Interleukin 18–independent engagement ofinterleukin 18 receptor-a is required forautoimmune inflammation

Ilona Gutcher1, Eduard Urich1, Karina Wolter2, Marco Prinz2 & Burkhard Becher1

T helper type 1 (TH1) lymphocytes are considered to be the main pathogenic cell type responsible for organ-specific

autoimmune inflammation. As interleukin 18 (IL-18) is a cofactor with IL-12 in promoting TH1 cell development, we examined

the function of IL-18 and its receptor, IL-18R, in autoimmune central nervous system inflammation. Similar to IL-12-deficient

mice, IL-18-deficient mice were susceptible to experimental autoimmune encephalomyelitis. In contrast, IL-18Ra-deficient mice

were resistant to experimental autoimmune encephalomyelitis, indicating involvement of an IL-18Ra ligand other than IL-18

with encephalitogenic properties. Moreover, engagement of IL-18Ra on antigen-presenting cells was required for the generation

of pathogenic IL-17-producing T helper cells. Thus, IL-18 and TH1 cells are dispensable, whereas IL-18Ra and IL-17-producing

T helper cells are required, for autoimmune central nervous system inflammation.

Tissue-specific, cell-mediated autoimmune diseases such as rheuma-toid arthritis, type I diabetes or multiple sclerosis are widelybelieved to be mediated by autoreactive T helper lymphocytes.Until recently, multiple sclerosis and rheumatoid arthritis were recog-nized as being mediated by T helper type 1 (TH1) cells, an ideasupported by findings obtained with their respective animalmodels, experimental autoimmune encephalomyelitis (EAE) andcollagen-induced arthritis1,2. In EAE, for example, disease can beinduced in susceptible rodents by adoptive transfer of myelin-reactiveTH1 cells as well as by active immunization with myelin antigens.Nevertheless, the main evidence supporting the idea that pathogenicautoimmune cells are polarized TH1 cells is based on the factthat tissue-invading T cells usually express interferon-g (IFN-g)3,4.Furthermore, autoaggressive TH2 cells were believed to haveanti-encephalitogenic properties, and TH2 immune deviation iswidely considered a promising therapeutic strategy for the treatmentof multiple sclerosis5–7.

However, a paradigm shift has occurred because of several keyfindings. First, IFN-g-deficient and tumor necrosis factor–deficientmice, for example, have been found to be susceptible to EAE8–11. Infact, IFN-g deficiency even renders EAE-resistant mouse strainssusceptible to EAE9. Second, although the p40 subunit of IL-12 iscritical for the development of autoimmunity, its binding partner p35is irrelevant12. That result is explained by the fact that p40 is also thelarge subunit of IL-23, which, in contrast to IL-12, is vital forautoimmunity12. Third, in contrast to IL-12, IL-23 is not a principalTH1-inducing cytokine but instead drives the population expansion of

IL-17-polarized T cells. IL-17, in contrast to IFN-g, is indeed firmlylinked with encepahlopathogenicity13,14.

Because IL-12 is no longer considered an essential factor for EAE,we wanted to assess the involvement of IL-18 in TH1 cell polarizationand EAE development. IL-18 is a proinflammatory cytokine and amember of the IL-1 ‘superfamily’ of cytokines. Its main functionduring inflammation seems to be the polarization of TH1 cells15–17.IL-18 is secreted by antigen-presenting cells (APCs) and signalsthrough the IL-18 receptor (IL-18R), a heterodimer consisting of aligand-binding IL-18Ra subunit and a signaling IL-18Rb subunit (alsocalled IL-1RAcPL and IL-1R7). Signaling ‘downstream’ of IL-18R, likethat of other IL-1 and Toll-like receptors, activates the kinase IRAK4and adaptor molecule MyD88 (refs. 18–20). IL-18R is expressed onlymphocytes as well as on accessory cells21–24. Although it is firmlyestablished that IL-18 can bind to the IL-18R complex, its affinityfor IL-18Ra is weak25,26. Given the protective activity of IL-12(refs. 12,27,28) and IFN-g29, we sought to identify the function ofIL-18 and IL-18Ra in central nervous system (CNS) autoimmunity.

Here we show that IL-18 did not exert a pathogenic effect in thedevelopment of EAE, as Il18–/– mice were fully susceptible to EAE.However, deletion of its receptor (IL-18Ra) resulted in completeresistance to EAE, suggesting the presence of an alternative ligandwith encephalitogenic properties. Loss of IL-18Ra affected neither thepriming nor the population expansion of antigen-driven T cells.However, engagement of IL-18Ra on APCs was critical for thegeneration of IL-17-producing T helper cells (TH-17 cells) throughan IL-23-dependent mechanism.

Received 21 March; accepted 17 July; published online 13 August 2006; doi:10.1038/ni1377

1Neuroimmunology Unit, Neurology Clinic, University of Zurich, 8057 Zurich, Switzerland. 2Institute of Neuropathology, Georg-August-University, D-37075 Gottingen,Germany. Correspondence should be addressed to B.B. (burkhard.becher@neuroimm.unizh.ch).

946 VOLUME 7 NUMBER 9 SEPTEMBER 2006 NATURE IMMUNOLOGY

A R T I C L E S©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

RESULTS

Mice deficient in both IL-12p35 and IL-18 develop EAE

EAE is generally referred to as a TH1 cell–mediated disease, yetdeletion of IL-12, a chief TH1-promoting cytokine, does not lead toresistance to EAE12. To assess whether IL-18 is capable of compensat-ing for the loss of IL-12p35 and thus is responsible for the suscept-ibility of IL-12p35-deficient mice to EAE, we generated mice deficientin both IL-12p35 (Il12a–/–) and IL-18 (Il18–/–). Immunization of micewith myelin oligodendrocyte glycoprotein peptide consisting of aminoacids 35–55 (MOG(35–55)) emulsified in complete Freund’s adjuvant(CFA) showed that Il12a–/–Il18–/– mice were susceptible to EAE andhad a disease severity and development similar to that of wild-typemice (Table 1). Immunization of wild-type and single-mutant Il18–/–

mice with MOG(35–55) showed that Il18–/– mice were also susceptibleto EAE (Fig. 1 and Table 1). Those results demonstrated that IL-18 isnot responsible for the EAE susceptibility of Il12a–/– mice andindicated that IL-18 itself is a cytokine with little or no effect onEAE pathogenesis.

IL-18 is required for mitogen- but not antigen-driven immunity

To rule out the possible expression of a truncated yet functionalform of IL-18, we extensively verified the targeting strategy andgenotype of the mice. We established that Il18–/– mice did notproduce IL-18 mRNA by doing RT-PCR with primers outside thetarget of deletion. We also analyzed by enzyme-linked immunosorbentassay (ELISA) whether we could detect IL-18 secreted from activatedsplenocytes derived from wild-type and Il18–/– mice; this showedthat Il18–/– mice were indeed completely deficient in IL-18 (Supple-mentary Fig. 1 online). As noted in many experimental systems,deletion of IL-18 consistently results in the paucity of an IFN-gresponse30,31. We thus stimulated lymphocytes derived from naivewild-type, Il18–/– and Il18r1–/– (IL-18Ra-deficient) mice in vitrofor 16 h with the lectin concanavalin A and measured IFN-g produc-tion by ELISA. Consistent with the idea thatIL-18 polarizes TH1 cells, lymphocytes fromIl18–/– and Il18r1–/– mice did not secrete IFN-g, in contrast to wild-type lymphocytes(Fig. 2a). We further verified that T cellfunction and activation during polyclonalstimulation was not directly impaired, asthere was no difference in IFN-g secretionby stimulated wild-type, Il18–/– and Il18r1–/–

purified CD4+ T cells (SupplementaryFig. 2 online).

