A

mpptDimvsrows©

K

1

attecIimbcids

1d

Joint Bone Spine 76 (2009) 464–473

Review

Interferons and autoimmune disorders

Olivier MeyerRheumatology Department, Bichat Teaching Hospital, 46, rue Henri-Huchard, 75018 Paris, France

Accepted 31 March 2009Available online 20 September 2009

bstract

Interferons are ubiquitous cytokines produced by all mononuclear cell types in response to infection by a DNA or RNA virus. There are threeajor classes of interferons: type I or nonimmune interferons consist chiefly of interferons alpha produced by leukocytes and of interferon beta

roduced by fibroblasts, although there are several other less important variants; type II or immune interferon is interferon gamma, which is mainlyroduced by NK cells and T cells; and type III consists of the lambda interferons. Each type is characterized by a specific receptor and signalransduction pathway. Toll-like receptors (TLRs) on the cell membrane and endosomes recognize viruses and other microorganisms. Binding ofNA or RNA to endosomal TLRs generates a signal whose transduction pathways lead to molecules capable of binding to genes for various

nterferons, interleukin-1, and TNF�. Interferons can stimulate or inhibit up to 300 different genes encoding proteins involved in antiviral defenseechanisms, inflammation, adaptive immunity, angiogenesis, and other processes. The properties of interferons are used to treat a number of

iral infections (e.g., hepatitis B and hepatitis C), inflammatory diseases (interferon beta for multiple sclerosis and interferon gamma for systemicclerosis), and malignancies. Overactivation of the interferon pathways has been demonstrated in patients with systemic lupus erythematosus. Theesult is a characteristic pattern of mRNA expression known as the interferon signature. Interferon overactivation is related to inadequate clearance

f apoptotic particles with accumulation of apoptosis products (DNA-CpG motifs and U-RNA). Similar abnormalities have been found in patientsith primary Sjögren’s syndrome, systemic sclerosis, and polymyositis, as well as in some cases of rheumatoid arthritis. Immunomodulation

trategies designed to decrease interferon overactivity are being evaluated in patients with systemic lupus erythematosus.2009 Published by Elsevier Masson SAS on behalf of the Société Française de Rhumatologie.

ty; Au

tdIte

2

Mf

eywords: Interferon alpha; Interferon beta; Interferon gamma; Innate immuni

. Introduction

Interferon (IFN) was first identified in 1957. A protective rolegainst RNA viruses was inferred from the ability of viral infec-ions to induce IFN production [1]. Subsequently, IFN was foundo decrease tumor growth, inflammation, and angiogenesis. Thexistence of many different IFNs was established. IFNs are nowonsidered to play key roles in both innate immunity (IFN� andFN�) and adaptive immunity (IFN�). The IFN pathways arencreasingly well understood. These pathways are activated inany situations such as defense mechanisms against viral and

acterial infections, solid tumors, and hematological malignan-ies. They are also involved in several autoimmune diseases

ncluding systemic lupus erythematosus (SLE), Sjögren’s syn-rome, adult-onset rheumatoid arthritis (RA), polymyositis, andystemic sclerosis [2]. Insights gained into the effects of IFNs,

E-mail address: olivier.meyer@bch.aphp.fr.

2

diIa

297-319X/$ – see front matter © 2009 Published by Elsevier Masson SAS on behaloi:10.1016/j.jbspin.2009.03.012

toimmune disease; Interferon signature

ogether with the ability to clone the main IFNs, have led to theevelopment of new treatments designed either to support theFN pathways (hepatitides C and B, adjuvant cancer treatment,reatment of some forms of multiple sclerosis) or to block theffects of IFNs (SLE).

. The main interferons and their cellular sources

IFNs are a family of proteins that share similar properties.any IFNs also have similar amino acid compositions. IFNs

all into three main classes (Table 1).

.1. Type I IFNs

Type I IFNs in humans consist of 13 IFN� molecules pro-

uced by leukocytes; IFN� produced by fibroblasts and lessmportantly, IFN�, IFN�, and IFN� molecules. These type IFNs are encoded by 17 nonallelic genes that lack introns andre located on chromosome 9 in humans. Type I IFNs are

f of the Société Française de Rhumatologie.

O. Meyer / Joint Bone Spine 76 (2009) 464–473 465

Table 1The main interferons.

IFN type Class Genes (functional) Receptor Main source cells

I �

�

�, �, �*

17 (13)1

IFNAR1IFNAR2

pDCsFibroblasts and pDCs

II (immune) � 1 IFNGR1IFNGR2

Activated T cellsNK cellsNK T cellsMacrophagesmDCs

III � 3 IFNLR1IL10R2

pDCs

*INF only in swine; INF only in ruminants.pDCs: plasmacytoid dendritic cells; mDCs: macrophage-derived dendritic cells; NK:

Fig. 1. Toll-like receptors (TLRs). TLRs are membrane-spanning proteinslocated either on the plasma membrane (TLRs 1, 2, 4, 5, 6, and 10) or onthe endosomal membrane (TLRs 3, 7, 8, and 9). Naturally occurring TLR lig-ands are found in microorganisms (bacteria, parasites, RNA viruses, and DNAviruses). Stimulation of TLRs 3, 4, 7, and 9 leads to the production of IFN�/�.TLR: Toll-like receptor; MyD88: myeloid differentiation primary-response gene88; TIRAP: TIR-domain containing adapter protein; TRIF: TIRAP-inducingITe

gsspl(cpsf

gacaTo6opto a ubiquitous surface receptor composed of two membrane-spanning proteins IFN�R1 and IFN�R2 that form a ternarycomplex with the ligand (Fig. 2).

