation
and presentation of ‘altered self’ structures, which are not
part of the natural processed and presented self antigen
repertoire. Several lines of experiments have demonstrated that such modified self structures provide efficient
immunogens, and they have been suggested to play a
central role in autoimmune disease pathogenesis provided that the cryptic antigens are accessible as targets
in vivo. For instance, it has been shown that immunization
of healthy mice with modified antigens, such as
de-methylated syngeneic chromatin, triggers autoimmune
disease, not seen by immunization with the native antigen
[33, 34]. Acetylated DNA similarly accelerates disease
manifestation in lupus-prone mice [35].
Representing a class of antigens towards which the
immune system is not normally tolerized, the functional
immunogenicity of ‘altered self’ antigens when introduced
with adjuvants is no surprise. By themselves, however,
most altered self structures are not known to be immunogenic (i.e. serve as danger signals). The crucial factor
would therefore be the process by which they are brought
about. An obvious source of such alterations would be an
inflammatory process, with oxidative damage and release
and activation of potent proteases, for example, a burn
injury or a chronic infectious process. The accompanying
release of danger signals would then activate APCs, and
offer both native and altered self structures for T-cell activation. In some instances, the altered antigen could
in itself serve as a danger signal, such as might be the
case for hypomethylated DNA, which likely obtains the
ability to function as a TLR9 ligand. Likewise, loss of
a mannosidase in mice causes lupus-like disease likely
precipitated by the unmasking of TLR ligands [36].
In this case, danger signals and target antigens differ,
demonstrating that the underlying defect does not necessarily dictate the targets of the ensuing autoimmune
response. For most antigens, the potential for
‘de-masking’ of danger signals by biochemical modification remains unknown. Consequently, rather than offering
a complete scenario for the origin of autoimmunity, the
altered self model circles out a subset of epitopes towards
which tolerance is easily broken in the face of
danger-driven APC activation. Although these antigens
might not always be a pathogenetic factor (i.e. the
danger signal), there is still a possibility, therefore, that
they serve as a permissive factor in the maintenance of
autoreactivity analogous to antigens with limited tissue
distribution.
Scenario 4: autoantigens in complexes. The hapten-carrier model was proposed to explain how poor
immunogens, such as DNA, can become targets of an
immune response when in complex with a more potent
immunogenic structure. According to this model, by
635
Anders A. Tveita
virtue of its physical association with the carrier protein,
the hapten follows along upon antibody-mediated internalization of the carrier in APCs or carrier-specific B cells.
The hapten is then presented alongside carrier epitopes,
and serves as a T-cell target in a process initiated by an
anti-carrier response. There is no contradiction between
the danger model and the hapten-carrier principle. The
fundamental property of a hapten carrier is that it introduces a protein antigen that is not effectively displayed on
host cells, analogous to ‘altered self’ antigens, presented
in a dangerous context, for example, through the use of
adjuvants. Also, there is the possibility that certain hapten
structures may represent ligands to yet undescribed
receptors, thus serving as danger signals in their own
right [37]. For instance, HMGB1, a well-recognized endogenous danger signal, has been shown to serve as an
effective adjuvant, boosting the immune response
against nucleosomes, which by themselves are poor
immunogens. HMGB1-containing nucleosome complexes
activate dendritic cells in vitro, in a fashion that is competitively inhibited by addition of a recombinant form of a
DNA-binding domain of HMGB1 [38]. One possible explanation for these findings would be that HMGB1 serves as
a hapten structure by (i) its ability to bind nucleosomal
DNA and (ii) to serve as a danger signal. This effectively
co-localizes the danger signal with a number of nuclear
antigens. Furthermore, HMGB1 has been shown to induce
cross-presentation of ingested antigens by APCs, thus
also facilitating a cytotoxic T-cell response [39]. Similar
mechanisms might explain the effect of using heat-shock
protein complexes as a means of boosting vaccination
against various tumour antigens [40].
