TLR-dependent and TLR-independent pathways of type I interferon induction in systemic autoimmunity

NATURE MEDICINE VOLUME 13 | NUMBER 5 | MAY 2007 543

TLR-dependent and TLR-independent pathways of type I interferon induction in systemic autoimmunityRoberto Baccala, Kasper Hoebe, Dwight H Kono, Bruce Beutler & Argyrios N Theofilopoulos

We formulate a two-phase paradigm of autoimmunity associated with systemic lupus erythematosus, the archetypal autoimmune disease. The initial Toll-like receptor (TLR)-independent phase is mediated by dendritic cell uptake of apoptotic cell debris and associated nucleic acids, whereas the subsequent TLR-dependent phase serves an amplification function and is mediated by uptake of TLR ligands derived from self-antigens (principally nucleic acids) complexed with autoantibodies. Both phases depend on elaboration of type I interferons (IFNs), and therapeutic interruption of induction or activity of these cytokines in predisposed individuals might have a substantial mitigating effect in lupus and other autoimmune diseases.

What causes autoimmunity? Despite many advances over the years, this question remains largely unanswered. One might say that autoimmune diseases are genetically complex, depend upon a variety of stochastic environmental factors, and are so protean that each may really represent a collection of diseases with very different origins leading to a similar end. The striking feature, however, is that disease expression depends on the integrity of the immune system, and the challenge is thus to precisely define what is dysfunctional when most of this highly complex system operates exactly as it should.

We know that autoimmunity is strongly heritable, yet classical genetic analysis, which often illuminates more narrowly defined puzzles in biol-ogy, has primarily provided islands of fact in a sea of uncertainty. Many mutations affecting components of the immune system can prevent organ-specific and systemic autoimmunity, whereas other mutations can enhance disease in predisposed mice and even induce autoimmu-nity in healthy mice (reviewed in refs. 1,2). Although these findings have pointed out the multiplicity of pathways affecting autoimmunity, what is uniquely ‘wrong’ in predisposed individuals and the exact nature of the mechanisms by which susceptibility and effector genes promote disease largely elude us.

With time, a set of unifying principles has emerged from fields once seemingly unconnected to one another. Our growing understanding of innate immune signaling pathways has played a major part in this and has provided new paradigms to explain autoimmunity, particularly systemic autoimmunity associated with lupus. Although this disease may originate in several different ways, its pathogenesis clearly depends on events causing activation of antigen-presenting cells (APCs), which

lie at the interface of the innate and adaptive immune systems. On the basis of recent developments, we deduce that two fundamental phases exist in systemic autoimmunity. Each phase depends upon a different pathway for APC activation, and each revolves around the production of a key class of cytokines: the type I interferons (IFNs).

Interferons as a focal point: how and whyA relevant critical event is needed to analyze any biological process, and, for lupus, one such event seems to be type I interferon production. Type I IFNs constitute a large cytokine family, of which the multiple IFN-α and the single IFN-β are the most immunologically relevant. Induction of IFN-α/β is primarily controlled by IFN regulatory factors (IRFs), and signaling through a common two-chain receptor (IFNAR) leads to the activation of a large set of genes that regulate many biologi-cal functions in a cell-specific manner (reviewed in refs. 3,4). Although virtually any cell type can elaborate type I IFNs, the major producers are a phenotypically and functionally distinct subset of DCs, known as plasmacytoid dendritic cells (pDCs) (reviewed in ref. 5). Redistribution of pDCs from blood to skin occurs in individuals with lupus, and implies the acquisition of chemokine receptors by these cells, induced by IFN-α/β production.

Lupus is a highly heterogeneous, multiorgan disorder primarily afflicting women and is characterized by the production of diverse autoantibodies, predominantly against nucleosomal and spliceosomal antigens. Type I IFNs, central to both innate and adaptive immunity, have received particular attention for their role in the development of autoimmune responses, and a preponderance of evidence supports their disease-promoting activity in lupus (reviewed in refs. 6,7). In humans with active disease, levels of IFN-α are increased in serum and affected tissues. Peripheral blood lymphoid cells and kidneys express a so-called ‘interferogenic signature’ (in that IFN-controlled genes are upregulated), and IFN-α in lupus sera promotes monocyte maturation to antigen-presenting dendritic cells (DCs). In addition, lupus-like manifestations

The Scripps Research Institute, Department of Immunology, La Jolla, California

93037, USA.

E-mail: [email protected]

Published online 3 May 2007; doi:10.1038/nm1590

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

544 VOLUME 13 | NUMBER 5 | MAY 2007 NATURE MEDICINE

are observed in some individuals with trisomy of the type I IFN gene cluster8, and single nucleotide polymorphisms in two IFN-related genes, encoding IFN regulatory factor 5 (IRF5) and tyrosine kinase 2 (TYK2), are associated with human lupus9,10. Furthermore, IFN-α therapy for unrelated disorders is frequently associated with various autoimmune manifestations, including lupus-like disease. In NZB and (NZB × W)F1 lupus-predisposed mice, in which the disease most closely resembles the human disorder, manifestations are accelerated and enhanced by type I IFNs and their inducers. Moreover, disease is reduced in NZB mice homozygous and even heterozygous for Ifnar1 gene deletion11, whereas in lupus-prone mice with the Faslpr mutation (uncommon in human lupus), the absence of IFNAR has variable effects, with amelioration in the mild disease–developing 129xB6-Faslpr mice12 and exacerbation in the severe disease–developing MRL-Faslpr mice13. Moreover, IFN-β treatment reduces serologic and histologic disease manifestations in MRL-Faslpr mice14. Interestingly, however, blocking IFN-γ reduces dis-ease in MRL-Faslpr mice (reviewed in ref. 15). The differential effects of IFNs are thus clearly contextual, depending upon the underlying genetic defect(s) driving the autoimmune process.

IFN-α/β promote normal and abnormal (autoimmune) responses by a multitude of effects on both the innate and adaptive immune systems. A major effect is DC activation and upregulation of MHC and costimu-latory molecules (CD40, CD80, CD86), leading to efficient self-antigen presentation to previously quiescent low-affinity autoreactive T cells. Relevantly, DCs of C57BL/6 mice congenic for a major lupus-predispos-ing locus (Sle3) display enhanced activation and antigen presentation compared to wild-type controls16. IFN-α/β signaling also increases DC survival, which may also contribute to disease as DCs rendered apop-tosis-resistant can induce lupus-like manifestations in normal and pre-disposed mice17. DCs generated in the presence of IFN-α/β–containing lupus sera also promote maturation of effector CD8+ T cells that kill target cells and generate nucleosomes and other lupus autoantigens7. In addition, IFN-α/β–activated DCs produce several cytokines, notably BAFF (BLyS) and APRIL, which promote B-cell survival, differentiation and isotype switching, characteristics that are associated with lupus. Thus, humans and mice with lupus have high serum levels of BAFF that correlate with autoantibody titers. Moreover, normal mice transgenic for BAFF or lacking the BAFF inhibitor TACI have vastly increased numbers of mature B cells and develop lupus-like disease, whereas lupus mice treated with recombinant TACI-Ig or receptors for BAFF and APRIL have reduced disease6,7.

IFN-α/β also exert several direct effects on T and B cells relevant to the pathogenesis of systemic autoimmunity, including increased T-cell activation, survival, cross-priming and T helper type 1 (TH1) deviation, as well as B-cell activation and Ig isotype switching6,18. IFNAR signaling also promotes expansion of B1 cells, thought to be major producers of autoantibodies11. Moreover, type I IFNs inhibit B- and T-cell lympho-poiesis19, thereby limiting dilution of expanded antigen-specific clones. This would be useful during foreign antigen responses, but harmful in autoimmunity. There is also cross-talk between type I IFNs and other cytokines, most notably the lupus-promoting IFN-γ (ref. 6). Moreover, cross-regulation between IFN-α/β and tumor necrosis factor (TNF)-α has been postulated to play a role in the induction of lupus and rheu-matoid arthritis, if IFN-α/β predominate, and rheumatoid arthritis, if TNF-α predominates7. Thus, although the many immune response–enhancing effects of type I IFNs are normally beneficial, they can be detrimental in autoimmune-predisposed individuals.