FN�; TRAM: TRIF-related adapter molecule; LPS: lipopolysaccharide; TIR:oll/IL-1R; membrane cellulaire: plasma membrane; séquences répétées richesn leucine: leucine-rich repeats; domaine TIR: TIR domain.

lycosylated proteins containing 160 to 200 amino acids andharing 30% to 55% homology. In humans, only IFN� and IFN�eem to exhibit tissue specificity. The other type I IFNs can beroduced by all cell types (and not only by leukocytes and fibrob-asts). After appropriate stimulation via the toll-like receptorTLR) TLR3 or TLR4, dendritic cells originating from mono-

ytes chiefly produce IFN�1 and IFN� (Fig. 1). IFN� is mainlyroduced by immature plasmacytoid dendritic cells (pDCs) thatelectively express TLR7 and TLR9, two endosomal receptorsor single-strand RNA and hypomethylated DNA (at cytosine

FItc

natural killer.

uanine [CpG] dinucleotide sites). These cells differ from otherntigen-presenting cells, such as macrophage-derived dendriticells (mDCs), which preferentially express TLR1, TLR2, TLR3,nd TLR8; or monocytes, which express TLR1, TLR2, TLR4,LR5, and TLR8. IFN� production by pDCs is rapid, with 60%f the newly induced transcriptome genes being activated withinh after stimulation. Within 24 h, each pDC produces 3 to 10 pgf IFN�, which is 200 to 1000 times more than the amountroduced by any other circulating cell type. Type I IFNs bind

ig. 2. Receptors of three types of interferons and signal transduction proteins.SRE: IFN-stimulated response element; STAT: signal transducers and activa-ors of transcription; ISGF3: IFN-stimulated gene factor 3, STAT1-STAT2-IRF9omplex; GAS: IFN�-activated site; noyau: nucleus; cytoplasme: cytosol.

4 Spine 76 (2009) 464–473

2

wpIiaca

2

Ibrs

kIttrcpip

3

cpl(siOwoncerctd

p(uamowp

Fig. 3. The two interferon activation pathways: viral RNA induces activation ofendosomal TLR3 and of the rig-like-helicases (RLHs) RIG-1 and MDA5. RIP1:receptor-interacting protein 1.DD: death-domain; TIR: Toll/IL-1 receptor; TM: transmembrane region; RIG1:retinoic-acid-inducible gene I; MDA5: melanoma differentiation-associatedantigen 5; CARD: caspase recruitment domain TRIF: Toll/IL-1R (TIR) domain– containing adapter protein inducing IFN�; TBK1: (tank-binding kinase 1);Ian

rt[

dp

bS

4

iamgai

66 O. Meyer / Joint Bone

.2. Type II IFN

There is only one type II IFN, called immune IFN or IFN�,hich is produced by NK cells and T cells. The protein is com-osed of 140 amino acids and shares no homology with type IFNs. The IFN� receptor is distinct from the type I IFN receptors:t is a heterodimer of two membrane-spanning proteins, IFN�R1nd IFN�R2, of which the first is constitutively expressed on allell types and the second is tightly regulated and expressed onnarrower spectrum of cell types (Fig. 2).

.3. Type III IFNs

Type III IFNs consist of three IFN� molecules also known asL-28A, IL-28B, and IL-29. They are co-produced with IFN�ut act by binding to a different receptor from the type I IFNeceptors. The IFN� receptor is composed of two membrane-panning proteins, IFNLR1 and IL-10R2 (Fig. 2).

Type I and III IFN receptors are associated with the Janusinase JAK1 and the tyrosine kinase TYK2, whereas type IIFN receptors are associated with JAK1 and JAK2. Transduc-ion of the IFN signal requires a cascade of phosphorylations ofransduction proteins such as STAT1 and STAT2 and involvesegulatory proteins such as IFN regulatory factor 9 (IRF9), whichan bind to the IFN-stimulated response element (ISRE) at theromoter sites of IFN-responsive genes [3]. A complex regulat-ng system whose description would be beyond the scope of thisaper contributes to inhibit or to enhance IFN-generated signals.

. Inducers of interferon production

Viruses, a number of bacteria, and bacterial lipopolysac-haride (LPS) can induce the production of type I IFNs. Theroduction of immune IFN (IFN�) is stimulated by severalectins that are mitogenic for T cells, such as phytohemagglutininPHA). At the molecular level, double-stranded RNAs (or theynthetic double-stranded polyribonucleotide poly(I) poly(C))nduces IFN production by binding to TLR3 in the endosomes.ther potent IFN inducers include single-stranded viral RNA,hich binds to TLR7 and TLR8 and hypomethylated bacterialr viral DNA, which binds to TLR9. LPS binds to TLR4. Aumber of viral glycoproteins and other bacterial componentsan induce IFN production via mechanisms that remain to belucidated (Fig. 1). In addition to the endosomal TLRs, whichespond to viral RNAs and DNAs [4], other cytosolic receptorsan bind viral RNAs, thereby inducing the production of IFNs;hese receptors are helicases containing a caspase recruitmentomain (CARD) and known as RIG-1 and MDA5 [5] (Fig. 3).

Signal transduction from the TLRs requires either junctionroteins such as MyD88 and TRAF6 (TLR5/7/8/9) or TRIFTLR3/4) and transcription factors belonging to the IFN reg-latory factor (IRF) family such as ITRF3 for TLR3/4; RIG-1nd MDA-5; IRF7 for TLR7/8/9 [6]; and IRF1 for TLR9 in

acrophages and mDCs. Several studies point to involvement

f specific allelic IRF5 variants in SLE [7]. IRF5, togetherith NFkB, participates in the transcription of genes encodingroinflammatory cytokines (TNF�, IL-6, and IL-12p40). IRF5

faut

KK�: inhibitor of NK-kB kinase E; TRAF: tumor necrosis factor receptor-ssociated factor; NS3-4A: hepatitis C virus protease; hélicase: helicase; noyau:ucleus.

ecruitment via MyD88 activation follows a pathway parallel tohe IRF1 pathway on mDCs and to the IRF7 pathway on pDCs8].

The RIG-1 and MDA-5 pathway seems crucial to IFN pro-uction by fibroblasts and by macrophages and mDCs. The TLRathways are involved in IFN production by pDCs [9].

Several compounds that inhibit the TLR pathways responsi-le for type I IFN production have been identified and designatedT2, SIGIRR, SARM, and SGLS [6,10].