To establish the function of IL-18 and IL-18Ra in adaptive antigen-driven immuneresponses, we immunized wild-type, Il18–/–

and Il18r1–/– mice subcutaneously with

MOG(35–55) in CFA or with keyhole limpet hemocyanin (KLH),as a large immunogenic protein antigen (of about 400 kilodaltons),in CFA and 7 d later isolated and restimulated lymphocyteswith MOG or KLH in vitro. When we used KLH as immunogen,we did not find a significant difference in IFN-g productionby lymphocytes from wild-type, Il18–/– or Il18r1–/– mice (Fig. 2b).Immunization with MOG-peptide, however, resulted in Il18r1–/–

cells that produced significantly less IFN-g than did lymphocytesfrom wild-type or Il18–/– mice. Those data supported theconclusion that although IL-18 and IL-18Ra are critical cofactorsfor the early IFN-g response of mitogen- or lectin-activatedT cells, MOG(35–55)-induced responses showed discordant activityby lymphocytes lacking IL-18 or IL-18Ra. In contrast, activation andIFN-g production via the complex large protein antigenKLH seemed to be independent of IL-18 or IL-18Ra, a resultin agreement with a published report demonstrating theredundant function of IL-18 in IFN-g production32.

Il18r1–/– mice are resistant to EAE

Mice deficient in IL-18Ra have been described as having an immu-nological phenotype similar to that of Il18–/– mice33. However, wefound that in contrast to wild-type and Il18–/– mice, Il18r1–/– micewere resistant to EAE induction (Fig. 3 and Table 1). Histologicalanalysis of spinal cords obtained 28 d after EAE induction showed thatsusceptible wild-type and Il18–/– mice had considerable inflammationand demyelination of the CNS (Fig. 4). More detailed analysisdemonstrated infiltration of T cells, macrophages and B cells andaxonal damage in the spinal cords of the mice (Fig. 4 and Supple-mentary Fig. 3 online). In contrast, Il18r1–/– mice showed noleukocyte infiltration or demyelination (Fig. 4 and SupplementaryFig. 3). Quantitative RNA analysis of spinal cords from wild-type andIl18–/– mice with EAE showed that in accordance with the presence of

Table 1 EAE susceptibility of mice

Genotype Incidence Disease onset

Maximum

clinical score

WT 25 of 30 (83.33) 12.5 d 2.56 ± 0.12

IL12a–/–IL18–/– 7 of 10 (70) 14.1 d 3.20 ± 0.08

IL18–/– 20 of 22 (91) 12.8 d 2.35 ± 0.13

IL18r1–/– 2 of 20 (10) 18.5 d 2.60 ± 0.12

EAE disease incidence (%), mean time of disease onset (d) and mean maximum clinical score(± s.e.m.) of diseased mice, for wild-type, IL12a–/–IL18–/–, IL18–/– and IL18r1–/– mice in whichEAE was induced by subcutaneous immunization with MOG(35–55).

WT

ll18 –/–

Clin

ical

sco

re

0

1

2

3

9 13 19 24Time after immunization (d)

Figure 1 IL-18 is not required for EAE induction. EAE progression in wild-

type mice (WT; n ¼ 30) and Il18–/– mice (n ¼ 22). Data are from one

representative of three individual experiments.

+ Con A – Con A0

0.2IFN

-γ (

ng/m

l)

IFN

-γ (

ng/m

l)

IFN

-γ (

ng/m

l)

0.4

0.6

0.8

1.0 P < 0.002 P = 0.001

0

2

4

6

8

10

+ KLH– MOG0

200

400

600

800WTll18–/–

ll18r1–/–

+ MOG – KLH

a b

Figure 2 IL-18 is required for mitogen- but not antigen-driven TH1 development. (a) ELISA of IFN-gsecretion by naive wild-type, Il18–/– and Il18r1–/– lymph node cells, stimulated for 16 h with 5 mg/ml of

concanavalin A (+ Con A) or medium (– Con A). (b) ELISA of IFN-g in supernatants of cells from mice

immunized with 200 mg MOG(35–55) or KLH in CFA; 7 d after immunization, lymph node cells were

isolated and restimulated in duplicate for 48 h with 50 mg/ml of MOG(35–55) or KLH or medium

(– MOG; – KLH). Data are representative of at least two individual experiments with two mice

per group.

NATURE IMMUNOLOGY VOLUME 7 NUMBER 9 SEPTEMBER 2006 947

A R T I C L E S©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

inflammatory infiltrates, they had increasedlocal expression of chemokines and inflam-matory cytokines (Supplementary Fig. 3).

The different susceptibilities of Il18–/– andIl18r1–/– mice to EAE induction (Fig. 3a)indicates the possibility of an IL-18Ra ligandother than IL-18 with encephalitogenic prop-erties. To verify that IL-18Ra and IL-18 haveindependent biological functions, we blockedIL-18Ra in EAE-susceptible Il18–/– mice.Treatment of Il18–/– mice with a monoclonalantibody to IL-18Ra 1 d before immuniza-tion and every 3 d thereafter until the end ofthe experiment significantly reduced diseasedevelopment (Fig. 3b) without causing dele-tion of IL-18Ra-expressing cells or alteringthe composition of peripheral leukocytes in the blood, lymph nodes orspleens (Supplementary Table 1 online). Treating Il18–/– mice withthe monoclonal antibody to IL-18Ra on day 10 after immunizationalso abrogated EAE progression (Fig. 3c), suggesting that IL-18Raengagement is important during the effector phase of EAE. Therefore,blockade of IL-18Ra prevents EAE even in mice in which its proposedligand is completely removed by gene targeting, and indeed IL-18reportedly has a low affinity for IL-18Ra. Thus, we propose that anIL-18Ra ligand other than IL-18 is responsible for the engagement,signaling and immune development mediated by IL-18Ra.