This overwhelming evidence has prompted inquiry into the critical question of what drives IFN-α/β production, as the answer should yield insight into the etiology of lupus and other autoimmune diseases, and elucidate pathways that might be targeted for therapy. Fortunately, we now know quite a bit about how type I IFN can be produced and there-fore have a fairly restricted number of avenues to explore. The stimuli for type I IFN production can be exogenous or endogenous, and in both cases TLR dependent or TLR independent (Fig. 1). All four categories are considered here.

Type I IFN induction by exogenous stimuliRecent advances have provided a clearer picture of the pathways by which the innate immune system senses bacterial and viral pathogens, and how this leads to production of type I IFNs. This process is primarily mediated by the engagement of specific TLRs, but also by TLR-indepen-dent pathways (reviewed in refs. 3,20–24). Deciphering these pathways has advanced our understanding of the mechanisms by which microbial pathogens can potentially contribute to autoimmunity and has triggered parallel studies on pathways engaged by endogenous stimuli.

TLR-dependent pathways. Currently, 13 mammalian TLR paralogs, expressed on cell surfaces or in endoplasmic reticulum (ER) and endo-somal compartments, have been identified. Cell surface TLRs 1, 2, 4, 5, 6 and 11 sense lipids, lipoproteins or peptidoglycans from bacteria, fungi or protozoa. Intracellular TLRs 3, 7, 8 and 9 detect bacterial and viral nucleic acids. The ligands for TLRs 10, 12 and 13 remain unidentified. TLR10 is expressed in humans but not mice, TLR8 is not functional in mice, and TLRs 11, 12, and 13 are expressed in mice but not humans. TLRs involved in the production of IFN-α/β include TLRs 7 and 9, strongly expressed by human and mouse pDCs, and TLRs 3, 4 and 9, expressed by conventional DCs and macrophages. Interestingly, only the TLR2 complexes (TLR2/TLR6 and TLR1/TLR2) are considered inca-pable of activating type I IFN responses. No information on the type I IFN–inducing activity of TLRs 5 and 10 (in humans) or TLRs 11, 12 and 13 (in mice) is presently available. Overall, the various hematopoietic cell types display considerably different TLR expression profiles and responses, but the biological significance of these differences has not yet been fully defined.

The specific ligands recognized by IFN-α/β–inducing TLRs include dsRNA for TLR3, LPS for TLR4, ssRNA for TLRs 7 and 8, and unmeth-ylated CpG-DNA for TLR9. Ligand binding leads to TLR dimerization and conformational changes in both the receptor ectodomains and the cytoplasmic TLR–interleukin-1 receptor (TIR) domains. TIR domains bind specific adaptors with like domains: that is, Trif (TICAM1) for TLR3; TRAM (TICAM2), Trif, TIRAP (Mal) and MyD88 for TLR4; and MyD88 for TLRs 7, 8 and 9. These adaptors propagate signaling by

Endogenous stimuli

Exogenous stimuli

Dendritic cellsActivationSurvival

Ag presentationBLyS, APRILChemokine R

T cellsActivationSurvival

Cross-primingTH1 deviationChemokine R

B cellsActivation

Ig switchingAg presentation

B1 cell expansion

TissuesUpregulation of MHC and chemokines by mesangial

cellsTubular epithelial cells

MacrophagesNeuronal cells

IFN-α/β

TLR-independentTLR-dependent

Figure 1 Modes of type I IFN induction and effects on immunocytes and parenchymal cells that promote systemic autoimmunity.

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

NATURE MEDICINE VOLUME 13 | NUMBER 5 | MAY 2007 545

recruiting kinases that mediate activation of the transcription factors AP-1 (heterodimer of ATF2 and c-Jun), NF-κB, IRF3 and IRF7, which are all required for IFN-α/β production. Negative regulators of TLR and type I IFN signaling (ST2, SIGIRR, SARM and SOCS, to name a few) have also been identified24,25. The TLR-associated signaling pathways have been extensively reviewed3,20–24 and, together with more recent findings, are summarized here (Fig. 2).

TLR-independent pathways. Two cytosolic RNA helicases, RIG-I (retinoic acid–inducible gene I) and MDA5 (melanoma differentiation–associated gene 5, also called Helicard), mediate TLR-independent IFN-α/β induction by actively replicating RNA viruses and synthetic RNA (reviewed in refs. 20,26; Fig. 3). RIG-I and MDA5 bind through caspase recruitment domains (CARD) to mitochondrial IPS-1 (also known as MAVS, VISA or Cardif), initiating signaling cascades that lead to IRF3, IRF7, NF-κB and AP-1 activation and IFN-α/β expression. RIG-I and MDA5 are upregulated by IFN-α/β through a positive feedback

mechanism and are differentially engaged by various viruses and synthetic RNAs (ref. 27). RIG-I recognizes cytosolic uncapped ssRNA bearing 5′-triphosphates generated by some viral polymerases28,29, whereas MDA5, rather than TLR3, appears to be the dominant recep-tor for polyI:C in vivo27,30. A third helicase, LGP2, lacks CARD domains and inhibits RIG-I and MDA5 responses, probably by sequester-ing stimulatory RNA (refs. 31,32). The RIG-I/MDA5 pathway appears essential for IFN-α/β production by fibroblasts, conventional DCs and macrophages, but not pDCs that operate through the TLR pathways. In addition to viral cytosolic RNA, bacterial cytosolic DNA also triggers TLR-independent type I IFN produc-tion through an as yet not fully defined path-way that requires IRF3, but not activation of NF-κB and MAPKs (ref. 33).

Type I IFN induction by endogenous stimuliUntil recently, self-tolerance and pathogenic autoimmunity were mainly addressed from the perspective of the adaptive immune system. The discovery that endogenous materials, par-ticularly self–nucleic acids and related immune complexes, can stimulate the innate immune system by TLR-dependent and TLR-indepen-dent pathways has added a new dimension that requires reappraisal of how self-tolerance is circumvented.

TLR-dependent pathways. The TLRs that might exhibit dangerous cross-reactivity with self-constituents are primarily TLRs 3, 7, 8 and 9, which recognize nucleic acids, though, to be sure, other TLRs may have endogenous ligands (reviewed in ref. 34). These TLRs are confined to endosomal compartments, in which micro-bial, but not mammalian, nucleic acids are gen-erally concentrated by phagocytosis. Additional safety mechanisms include the methylation of stimulatory DNA and RNA motifs, and the rapid degradation of nucleic acids by abun-dant nucleases. These barriers, however, are not

infallible and can be breached, resulting in direct or indirect stimulation of autoreactive T and B cells via two novel pathways (Fig. 4).

The first mechanism identified in breaching these barriers involves human FcγRIIa (CD32, corresponding to mouse FcγRIII), which mediates uptake by pDCs of apoptotic/necrotic materials complexed with lupus serum IgG autoantibodies, leading to secretion of large amounts of IFN-α (reviewed in ref. 35). Attenuation of this activity primarily by RNase sug-gested that the activating principle is RNA in the small nuclear ribonu-cleoprotein (snRNP) particles, which are common autoantigenic targets in lupus. Indeed, several recent studies have demonstrated that U-rich small nuclear RNAs (snRNAs), particularly U1, which are tightly bound to Sm and other snRNP components, engage TLR7/8 when delivered to endosomes via autoantibodies or liposomes34,36–40. Insofar as the uptake of immune complexes containing DNA or chromatin by mouse DCs is concerned, FcγRIII- and TLR9-dependent pathways of cytokine induction have been described37,41. Likewise, DNA-containing immune