. Biological effects of interferons

IFNs induce the expression of hundreds of genes involvedn many biological functions [11] such as defense mechanismsgainst viruses and bacteria, apoptosis, the cell cycle, inflam-ation, innate immunity, and adaptive immunity. Some of these

enes are regulated by both type I and type II IFNs and others bysingle IFN type. For instance, the IRF1 gene is preferentially

nduced by IFN�, whereas the gene encoding hypoxia-inducible

actor 1 (HIF-1) is selectively induced by IFN�. Transcriptomenalysis by mRNA amplification on hybridization chips can besed to determine the profile of gene activation in response tohe IFNs produced by circulating mononuclear cells in various

Spin

d[appf

4

IrrwmicTte

4

4

i

•

••

•

•

•

•

4

iNl

T1IettapaIHagctredB

4

lIfg

5

eSgs

5

5

wCta[a6Ctcaf

O. Meyer / Joint Bone

isease states. This profile is known as the IFN signature12,13]. The IFN signature includes genes encoding cytokinesnd chemokines, membrane receptors, signal transductionroteins, growth factors or apoptosis factors, antimicrobialroteins, adhesion molecules, and other compounds responsibleor the biological effects of IFNs.

.1. Antiviral and pro-apoptotic effects

Induction of antiviral effectors was the first known effect ofFNs. At least four pathways by which IFNs mediate the antivi-al response have been identified: the MxGTPase pathway, theibonuclease l-regulating 2,5′-oligoadenylate-synthase path-ay, the protein kinase A pathway, and the ISG15 ubiquitin-likeodifier pathway. An explanation of the complex mechanisms

nvolved would be far beyond the scope of this update [14]. IFNsan activate an apoptosis pathway involving the production ofNF-related apoptosis-inducing ligand (TRAIL). This ability

o induce apoptosis contributes not only to the anti-microbialffects of IFNs [15], but also to their anti-tumor effects [16].

.2. Immunomodulating effects of interferons

.2.1. IFNα/βThe multiple direct and indirect effects of IFN�/� on adaptive

mmunity are listed below:

DC activation with increased expression of Class I majorhistocompatibility complex (MHC) molecules and of co-activation molecules such as CD40, CD80, and CD86.IFN�/� induce the production of various chemokines includ-ing CXCL8 (IL-8), CXCL9 (MIG), CXCL10 (IP10), andCXCL11 (I-TAC), and of their receptors, thereby influencingDC targeting in the peripheral lymphoid organs and T-cellactivation via production by mDCs of the cytokines IL-12,IL-15, IL-18, and IL-23;enhanced NK cell activity;stimulation of macrophage development and activationand stimulation of inducible NO synthase expression bymacrophages;enhanced T-cell proliferation, via a direct mechanism orvia IL-15 induction by antigen-presenting cells, which con-tributes to Th1 cell maturation;pro-apoptotic and anti-proliferating effects on T cells(increased expression of apoptotic molecules and activationof pro-caspases);enhanced B-cell proliferation via increased expression of thecytokines BAFF (Blys) and APRIL by DCs and plasmablastdifferentiation to plasma cells; enhanced immunoglobulinclass switching;anti-angiogenic and anti-proliferating effects [17].

.2.2. INFγ

IFN� is produced by pDCs and mDCs during the innatemmune response, under the influence of IFN�/� and IL-12.K cells, NK T cells, and cytotoxic Th1 cells that are stimu-

ated via their respective activating receptors can produce IFN�.

wpDI

e 76 (2009) 464–473 467

h1 cell maturation, under the influence of IFN�, IFN�/�, IL-2, and other cytokines such as IL-27 and IL-23, increasesFN� production by activating STAT4. IFN� stimulates TLR4xpression on immature DCs, thereby amplifying their responseo LPS; exerts pro-apoptotic effects, restoring the balance ofhe T-cell population expanded by the antigenic stimulation;nd induces the expression of T-bet, a transcription factor thatromotes immunoglobulin class switching to IgG2a (in mice)nd antagonizes the effect of IL-4 on switching to IgG1 andgE. Furthermore, IFN� increases the expression of Class IILA molecules and promotes the ThCD4+ response. IFN�

lso increases the expression of Class I HLA molecules, anti-en priming, and antigen presentation. Other effects of IFN�onsist of effector cell recruitment at sites of inflammation viahe production by macrophages of chemokines and chemokineeceptors, stimulation of the macrophage oxidative burst vianhanced expression of the enzymes iNOS and NADPH oxi-ase [1,18], and potentiation of TGF�1-induced stimulation ofAFF expression by macrophages [19].

.3. Anti-angiogenic effects of interferons

INF� is responsible for the decrease in circulating endothe-ial progenitor cells seen in patients with SLE. Another effect ofNF� in SLE is impairment of the ability of myeloid cells to dif-erentiate into angiogenic cells producing vascular endothelialrowth factor (VEGF) and hepatic growth factor (HGF) [20].

. Interferons and connective tissue diseases

The first evidence that IFNs were involved in the pathogen-sis of connective tissue diseases came from studies of SLE.ubsequently, IFNs were found to play a role in primary Sjö-ren’s syndrome, polymyositis, some forms of RA, and systemicclerosis.

.1. Systemic lupus erythematosus

.1.1. Main arguments implicating IFNα

In 1979, high serum levels of IFN� were reported in patientsith active SLE, and an acid-labile form of IFN� was identified.irculating IFN� levels correlate with disease activity parame-

ers such as the SLEDAI, number of organs involved, titer ofnti-dsDNA antibodies, and degree of hypocomplementemia21]. Furthermore, patients treated with IFN� often produceutoantibodies, most notably antinuclear antibodies (30% to0% of patients) and anti-dsDNA antibodies (10% of patients).linical lupus develops in about 1% of these patients. Although

hese rates of occurrence are lower than those reported for clini-al thyroiditis and antithyroid antibodies (anti-TPO in 30% andnti-TG in 15% of patients), they confirm the propensity of IFN�or inducing autoimmunity.

The mechanism leading to IFN� overproduction in SLE

as elucidated at the turn of the century. The serum of SLEatients was found to contain immune complexes composed ofNA or RNA capable of inducing the production of IFN� [18].

mmature pDCs were shown to produce large amounts of IFN�

468 O. Meyer / Joint Bone Spine 76 (2009) 464–473

Fig. 4. Model for sensitization to endogenous DNA: endogenous DNA bindsto the antimicrobial peptide LL37 and to the nuclear protein HMGB1, and theresulting heteropolymer is routed to the endosomes.ITAM: immunoreceptor tyrosine-based activation motif; HMGB1: high-mobility group box1 protein; RAGE: receptor of advanced glycationeb1

w[

5

psgPsnTtItmioIcauo

eh

Fig. 5. Routing of DNA and RNA autoantigens to the endosomal TLR7 andTrR

EgsbkcpnDtoeaFrpc

ndproducts (receptor for the HMGB1 protein); LL37: 37-amino acid antimicro-ial peptide LL-37 derived from hCAP18 (human cationic antimicrobial protein8 kD).

ithin tissues. IFN� can induce monocyte maturation into DCs22].