Persistence of CNS inflammation and TH-17 cells requires

IL-18RaEAE is characterized by a massive influx of inflammatory cells into theCNS at the peak of disease, yet immune cells also invade the CNSbefore the onset of clinical symptoms34,35. For example, recruitment ofCD4+ T cells into the CNS is critical for the initiation of the effectorphase of EAE, whereas invasion of polymorphonuclear leukocytesseems to be involved in orchestrating these events36. Therefore, toestablish the effect of IL-18Ra on the capacity of inflammatory cells toinvade the CNS, we immunized mice with MOG(35–55) and analyzedtheir CNS for inflammatory infiltrates. In contrast to peak disease,

when the Il18r1–/– CNS was devoid of inflammatory infiltrates(Fig. 4), on day 7 after immunization, the number of CNS-infiltratingleukocytes (CD45hi cells) was similar in Il18r1–/–, wild-type and Il18–/–

mice (Fig. 5a). Detailed analysis of the invading cells showed thatcomparable numbers of CD4+ T cells, granulocytes, macrophages andB cells were present in all three strains (Supplementary Fig. 3).Because inflammatory cells are absent from the CNS during clinicaldisease, it seems that Il18r1–/– inflammatory cells do not persist duringthe effector phase of the disease. Such results resemble those obtainedwith IL-23p19-deficient (Il23a–/–) mice and IL-23p40-deficient(Il12b–/–) mice, which are also resistant to MOG(35–55)-inducedEAE and in which inflammatory cells are found in the CNS earlyafter immunization but before disease begins13,37.

Given the similarities between Il18r1–/– and Il23–/– mice regardingtheir EAE resistance with concomitant inflammatory cell invasioninto the CNS, we assessed the effect of IL-18Ra on IL-17 production.TH-17 cells are now widely believed to be the main pathogenicpopulation during autoimmune CNS inflammation13,14. We thereforequantified IL-17-producing MOG-reactive T cells invading the CNSbefore and after clinical disease onset in mice immunized withMOG(35–55). Enzyme-linked immunospot analysis showed thatimmediately before disease onset (9 days after immunization), the

a b c

9 14 19 24

Clin

ical

sco

re

Time after immunization (d)

0

1

2

3

0

1

2

3

10 13 17 21 25

* * * * * *Clin

ical

sco

re

0

1

2

3

Clin

ical

sco

re

IgG

* * * * *10 13 18 21

IgG

Time after immunization (d) Time after immunization (d)

WTll18–/– Anti-IL-18Rα Anti-IL-18Rαll18r1–/–

Figure 3 Discordant EAE progression in Il18–/– and Il18r1–/– mice. (a) EAE progression in wild-type,Il18–/– and Il18r1–/– mice. Data are from one representative of three individual experiments. (b) EAE

progression in Il18–/– mice treated with 450 mg anti-IL-18Ra or control immunoglobulin G (IgG) 1 d

before immunization with MOG(35–55) and with 300 mg anti-IL-18Ra every 3 d thereafter. (c) EAE

progression in Il18–/– mice immunized and treated with 300 mg anti-IL-18Ra or control immuno-

globulin G at the first sign of disease. *, P o 0.05. Data are representative of at least two individual

experiments (n Z 5 mice/group).

H&E

WT

Il18

–/–

Il18r

1–/

–

LFB-PAS MAC3 CD3 B220 APP

Figure 4 Histopathology of Il18–/– and Il18r1–/– mice. Immunohistochemistry of paraffin-embedded spinal cord sections from PBS-perfused wild-type, Il18–/–

and Il18r1–/– mice on day 28 after induction of EAE with MOG(35–55); sections were stained with hematoxylin and eosin (H&E), Luxol fast blue–periodic

acid Schiff (LFB-PAS), anti-MAC3 (MAC3), anti-CD3 (CD3), anti-B220 (B220) or antibody to amyloid precursor protein (APP). Arrows (insets) indicate

positive cells. Scale bars, 100 mm (main images), 50 mm (insets, CD3 and B200) and 25 mm (inset, APP).

948 VOLUME 7 NUMBER 9 SEPTEMBER 2006 NATURE IMMUNOLOGY

A R T I C L E S©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

number of MOG-specific IFN-g-secreting cells invading the CNSwas similar in EAE-susceptible Il18–/– and EAE-resistant Il18r1–/–

mice, whereas TH-17 cells were nearly completely absent from theIl18r1–/– mice only (Fig. 5b). After disease onset, moreover, althoughwe detected increased numbers of IFN-g- and IL-17-secreting cells inthe susceptible Il18–/– and wild-type mice, these cells were almostcompletely absent from the Il18r1–/– mice (Fig. 5c).

To determine whether Il18r1–/– mice are defective in generatingIL-17-producing T cells, we analyzed IL-17 expression by freshlyprimed lymphocytes in vitro. We immunized wild-type, Il18–/– andIl18r1–/– mice with either MOG(35–55) or KLH and collected lym-phocytes from the draining lymph nodes after 7 d. By real-time PCRanalysis of mRNA obtained from lymphocytes after restimulation withMOG antigen, we found that the expression of IL-17 mRNA wassignificantly lower in Il18r1–/– cells than in wild-type or Il18–/– cells(Fig. 6a). We corroborated those results by ELISA of cell culture mediaof lymphocytes rechallenged with MOG(35–55) in vitro, whichdemonstrated that IL-17 production was sig-nificantly lower for Il18r1–/– but not Il18–/– orwild-type lymphocytes (Fig. 6b). In contrastto the production of IFN-g, which showedsome dependence on the immunogenicity ofthe stimulating antigen, IL-17-secretion wasconsistently lower in the Il18r1–/– leukocytes,even when KLH was used to immunize andrestimulate the cells. Thus, the resistance ofIl18r1–/– mice to EAE could certainly derivefrom their inability to generate sufficientnumbers of inflammatory TH-17 cells.

IL-18Ra expression on accessory cells

The results thus far have shown that IL-18Radeficiency completely prevents the develop-ment of EAE and TH-17 cell polarization,whereas IL-18 seems to be irrelevant for bothphenotypes. The cell type on which theIL-18Ra exerts its main effects remainsunknown. This is due mainly to the factthat IL-18Ra is expressed by various celltypes and tissues21,38,39. To identify theIL-18Ra-expressing cell type needed forEAE induction, we selectively expressed IL-18Ra on cells in the leukocyte compartment

using irradiation bone marrow chimeras. After irradiation and recon-stitution, the immune compartment in secondary lymphoid tissues ofrecipient mice is composed of hematopoietic cells derived from donormice40,41. We generated bone marrow chimeras by transferring either a4:1 ratio of recombination-activating gene 1–deficient (Rag1–/–) andIl18r1–/– bone marrow into wild-type recipients (Rag1–/– + Il18r1–/–

- wild-type) or Il18r1–/– bone marrow only into wild-type recipients(Il18r1–/– - wild-type); as a control, we also transferred wild-typebone marrow into wild-type recipients (wild-type - wild-type).Because Rag1–/– mice do not have lymphocytes, all lymphocytesfrom the Rag1–/– + Il18r1–/– - wild-type chimera were IL-18Radeficient, whereas most other nonlymphocyte leukocytes (accessorycells such as dendritic cells, macrophages, neutrophils and so on)expressed IL-18Ra.