cDCMacrophages pDC

Endosome

Nucleus

TRAF6 IRAK4

IRAK1

TAK1

MAPK

TBK1

TRAF3

NAP1

LPS

IRF7

TRAF6

Irf7 IfnaIfnb1 Irf7IfnaIfnb1

AP-1

TRAF6 IRAK4

IRAK1

TAK1

NF-kB

IRF

3

IRF3

NF

-kB

NF

-κB

NF-κBAP-1

AP

-1

AP

-1

IRF

7

IRF

7

IRF

7

IRF7

IKKε IKKαβγ

MyD88MyD88 Trif

ISG

F3

ISG

F3

IKKαβγ

IKKαOpn

TRAF3IRF5

NF-κB

RIP-1

dsRNA ssRNA CpG-DNATRAM

IFNARIFNAR

TLR4

TLR3 TLR7 TLR9

PKCε

NF-κB

MyrPIP2

IκBIκB

Figure 2 TLR-dependent pathways of IFN-α/β induction. Left, in conventional DCs and macrophages, engagement of TLR3 by dsRNA in endosomes and of TLR4 by LPS at the cell surface mediates IFN-α/β production. TLR3 induces a Trif-dependent pathway, whereas TLR4 induces both Trif- and MyD88-dependent pathways, of which the former requires TRAM and the latter TIRAP. TRAM and TIRAP are phosphorylated by PKCε and the Bruton tyrosine kinase, respectively. Also, TRAM contains an N-terminal myristoylation site (Myr) and TIRAP a phosphatidylinositol 4,5-biphosphate (PIP2)-binding domain targeting these ‘sorting adaptors’ to the plasma membrane for interaction with the ‘signaling adaptors’ Trif and MyD88. With the participation of TRAF3, NAP1, RIP-1 and TRAF6 (a factor with ubiquitin (Ub) ligase activity), Trif recruits kinases (TBK1, IKKε, TAK1, MAPK and IKKαβγ) that mediate activation of the transcription factors IRF3, AP-1 and NF-κB. In addition, with the participation of IRAK1 and IRAK4, MyD88 can also induce TRAF6-dependent activation of AP-1 and NF-κB, but not IRF3. IRF3, AP-1 and NF-κB translocate into the nucleus and bind to positive regulatory domains on the Ifnb1 gene promoter, leading to chromatin remodeling and Ifnb1 transcription. IRF3 (without AP-1 and NF-κB) can also drive expression of Ifna4, but not other Ifna genes (not shown). Secreted IFN-β (and IFN-α4) then signals through the IFNAR and stimulates ISGF3-mediated production of IRF7, which, upon activation by TBK1/IKKε, induces expression of several Ifna genes. Right, in pDCs, engagement of endosomal TLR7 by ssRNA and TLR9 by CpG-DNA induces a MyD88-dependent pathway that leads to activation of the constitutively expressed IRF7 and transcription of several Ifna genes. IRAK1, IRAK4, TRAF6, osteopontin (Opn), IKKα and TRAF3 are involved in IRF7 activation. IRF5 also contributes to TLR7 signaling in human but not mouse cells. In addition, MyD88 promotes activation of AP-1 and NF-κB, which, together with IRF7, initiate Ifnb1 transcription. TLR9 is also present in myeloid DCs and macrophages and induces IFN-β through MyD88 and IRF1, instead of IRF3/7.

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

546 VOLUME 13 | NUMBER 5 | MAY 2007 NATURE MEDICINE

complexes from lupus sera induce IFN-α by pDCs through cooperation with FcγRIIa and TLR9 (ref. 42). Thus, uptake of immune complexes through FcγR facilitates the trafficking of nucleic acids to TLR-contain-ing endosomal compartments. Moreover, nucleic acids may be protected from nuclease activity when contained within nucleosomal or RNP parti-cles, and protection may be enhanced when such particles are complexed with autoantibodies. Nonetheless, hypomethylated nucleic acid motifs will still be required, implying that the few stimulatory CpG islands in mammalian DNA are sufficient for TLR engagement. Such islands may be preferentially released during apoptosis and appear to be enriched in DNA-containing immune complexes isolated from lupus sera43.

A second pathway that can lead to a breach in the protective barrier encompasses the uptake of chromatin, snRNPs or related immune com-plexes by B cells expressing antigen receptors (BCR) that bind either the antigenic moieties of the macromolecule (BCR specific for DNA/histone or Sm/RNP) or the Fc moiety of the autoantibody (BCR with rheuma-toid factor activity)34. These materials are transported into endosomal compartments where DNA interacts with TLR9 and RNA with TLR7/8. The simultaneous engagement of a BCR and a TLR results in efficient B-cell activation and proliferation. Remarkably, addition of IFN-α to transgenic rheumatoid factor B-cell cultures significantly increases the response to Sm/RNP-containing immune complexes, presumably due to upregulation of TLR7, suggesting that optimal stimulation of B cells in vivo requires the concerted uptake of these complexes by pDCs and the positive feedback effects of the pDC-produced IFN-α/β on B cells.

TLR-independent pathways. Endogenous stimuli that can induce type I IFNs by TLR-independent mechanisms have also been iden-tified, and might play a role in the pathogenesis of systemic auto-immunity (Fig. 4). As noted earlier, one form of such induction is mediated by cytosolic RNA sensed by the RIG-I or MDA5 helicases.

Because most eukaryotic RNA species are known to lack 5′-triphosphate groups, which are required for sensing by RIG-I, it has been suggested that this feature allows discrimi-nation between self-RNA and viral RNA. However, RNA transcripts in the nucleus and some RNA species (for example, 7SL RNA) in the cytosol of eukaryotic cells display 5′- triphosphates, and therefore uptake of apop-totic materials containing such RNA species by conventional DCs or macrophages may lead to RIG-I engagement and type I IFN production.

Intracellular administration of right-handed B-form dsDNA, the most common conformation of mammalian DNA, also trig-gers TLR-independent production of type I IFNs (refs. 44,45), and such DNA may be pres-ent in phagocytosed apoptotic debris. Purified nucleosomes have also been found to directly induce MyD88-independent DC maturation and production of inflammatory cytokines and chemokines, although type I IFNs were not assessed in this study46. In addition, spon-taneous internalization of plasmid DNA by B cells promotes TLR9-independent upregula-tion of costimulatory molecules and enhances antigen presentation47. IFN-α/β induction by cytosolic B-form DNA involves activation of IRF3 and NF-κB but is independent of RIG-I/MDA5 and IPS-1 (refs. 44,45,48). The

specific sensors and adaptors recognizing self-cytosolic DNA, as is the case with bacterial DNA, remain to be defined.

Induction of type I IFNs by uptake of mammalian nucleic acids derived from apoptotic materials may amplify immunologic responses not only against foreign pathogens, but also against self-antigens in predisposed individuals. The potential relevance of this mechanism in autoimmunity is suggested by the finding that inadequate digestion of extracellular DNA in DNase I-deficient mice leads to a lupus-like disease49, and function-impairing DNase I gene mutations have been found in a few individuals with lupus50,51. In addition, nondegraded intracellular DNA in DNAse II-deficient macrophages mediates TLR-independent induction of inflammatory cytokines, including IFN-β, and causes severe anemia that can be prevented in IFNAR-deficient mice52,53. A chronic polyarthritis syndrome resembling human rheumatoid arthritis and mediated by a TLR-independent path-way has also been observed in mice with an induced deletion of the DNase II gene54.

Recent experiments have shown that some forms of cell death trigger a strong TLR-independent signal that leads to the induction of type I IFNs and in vivo priming of T cells with specificity for antigens in dying cells55. Programmed cell death induced by γ- or ultraviolet irradiation and, immunologically more relevant, by Fas ligation, triggers such responses. The cells capable of sensing cell death–derived materials are represented within Flt3L-derived DCs, display markers of immature lymphoid B220–PDCA–CD8– DCs and convert into mature lymphoid B220–PDCA–CD8+ DCs upon stimulation. In contrast, myeloid DCs, pDCs and macrophages are unresponsive to this stimulus. The exact mechanisms by which uptake of cell death–derived materials induce IFN-α/β in this DC subset and the cellular compartments where sens-ing occurs remain to be defined, but candidates may include nucleic

CARDCARD

CARD CARD CARD

CARD CARD

cDCMacrophages

TRAF6

MAPKTBK1

TRAF3

IKKεIKKαβγ

FADD

RIP1

AP-1 IRF3

NF-κB

IκB

NF

-κB

AP

-1

IRF

3

IRF

7

IRF7

Irf7 IfnaIfnb1

Nucleus

IFNAR

Mitochondrion B-DNALGP2 RNA

Adaptor?

DNA sensor?