.1.2. IFNα and plasmacytoid dendritic cells (pDCs)IFN�-producing DCs in the bloodstream are decreased in

atients with active SLE, probably because they migrate to tissueites of inflammation. The mRNAs encoded by IFN-responsiveenes can be analyzed using cDNA chips with quantitative RT-CR. This technique reveals a specific gene profile, or IFNignature, typical for active SLE with hematological, renal, oreurological involvement [23], particularly in pediatric patients.he results show not only that some genes are activated, but also

hat other genes are suppressed [12,13,24]. They indicate thatFN�/�, produced chiefly by immature pDCs, is responsible forhe increase in autoantibody production during SLE flares. The

echanisms that lead to IFN overproduction probably involve anncrease in apoptosis products with the release of large amountsf endogenous molecules capable of stimulating TLR3/4/7/8/9.mpaired clearance of apoptosis products (due to a deficiency inirculating DNAse and to absence of CRP and other pentraxins)nd partial or complete deficiency in C4, C2, or C1q (which sol-bilize the immune complexes) contribute to elevate the levels

f DNA and of endogenous ribonucleoproteins.

In healthy individuals, endogenous DNA does not enter thendosomal compartment, which contains TLRs responsive toypomethylated bacterial or viral DNA (DNA CpG) (Fig. 4).

p

ew

LR9 receptors in pDCs (6A, 6B) and B cells (6C, 6D, 6E) via Fc�R surfaceeceptors that bind immune complexes or the B-cell receptor (BCR) to DNA orNA or IgG (rheumatoid factor).

ndogenous DNA contains hypomethylated sequences analo-ous to those found in bacterial and viral DNA and can thereforetimulate specific TLRs. Furthermore, endogenous DNA canind to proteins such as the antimicrobial peptide LL37 (alsonown as CAMP), which is released by damaged epithelialells or secreted by neutrophils as protein hCAP18 [25]. LL37rotects DNA from extracellular degradation and allows endoge-ous DNA to enter the early endosomal compartment. LL37,NA (isolated or bound to anti-DNA antibodies), and the pro-

ein HMGB1 form a complex that is taken up by the endosomesf pDCs [26] (Fig. 5). Free DNA bound to LL37 enters thendosomes via membrane lipid rafts, whereas DNA bound tonti-DNA IgG combined with LL37 and HMGB1 enters via thec�RIIA receptor [27]. In the endosome, HMGB1 binds to theeceptor for advanced glycation endproducts (RAGE), therebyromoting persistence of the DNA within the early endosomalompartment. Hypomethylated DNA binds to TLR9, leading to

rolonged activation of this receptor.

Endogenous RNAs can stimulate endosomal TLR7 if theyscape extracellular degradation by RNAses. This is the caseith uridine- or uridine/guanine-rich RNAs such as uridine-rich

O. Meyer / Joint Bone Spin

Fs

ssomtaBXcilt

u(otBiop

5

iRtTisToaFI

cssis

5l

aTatirOltpmotcpTnoavoppl

5

5

nta5llRIfs

pwms

ig. 6. Two-stage model for the pathophysiology of systemic lupus erythemato-us.

mall nuclear ribonucleoproteins (U-snRNPs) [28] (Fig. 6). U-nRNPs enter the endosomal compartment either via liposomesr as antigen-antibody immune complexes. B cells specific forotifs in U-snRNPs undergo activation after the antigen binds

o the B-cell receptor (BCR), generating a signal that is potenti-ted by RNA-induced stimulation of TLR7. In the autoimmunexSB mouse carrying the Yaa mutation, the tlr7 gene on thechromosome is duplicated and the X-chromosome fragment

arrying tlr7 is translocated to the Y chromosome. The results TLR7 overexpression, which may explain the high rate of aupus-like disease in the males [29]. Other genes located nearlr7 may contribute to the lupus phenotype [30].

Activation of pDCs leading to IFN� overproduction is reg-lated by various receptors, including blood DC antigen 2BDCA2), a C-type lectin receptor expressed at the surfacef pDCs. Stimulation of BDCA2 activates the ITAM pathway,hereby slowing IFN� production by pDCs [27]. The level ofDCA2 expression on circulating mononuclear cells is lower

n patients with SLE than in healthy controls [31]. Stimulationf these receptors to decrease IFN� production may hold thera-eutic potential in SLE and other connective tissue diseases.

.1.3. IFNα and other antigen-presenting cellsIFN� production can be induced by apoptotic products,

ndependently from TLR stimulation. Cytosolic 5′-triphosphateNA binds to and activates the helicases RIG-1 and MDR5,

wo members of the NOD-like receptor (NLR) family (Fig. 3).his mechanism, which occurs in mDCs and macrophages,

ncreases the production of IFN�. Double-stranded B-DNAtimulates the production of inflammatory cytokines via aLR-independent mechanism and also enhances the production

f co-stimulation molecules on B cells. These effects potentiatentigen presentation. Programmed cell death acting through theas system via a TLR-independent mechanism induces type IFN production and stimulates T cells specific of the antigens

gltt

e 76 (2009) 464–473 469

ontained in apoptotic particles. The DCs that respond to thistimulation carry the immature lymphoid markers B290−,DCA−, CD8− and, after stimulation, CD8+. The mechanisms believed to involve non-TLR nucleic acid-sensitive receptorsuch as the NLRs [9,32] (Fig. 3).