As anticipated from the data above, Il18r1–/– - wild-type chimericmice were resistant to EAE after immunization with MOG(35–55)(Fig. 7a). However, the addition of Rag1–/– bone marrow to the

IL-17

IFN-γ

IL-17

IFN-γ

CD45

Eve

nts

42% 34% 32% 13%WT EAE

WT

Il18–/– EAE

Il18–/–

Il18r1–/– EAE

Il18r1–/– WT Il18–/– Il18r1–/–

Naive WT

Spo

ts/w

ell

P = 0.015

P = 0.002

0

200

400

600

800

IFN-γ IL-17 IFN-γ IL-17

P = 0.02

P = 0.004

0

200

400

600

800

Spo

ts/w

ell

256

128

0

WT Il18–/–

NaiveII18r1–/–

a

b c

Figure 5 Tissue invasion at preclinical stages is

not affected by IL-18Ra. (a) Flow cytometry of

CNS-derived leukocytes from Il18–/– and Il18r1–/–

mice immunized with MOG(35–55) 7 d after

immunization; cells were stained with anti-CD45

and anti-CD11b. Naive, nonimmunized. Numbers

above bracketed lines indicate percent CD45hi

cells in the CNS. Data are from onerepresentative of three individual experiments

with two mice per experiment. (b,c) Enzyme-

linked immunospot analyses of IFN-g and IL-17

production by CNS-derived MOG-reactive

lymphocytes restimulated for 18–20 h with

MOG(35–55). (b) 9 days after immunization,

before disease onset. (c) 14 d after

immunization, after disease onset. Graphed data

(mean + s.e.m.) represent at least two individual

experiments with two mice per group.R

elat

ive

IL-1

7 m

RN

AIL

-17

(ng/

ml)

0

2

4

6

Rel

ativ

e IL

-17m

RN

A

0

10

20

30

40

50

+ KLH – KLH

+ KLH – KLH

0

20

40

60

80

+ MOG

a

b– MOG

+ MOG – MOG

P = 0.016

P = 0.001

P = 0.0099

P = 0.028

WTIl18–/–

Il18r1–/–

IL-1

7 (n

g/m

l)

0

10

20

30

40

Figure 6 TH-17 cell induction depends on IL-18Ra but not IL-18. (a) Real-time PCR analysis of

IL-17 mRNA expression by wild-type, Il18–/– and Il18r1–/– lymphocytes obtained from mice at 7 d afterimmunization with MOG(35–55) or KLH and restimulated for 48 h with 50 mg/ml of MOG(35–55) or

KLH or with medium alone (– MOG; – KLH). Results are normalized to b-actin expression, and samples

were analyzed in duplicate. Data (mean + s.e.m.) are representative of two individual experiments

with at least two mice per group. (b) ELISA of IL-17 expression by lymphocytes restimulated with

MOG(35–55) or KLH, analyzed in duplicate.

NATURE IMMUNOLOGY VOLUME 7 NUMBER 9 SEPTEMBER 2006 949

A R T I C L E S©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

Il18r1–/– bone marrow (Rag1–/– + Il18r1–/– - wild-type) resulted insusceptibility to EAE (Fig. 7a). Thus, IL-18Ra expression on accessorycells is sufficient for EAE induction even when lymphoctyes areIL-18Ra deficient. Because IL-18Ra is present on myeloid cells42

(such as APCs), we next tested the capacity of IL-18Ra-deficientAPCs to prime naive T cells. We purified CD4+ T cells from 2d2transgenic mice, which express a T cell receptor (TCR) specific forMOG(35–55), and cultured them together with mature, MOG(35–55)-pulsed, wild-type, Il18–/– or Il18r1–/– bone marrow–derived den-dritic cells (DCs). We found no difference in the ability of the differentDCs to stimulate proliferation of the 2d2 T cells (Fig. 7b). To confirmthat result in an in vivo setting, we injected carboxyfluoresceindiacetate succinimidyl diester–labeled 2d2 cells into wild-type,Il18–/– and Il18r1–/– mice, immunized the mice with MOG(35–55)and then analyzed proliferation of the 2d2 cells. Again, we foundno difference in proliferation of the 2d2 cells in wild-type, Il18–/– andIl18r1–/– mice (Fig. 7c). Additional experiments with adoptive transferof TCR-transgenic T cells confirmed that IL-18Ra deficiency on

nonlymphocytic leukocytes reduced the generation of TH-17 cells(Supplementary Fig. 4 online).

To assess the ability of myeloid cells to reach the CNS tissue, wegenerated mixed bone marrow chimeras by transferring CD45 con-genic wild-type and Il18r1–/– bone marrow into wild-type recipientsand evaluated the capacity of myeloid populations to invade the CNSduring disease. We first confirmed that the mice had a 1:1 ratio ofwild-type and Il18r1–/– hematopoietic cells in peripheral blood andspleens. After EAE induction, we found both myeloid populations inthe CNS at a ratio of 1:1, indicating that there was no migratorydifference between the two genotypes of myeloid cells (data notshown). Expression of activation markers and costimulatory mole-cules by lipopolysaccharide-matured DCs from wild-type, Il18 –/– andIl18r1–/– mice showed no difference in expression of CD80, CD86 andCD40 (data not shown). To confirm the function of IL-18Ra signalingin accessory cells during the effector phase of EAE, we adoptivelytransferred encephalitogenic MOG-reactive T cells derived from wild-type donor mice into both wild-type and Il18r1–/– recipient mice.

0

1

2

3

4

8 11 14 17

Clin

ical

sco

re

* * *# # #

Time after immunization (d)

WT

Il18r1–/–Rag1–/– + Il18r1–/–

WT

Il18r1–/–

WT

Il18r1–/–Il18 –/–

0

1

2

3

5 8 11 14 17

Clin

ical

sco

re

* # # # # * *# # #

Time after immunization (d)

Pro

lifer

atio

n(1

03 c.p

.m.)

3 × 1

03

10 ×

103

T cells

only

DCs only

80

60

40

20

0

WT Il18 –/– Il18r1–/– Not immunized

Vα3

.2

CFSE

a

c

b d

Figure 7 IL-18Ra deficiency specifically affects nonlymphocytic leukocytes. (a) EAE progression in Il18r1–/– - wild-type (filled triangles), Il18r1–/–

+ Rag1–/– - wild-type (open squares) and wild-type - wild-type (filled diamonds) bone marrow chimeras actively immunized with MOG(35–55).

*, P o 0.05; #, P o 0.01. Data are from one representative of two individual experiments (n Z 5 mice/group). (b) Proliferation of TCR-transgenic (2d2)

CD4+ T cells stimulated with bone marrow–derived DCs from wild-type, Il18–/– and Il18r1–/– mice; cells were matured with 10 mg/ml of lipopolysaccharide

and were then pulsed with 1 mg/ml of MOG(35–55). Error bars, s.e.m.; n ¼ 2 mice per group. (c) Flow cytometry of the in vivo proliferation of CFSE

(carboxyfluorescein diacetate succinimidyl diester)–labeled 2d2 cells transferred into MOG(35–55)-immunized wild-type, Il18–/– and Il18r1–/– miceand a nonimmunized wild-type control mouse (Not immunized). Data are representative of experiments repeated at least twice with at least two mice

per experiment. (d) EAE progression in recipient Il18r1–/– and wild-type mice after induction of EAE by adoptive transfer of 25 � 106 MOG-reactive

lymphocytes. *, P o 0.05; #, P o 0.01. Data are one representative of two individual experiments (n Z 5 mice/group).

p40

(pg/

ml)