IPS-1

ISG

F3

MDA5

RIG-I

NF-κB

Figure 3 TLR-independent pathways of IFN-α/β induction. Engagement of the cytosolic RNA helicases RIG-I (for example, by 5′-triphosphate RNA) and MDA5 (for example, by polyI:C) induces conformational changes that allow CARD-mediated homotypic interaction between RIG-I/MDA5 and the mitochondrial adaptor IPS-1. The ensuing TBK1/IKKε, IKKαβγ and MAPK signaling induces IRF3, AP-1 and NF-κB translocation, and Ifnb1 transcription. It has also been shown that through the engagement of unknown sensor(s) and adaptor(s), cytosolic B-form dsDNA mediates activation of TBK1/IKKε and IKKαβγ, translocation of IRF3 and NF-κB, and expression of IFN-α/β.

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

NATURE MEDICINE VOLUME 13 | NUMBER 5 | MAY 2007 547

acid sensors and possibly members of the recently described family of NOD-like receptors (reviewed in ref. 56).

Cumulatively, the findings suggest that apoptotic materials, in addi-tion to carrying a cargo of autoantigens57, also contain nucleic acids and other undefined ligands that induce IFN-α/β production, which acting in concert provoke autoimmune responses. Therefore, defects in the generation, removal or degradation of apoptotic debris may be primary contributors to the pathogenesis of systemic autoimmunity. Indeed, human and mouse lupus are characterized by increased apop-tosis and defective clearance of apoptotic materials58,59, as is the case in humans and some mouse strains deficient in C1q (ref. 60), as well as in mice lacking members of the Tyro 3 family of receptor tyrosine kinases (Tyro 3, Axl and Mer)61. Furthermore, injections of apoptotic cells or apoptosis-promoting agents into normal and lupus mice induce autoantibodies and kidney disease62,63, whereas increased clearance of apoptotic bodies by administration of calreticulin-binding adiponec-tin reduces autoimmunity in MRL-Faslpr mice64. In this regard, it is well established that apoptotic materials can be perceived as a ‘danger signal’65 and trigger adaptive immune responses against microbial and tumor antigens55,66–68.

Type I IFN induction pathways and lupus pathogenesisWe turn then to the question of which of the four general pathways for type I IFN induction predominates in the pathogenesis of lupus. On the basis of the evidence discussed above, it is likely that all of them con-tribute, depending upon which phase of the disease is being considered.

Nonetheless, direct experimental documentation has been provided pri-marily for the role of the TLR-dependent pathways.

Exogenous TLR stimuli and autoimmunity. Microbial infections are known to precipitate or aggravate autoimmune diseases, and, although antigenic mimicry has often been suggested as an explanation, activation of the immune system by TLR engagement may be another important contributing factor. For instance, B10.S mice transgenic for a T-cell recep-tor specific for an encephalitogenic myelin proteolipid protein peptide are normally resistant to experimental allergic encephalomyelitis, but tolerance and disease resistance can be broken after APC activation by TLR9 or TLR4 ligands69. Similarly, mice with transgenic expression of lymphocytic choriomeningitis virus (LCMV)-gp in the pancreas or liver develop overt diabetes or autoimmune hepatitis, respectively, only when immunized with an immunodominant LCMV-gp peptide coupled with TLR7 or TLR3 ligation, leading to production of IFN-α/β and other inflammatory cytokines, MHC class I upregulation, and tissue release of chemokines that attract self-reactive CD8+ T cells70,71. An important dis-tinction in this model is that, unlike peptide immunization that requires the addition of a TLR ligand, infection with LCMV is sufficient to promote disease. The effect of the virus could be attributed to direct TLR triggering; however, release of cell debris and nucleic acids following virus-induced tissue damage may also lead to TLR-independent IFN-α/β production. TLR signaling also appears to play a role in the pathogenesis of other organ-specific autoimmune models72. Notably, despite the potential for pathogenic consequences, TLR stimulation has been a means to promote beneficial autoimmune responses against cancer cells73,74.

IFN-α/β Proliferation/Ag presentation

DNADNA

DNA

RNA

RNA

TLR

-dep

ende

ntT

LR-in

depe

nden

t

Chromatin

TLR7 TLR9TLR7 TLR9

Cytosolic sensors(s)

RIG-IorMDA5

Transport? Transport?

Anti-Sm/RNP Anti-DNA Anti-Sm/RNP RF Anti-DNA RF

pDC

B cell B cell

LymphoidDC

MyeloidDC

Apoptotic cell

Self-Ag

snRNP

BCRFcγRlla

B cell

Figure 4 Endogenous stimuli promoting IFN-α/β production by pDCs and conventional DCs, and activation of B cells. Top, TLR-dependent pathways. Nuclear autoantigens (snRNP, chromatin and their subcomponents) complexed with autoantibodies are taken up via FcγRIIa and induce IFN-α/β production by pDCs through engagement of endosomal TRL7 and TLR9 by RNA and DNA, respectively. Similar complexes can also promote TLR-dependent activation of B cells expressing antigen receptors (BCR) specific for either autoantigenic epitopes (Sm/RNP, DNA, histones) or the Fc portion of the autoantibodies in the immune complex (rheumatoid factor, RF). The dual engagement of BCR and TLR7/9 results in efficient B-cell proliferation. The activated B cells can also present peptides from the endocytosed macromolecules to T helper cells. Bottom, TLR-independent pathways. RNA and DNA, or macromolecules containing these nucleic acids, and other undefined, apoptotic cell-derived materials induce TLR-independent IFN-α/β production by myeloid or lymphoid DC subsets. In addition, uptake of DNA by B cells with diverse self-antigen specificities can promote antigen presentation.

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

548 VOLUME 13 | NUMBER 5 | MAY 2007 NATURE MEDICINE

Turning to systemic autoimmunity, flares of human lupus are fre-quently associated with, or follow, viral infections, and Epstein-Barr virus (EBV) is considered a major environmental risk factor for this disease75. Moreover, the disease is enhanced in lupus-predisposed mice repeatedly injected with bacterial or viral TLR ligands or mimics, such as polyI:C (in 129xB6-Faslpr mice), LPS/lipid A (in female MRL+/+, BXSB and NZW mice), imiquimod (in MRL-Faslpr mice) and hypo-methylated bacterial or synthetic CpG-DNA (in (NZB × W)F1 mice)6. Although engagement of TLRs expressed in tissues (particularly in the kidney) may also contribute to end-organ inflammation and damage (reviewed in ref. 76), the effects of TLRs on lupus are primarily attributed to systemic immune system activation. Type I IFNs are the probable mediators of these effects, as disease exacerbation is not observed in similarly injected IFNAR-deficient 129xB6-Faslpr mice12; also, repeated administration of IFN-α or a plasmid encoding this cytokine accelerates disease in NZB and (NZB × W)F1 mice77, and type I IFNs upregulate TLRs, including TLRs 7 and 9, in DCs and B cells78,79. It should be noted, however, that mice of normal background treated with TLR agonists, despite displaying some autoantibody specificities, lack autoantibodies to dsDNA and histologic disease manifestations77,80. Thus, defects in B-cell tolerance and other genetic contributions are required for full disease development.