Immunization with a polyomavirus transcription factor in
complex with mammalian DNA induces transient
anti-dsDNA production in healthy mice. Induction of transgenic expression of this transcription factor in kidneys
following such immunization causes sustained production
of anti-dsDNA in the absence of a viral infection [41]. The
hapten-carrier model demonstrates that autoreactivity
towards DNA can emerge from the initial breech of
tolerance to a smaller set of antigens within a chromatin
complex. Although artificial, this system demonstrates an
expansion of autoreactivity that might be relevant to the
initiation of autoimmune disease.
If one accepts that transient autoreactivity is a natural
aspect of the response to tissue injury, the appearance of
autoantibodies would be expected during the course of
substantial inflammatory processes. Such antibodies
would react with their antigens to form immune complexes, and facilitate Fc-mediated clearance by APCs. If
formation of immune complexes is increased, or if their
clearance is insufficient, a ‘hapten-carrier’-like situation
could emerge, with the antibody functioning as a ‘carrier’,
facilitating Fc-mediated uptake and high-level presentation of complex-contained antigens [42]. Uptake of
immune complexes can induce different responses in
various APCs, depending on contrasting effects by
uptake via particular subtypes of Fc receptors, referred
to as ‘inhibitory’ or ‘stimulatory’ depending on the
636
response of the APC upon cross-linking of receptors. Of
note, the inhibitory Fc receptor, FcRIIb, seems to sequester immune complexes on APCs in a way that prevents
their presentation in MHC. It has been demonstrated that
retroviral delivery of FcRIIb to lupus-prone mice can eliminate autoantibody production and prevent the disease
[43]. If the balance between stimulatory and inhibitory
Fc receptors present on APCs is somehow shifted in
favour of the former, this could provide a driving force
for persisting autoimmunity by promoting the presentation
of immune complex-captured autoantigen(s) in a dangerous context. According to this scenario, the lupus-defining
defect is ascribed to a hyper-reactive response to
‘physiological autoreactivity’, with immunogenic uptake
of immune complexes serving as a sustainment
mechanism, providing a constant source of danger
signals. If this is the case, looking at the phenotype of
APCs upon challenge with immune complexes would
seem a logical first step in revealing the underlying
mechanism(s).
Systemic autoimmunity: is there a unifying concept?
The above scenarios describe conditions under which a
persisting immune response against tissue antigens could
be initiated and maintained. They all relate to defects or
aberrancies in the clearance of cellular debris, which leads
to the retention of cellular remnants. The continued presence of such cellular constituents could, by itself,
provide the danger signals needed to initiate and maintain
a cellular and humoral immune response. There would
thus not be any absolute requirement for an external
event to trigger autoimmunity, although exogenously
inflicted tissue injury (ultraviolet light, hypoxia, infections,
trauma, etc.) could conceivably accelerate or expand the
process, hastening the appearance of clinical manifestations (e.g. lupus flares). A cautious reservation is the likely
possibility that some autoimmune diseases might actually
evolve from unappreciated persistent infections. In such a
situation, the continued infliction of tissue injury by the
infectious agent would provide the driving force of
the immune response. In contrast to ‘true’ autoimmune
diseases, the autoimmunity of persisting infections is an
aspect of a biologically meaningful but inadequate
immune response. However, studies on autoimmunityprone animals bred in strictly germ-free environments
confirm that autoimmunity exists also in the absence of
infectious agents, that is, ‘true’ sustained autoimmunity
is a reality [44, 45]. Importantly, neither of the abovementioned scenarios is exclusive. Rather, they complement each other and could be aspects of deficiencies
affecting both the processing of the antigen in the dying
cell and its uptake, digestion and presentation by APCs.
A number of specific molecular defects that relate to these
processes have been proposed as possible contributors
to SLE and other systemic autoimmune diseases, with
the list still growing. The idea that deficiencies related to
apoptosis or clearance of dead cells form the basis of SLE
pathogenesis is now widely held within the research community. Figure 1 is an attempt to summarize factors that
www.rheumatology.oxfordjournals.org
Danger signalling and autoimmunity
FIG. 1 Key elements in experimental and human lupus pathogenesis.