.1.4. Two-stage model of IFN α/β induction in systemicupus erythematosus

The first stage may be due to the release of large amounts ofpoptotic material containing nucleic acids (DNA and RNPs).he cause of the cell death may be a microorganism, a physicalgent (e.g., ultraviolet radiation), or another source of cell stresshat activates the Fas/FasL system. Early apoptotic material isnternalized by specialized DCs that transport antigenic mate-ial and produce IFN�/� via a TLR-independent mechanism.nce this process is initiated, DCs activated by IFN�/� stimu-

ate autoreactive T and B cells. IFN�/� can, either directly or viahe production of B-cell trophic factors (e.g., BAFF), induce theroliferation, maturation, and survival of autoreactive B cells,ost notably those stimulated by the simultaneous engagement

f their BCR and TLRs (Fig. 5C, D, and E). The first autoan-ibodies to be produced bind to the antigen to form immuneomplexes containing nucleoproteins released during the latehases of apoptosis (or necrosis), thereby initiating the second,LR-dependent amplification stage. This stage involves inter-alization of immune complexes via the Fc�R of pDCs andther DCs, together with increased IFN �/� production, as wells increased B-cell stimulation. The result is a self-perpetuatingicious cycle in which the immune complexes formed becausef excessive apoptosis (late apoptosis) maintain a continuousroduction of IFN�/� (Fig. 6) also produced as a result of aositive interaction with type I IFN�/�, most notably during theate amplification phase.

.2. Other connective tissue diseases

.2.1. Primary Sjögren’s syndromePrimary Sjögren’s syndrome (pSS) shares with SLE a

umber of features including the female predominance, sys-emic manifestations, hyperimmunoglobulinemia, production ofntinuclear autoantibodies (most notably anti-snRNPs such as2 kD anti-SSA/Ro), lymphoid hyperplasia in target organs witharge numbers of autoreactive B cells producing autoantibodiesocally, and elevation of BAFF (Blys) cytokine levels (Fig. 7).ecent research has established a major role for the IFNs and

FN pathways in pSS. Many similarities with SLE have beenound, and pSS has been described as localized lupus of thealivary glands.

Genetic studies have established a role for several IFN-athway genes. For instance, associations have been reportedith polymorphisms of the IRF5 gene [33], whose transcriptay stimulate the production of IFN�. Immunohistochemistry

tudies showed that pDCs were present within the salivary

lands of patients with pSS, [34] and mRNA analyses indicatedocal IFN� production. Circulating IFN� levels were consis-ently elevated in some studies but not in others, in contrasto findings in SLE. In one study, circulating IFN� levels were

470 O. Meyer / Joint Bone Spin

hpblotca

partltfiCvsgCBulcldp

oEiht[ipb

pfsvptmarttaulutSivt

5

afialoosfma

tItIs

fmcimotwp

5

Fig. 7. Pathophysiological model for primary Sjögren’s syndrome.

igh and increased further after etanercept therapy [35]. IFN�roduction, occurring chiefly at the local level, is stimulatedy material released from apoptotic endothelial cells and byocally formed immune complexes (RNP complexes composedf Ro52 kD/SSA) via some of the TLRs and the Fc�RIIA recep-ors of pDCs. Patients with pSS do not have serum immuneomplexes of DNA and anti-DNA, and only anti-SSA (± SSB)ntibodies induce IFN�.

Studies of the salivary gland transcriptome in patients withSS showed that many mRNAs produced in larger or smallermounts than in controls encoded proteins whose genes wereegulated by IFN�/�. Thus, the inflammatory salivary glandissue has an IFN signature similar to the one found in circu-ating mononuclear cells from SLE patients [36]. Studies ofhe transcriptome of circulating monocytes produced similarndings [37]. The gene encoding CXCL10 is overexpressed.XCL10 is a chemokine whose receptor is found on acti-ated T cells, of which a large number are present in thealivary glands of patients with pSS. Another overexpressedene encodes CXCL12 (SDF-1), a chemokine that binds to theXCR3 receptor on IFN�-producing pDCs and to CXCR4 on-cell progenitors. When salivary-gland epithelial cells are stim-lated in vitro by an RNA virus or TLR3 ligands, they producearge amounts of BAFF mRNA and protein. This BAFF increasean be partially blocked by a type I IFN inhibitor, suggesting ateast a contributory role for the TLR/IFN pathway in the pro-uction of BAFF, which plays a key role in B-cell proliferation,lasma cell differentiation, and in situ antibody production [38].

The production of apoptotic particles releases large amountsf snRNPs, including those carrying the Ro52-kD protein, an3 ligase with antiproliferative and pro-apoptotic effects. IFN�

ncreases the production of Ro52-kD. Cellular biology studiesave established that Ro52-kD translocation from the cytosolo the nucleus under the influence of IFN� precedes cell death

39]. Ro52-kD contributes to stimulate macrophage-mediatednnate immune responses [40]. Patients with pSS and Raynaud’shenomenon have larger numbers of IFN�-producing peripherallood mononuclear cells than do controls [41].

osdh

e 76 (2009) 464–473

Taken together, the available experimental data suggest aathophysiological model for pSS similar to that put forwardor SLE. Under the influence of an environmental factor thattimulates innate immunity; for instance a virus (Epstein-Barrirus?) colonizing the salivary-gland epithelial cells, these cellsroduce type I IFNs and express chemokines and cytokineshat attract inflammatory cells, particularly pDCs carrying the

atching chemokine receptors. Epithelial cells, which becomepoptotic when stimulated by IFN�, release large amounts ofibonucleic material (SSA snRNPs), which stimulate the adap-ive response of T and B cells. The T-cell response results inhe release of IFN�, which further amplifies the IFN responsend the autoreactive B-cell response. Autoreactive B cells, stim-lated by BAFF produced by antigen-presenting cells (mDCs),ocally produce anti-SSA autoantibodies, as well as type I IFNsnder the influence of simultaneous TLR and Fc�RIIa stimula-ion. The result is a further increase in IFN� production by pDCs.SA-Ro52-kD/anti-SSA immune complexes act as endogenous

nducers of IFN� and IL-12 (which induces IFN� productionia STAT4), thereby initiating a vicious circle that perpetuateshe inflammation.

.2.2. Systemic sclerosisType I and II IFNs inhibit collagen production both in vivo

nd in vitro when they are added to normal or sclerodermabroblasts. This effect prompted a number of therapeutic tri-ls in patients with diffuse systemic sclerosis, which met withittle success. Oddly enough, there have been several reportsf systemic sclerosis induced by IFN� therapy for hepatitis Cr myeloproliferative diseases or by IFN� therapy for multipleclerosis. A study of the transcriptome of circulating leukocytesrom patients with systemic sclerosis showed amplification ofRNAs for a few genes involved in the IFN pathway [42],

lthough the IFN signature was less typical than in SLE.Serum from patients with systemic sclerosis and anti-

opoisomerase I (Scl70) antibodies induces a higher level ofFN� production by normal peripheral blood mononuclear cellshan does serum from patients with anticentromere antibody.FN� production is higher in patients with diffuse systemicclerosis and in those with interstitial lung disease [43].