+ α-CD40 – α-CD40 + MOG – MOG + KLH – KLH0

50

100

150

200

250P = 0.0044

P = 0.01 P = 0.0238WTIl18–/–

Il12b–/–Il18r1–/–

0

100

200

300

400

500

p40

(pg/

ml)

0

100

200

300

400

500

600

p40

(pg/

ml)

a b c

Figure 8 IL-18Ra engagement promotes the production of IL-23p40. (a) ELISA of IL-12,IL-23p40 (p40) in the supernatants of T cell–depleted splenocyte

samples from naive wild-type, Il18–/– and Il18r1–/– mice; cells were stimulated for 36 h with 5 mg/ml of monoclonal antibody to CD40 (+ a-CD40) or

were left unstimulated (– a-CD40). (b) ELISA of p40 in culture medium of lymph node cells from wild-type, Il18–/– and Il18r1–/– mice immunized with

MOG(35–55) in CFA (b) or KLH in CFA (c); cells were isolated 7 d later and were restimulated for 48 h with 50 mg/ml of MOG(35–55) or medium (– MOG; b)

or with 50 mg/ml KLH or medium (– KLH; c). Data (mean + s.e.m.) are representative of two individual experiments with at least two mice per group.

950 VOLUME 7 NUMBER 9 SEPTEMBER 2006 NATURE IMMUNOLOGY

A R T I C L E S©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

Fully primed and activated encephalitogenic T cells derived fromwild-type mice induced EAE in wild-type recipient mice, yet theywere incapable of inducing clinical EAE in IL-18Ra-deficienthosts (Fig. 7d).

IL-18Ra-deficient APCs secrete limited IL-12,IL-23p40

As our results demonstrated a substantial decrease in IL-17 productionby T cells after restimulation, we analyzed the capacity of Il18r1–/–

APCs to secrete IL-12,IL-23p40, which is necessary for the populationexpansion of pathogenic TH-17 cells. To measure p40 secretion,we stimulated T cell–depleted splenocyte samples with antibody toCD40 (anti-CD40) for 36 h before measuring p40 productionby ELISA. We found that Il18r1–/– APCs, like IL12 p40–deficient(Il12b–/–) APCs, had impaired p40 production, in contrast to highp40 production by wild-type and Il18–/– APCs (Fig. 8a). Finally,we confirmed the deficient IL-12,IL-23p40 production in an adaptiveimmune response by restimulating lymph node preparations fromMOG- and KLH-immunized mice in vitro and analyzing p40 produc-tion after 2 d. As expected, there was significantly less production ofp40 by Il18r1–/– lymphocytes than by wild-type and Il18–/– lympho-cytes (Fig. 8b,c). We failed to find significantly lower p19 expressionby stimulated APCs, which indicated that the effect of IL-18Radeficiency on TH-17 cells was probably regulated by the amountof p40 rather than p19 expression. In addition, we did not findlower transforming growth factor-b expression in Il18r1–/– APCs (datanot shown).

DISCUSSION

Organ-specific inflammatory diseases generally result from inap-propriate population expansion and activation of effector T lympho-cytes, which are capable of escaping peripheral tolerance and reactingto expressed self antigens. Until recently, TH1 lymphocytes werethought to represent the autoreactive T cells responsible for inducingcellular autoimmunity3–7. However, mounting evidence indicates thatTH-17 cells, not TH1 cells, are the pathogenic T cell effectors inautoimmunity and that their development is negatively regulated byboth TH1 and TH2 cytokines12–14,27,43–45.

We were interested in characterizing the function of IL-18 in EAE.IL-18 acts in synergy with IL-12 in the differentiation of naive TH cellsinto TH1 cells. Our initial goal was to determine if the EAE suscept-ibility of IL-12-deficient mice occurs as a result of redundancy,whereby the continued presence and activity of IL-18 in these miceis sufficient for EAE induction. Our data challenge that hypothesis, asIl12a–/–Il18–/– double-deficient mice were fully susceptible to EAE. Wealso found that Il18–/– mice were susceptible to EAE. That resultcontrasts with published data showing that Il18–/– mice are resistant toEAE38. The discrepancy between our data and that earlier study couldbe explained by different health status of the mice or their degree ofbackcrossing onto the C57BL/6 background. Our results with Il18–/–

mice are also in agreement with other published data obtained withindependently derived Il18–/– mice showing susceptibility to experi-mental autoimmune uvitis using a methodology similar to that usedin our study here46,47. Thus, Il18–/– mice are susceptible to EAE andexperimental autoimmune uvitis by the standard methodology ofinduction of experimental autoimmunity. Finally, another reportdemonstrating that IL-18 is redundant in an antigen-induced arthritismodel32 directly supports our finding that IL-18 was unnecessaryduring antigen-induced TH1 responses.

Despite the susceptibility of Il18–/– mice, we found that Il18r1–/–

mice were resistant to EAE, which indicates that an IL-18Ra-bindingligand other than IL-18 has encephalitogenic properties. There are

many ‘orphan receptors’ in the IL-1R ‘superfamily’, and given thatmembers of this family of receptor subunits form heterodimers withone another48, it is possible that IL-18Ra not only has differentbinding partners but also different ligands. We demonstrated theimportance of IL-18Ra by significantly attenuating disease develop-ment in Il18–/– mice using antibodies to IL-18Ra. Given that theknown IL-18Ra ligand, IL-18, was not present in those mice, thatresult provides persuasive evidence for the existence of an alternativeIL-18Ra ligand. Future studies should be aimed at identifying such anIL-18Ra-binding ligand as well as the composition of the ligand’s fullreceptor complex.

The development of EAE in wild-type mice is dependent on theinfiltration of activated CD4+ T cells into the CNS, an event that isaccompanied by the influx of other immune cells, including B cells,macrophages and granulocytes34–36. In Il18r1–/– mice, inflammatorycells were absent from the spinal cords at the peak of clinical EAE.However, we detected comparable CNS-invading leukocyte infiltra-tion in both Il18r1–/– and wild-type mice before disease onset.Therefore, IL-18Ra deficiency affects not the initial migratoryproperties of leukocytes but their persistence in the CNS. The presenceof inflammatory infiltrates in the Il18r1–/– CNS early after immun-ization, before the onset of clinical disease, is similar to the pheno-type of Il23–/– mice37, whose resistance to EAE probably results fromtheir inability to expand populations of and maintain IL-17-producingTH cells13,37.

One of our goals was to determine the mechanism of action ofIL-18Ra activation and signaling during EAE. Published data haveshown that the expression of IL-18Ra is broadly distributed21,38,39. Inexperiments with bone marrow chimeras, we demonstrated that theloss of IL-18Ra from an accessory cell was responsible for the loss ofTH-17 cells. The importance of IL-18Ra on accessory cells wasemphasized by adoptive transfer of otherwise encephalitogenic wild-type T cells that did not induce EAE in Il18r1–/– mice. We generallyfound that T cell activation and population expansion were notaffected by the loss of IL-18Ra and that IL-18Ra-deficient APCshad no difference in activation status. Finally, we demonstrated thatIL-18Ra signaling on APCs was critical for the secretion of IL-23p40and the subsequent maintenance of IL-17-secreting T cells. Our datathus support the hypothesis that IL-17 secretion is dependent on thecontinuous support of TH-17-promoting APCs.