This illustration is an attempted summary of elements implicated in lupus pathogenesis within a danger-signalling
framework. Bracketed references [19–22, 30, 31, 34, 46–61] point to studies relating the particular factor to the pathogenesis. The stage is divided into four main elements: (i) ‘origo of autoimmunity’, which is the site of primary danger, the
stage upon which the autoantigens are released. (ii) Illustrates the ‘priming mechanism’, in which activated APCs offer
signal 1 in the presence of signal 2, effectively priming T cells to initiate an immune response against antigens revealed
within the origo. (iii) Is an example of an ‘amplificatory process’ that contributes to sustained danger signalling, effectively
maintaining autoimmunity. TLR signalling is likely only one of many such mediators, but was chosen due to the central
role of anti-chromatin autoreactivity in lupus, and the likely importance of TLR signalling in mediating autoreactivity
against these structures. (iv) Is the ‘effector side’, where primed cytotoxic T lymphocytes (CTLs) and ‘pathogenic’
autoantibodies–immune complexes inflict end-organ damage. At this level, several adversaries come into play,
including complement, neutrophils, NK cells and proteolytic enzymes.
have been postulated to be important contributors to
lupus pathogenesis, and show how they all fit together,
with danger signalling being the central element. The
emerging picture is one in which this network of dying
cells, their phagocytosis and presentation by APCs and
the accompanying danger signals forms the basis of the
lupus condition.
Concluding remarks
Is there anything to suggest that systemic and
organ-specific autoimmunity have a common basis?
Following the reasoning of the danger model, the maintenance of tolerance is a common responsibility of cells
of all tissues, and the breach of tolerance tells us
that some of the cells containing the autoantigen(s)
in question are not doing their job correctly or are
www.rheumatology.oxfordjournals.org
inappropriately processed and cleared upon dying. By
considering that transient autoimmunity could be a
biologically meaningful or at least acceptable part of an
immune response, the focus in understanding autoimmune diseases is shifted from the process of lymphocyte
activation to the recruitment and activation of APCs
by danger signals released by cells in the involved
tissue(s).
This text does not provide a definite answer to the
fundamental question—what triggers autoimmune
diseases? I have presented several scenarios that I imagine could touch upon the answer. More such elements
probably exist, and new ones might emerge as our knowledge of the interplay between the innate immune system
and the tissues increases. The key to answering such
questions may lie very close, it is just a matter of getting the context right.
637
Anders A. Tveita
Rheumatology key messages
Aberrancies of apoptosis can release danger signals, serving as a potential driving force for
autoimmunity.
. Autoimmune diseases may relate to tissuerestricted pathology rather than dysfunctions in
the immune system itself.
. The continued presence of danger signals separates autoimmune diseases from ‘physiological’
autoreactivity.
.
14 Rahman A, Isenberg DA. Systemic lupus erythematosus.
N Engl J Med 2008;358:929–39.
15 Mortensen ES, Fenton KA, Rekvig OP. Lupus nephritis:
the central role of nucleosomes revealed. Am J Pathol
2008;172:275–83.
16 Kalaaji M, Mortensen E, Jorgensen L, Olsen R, Rekvig OP.
Nephritogenic lupus antibodies recognize glomerular
basement membrane-associated chromatin fragments
released from apoptotic intraglomerular cells. Am J Pathol
2006;168:1779–92.
17 Voll RE, Herrmann M, Roth EA, Stach C, Kalden JR,
Girkontaite I. Immunosuppressive effects of apoptotic
cells. Nature 1997;390:350–1.
Acknowledgements
I would like to thank Dr Ole Petter Rekvig, Polly Matzinger,
Tony Marion and Svetlana N. Zykova for helpful discussions and critical review of the manuscript.
Funding: This work was supported by internal funding
from the University of Tromsø.