Among the genes induced by IFN, IFI16 encodes a proteinound in large amounts in the epidermis and inflammatory der-is of scleroderma lesions. IFI16 plays a role in endothelial

ell proliferation. An immunohistochemistry study showed annfiltrate of CD123+ pDCs in scleroderma-affected skin speci-

ens [44]. In a study of candidate genes, increased frequenciesf some IRF5 allelic variants were found in patients with sys-emic sclerosis, and these variants were significantly associatedith interstitial lung disease [45]. IRF5 encodes a transductionrotein involved in the pro-inflammatory cytokine pathway.

.2.3. DermatomyositisSeveral myxoviruses have been incriminated in the devel-

pment of dermatomyositis. Published data remain sparse. Atudy of the transcriptome of muscle cells from patients withermatomyositis showed an IFN�/� profile, and an immuno-istochemistry study detected the myxovirus resistance protein

Spin

AabtaslsRab(

5

sccesaawtT[srcT

5

aorImrimi

5m

iattcgoSImS

catiaastpu

5

mmtspsiaIhabeTdIls

iiitti

6

bmsmbmadTstudy.

O. Meyer / Joint Bone

(MxA) induced by IFN�/� in the perifascicular muscle fibersnd in some capillary cells. CD4+pDCs were found in large num-ers within the muscle [46]. A number of autoantigens that areargeted by anti-synthetase antibodies, specific of polymyositis,nd associated with interstitial lung disease (e.g., histidyl tRNAynthetase and asparaginyl tRNA synthetase) exert a chemokine-ike effect by binding to the CCR5 receptors of immature DCs,imilar to the HIV and to chemokines such as MIP1� (CCL3),ANTES (CCL5), and MCP2 (CCL8) [47]. Thus, selection ofn autoantigen as a target for the autoantibody response maye related to pro-inflammatory properties of the autoantigenincluding stimulation of IFN production).

.2.4. Rheumatoid arthritisGrowing evidence indicates that adult-onset RA is a

yndrome. In patients who produce rheumatoid factor and anti-itrullinated protein antibodies (ACPA or anti-CCP) and whoarry a susceptibility allele of the PTPN22 gene and the sharedpitope of the HLADRB1* alleles, the disease is often ero-ive and involves TNF� as the key inflammatory cytokine. Innother form of RA, there are no rheumatoid factors or erosionsnd the HLA phenotype is often DR3; this form is associatedith several polymorphisms of the IRF5, STAT4 genes encoding

ransduction proteins involved in IFN pathways [48,49] and ofRAF1, which encodes a negative regulator of TNF� signaling50]. This form of RA shares similarities with SLE and Sjögren’syndrome, including involvement of IFNs in the inflammatoryesponse. The two forms of RA correspond to the model thatontrasts two categories of autoimmune diseases, one driven byNF� and the other by IFN� [51].

.2.5. Other autoimmune diseasesIFN�/� exert stimulating or inhibitory effects in various

utoimmune diseases that fall outside the scope of rheumatol-gy. For instance, in type I insulin-dependent diabetes mellitus,eliable animal models indicate a deleterious role for type IFNs. IFNs are also harmful in multiple sclerosis and its ani-al model of acute encephalitis, and IFN� is used to treat

elapsing/remitting forms of multiple sclerosis. Other examplesnclude thyroiditis, some forms of autoimmune hemolytic ane-

ia, and some forms of uveitis such as Behcet’s disease uveitisn which IFN� is used to treat flares [18].

.2.6. IFN pathway genes and gene products – clinicalonitoringThe identification of an IFN signature by transcriptome stud-

es of peripheral blood mononuclear cells from patients withctive SLE (most notably with renal and neurological manifes-ations [12]) prompted interest in longitudinal changes affectinghe mRNA profile and the serum levels of proteins exhibitingytokine, chemokine, or cell growth-stimulating effects. Theoal is to identify new biomarkers that reliably reflect the coursef SLE. The IFN signature was evaluated in 81 patients with

LE, who were then divided into two groups based on whetherFN upregulation compared to controls was very marked or lessarked. IFN upregulation showed positive correlations with theLEDAI and SLAM-R. However, when the IFN profiles were

C

e 76 (2009) 464–473 471

ompared to the serum levels of chemokines encoded by IFN-ctivated genes, no significant difference was found between thewo patient groups. Possible explanations to this discordancenclude the short half-life of cytokine and chemokine mRNAsnd an extravascular location of the primary sites of cytokinend chemokine production. The SLEDAI and SLAM-R showedtronger correlations with the IFN-regulated chemokine pro-ein score than with the IFN-regulated mRNA score [52]. Thus,rotein biomarkers (in the proteome) may have greater clinicaltility than nucleic biomarkers (in the transcriptome).

.2.7. IFNs and therapeutic immunointerventionType I IFNs have met with considerable success as treat-

ents for various viral diseases (hepatitis C, hepatitis B, andacroscopic polyarteritis nodosa associated with the hepati-

is B virus), hematological malignancies (myeloproliferativeyndromes and multiple myeloma), relapsing/remitting multi-le sclerosis, and severe Behcet’s uveitis. Immunomodulationtrategies designed to block the effects of type I IFNs on thenflammatory response are at the early stages of development innimal models and phase II clinical trials. A monoclonal anti-FN� antibody that neutralizes the biological effects of IFN�as been used in a few patients with SLE [53]. Monoclonalntibodies that stimulate the BDCA2 and BDCA4 receptors,oth of which inhibit IFN�/� production by pDCs, are undervaluation [31]. Immunomodulation strategies target TLR7 orLR9 by using synthetic oligonucleotides to block the trans-uction of signals responsible for the production of type IFNs [54,55]. In New Zealand F1(BxW) mice, this approachessened the severity of the renal involvement and prolongedurvival.

Great caution is in order given the role for IFNs in fightingnfection. Thus, complete IFN blockade may lead to high rates ofnfection. Patients with genetic deficiencies in STAT1 (a proteinnvolved in the type I IFN pathway) have severe infections andhose with genetic or acquired deficiencies in the IFN� recep-or develop disseminated tuberculosis or atypical mycobacterialnfections [56].