Endogenous TLR stimuli and autoimmunity. The relevance of endogenous stimuli of IFN-α/β in the pathogenesis of systemic autoim-munity has recently been addressed (Table 1). Despite the early in vitro finding that chromatin-containing immune complexes promote B-cell proliferation through TLR9 engagement, in vivo results with TLR9-defi-cient lupus-predisposed mice are not fully congruent with expectations. Thus, studies in incompletely backcrossed TLR9-deficient MRL-Faslpr

lupus mice showed that autoantibodies to chromatin and dsDNA are decreased, but survival, immune deposits and kidney disease are unaf-fected, probably due to TLR7-mediated induction of autoantibodies against Sm/RNP (ref. 81). The same study showed that absence of TLR3 has no significant effects. Another study with lupus-prone C57BL/6 mice transgenic for a DNA-specific IgM heavy chain that lacked the inhibi-tory FcγRIIB receptor demonstrated that TLR9 signaling is also required for class switching to pathogenic polyreactive and DNA-specific IgG2a and IgG2b autoantibodies and kidney disease, but not for generation of immature autoreactive B cells82. However, other groups found that TLR9 is not required for the production of dsDNA-specific autoantibodies and that, in fact, its absence leads to more severe disease. Thus, truly con-genic TLR9-deficient MRL-Faslpr mice show increases in dsDNA-specific autoantibodies (defined by ELISA), IgG levels (including the patho-genic IgG2a and IgG2b isotypes), lymphadenopathy, kidney damage and mortality83. Likewise, TLR9-deficient C57BL/6-Faslpr mice have higher titers of autoantibodies to dsDNA and nucleolar antigens, and kidney mesangial proliferation, but no nucleosome-specific autoantibodies84. Moreover, deletion of TLR9 causes aggravation of humoral, cellular and histologic disease characteristics, with concomitant increases in nucleolus-specific autoantibodies in a model of lupus caused by B-cell hyperactivity mediated by a gain-of-function mutation in the phospho-lipase Cg2 gene85. A recent detailed study, however, although concurring with overall enhanced disease manifestations in TLR9-deficient MRL-Faslpr, showed significant decreases in autoantibody titers to dsDNA measured by the highly specific Crithidia luciliae staining assay instead of the less specific ELISA (ref. 86). These results are compatible with previ-ous genetic studies in lupus-predisposed mice showing low association between loci affecting dsDNA-specific autoantibodies and nephritis87.

Why deletion of TLR9 enhances overall disease is unclear, but there are several possibilities. Thus, TLR9 may be involved in the generation of polyspecific autoantibodies akin to natural autoantibodies, which may promote uptake and removal of apoptotic materials by the phago-cytic cells of the reticuloendothelial system (reviewed in ref. 88). Other potential mechanisms include decreases in DNA-specific B cells, leading to availability of niches for expansion of presumably more pathogenic Sm/RNP-specific B cells, enhanced availability of nuclear particles for TLR7 engagement, diversion of DNA trafficking to the cytosol and IFN-α/β induction by TLR-independent pathways, or ineffectiveness of TLR9-binding inhibitory sequences in mammalian DNA. Disease exac-erbation in TLR9-deficient MRL-Faslpr mice might also be attributed to decreased production of IFN-α/β, which appears to be protective in this lymphoproliferation/autoimmunity model13,14.

Additional evidence for the involvement of TLRs in the broader nuclear antigen-specific autoantibody (ANA) response is the find-ing that MyD88 deficiency causes reductions in autoantibodies to chromatin, Sm and IgG (rheumatoid factors) in MRL-Faslpr mice89. Similarly, in the absence of MyD88 signaling, lupus-prone FcγRIIB−/− C57BL/6 mice transgenic or not for a DNA-specific IgM heavy chain show reduced switching to pathogeneic IgG2a and IgG2b autoantibod-ies82. Interpretation of these findings, however, is complicated by the concurrent decrease in the function of several MyD88-signaling cyto-kines (IL-1, IL-18, IL-33). Nonetheless, several recent reports support the role of TLR7 in the pathogenesis of lupus. Thus, TLR7-deficient MRL-Faslpr mice have low titers of autoantibodies to Sm/RNPs and, importantly, less severe disease, as evidenced by fewer activated lym-phocytes and pDCs, and a modest decrease in renal pathology86. Loss of B-cell tolerance in Ig knock-in C57BL/6 mice expressing heavy and light chains of an autoantibody recognizing ssDNA, RNA and nucleo-somes has also been shown to be TLR7 dependent90. Furthermore, the Y chromosome–linked autoimmunity accelerator (Yaa) locus in male

Stress

Fas/FasLUV

Amplification phase(TLR-dependent)

Immunecomplex

Late apoptosisnecrosis

Auto-antibodies

DC

DC

pDC

Early apoptosis

BLySAPRIL

Help

FcγRlla

IFN-α/β

CD80CD86MHC

Initiation phase(TLR-independent)

Figure 5 The paradigm of the two-phase IFN-α/β induction in systemic autoimmunity. The initiation phase is TLR independent and mediated by apoptotic cell material taken up by a specialized subset of lymphoid DCs, leading to IFN-α/β production. Under the effect of IFN-α/β, lymphoid and myeloid DCs upregulate MHC and costimulatory molecules and differentiate into efficient self-antigen–presenting cells, leading to activation of quiescent autoreactive T helper cells (for example, those specific for RNP-derived peptides). As a result of T-cell help, and the effects of BLys, APRIL and IFN-α/β, autoreactive B cells (for example, Sm/RNP-specific) proliferate and differentiate into plasma cells. Subsequent to the formation of nucleic acid–containing immune complexes, the amplification phase is induced, which encompasses TLR-dependent IFN-α/β induction in pDCs and DCs, and enhanced B-cell proliferation and autoantibody production.

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

NATURE MEDICINE VOLUME 13 | NUMBER 5 | MAY 2007 549

BXSB mice appears to result from an X to Y chromosome transloca-tion of a cluster of genes that intriguingly includes Tlr7 (refs. 91,92). Therefore, duplication of this gene and increased expression in B cells are probably responsible for the early-onset severe disease in male BXSB mice. The TLR7 duplication alone, however, is insufficient for disease, as the BXSB Y chromosome promotes lupus in predisposed strains, but not in consomic normal strains unless complemented with other autoimmunity-predisposing genetic abnormalities91,92. Finally, C57BL/6-Faslpr mice congenic for the 3d mutation of the Unc93b1 gene, which impairs signaling via TLRs 3, 7 and 9 as well as cross-presentation93, also show significantly decreased autoantibody levels, lymphoaccumulation and mortality.

The results as a whole provide considerable evidence that TLR engagement is involved in the induction of autoantibodies against macromolecular complexes encompassing nucleic acids, which pre-dominate and are pathognomonic in lupus. However, in some cases, such as with TLR9, TLR engagement may have protective effects, probably by inducing autoantibodies that remove subcellular struc-tures with broad TLR-engaging capacity. Studies correlating genetic backgrounds, nature of the stimuli and autoantibody specificities, as well as conditional deletions of TLRs in different hematologic cell types and tissues, will further clarify the exact role of TLRs in lupus pathogenesis. Nonetheless, type I IFN induction by self–nucleic acids, acting via TLR-dependent and TLR-independent pathways, appears to be the central event that drives expansion of nontolerant T and B cells and production of the predominant autoantibodies specific for nuclear antigens. By enhancing antigen presentation, this event may also be responsible for the production of lupus-associated autoanti-bodies against a wide array of self-constituents (for example, eryth-

rocytes, lymphocytes, neurons, glomeruli and phospholipids) that are not known to engage TLRs.

The two-phase paradigm of IFN-α/β induction in lupusAlthough microbial organisms and other environmental factors can induce type I IFNs and precipitate lupus, the fact that the disease devel-ops in predisposed mouse models in a consistent manner and even under germ-free conditions94 suggests that endogenous stimuli can be sufficient triggers. Indeed, apoptotic materials or nucleic acids when complexed with autoantibodies have been shown to induce TLR-depen-dent production of type I IFNs by pDCs. In view of our current know-ledge of TLR biology, this finding makes sense, as mammalian nucleic acids are normally excluded from the endosomes where the specific TLRs reside but can be targeted to these subcellular compartments upon autoantibody/FcγR-mediated uptake. However, this mechanism cannot provide the primary stimulus for IFN-α/β production as it is predi-cated on the pre-existence of specific autoantibodies. On the other hand, B cells displaying receptors that recognize nucleic acid–containing self-antigens (either directly or indirectly via rheumatoid factor activity) can be induced to proliferate through the simultaneous engagement of BCRs and TLRs. However, sustained autoreactive B-cell proliferation, differentiation and isotype switching to pathogenic Ig subclasses cannot solely be mediated by this mechanism as activated T helper cells are also required. Although B cells may act as APCs, efficient priming of auto-reactive T cells is likely to be mediated by type I IFN–activated DCs.