Disclosure statement: The author has declared no conflicts of interest.
References
1 Matzinger P. An innate sense of danger. Ann N Y Acad Sci
2002;961:341–2.
2 Moseley P. Stress proteins and the immune response.
Immunopharmacology 2000;48:299–302.
3 Matzinger P. Tolerance, danger, and the extended family.
Annu Rev Immunol 1994;12:991–1045.
18 Gaipl US, Beyer TD, Baumann I et al. Exposure of anionic
phospholipids serves as anti-inflammatory and immunosuppressive signal–implications for antiphospholipid
syndrome and systemic lupus erythematosus.
Immunobiology 2003;207:73–81.
19 Herrmann M, Voll RE, Zoller OM, Hagenhofer M,
Ponner BB, Kalden JR. Impaired phagocytosis of apoptotic cell material by monocyte-derived macrophages
from patients with systemic lupus erythematosus. Arthritis
Rheum 1998;41:1241–50.
20 Bickerstaff MC, Botto M, Hutchinson WL et al. Serum
amyloid p component controls chromatin degradation and
prevents antinuclear autoimmunity. Nat Med 1999;5:
694–7.
21 Bijl M, Reefman E, Horst G, Limburg PC, Kallenberg CG.
Reduced uptake of apoptotic cells by macrophages in
systemic lupus erythematosus: correlates with decreased
serum levels of complement. Ann Rheum Dis 2006;65:
57–63.
4 Matzinger P. An innate sense of danger. Semin Immunol
1998;10:399–415.
22 Botto M, Dell’Agnola C, Bygrave AE et al. Homozygous
C1q deficiency causes glomerulonephritis associated with
multiple apoptotic bodies. Nat Genet 1998;19:56–9.
5 Matzinger P. The danger model: a renewed sense of self.
Science 2002;296:301–5.
23 Zitvogel L, Kroemer G. Introduction: the immune response
against dying cells. Curr Opin Immunol 2008;20:501–3.
6 Gallucci S, Matzinger P. Danger signals: SOS to the
immune system. Curr Opin Immunol 2001;13:114–9.
24 Cvetanovic M, Mitchell JE, Patel V et al. Specific recognition of apoptotic cells reveals a ubiquitous and unconventional innate immunity. J Biol Chem 2006;281:
20055–67.
7 Hansen KE, Arnason J, Bridges AJ. Autoantibodies and
common viral illnesses. Semin Arthritis Rheum 1998;27:
263–71.
8 Rock KL, Kono H. The inflammatory response to cell
death. Annu Rev Pathol 2008;3:99–126.
25 Obeid M, Tesniere A, Ghiringhelli F et al. Calreticulin
exposure dictates the immunogenicity of cancer cell
death. Nat Med 2007;13:54–61.
9 Oppenheim JJ, Yang D. Alarmins: chemotactic activators
of immune responses. Curr Opin Immunol 2005;17:
359–65.
26 Koh JS, Wang Z, Levine JS. Cytokine dysregulation
induced by apoptotic cells is a shared characteristic of
murine lupus. J Immunol 2000;165:4190–201.
10 Taneja VD, David CS. Genetics and autoimmunity: HLA
and MHC genes. In: Rose NR MI, ed. The autoimmune
diseases. St. Louis, Mo: Elsevier Academic Press,
2006:261–72.
27 Leadbetter EA, Rifkin IR, Hohlbaum AM, Beaudette BC,
Shlomchik MJ, Marshak-Rothstein A. Chromatin-IgG
complexes activate B cells by dual engagement of IgM
and toll-like receptors. Nature 2002;416:603–7.
28 Stetson DB, Ko JS, Heidmann T, Medzhitov R. Trex1
prevents cell-intrinsic initiation of autoimmunity. Cell 2008;
134:587–98.
11 Mellins ED. The role of the MHC in autoimmunity: an
overview. J Rheumatol 1992;33:63–9.