In summary, we have provided evidence challenging the TH1paradigm of autoimmunity by demonstrating a nonpathogenic func-tion for IL-18 in EAE. In contrast, however, we have shown thatIL-18Ra is essential for the development of EAE, thus indicating thepresence of an alternative IL-18Ra-binding ligand. Identification ofthe alternative IL-18Ra-binding ligand could provide a potent newtherapeutic approach for the treatment of organ-specific inflamma-tory diseases such as multiple sclerosis. The attractive feature of thisIL-18Ra-dependent pathway for therapeutic targeting is that its lossdoes not completely suppress immunity but instead abrogates thedevelopment of pathogenic autoimmune effector T cells.

METHODSMice. Female C57BL/6 mice were purchased from Harlan Laboratories.

IL-12p35-deficent (Il12a–/–) mice and IL-12p40-deficient (Il12b–/–) mice were

purchased from Jackson Laboratories. Homozygous Il18–/– and Il18r1–/– mice

(backcrossed onto the C57BL/6 background for more than eight generations)

were provided by S. Akira (Osaka University, Osaka, Japan) and Rag1–/– mice

were provided by R. Zinkernagel (University Hospital Zurich, Zurich, Switzer-

land) and were bred ‘in-house’ in specific pathogen–free conditions. The 2d2

(MOG-TCR-Tg) mice were provided by V. Kuchroo (Harvard Medical School,

Boston, Massachusetts). Animal experiments were approved by the Swiss

NATURE IMMUNOLOGY VOLUME 7 NUMBER 9 SEPTEMBER 2006 951

A R T I C L E S©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

Veterinary Office (69/2003 and 70/2003; Zurich, Switzerland). Irradiation bone

marrow–chimeric mice were generated as described41. Bone marrow donor

mice were killed with CO2 and bone marrow cells were isolated by flushing of

femur, tibia, radius and hip bones with PBS. Bone marrow cells were then

passed through a cell strainer with a pore size of 100 mm and cells were washed

with PBS. Recipient mice were lethally irradiated with 1,100 rads (split dose)

and were injected intravenously with 12 � 106 to 25 � 106 bone marrow cells.

Engraftment took place over 8 weeks of recovery.

Induction of EAE. Mice were immunized subcutaneously with 200 mg of

MOG(35–55) (MEVGWYRSPFSRVVHLYRNGK; obtained from GenScript)

emulsified in CFA (Difco). Mice received 200 ng pertussis toxin (Sigma-

Aldrich) intraperitoneally at the time of immunization and 48 h later.

For adoptive transfer, MOG-reactive T cells were generated as described40.

Monoclonal antibody to IL-18Ra (clone 112624; R&D Systems) was adminis-

tered either 1 d before immunization (450 mg/mouse) and every 3 d thereafter

(300 mg/mouse) or every 3 d beginning from disease onset (300 mg/mouse).

Mice were assigned scores daily as follows: 0, no detectable signs of EAE;

0.5, distal tail limp; 1, complete tail limp; 2, unilateral partial hindlimb

paralysis; 2.5, bilateral partial limb paralysis; 3, complete bilateral hind-

limb paralysis; 3.5, complete hindlimb paralysis and unilateral forelimb

paralysis; 4, total paralysis of forelimbs and hindlimbs (mice with a score

above 4 to be killed); 5, death. Each time point presents the average disease

score of each group. Statistical significance was assessed with an unpaired

Student’s t-test.

Histology and flow cytometry. Mice were perfused with PBS and then with 4%

(weight/volume) paraformaldehyde in PBS. Spinal columns were removed and

were fixed in 4% (weight/volume) paraformaldehyde in PBS. Spinal cords were

then dissected and were embedded in paraffin before being stained with either

hematoxylin and eosin or with anti-CD3 (145-2C11), anti-B220 (RA3-6B2) or

anti-Mac-3 (M3/84; BD Pharmingen) to assess infiltration of inflammatory

cells. Luxol fast blue stain was used to determine the degree of demyelination or

amyloid precursor protein was used to assess the extent of axonal damage.

For cytofluorometry, anti-CD45 (30/F11), anti-CD4 (RM4-5), anti-Va3.2

(RR3-16), anti-CD11b (M1/70), anti-Gr-1 (RB6-8C5) and anti-B220 (RA3-

6B2; BD Pharmingen) were used. For analysis of CNS-invading cells, mice were

killed with CO2 and were perfused intracardially with PBS as described41.

Spinal cords were flushed out with PBS and brains were dissected to isolate the

brainstem. Tissues were homogenized and strained through a nylon filter with a

pore size of 100 mm (Fisher). After centrifugation, cell suspensions were

resuspended in 30% Percoll (Pharmacia) and were centrifuged at 18,500g for

30 min at 4 1C. Interphase cells were collected and were washed extensively

before being stained. For flow cytometry, antibodies were incubated with cells

for 20 min at 4 1C and then cells were analyzed with a FACSCalibur (BD

Pharmingen) and CellQuest software. Post-acquisition analysis was done with

WinMDI 2.8 software (The Scripps Research Institute).

Proliferation and cytokine assays. Spleens and axillary and inguinal lymph

nodes were isolated from naive mice or mice primed by injections of 100 mg/

flank of MOG(35–55) or KLH (Sigma) emulsified in CFA 7 d earlier. Cells

(2 � 105 cells/well) were plated in triplicate in 96-well plates. For naive cells,

5 mg/ml of concanavalin A was used for 16 h of stimulation and 5 mg/ml of anti-

CD40 (FGK) was used for 36 h of stimulation before analysis of cytokine

production by ELISA. CD4+ T cells were purified from naive splenocyte

samples with BD IMag Magnetic Beads (BD Pharmingen) and were stimulated

for 36 h with 5 mg/ml of anti-CD3 (2c11) and 5 mg/ml of anti-CD28 (37N)

before analysis of IFN-g production by ELISA. MOG- or KLH-reactive cells

were stimulated in triplicate for 48 h with either 50 mg/ml of MOG(35–55) or

KLH, 5 mg/ml of concanavalin A, or medium, and 0.5 mCi/ml of [3H]thymidine

was added after 24 h for assessment of proliferative responses. Thymidine

incorporation was assessed with a Filtermate Collecter (Applied Biosystems)

and a scintillation and luminescence counter. For cytokine analysis, culture

supernatants of identical ‘sister’ cultures were collected after 48 h and were

analyzed in duplicate for IFN-g, IL-17 and IL-23p40 production by ELISA

(Pharmingen). For real-time PCR analysis, RNA was isolated from restimulated

cells by TRIzol extraction (Invitrogen).

Enzyme-linked immunospot analysis. Lymphocytes were isolated from the

CNS of MOG(35–55)-immunized mice on days 9 and 14 after immunization

by differential Percoll centrifugation as described above. Cells (2 � 105

cells/well) were plated in complete RPMI medium containing 50 mg/ml of

MOG(35–55) in 96-well plates (Millipore) coated with 7.5 mg/ml of anti-IFN-g(AN18; Mabtech) or 2 mg/ml of anti-IL-17 (TC11-18H10; BD Pharmingen).