On the basis of the reviewed evidence, we therefore propose that the effector stage of lupus-associated systemic autoimmunity can be for-mulated as a two-phase process, in which type I IFN production is a central event (Fig. 5). The initiation phase is mediated by early apoptotic

Table 1 Evidence for the in vivo role of TLRs in mouse models of lupusGenotype Model* Autoantibodies GN Mortality Reference

Tlr9–/– Mixed MRL-Faslpr ↓ dsDNA-specific

↑ Sm-specific

↔ ND 81

C57BL/6 (Fcgr2b–/–, 56R transgene encoding polyspecific antibody)

↓ Switch in polyspecific IgG2a and IgG2b

↓ ↓ 82

MRL-Faslpr ↑ dsDNA-specific ↑ ↑ 83

MRL-Faslpr ↓ dsDNA-specific

↑ Sm-specific

↑ ↑ 86

C57BL/6-Faslpr ↑ dsDNA-specific

↑ Nucleolus-specific

↓ Nucleosome-specific

↑ ↔ 84

C57BL/6-Ali5 (gain-of-function mutation in Cg2) ↑ Nucleolus-specific ↑ ND 85

Tlr3–/– Mixed MRL-Faslpr ↔ ↔ ↔ 81

Myd88–/– Mixed MRL-Faslpr ↓ dsDNA-specific

↓ Sm-specific

ND ND 89

C57BL/6 (Fcgr2b–/–) ↓ Switch in polyspecific IgG2a and IgG2b

↓ ↓ 82

C57BL/6 (Fcgr2b–/–, 56R transgene encoding polyspecific antibody)

↓ Switch in polyspecific IgG2a and IgG2b

ND ND 82

Tlr7–/– MRL-Faslpr ↓ Sm-specific ↓ ND 86

C57BL/6 (564 transgene encoding nucleic acids-specific antibody)

↓ 564 nucleic acids–specific ND ND 90

Yaa

(Tlr7 duplication)

BXSB ↑ Diverse specificities ↑ ↑ 91,92

C57BL/6 (Fcgr2b–/–) ↑ Nucleolus-specific ↑ ↑ 91

C57BL/6 (Sle1)

↑ dsDNA-specific ↑ snRNP-specific ↑ Glomerulus-specific

↑

↑

92

*See full description of the model in text. ↓, decrease; ↑, increase; ↔, no effects; GN, glomerulonephritis; ND, not determined.

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

550 VOLUME 13 | NUMBER 5 | MAY 2007 NATURE MEDICINE

materials and associated nucleic acids that are taken up by specialized DC subsets, providing both the antigenic cargo and the stimulus for TLR-independent production of IFN-α/β. Once this process is ignited, IFN-α/β–induced DC activation leads to stimulation of previously qui-escent autoreactive T and B cells. IFN-α/β may also, either directly or indirectly through DC-produced B-cell trophic factors, further promote proliferation, maturation and survival of autoreactive B cells, includ-ing those stimulated by the combined engagement of BCRs and TLRs. Once autoantibodies are made and immune complexes are formed with nucleic acid–containing macromolecules derived from late apoptotic/necrotic cells, the TLR-dependent amplification phase is initiated, which encompasses FcγR-mediated uptake of such complexes by pDCs and DCs, increased IFN-α/β production and further B-cell stimulation. These combined effects lead to perpetuation of the autoimmune process and enhanced pathogenicity. Type II IFN (IFN-γ) may also be induced due to cross-talk with type I IFNs, particularly at the late amplification phase, and contribute to disease progression.

ConclusionsDemarcation of the immune system into innate and adaptive arms and recognition that endosomal TLRs and certain cytosolic receptors sense nucleic acids have shaped our views about how immune responses can sometimes be initiated. These developments have had a major impact on understanding the pathogenesis of lupus-associated systemic autoim-munity. Within this framework, TLR-dependent and TLR-independent induction of IFN-α/β by apoptotic materials and nucleic acids appears to be a central pathogenic event. Future challenges are to identify specific defects associated with IFN-α/β production and signaling, abnormali-ties in the generation, disposal and sensing of endogenous stimuli, and interactions between these effector cytokines and other disease-predis-posing genes. Nonetheless, blockade of IFN-α/β (and/or the intercon-nected IFN-γ) with antibodies and recombinant receptors, or use of DNA sequences and other means to inhibit TLR engagement37,95,96, appear to be promising approaches to treat lupus. Moreover, with few exceptions (for example, multiple sclerosis), such treatments are likely to be appli-cable to other autoimmune/inflammatory diseases, including Sjögren syndrome97, type I diabetes98, psoriasis99 and dermatomyositis100. The prediction is that antagonists for TLRs or IFNs (even with partial inhibi-tory activity) will be as beneficial as antagonists for TNF-α, although the spectrum of diseases may differ. Such therapeutic strategies, however, should be applied with considerable caution because of the potential for life-threatening innate and/or adaptive immune deficiencies or untoward TH2-mediated allergic responses (reviewed in ref. 101). Alternatively, as the two phase paradigm posits, removing or reducing sensing of endoge-nous macromolecular structures that initiate production of IFNs may be a novel therapeutic approach with the potential advantage of not com-promising the beneficial functions of TLRs and IFNs.

ACKNOWLEDGMENTSWe apologize that space limitations preclude citation of all original articles. We thank K. Occhipinti-Bender for editorial assistance. The work of the authors has been supported by the US National Institutes of Health.

COMPETING INTERESTS STATEMENTThe authors declare no competing financial interests.

Published online at http://www.nature.com/naturemedicineReprints and permissions information is available online at http://npg.nature.com/reprintsandpermissions

1. Kono, D.H., Baccala, R. & Theofilopoulos, A.N. Genes and genetics of murine lupus. in Systemic Lupus Erythematosus (ed. Lahita, R.G.) 225–263 (Academic Press, San Diego, 2004).

2. Goodnow, C.C., Sprent, J., Fazekas de St Groth, B. & Vinuesa, C.G. Cellular and genetic mechanisms of self tolerance and autoimmunity. Nature 435, 590–597 (2005).

3. Honda, K. & Taniguchi, T. IRFs: master regulators of signalling by Toll-like receptors and cytosolic pattern-recognition receptors. Nat. Rev. Immunol. 6, 644–658 (2006).

4. Platanias, L.C. Mechanisms of type-I- and type-II-interferon-mediated signalling. Nat. Rev. Immunol. 5, 375–386 (2005).

5. Liu, Y.J. IPC: professional type 1 interferon-producing cells and plasmacytoid dendritic cell precursors. Annu. Rev. Immunol. 23, 275–306 (2005).

6. Theofilopoulos, A.N., Baccala, R., Beutler, B. & Kono, D.H. Type I interferons (�/�) in immunity and autoimmunity. Annu. Rev. Immunol. 23, 307–335 (2005).

7. Banchereau, J. & Pascual, V. Type I interferon in systemic lupus erythematosus and other autoimmune diseases. Immunity 25, 383–392 (2006).

8. Zhuang, H. et al. Lupus-like disease and high interferon levels corresponding to trisomy of the type I interferon cluster on chromosome 9p. Arthritis Rheum. 54, 1573–1579 (2006).

9. Sigurdsson, S. et al. Polymorphisms in the tyrosine kinase 2 and interferon regulatory factor 5 genes are associated with systemic lupus erythematosus. Am. J. Hum. Genet. 76, 528–537 (2005).

10. Graham, R.R. et al. A common haplotype of interferon regulatory factor 5 (IRF5) regu-lates splicing and expression and is associated with increased risk of systemic lupus erythematosus. Nat. Genet. 38, 550–555 (2006).

11. Santiago-Raber, M.L. et al. Type-I interferon receptor deficiency reduces lupus-like dis-ease in NZB mice. J. Exp. Med. 197, 777–788 (2003).

12. Braun, D., Geraldes, P., Demengeot, J. Type I interferon controls the onset and severity of autoimmune manifestations in lpr mice. J. Autoimmun. 20, 15–25 (2003).

13. Hron, J.D. & Peng, S.L. Type I IFN protects against murine lupus. J. Immunol. 173, 2134–2142 (2004).

14. Schwarting, A. et al. Interferon-β: a therapeutic for autoimmune lupus in MRL-Faslpr mice. J. Am. Soc. Nephrol. 16, 3264–3272 (2005).

15. Baccala, R., Kono, D.H. & Theofilopoulos, A.N. Interferons as pathogenic effectors in autoimmunity. Immunol. Rev. 204, 9–26 (2005).