. Conclusion

IFNs constitute a family of cytokines located at the interfaceetween the innate and adaptive immune systems. They exertultiple antiviral and pro-inflammatory effects and modulate

everal hundred genes whose products stimulate or inhibit theechanisms involved in inflammation. These mechanisms have

een well documented in clinical practice. In particular, inflam-ation can be induced by endogenous DNA or RNA, which

re targets for the autoimmune response in SLE and, to a lesseregree, in Sjögren’s syndrome and diffuse systemic sclerosis.reatments targeting IFN�/� or IFN-pathway proteins are under

onflicts of interest

None of the authors has any conflicts of interest to declare.

4 Spin

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

72 O. Meyer / Joint Bone

eferences

[1] Borden EC, Sen GC, Uze G, et al. Interferons at age 50: past, current andfuture impact on biomedicine. Nat Rev Drug Discov 2007;6:975–90.

[2] Vilcek J. Fifty years of interferon research: aiming at a moving target.Immunity 2006;25:343–8.

[3] Honda K, Takaoka A, Taniguchi T. Type I interferon [corrected] geneinduction by the interferon regulatory factor family of transcription factors.Immunity 2006;25:349–60.

[4] Takeshita F, Ishii KJ. Intracellular DNA sensors in immunity. Curr OpinImmunol 2008;20:383–8.

[5] Meylan E, Tschopp J, Karin M. Intracellular pattern recognition receptorsin the host response. Nature 2006;442:39–44.

[6] Roelofs MF, Abdollahi-Roodsaz S, Joosten LA, et al. The orchestra of Toll-like receptors and their potential role in frequently occurring rheumaticconditions. Arthritis Rheum 2008;58:338–48.

[7] Graham RR, Kozyrev SV, Baechler EC, et al. A common haplotype ofinterferon regulatory factor 5 (IRF5) regulates splicing and expression andis associated with increased risk of systemic lupus erythematosus. NatGenet 2006;38:550–5.

[8] Colonna M. TLR pathways and IFN-regulatory factors: to each its own.Eur J Immunol 2007;37:306–9.

[9] Baccala R, Hoebe K, Kono DH, et al. TLR-dependent and TLR-independent pathways of type I interferon induction in systemicautoimmunity. Nat Med 2007;13:543–51.

10] Liew FY, Xu D, Brint EK, et al. Negative regulation of toll-like receptor-mediated immune responses. Nat Rev Immunol 2005;5:446–58.

11] Der SD, Zhou A, Williams BR, et al. Identification of genes differentiallyregulated by interferon alpha, beta, or gamma using oligonucleotide arrays.Proc Natl Acad Sci U S A 1998;95:15623–8.

12] Baechler EC, Batliwalla FM, Karypis G, et al. Interferon-inducible geneexpression signature in peripheral blood cells of patients with severe lupus.Proc Natl Acad Sci U S A 2003;100:2610–5.

13] Bennett L, Palucka AK, Arce E, et al. Interferon and granulopoiesis sig-natures in systemic lupus erythematosus blood. J Exp Med 2003;197:711–23.

14] Sadler AJ, Williams BR. Interferon-inducible antiviral effectors. Nat RevImmunol 2008;8:559–68.

15] Qiu H, Fan Y, Joyee AG, et al. Type I IFNs enhance susceptibility toChlamydia muridarum lung infection by enhancing apoptosis of localmacrophages. J Immunol 2008;181:2092–102.

16] Kayagaki N, Yamaguchi N, Nakayama M, et al. Type I interferons (IFNs)regulate tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)expression on human T cells: A novel mechanism for the antitumor effectsof type I IFNs. J Exp Med 1999;189:1451–60.

17] Lee PY, Li Y, Richards HB, et al. Type I interferon as a novel risk factorfor endothelial progenitor cell depletion and endothelial dysfunction insystemic lupus erythematosus. Arthritis Rheum 2007;56:3759–69.

18] Baccala R, Kono DH, Theofilopoulos AN. Interferons as pathogenic effec-tors in autoimmunity. Immunol Rev 2005;204:9–26.

19] Kim HA, Jeon SH, Seo GY, et al. TGF-beta1 and IFN-gamma stimulatemouse macrophages to express BAFF via different signaling pathways. JLeukoc Biol 2008;83:1431–9.

20] Denny MF, Thacker S, Mehta H, et al. Interferon-alpha promotes abnormalvasculogenesis in lupus: a potential pathway for premature atherosclerosis.Blood 2007;110:2907–15.

21] Bengtsson AA, Sturfelt G, Truedsson L, et al. Activation of type I interferonsystem in systemic lupus erythematosus correlates with disease activity butnot with antiretroviral antibodies. Lupus 2000;9:664–71.

22] Blanco P, Palucka AK, Gill M, et al. Induction of dendritic cell dif-ferentiation by IFN-alpha in systemic lupus erythematosus. Science2001;294:1540–3.

23] Baechler EC, Gregersen PK, Behrens TW. The emerging role of inter-

feron in human systemic lupus erythematosus. Curr Opin Immunol2004;16:801–7.

24] Rus V, Chen H, Zernetkina V, et al. Gene expression profiling in periph-eral blood mononuclear cells from lupus patients with active and inactivedisease. Clin Immunol 2004;112:231–4.

[

e 76 (2009) 464–473

25] Gilliet M, Lande R. Antimicrobial peptides and self-DNA in autoimmuneskin inflammation. Curr Opin Immunol 2008;20:401–7.

26] Tian J, Avalos AM, Mao SY, et al. Toll-like receptor 9-dependent activa-tion by DNA-containing immune complexes is mediated by HMGB1 andRAGE. Nat Immunol 2007;8:487–96.

27] Gilliet M, Cao W, Liu YJ. Plasmacytoid dendritic cells: sensing nucleicacids in viral infection and autoimmune diseases. Nat Rev Immunol2008;8:594–606.

28] Vollmer J, Tluk S, Schmitz C, et al. Immune stimulation mediated byautoantigen binding sites within small nuclear RNAs involves Toll-likereceptors 7 and 8. J Exp Med 2005;202:1575–85.

29] Fairhurst AM, Hwang SH, Wang A, et al. Yaa autoimmune phenotypes areconferred by overexpression of TLR7. Eur J Immunol 2008;38:1971–8.