Plates were incubated at 37 1C in 5% CO2 for 18 h (anti-IL-17) or 20 h (anti-

IFN-g), at which point cells were discarded and plates were washed with

PBS. Then, 0.5 mg/ml of biotin-conjugated anti-IFN-g (R4-6A2; Mabtech) or

1 mg/ml of biotin-conjugated anti-IL-17 (TC11-8H4.1; BD Pharmingen) was

added, followed by incubation at 25 1C for 2 h or 4 h, respectively. After plates

were washed, streptavidin–alkaline phosphatase (Mabtech) was added, followed

by incubation for 1 h at 25 1C. Plates were washed with PBS, and 100 ml of the

substrate solution BCIP/NBTplus (5-bromo-4-chloro-3-indolylphosphate–

nitro blue tetrazolium; Biosource) was added to the wells, which were

developed until distinct spots emerged. Plates were analyzed with an

enzyme-linked immunospot reader (ImmunoSpot; CTL).

Generation of bone marrow–derived DCs. Bone marrow–derived DCs

were generated as described52. Femurs from bone marrow donor mice were

removed and bone marrow cells were isolated by flushing of the bones with

PBS; cells were then filtered through a cell strainer with a pore size of 100 mm.

Cells (2 � 106 to 2.5 � 106 cells in 10 ml) were cultured in RPMI medium

containing 10% FCS with the addition of 10% conditioned medium obtained

from X-63 cells transfected with a plasmid containing granulocyte-macrophage

colony-stimulating factor (obtained from A. Rolink, University of Basle,

Switzerland). After at least 6 d, bone marrow–derived DCs were matured with

1 mg/ml of lipopolysaccharide overnight; immature bone marrow–derived DCs

were maintained in medium containing granulocyte-macrophage colony-

stimulating factor. Bone marrow–derived DCs were used from day 7 to day 9.

Transgenic T cell proliferation and polarization. For in vitro proliferation of

transgenic T cells, spleens were collected from naive 2d2 mice and CD4+ T cells

were purified with BD-IMag magnetic beads (BD Pharmingen). The purity of

T cell isolation was verified by flow cytometry. The 2d2 T cells (1 � 105) were

cultured in a 96-well plate together with immature or mature bone marrow

DCs (3 � 103 to 10 � 103). Before coculture, bone marrow DCs were pulsed for

3 h with 1 mg/ml of MOG(35–55) in RPMI medium, then were washed and

were irradiated with 2,000 rads. Unpulsed DCs were used as a control, as were

T cells cultured alone. Cells were incubated for 4 d and [3H]thymidine was

added for the last 18 h of culture. For in vivo proliferation of TCR-transgenic

cells, mice were injected intravenously with 25 � 106 2d2 cells labeled with

10 mM carboxyfluorescein diacetate succinimidyl diester (Invitrogen–Molecular

Probes) and were immunized subcutaneously by bilateral flank injection of

100 mg MOG(35–55) emulsified in CFA. After 4 d, mice were killed and lymph

node cells were isolated and stained with antibody to the MOG transgenic TCR

(anti-Va3.2) before cytofluorometry.

Note: Supplementary information is available on the Nature Immunology website.

ACKNOWLEDGMENTSWe thank H. Hofstetter (University of Wurzburg, Wurzburg, Germany) andV. Woertmann (University of Zurich, Zurich, Switzerland) for technicalassistance; and E. Saller and T. Buch (University of Zurich, Zurich, Switzerland)for critical review of this manuscript. Supported by the Swiss National ScienceFoundation (B.B.), the National Center for Competence in Research (B.B.), theSwiss Multiple Sclerosis Society (B.B.), the Hertie Foundation (B.B. & M.P.),Serono Pharmaceuticals Geneva (B.B.), the Center for Neuroscience Research inZurich (I.G.), Roche Research Foundation (EU) and the National MultipleSclerosis Society (Harry Weaver Neuroscience scholar; B.B.).

AUTHOR CONTRIBUTIONSI.G. did all the experiments unless stated otherwise; E.U. helped with thegeneration of experimentation and analysis of bone marrow–chimeric mice;K.W. and M.P. analyzed the histopathological data; and B.B. and I.G. designedall the experiments and prepared the manuscript.

COMPETING INTERESTS STATEMENTThe authors declare that they have no competing financial interests.

952 VOLUME 7 NUMBER 9 SEPTEMBER 2006 NATURE IMMUNOLOGY

A R T I C L E S©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

Published online at http://www.nature.com/natureimmunology/

Reprints and permissions information is available online at http://npg.nature.com/

reprintsandpermissions/

1. O’Garra, A., Steinman, L. & Gijbels, K. CD4+ T-cell subsets in autoimmunity. Curr.Opin. Immunol. 9, 872–883 (1997).

2. Brand, D.D., Kang, A.H. & Rosloniec, E.F. Immunopathogenesis of collagen arthritis.Springer Semin. Immunopathol. 25, 3–18 (2003).

3. Renno, T., Krakowski, M., Piccirillo, C., Lin, J.Y. & Owens, T. TNF-a expression byresident microglia and infiltrating leukocytes in the central nervous system of mice withexperimental allergic encephalomyelitis. Regulation by Th1 cytokines. J. Immunol.154, 944–953 (1995).

4. Merrill, J.E. et al. Inflammatory leukocytes and cytokines in the peptide-induceddisease of experimental allergic encephalomyelitis in SJL and B10.PL mice. Proc.Natl. Acad. Sci. USA 89, 574–578 (1992).

5. Racke, M.K. et al. Cytokine-induced immune deviation as a therapy for inflammatoryautoimmune disease. J. Exp. Med. 180, 1961–1966 (1994).

6. van der Veen, R.C. & Stohlman, S.A. Encephalitogenic Th1 cells are inhibited by Th2cells with related peptide specificity: relative roles of interleukin (IL)-4 and IL-10.J. Neuroimmunol. 48, 213–220 (1993).

7. Chen, Y., Kuchroo, V.K., Inobe, J., Hafler, D.A. & Weiner, H.L. Regulatory T cell clonesinduced by oral tolerance: suppression of autoimmune encephalomyelitis. Science265, 1237–1240 (1994).

8. Frei, K. et al. Tumor necrosis factor a and lymphotoxin a are not required for inductionof acute experimental autoimmune encephalomyelitis. J. Exp. Med. 185, 2177–2182(1997).

9. Willenborg, D.O., Fordham, S., Bernard, C.C., Cowden, W.B. & Ramshaw, I.A. IFN-gplays a critical down-regulatory role in the induction and effector phase of myelinoligodendrocyte glycoprotein-induced autoimmune encephalomyelitis. J. Immunol.157, 3223–3227 (1996).

10. Chu, C.Q., Wittmer, S. & Dalton, D.K. Failure to suppress the expansion of the activatedCD4 T cell population in interferon g-deficient mice leads to exacerbation of experi-mental autoimmune encephalomyelitis. J. Exp. Med. 192, 123–128 (2000).