16. Zhu, J. et al. T cell hyperactivity in lupus as a consequence of hyperstimulatory antigen-presenting cells. J. Clin. Invest. 115, 1869–1878 (2005).

17. Chen, M. et al. Dendritic cell apoptosis in the maintenance of immune tolerance. Science 311, 1160–1164 (2006).

18. Le Bon, A. et al. Cutting edge: enhancement of antibody responses through direct stimula-tion of B and T cells by type I IFN. J. Immunol. 176, 2074–2078 (2006).

19. Lin, Q., Dong, C. & Cooper, M.D. Impairment of T and B cell development by treatment with a type I interferon. J. Exp. Med. 187, 79–87 (1998).

20. Kawai, T. & Akira, S. TLR signaling. Cell Death Differ. 13, 816–825 (2006).21. Beutler, B. et al. Genetic analysis of host resistance: Toll-like receptor signaling and

immunity at large. Annu. Rev. Immunol. 24, 353–389 (2006).22. Iwasaki, A. & Medzhitov, R. Toll-like receptor control of the adaptive immune responses.

Nat. Immunol. 5, 987–995 (2004).23. Fitzgerald, K.A. & Chen, Z.J. Sorting out Toll signals. Cell 125, 834–836 (2006).24. Liew, F.Y., Xu, D., Brint, E.K. & O’Neill, L.A. Negative regulation of toll-like receptor-

mediated immune responses. Nat. Rev. Immunol. 5, 446–458 (2005).25. O’Neill, L.A. DisSARMing Toll-like receptor signaling. Nat. Immunol. 7, 1023–1025

(2006).26. Meylan, E. & Tschopp, J. Toll-like receptors and RNA helicases: two parallel ways to

trigger antiviral responses. Mol. Cell 22, 561–569 (2006).27. Kato, H. et al. Differential roles of MDA5 and RIG-I helicases in the recognition of RNA

viruses. Nature 441, 101–105 (2006).28. Hornung, V. et al. 5′-Triphosphate RNA is the ligand for RIG-I. Science 314, 994–997

(2006).29. Pichlmair, A. et al. RIG-I-mediated antiviral responses to single-stranded RNA bearing

5′ phosphates. Science 314, 997–1001 (2006).30. Gitlin, L. et al. Essential role of mda-5 in type I IFN responses to polyriboinosinic:poly-

ribocytidylic acid and encephalomyocarditis picornavirus. Proc. Natl. Acad. Sci. USA 103, 8459–8464 (2006).

31. Yoneyama, M. et al. Shared and unique functions of the DExD/H-box helicases RIG-I, MDA5, and LGP2 in antiviral innate immunity. J. Immunol. 175, 2851–2858 (2005).

32. Rothenfusser, S. et al. The RNA helicase Lgp2 inhibits TLR-independent sensing of viral replication by retinoic acid-inducible gene-I. J. Immunol. 175, 5260–5268 (2005).

33. Stetson, D.B. & Medzhitov, R. Recognition of cytosolic DNA activates an IRF3-dependent innate immune response. Immunity 24, 93–103 (2006).

34. Marshak-Rothstein, A. Toll-like receptors in systemic autoimmune disease. Nat. Rev. Immunol. 6, 823–835 (2006).

35. Ronnblom, L., Eloranta, M.L. & Alm, G.V. The type I interferon system in systemic lupus erythematosus. Arthritis Rheum. 54, 408–420 (2006).

36. Vollmer, J. et al. Immune stimulation mediated by autoantigen binding sites within small nuclear RNAs involves Toll-like receptors 7 and 8. J. Exp. Med. 202, 1575–1585 (2005).

37. Barrat, F.J. et al. Nucleic acids of mammalian origin can act as endogenous ligands for Toll-like receptors and may promote systemic lupus erythematosus. J. Exp. Med. 202, 1131–1139 (2005).

38. Savarese, E. et al. U1 small nuclear ribonucleoprotein immune complexes induce type I interferon in plasmacytoid dendritic cells through TLR7. Blood 107, 3229–3234 (2006).

39. Martin, D.A. & Elkon, K.B. Autoantibodies make a U-turn: the toll hypothesis for auto-antibody specificity. J. Exp. Med. 202, 1465–1469 (2005).

40. Wagner, H. Endogenous TLR ligands and autoimmunity. Adv. Immunol. 91, 159–173 (2006).

41. Boule, M.W. et al. Toll-like receptor 9-dependent and -independent dendritic cell acti-vation by chromatin-immunoglobulin G complexes. J. Exp. Med. 199, 1631–1640 (2004).

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

NATURE MEDICINE VOLUME 13 | NUMBER 5 | MAY 2007 551

42. Means, T.K. et al. Human lupus autoantibody-DNA complexes activate DCs through cooperation of CD32 and TLR9. J. Clin. Invest. 115, 407–417 (2005).

43. Sano, H. & Morimoto, C. Dna isolated from DNA/anti-DNA antibody immune complexes in systemic lupus erythematosus is rich in guanine-cytosine content. J. Immunol. 128, 1341–1345 (1982).

44. Ishii, K.J. et al. A Toll-like receptor-independent antiviral response induced by double-stranded B-form DNA. Nat. Immunol. 7, 40–48 (2006).

45. Martin, D.A. & Elkon, K.B. Intracellular mammalian DNA stimulates myeloid dendritic cells to produce type I interferons predominantly through a toll-like receptor 9-inde-pendent pathway. Arthritis Rheum. 54, 951–962 (2006).

46. Decker, P., Singh-Jasuja, H., Haager, S., Kotter, I. & Rammensee, H.G. Nucleosome, the main autoantigen in systemic lupus erythematosus, induces direct dendritic cell activa-tion via a MyD88-independent pathway: consequences on inflammation. J. Immunol. 174, 3326–3334 (2005).

47. Cortez-Gonzalez, X. et al. TLR9-independent activation of B lymphocytes by bacterial DNA. DNA Cell Biol. 25, 253–261 (2006).

48. Sun, Q. et al. The specific and essential role of MAVS in antiviral innate immune responses. Immunity 24, 633–642 (2006).

49. Napirei, M. et al. Features of systemic lupus erythematosus in Dnase1-deficient mice. Nat. Genet. 25, 177–181 (2000).

50. Tsukumo, S. & Yasutomo, K. DNaseI in pathogenesis of systemic lupus erythematosus. Clin. Immunol. 113, 14–18 (2004).

51. Bodano, A., Amarelo, J., Gonzalez, A., Gomez-Reino, J.J. & Conde, C. Novel DNASE I mutations related to systemic lupus erythematosus. Arthritis Rheum. 50, 4070–4071 (2004).

52. Yoshida, H., Okabe, Y., Kawane, K., Fukuyama, H. & Nagata, S. Lethal anemia caused by interferon-� produced in mouse embryos carrying undigested DNA. Nat. Immunol. 6, 49–56 (2005).

53. Okabe, Y., Kawane, K., Akira, S., Taniguchi, T. & Nagata, S. Toll-like receptor-indepen-dent gene induction program activated by mammalian DNA escaped from apoptotic DNA degradation. J. Exp. Med. 202, 1333–1339 (2005).

54. Kawane, K. et al. Chronic polyarthritis caused by mammalian DNA that escapes from degradation in macrophages. Nature 443, 998–1002 (2006).

55. Janssen, E. et al. Efficient T cell activation via a Toll-Interleukin 1 Receptor-indepen-dent pathway. Immunity 24, 787–799 (2006).

56. Fritz, J.H., Ferrero, R.L., Philpott, D.J. & Girardin, S.E. Nod-like proteins in immunity, inflammation and disease. Nat. Immunol. 7, 1250–1257 (2006).

57. Andrade, F., Casciola-Rosen, L. & Rosen, A. Apoptosis in systemic lupus erythematosus. Clinical implications. Rheum. Dis. Clin. North Am. 26, 215–227, v (2000).

58. Gaipl, U.S. et al. Inefficient clearance of dying cells and autoreactivity. Curr. Top. Microbiol. Immunol. 305, 161–176 (2006).