12 Sekine H, Graham KL, Zhao S et al. Role of MHC-linked
genes in autoantigen selection and renal disease in a
murine model of systemic lupus erythematosus. J
Immunol 2006;177:7423–34.
13 McHugh NJ, Owen P, Cox B, Dunphy J, Welsh K. MHC
class II, tumour necrosis factor alpha, and lymphotoxin
alpha gene haplotype associations with serological subsets of systemic lupus erythematosus. Ann Rheum Dis
2006;65:488–94.
638
29 Kawane K, Fukuyama H, Yoshida H et al. Impaired thymic
development in mouse embryos deficient in apoptotic
DNA degradation. Nat Immunol 2003;4:138–44.
30 Yasutomo K, Horiuchi T, Kagami S et al. Mutation of
DNASE1 in people with systemic lupus erythematosus.
Nat Genet 2001;28:313–4.
31 Napirei M, Karsunky H, Zevnik B, Stephan H,
Mannherz HG, Moroy T. Features of systemic lupus
www.rheumatology.oxfordjournals.org
Danger signalling and autoimmunity
erythematosus in DNase1-deficient mice. Nat Genet 2000;
25:177–81.
32 Wilber A, O’Connor TP, Lu ML, Karimi A, Schneider MC.
Dnase1l3 deficiency in lupus-prone MRL and NZB/W F1
mice. Clin Exp Immunol 2003;134:46–52.
33 Qiao B, Wu J, Chu YW et al. Induction of systemic lupus
erythematosus-like syndrome in syngeneic mice by
immunization with activated lymphocyte-derived DNA.
Rheumatology 2005;44:1108–14.
34 Wen ZK, Xu W, Xu L et al. DNA hypomethylation is crucial
for apoptotic DNA to induce systemic lupus
erythematosus-like autoimmune disease in
SLE-non-susceptible mice. Rheumatology 2007;46:
1796–803.
35 Dieker JW, Fransen JH, van Bavel CC et al.
Apoptosis-induced acetylation of histones is pathogenic in
systemic lupus erythematosus. Arthritis Rheum 2007;56:
1921–33.
36 Green RS, Stone EL, Tenno M, Lehtonen E, Farquhar MG,
Marth JD. Mammalian N-glycan branching protects
against innate immune self-recognition and inflammation
in autoimmune disease pathogenesis. Immunity 2007;27:
308–20.
37 Palm NW, Medzhitov R. Immunostimulatory activity of
haptenated proteins. Proc Natl Acad Sci USA 2009;106:
4782–7.
38 Urbonaviciute V, Furnrohr BG, Meister S et al. Induction of
inflammatory and immune responses by
HMGB1-nucleosome complexes: implications for the
pathogenesis of SLE. J Exp Med 2008;205:3007–18.
39 Apetoh L, Ghiringhelli F, Tesniere A et al. Toll-like receptor
4-dependent contribution of the immune system to anticancer chemotherapy and radiotherapy. Nat Med 2007;
13:1050–9.
40 Murshid A, Gong J, Calderwood SK. Heat-shock proteins
in cancer vaccines: agents of antigen cross-presentation.
Expert Rev Vaccines 2008;7:1019–30.
41 Bendiksen S, Van Ghelue M, Winkler T, Moens U,
Rekvig OP. Autoimmunity to DNA and nucleosomes in
binary tetracycline-regulated polyomavirus T-Ag transgenic mice. J Immunol 2004;173:7630–40.
42 Means TK, Latz E, Hayashi F, Murali MR, Golenbock DT,
Luster AD. Human lupus autoantibody-DNA complexes
activate DCs through cooperation of CD32 and TLR9. J
Clin Invest 2005;115:407–17.
43 McGaha TL, Sorrentino B, Ravetch JV. Restoration of
tolerance in lupus by targeted inhibitory receptor expression. Science 2005;307:590–3.
44 Maldonado MA, Kakkanaiah V, MacDonald GC et al. The
role of environmental antigens in the spontaneous development of autoimmunity in MRL-lpr mice. J Immunol
1999;162:6322–30.