30] Santiago-Raber ML, Kikuchi S, Borel P, et al. Evidence for genes in addi-tion to Tlr7 in the Yaa translocation linked with acceleration of systemiclupus erythematosus. J Immunol 2008;181:1556–62.

31] Blomberg S, Eloranta ML, Magnusson M, et al. Expression of the markersBDCA-2 and BDCA-4 and production of interferon-alpha by plasma-cytoid dendritic cells in systemic lupus erythematosus. Arthritis Rheum2003;48:2524–32.

32] Fritz JH, Ferrero RL, Philpott DJ, et al. Nod-like proteins in immunity,inflammation and disease. Nat Immunol 2006;7:1250–7.

33] Miceli-Richard C, Comets E, Loiseau P, et al. Association of an IRF5gene functional polymorphism with Sjogren’s syndrome. Arthritis Rheum2007;56:3989–94.

34] Bave U, Nordmark G, Lovgren T, et al. Activation of the type I inter-feron system in primary Sjogren’s syndrome: a possible etiopathogenicmechanism. Arthritis Rheum 2005;52:1185–95.

35] Mavragani CP, Niewold TB, Moutsopoulos NM, et al. Augmentedinterferon-alpha pathway activation in patients with Sjogren’s syndrometreated with etanercept. Arthritis Rheum 2007;56:3995–4004.

36] Gottenberg JE, Cagnard N, Lucchesi C, et al. Activation of IFN pathwaysand plasmacytoid dendritic cell recruitment in target organs of primarySjogren’s syndrome. Proc Natl Acad Sci U S A 2006;103:2770–5.

37] Wildenberg ME, van Helden-Meeuwsen CG, van de Merwe JP, et al.Systemic increase in type I interferon activity in Sjogren’s syndrome: aputative role for plasmacytoid dendritic cells. Eur J Immunol 2008;38:2024–33.

38] Ittah M, Miceli-Richard C, Gottenberg JE, et al. Viruses induce highexpression of BAFF by salivary gland epithelial cells through TLR-and type-I IFN-dependent and -independent pathways. Eur J Immunol2008;38:1058–64.

39] Strandberg L, Ambrosi A, Espinosa A, et al. Interferon-alpha induces up-regulation and nuclear translocation of the Ro52 autoantigen as detectedby a panel of novel Ro52-specific monoclonal antibodies. J Clin Immunol2008;28:220–31.

40] Kong HJ, Anderson DE, Lee CH, et al. Cutting edge: autoantigenRo52 is an interferon inducible E3 ligase that ubiquitinates IRF-8 andenhances cytokine expression in macrophages. J Immunol 2007;179:26–30.

41] Willeke P, Schluter B, Schotte H, et al. Interferon-gamma is increased inpatients with primary Sjogren’s syndrome and Raynaud’s phenomenon.Semin Arthritis Rheum 2008, doi:10.1016/j.semarthrit.2008.04.002.

42] Tan FK, Zhou X, Mayes MD, et al. Signatures of differentially regu-lated interferon gene expression and vasculotrophism in the peripheralblood cells of systemic sclerosis patients. Rheumatology (Oxford)2006;45:694–702.

43] Kim D, Peck A, Santer D, et al. Induction of interferon-alpha by scle-roderma sera containing autoantibodies to topoisomerase I: associationof higher interferon-alpha activity with lung fibrosis. Arthritis Rheum2008;58:2163–73.

44] Coelho LF, de Oliveira JG, Kroon EG. Interferons and scleroderma-a newclue to understanding the pathogenesis of scleroderma? Immunol Lett

2008;118:110–5.

45] Dieude P, Guedj M, Wipff J, et al. Association between theIRF5 rs2004640 functional polymorphism and systemic sclerosis. Anew perspective for pulmonary fibrosis. Arthritis Rheum 2009;60:225–33.

Spin

[

[

[

[

[

[

[

[

[

O. Meyer / Joint Bone

46] Greenberg SA, Pinkus JL, Pinkus GS, et al. Interferon-alpha/beta-mediated innate immune mechanisms in dermatomyositis. Ann Neurol2005;57:664–78.

47] Howard OM, Dong HF, Yang D, et al. Histidyl-tRNA synthetase andasparaginyl-tRNA synthetase, autoantigens in myositis, activate chemokinereceptors on T lymphocytes and immature dendritic cells. J Exp Med2002;196:781–91.

48] Dieguez-Gonzalez R, Calaza M, Perez-Pampin E, et al. Association of inter-feron regulatory factor 5 haplotypes, similar to that found in systemic lupuserythematosus, in a large subgroup of patients with rheumatoid arthritis.Arthritis Rheum 2008;58:1264–74.

49] Orozco G, Alizadeh BZ, Delgado-Vega AM, et al. Association of STAT4

with rheumatoid arthritis: a replication study in three European populations.Arthritis Rheum 2008;58:1974–80.

50] Zervou MI, Sidiropoulos P, Petraki E, et al. Association of a TRAF1 and aSTAT4 gene polymorphism with increased risk for rheumatoid arthritis ina genetically homogeneous population. Hum Immunol 2008;69:567–71.

[

[

e 76 (2009) 464–473 473

51] Banchereau J, Pascual V, Palucka AK. Autoimmunity throughcytokine-induced dendritic cell activation. Immunity 2004;20:539–50.

52] Bauer JW, Baechler EC, Petri M, et al. Elevated serum levels of interferon-regulated chemokines are biomarkers for active human systemic lupuserythematosus. PLoS Med 2006;3:e491.

53] Wallace J, Petri M, Olsen N, et al. MEDI-545, an anti-interferon alphamonoclonal antibody, shows evidence of clinical activity in systemic lupuserythematosus. Arthritis Rheum 2007;56:S526–7.

54] Dong L, Ito S, Ishii KJ, et al. Suppressive oligodeoxynucleotides delaythe onset of glomerulonephritis and prolong survival in lupus-proneNZB × NZW mice. Arthritis Rheum 2005;52:651–8.

55] Klinman DM. Immunotherapeutic uses of CpG oligodeoxynucleotides. NatRev Immunol 2004;4:249–58.

56] Kampmann B, Hemingway C, Stephens A, et al. Acquired predispositionto mycobacterial disease due to autoantibodies to IFN-gamma. J Clin Invest2005;115:2480–8.