59. Ogden, C.A. & Elkon, K.B. Role of complement and other innate immune mechanisms in the removal of apoptotic cells. Curr. Dir. Autoimmun. 9, 120–142 (2006).

60. Botto, M. et al. Homozygous C1q deficiency causes glomerulonephritis associated with multiple apoptotic bodies. Nat. Genet. 19, 56–59 (1998).

61. Lu, Q. & Lemke, G. Homeostatic regulation of the immune system by receptor tyrosine kinases of the Tyro 3 family. Science 293, 306–311 (2001).

62. Mevorach, D., Zhou, J.L., Song, X. & Elkon, K.B. Systemic exposure to irradiated apop-totic cells induces autoantibody production. J. Exp. Med. 188, 387–392 (1998).

63. Denny, M.F. et al. Accelerated macrophage apoptosis induces autoantibody formation and organ damage in systemic lupus erythematosus. J. Immunol. 176, 2095–2104 (2006).

64. Takemura, Y. et al. Adiponectin modulates inflammatory reactions via calreticulin receptor-dependent clearance of early apoptotic bodies. J. Clin. Invest. 117, 375–386 (2007).

65. Gallucci, S., Lolkema, M. & Matzinger, P. Natural adjuvants: endogenous activators of dendritic cells. Nat. Med. 5, 1249–1255 (1999).

66. Winau, F. et al. Apoptotic vesicles crossprime CD8 T cells and protect against tuber-culosis. Immunity 24, 105–117 (2006).

67. Albert, M.L., Sauter, B. & Bhardwaj, N. Dendritic cells acquire antigen from apoptotic cells and induce class I-restricted CTLs. Nature 392, 86–89 (1998).

68. Obeid, M. et al. Calreticulin exposure dictates the immunogenicity of cancer cell death. Nat. Med. 13, 54–61 (2007).

69. Waldner, H., Collins, M. & Kuchroo, V.K. Activation of antigen-presenting cells by microbial products breaks self tolerance and induces autoimmune disease. J. Clin. Invest. 113, 990–997 (2004).

70. Lang, K.S. et al. Toll-like receptor engagement converts T-cell autoreactivity into overt autoimmune disease. Nat. Med. 11, 138–145 (2005).

71. Lang, K.S. et al. Immunoprivileged status of the liver is controlled by Toll-like receptor

3 signaling. J. Clin. Invest. 116, 2456–2463 (2006).72. Eriksson, U. et al. Dendritic cell-induced autoimmune heart failure requires cooperation

between adaptive and innate immunity. Nat. Med. 9, 1484–1490 (2003).73. Krieg, A.M. CpG motifs in bacterial DNA and their immune effects. Annu. Rev.

Immunol. 20, 709–760 (2002).74. Seya, T. et al. Role of toll-like receptors and their adaptors in adjuvant immunotherapy

for cancer. Anticancer Res. 23, 4369–4376 (2003).75. Poole, B.D., Scofield, R.H., Harley, J.B. & James, J.A. Epstein-Barr virus and molecular

mimicry in systemic lupus erythematosus. Autoimmunity 39, 63–70 (2006).76. Pawar, R.D., Patole, P.S., Wornle, M. & Anders, H.J. Microbial nucleic acids pay a Toll

in kidney disease. Am. J. Physiol. Renal Physiol. 291, F509–F516 (2006).77. Mathian, A., Weinberg, A., Gallegos, M., Banchereau, J. & Koutouzov, S. IFN-� induces

early lethal lupus in preautoimmune (New Zealand Black x New Zealand White) F1 but not in BALB/c mice. J. Immunol. 174, 2499–2506 (2005).

78. Mohty, M. et al. IFN-� skews monocyte differentiation into Toll-like receptor 7-express-ing dendritic cells with potent functional activities. J. Immunol. 171, 3385–3393 (2003).

79. Bekeredjian-Ding, I.B. et al. Plasmacytoid dendritic cells control TLR7 sensitivity of naive B cells via type I IFN. J. Immunol. 174, 4043–4050 (2005).

80. Hang, L. et al. Induction of murine autoimmune disease by chronic polyclonal B cell activation. J. Exp. Med. 157, 874–883 (1983).

81. Christensen, S.R. et al. Toll-like receptor 9 controls anti-DNA autoantibody production in murine lupus. J. Exp. Med. 202, 321–331 (2005).

82. Ehlers, M., Fukuyama, H., McGaha, T.L., Aderem, A. & Ravetch, J.V. TLR9/MyD88 signaling is required for class switching to pathogenic IgG2a and 2b autoantibodies in SLE. J. Exp. Med. 203, 553–561 (2006).

83. Wu, X. & Peng, S.L. Toll-like receptor 9 signaling protects against murine lupus. Arthritis Rheum. 54, 336–342 (2006).

84. Lartigue, A. et al. Role of TLR9 in anti-nucleosome and anti-DNA antibody production in lpr mutation-induced murine lupus. J. Immunol. 177, 1349–1354 (2006).

85. Yu, P. et al. Toll-like receptor 9-independent aggravation of glomerulonephritis in a novel model of SLE. Int. Immunol. 18, 1211–1219 (2006).

86. Christensen, S.R. et al. Toll-like receptor 7 and TLR9 dictate autoantibody specific-ity and have opposing inflammatory and regulatory roles in a murine model of lupus. Immunity 25, 417–428 (2006).

87. Bagavant, H., Deshmukh, U.S., Gaskin, F. & Fu, S.M. Lupus glomerulonephritis revis-ited 2004: autoimmunity and end-organ damage. Scand. J. Immunol. 60, 52–63 (2004).

88. Avrameas, S. Natural autoantibodies: from ‘horror autotoxicus’ to ‘gnothi seauton’. Immunol. Today 12, 154–159 (1991).

89. Lau, C.M. et al. RNA-associated autoantigens activate B cells by combined B cell antigen receptor/Toll-like receptor 7 engagement. J. Exp. Med. 202, 1171–1177 (2005).

90. Berland, R. et al. Toll-like receptor 7-dependent loss of B cell tolerance in pathogenic autoantibody knockin mice. Immunity 25, 429–440 (2006).

91. Pisitkun, P. et al. Autoreactive B cell responses to RNA-related antigens due to TLR7 gene duplication. Science 312, 1669–1672 (2006).

92. Subramanian, S. et al. A Tlr7 translocation accelerates systemic autoimmunity in murine lupus. Proc. Natl. Acad. Sci. USA 103, 9970–9975 (2006).

93. Tabeta, K. et al. The Unc93b1 mutation 3d disrupts exogenous antigen presentation and signaling via Toll-like receptors 3, 7 and 9. Nat. Immunol. 7, 156–164 (2006).

94. Maldonado, M.A. et al. The role of environmental antigens in the spontaneous develop-ment of autoimmunity in MRL-lpr mice. J. Immunol. 162, 6322–6330 (1999).

95. Dong, L., Ito, S., Ishii, K.J. & Klinman, D.M. Suppressive oligodeoxynucleotides delay the onset of glomerulonephritis and prolong survival in lupus-prone NZB x NZW mice. Arthritis Rheum. 52, 651–658 (2005).

96. Rezaei, N. Therapeutic targeting of pattern-recognition receptors. Int. Immunopharmacol. 6, 863–869 (2006).

97. Nordmark, G., Alm, G.V. & Ronnblom, L. Mechanisms of disease: primary Sjogren’s syndrome and the type I interferon system. Nat. Clin. Pract. Rheumatol. 2, 262–269 (2006).

98. Stewart, T.A. Neutralizing interferon � as a therapeutic approach to autoimmune dis-eases. Cytokine Growth Factor Rev. 14, 139–154 (2003).

99. Nestle, F.O. et al. Plasmacytoid predendritic cells initiate psoriasis through interferon-� production. J. Exp. Med. 202, 135–143 (2005).

100. Wenzel, J. et al. Type I interferon-associated skin recruitment of CXCR3+ lymphocytes in dermatomyositis. Clin. Exp. Dermatol. 31, 576–582 (2006).

101. Ishii, K.J., Uematsu, S. & Akira, S. ‘Toll’ gates for future immunotherapy. Curr. Pharm. Des. 12, 4135–4142 (2006).

R E V I E W©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e