45 Mizutani A, Shaheen VM, Yoshida H et al.
Pristane-induced autoimmunity in germ-free mice. Clin
Immunol 2005;114:110–8.
46 Gaipl US, Beyer TD, Heyder P et al. Cooperation between
C1q and DNase I in the clearance of necrotic cell-derived
chromatin. Arthritis Rheum 2004;50:640–9.
www.rheumatology.oxfordjournals.org
47 Harley JB, Harley IT, Guthridge JM, James JA. The curiously suspicious: a role for Epstein–Barr virus in lupus.
Lupus 2006;15:768–77.
48 Barzilai O, Ram M, Shoenfeld Y. Viral infection can induce
the production of autoantibodies. Curr Opin Rheumatol
2007;19:636–43.
49 Caricchio R, McPhie L, Cohen PL. Ultraviolet B
radiation-induced cell death: critical role of ultraviolet dose
in inflammation and lupus autoantigen redistribution. J
Immunol 2003;171:5778–86.
50 Davis P, Percy JS. Role of ultraviolet light and UV DNA in
the induction of skin lesions in experimental animals. Br J
Dermatol 1978;99:201–5.
51 Lawley W, Doherty A, Denniss S et al. Rapid lupus autoantigen relocalization and reactive oxygen species accumulation following ultraviolet irradiation of human
keratinocytes. Rheumatology 2000;39:253–61.
52 Bijl M, Horst G, Bijzet J, Bootsma H, Limburg PC,
Kallenberg CG. Serum amyloid P component binds to late
apoptotic cells and mediates their uptake by
monocyte-derived macrophages. Arthritis Rheum 2003;
48:248–54.
53 Cohen PL, Caricchio R, Abraham V et al. Delayed apoptotic cell clearance and lupus-like autoimmunity in mice
lacking the c-mer membrane tyrosine kinase. J Exp Med
2002;196:135–40.
54 Bolland S, Ravetch JV. Spontaneous autoimmune disease
in Fc(gamma)RIIB-deficient mice results from
strain-specific epistasis. Immunity 2000;13:277–85.
55 Fukuyama H, Nimmerjahn F, Ravetch JV. The inhibitory
Fcgamma receptor modulates autoimmunity by limiting
the accumulation of immunoglobulin G + anti-DNA plasma
cells. Nat Immunol 2005;6:99–106.
56 Takahashi S, Fossati L, Iwamoto M et al. Imbalance
towards Th1 predominance is associated with acceleration of lupus-like autoimmune syndrome in MRL mice. J
Clin Invest 1996;97:1597–604.
57 Manolios N, Schrieber L, Nelson M, Geczy CL. Enhanced
interferon-gamma (IFN) production by lymph node cells
from autoimmune (MRL/1, MRL/n) mice. Clin Exp Immunol
1989;76:301–6.
58 Mathian A, Weinberg A, Gallegos M, Banchereau J,
Koutouzov S. IFN-alpha induces early lethal lupus in preautoimmune (New Zealand Black x New Zealand White)
F1 but not in BALB/c mice. J Immunol 2005;174:
2499–506.
59 Blanco P, Palucka AK, Gill M, Pascual V, Banchereau J.
Induction of dendritic cell differentiation by IFN-alpha in
systemic lupus erythematosus. Science 2001;294:1540–3.
60 Linker-Israeli M, Deans RJ, Wallace DJ, Prehn J,
Ozeri-Chen T, Klinenberg JR. Elevated levels of endogenous IL-6 in systemic lupus erythematosus. A putative
role in pathogenesis. J Immunol 1991;147:117–23.
61 Ishida H, Muchamuel T, Sakaguchi S, Andrade S,
Menon S, Howard M. Continuous administration of
anti-interleukin 10 antibodies delays onset of autoimmunity in NZB/W F1 mice. J Exp Med 1994;179:305–10.
639