11. Ferber, I.A. et al. Mice with a disrupted IFN-g gene are susceptible to the induction ofexperimental autoimmune encephalomyelitis (EAE). J. Immunol. 156, 5–7 (1996).

12. Becher, B., Durell, B.G. & Noelle, R.J. Experimental autoimmune encephalitisand inflammation in the absence of interleukin-12. J. Clin. Invest. 110, 493–497(2002).

13. Langrish, C.L. et al. IL-23 drives a pathogenic T cell population that inducesautoimmune inflammation. J. Exp. Med. 201, 233–240 (2005).

14. Park, H. et al. A distinct lineage of CD4 T cells regulates tissue inflammation byproducing interleukin 17. Nat. Immunol. 6, 1133–1141 (2005).

15. Okamura, H. et al. Cloning of a new cytokine that induces IFN-g production by T cells.Nature 378, 88–91 (1995).

16. Dinarello, C.A. IL-18: A TH1-inducing, proinflammatory cytokine and new member ofthe IL-1 family. J. Allergy Clin. Immunol. 103, 11–24 (1999).

17. Dinarello, C.A. Interleukin-18. Methods 19, 121–132 (1999).18. Adachi, O. et al. Targeted disruption of the MyD88 gene results in loss of IL-1- and

IL-18-mediated function. Immunity 9, 143–150 (1998).19. Suzuki, N. et al. IL-1 receptor-associated kinase 4 is essential for IL-18-mediated NK

and Th1 cell responses. J. Immunol. 170, 4031–4035 (2003).20. O’Neill, L.A. & Dinarello, C.A. The IL-1 receptor/toll-like receptor superfamily:

crucial receptors for inflammation and host defense. Immunol. Today 21, 206–209(2000).

21. Kaser, A. et al. Interleukin-18 attracts plasmacytoid dendritic cells (DC2s) andpromotes Th1 induction by DC2s through IL-18 receptor expression. Blood 103,648–655 (2004).

22. Tomura, M. et al. Differential capacities of CD4+, CD8+, and CD4–CD8– T cell subsetsto express IL-18 receptor and produce IFN-g in response to IL-18. J. Immunol. 160,3759–3765 (1998).

23. Xu, D. et al. Selective expression and functions of interleukin 18 receptor on T helper(Th) type 1 but not Th2 cells. J. Exp. Med. 188, 1485–1492 (1998).

24. Yoshimoto, T. et al. IL-12 up-regulates IL-18 receptor expression on T cells, Th1 cells,and B cells: synergism with IL-18 for IFN-g production. J. Immunol. 161, 3400–3407(1998).

25. Boraschi, D. et al. Cytokines in inflammation. Joint Workshop of the DeutscheGesellschaft fur Immunologie (DGfI) and the Gruppo di Cooperazione in Immunologia(GCI) Assergi (L’Aquila, Italy), February 8–11, 1998. Eur. Cytokine Netw. 9, 205–212(1998).

26. Torigoe, K. et al. Purification and characterization of the human interleukin-18receptor. J. Biol. Chem. 272, 25737–25742 (1997).

27. Cua, D.J. et al. Interleukin-23 rather than interleukin-12 is the critical cytokine forautoimmune inflammation of the brain. Nature 421, 744–748 (2003).

28. Gran, B. et al. IL-12p35-deficient mice are susceptible to experimental autoimmuneencephalomyelitis: evidence for redundancy in the IL-12 system in the induction ofcentral nervous system autoimmune demyelination. J. Immunol. 169, 7104–7110(2002).

29. Gran, B. et al. Early administration of IL-12 suppresses EAE through induction ofinterferon-g. J. Neuroimmunol. 156, 123–131 (2004).

30. Wei, X.Q. et al. Altered immune responses and susceptibility to Leishmania majorand Staphylococcus aureus infection in IL-18-deficient mice. J. Immunol. 163,2821–2828 (1999).

31. Kinjo, Y. et al. Contribution of IL-18 to Th1 response and host defense against infectionby Mycobacterium tuberculosis: a comparative study with IL-12p40. J. Immunol. 169,323–329 (2002).

32. Santos, L.L. et al. IL-18 is redundant in T-cell responses and in joint inflammation inantigen-induced arthritis. Immunol. Cell Biol. 84, 166–173 (2006).

33. Hoshino, K. et al. Cutting edge: generation of IL-18 receptor-deficient mice: evidencefor IL-1 receptor-related protein as an essential IL-18 binding receptor. J. Immunol.162, 5041–5044 (1999).

34. Hickey, W.F. Migration of hematogenous cells through the blood-brain barrier and theinitiation of CNS inflammation. Brain Pathol. 1, 97–105 (1991).

35. Wekerle, H., Sun, D., Oropeza-Wekerle, R.L. & Meyermann, R. Immune reactivity in thenervous system: modulation of T-lymphocyte activation by glial cells. J. Exp. Biol. 132,43–57 (1987).

36. McColl, S.R. et al. Treatment with anti-granulocyte antibodies inhibits the effectorphase of experimental autoimmune encephalomyelitis. J. Immunol. 161, 6421–6426(1998).

37. Becher, B., Durell, B.G. & Noelle, R.J. IL-23 produced by CNS-resident cells controlsT cell encephalitogenicity during the effector phase of experimental autoimmuneencephalomyelitis. J. Clin. Invest. 112, 1186–1191 (2003).

38. Shi, F.D., Takeda, K., Akira, S., Sarvetnick, N. & Ljunggren, H.G. IL-18 directsautoreactive T cells and promotes autodestruction in the central nervous system viainduction of IFN-g by NK cells. J. Immunol. 165, 3099–3104 (2000).

39. Leung, B.P. et al. A role for IL-18 in neutrophil activation. J. Immunol. 167,2879–2886 (2001).

40. Becher, B., Durell, B.G., Miga, A.V., Hickey, W.F. & Noelle, R.J. The clinical course ofexperimental autoimmune encephalomyelitis and inflammation is controlled by theexpression of CD40 within the central nervous system. J. Exp. Med. 193, 967–974(2001).

41. Greter, M. et al. Dendritic cells permit immune invasion of the CNS in an animal modelof multiple sclerosis. Nat. Med. 11, 328–334 (2005).

42. Vermot-Desroches, C. et al. Monoclonal antibodies specific for the IL-18 receptor. Cell.Immunol. 236, 101–104 (2005).

43. Oppmann, B. et al. Novel p19 protein engages IL-12p40 to form a cytokine, IL-23,with biological activities similar as well as distinct from IL-12. Immunity 13, 715–725(2000).

44. Aggarwal, S., Ghilardi, N., Xie, M.H., de Sauvage, F.J. & Gurney, A.L. Interleukin-23promotes a distinct CD4 T cell activation state characterized by the production ofinterleukin-17. J. Biol. Chem. 278, 1910–1914 (2003).

45. Harrington, L.E. et al. Interleukin 17–producing CD4+ effector T cells develop via alineage distinct from the T helper type 1 and 2 lineages. Nat. Immunol. 6, 1123–1132(2005).

46. Su, S.B. et al. Essential role of the MyD88 pathway, but nonessential roles of TLRs 2,4, and 9, in the adjuvant effect promoting Th1-mediated autoimmunity. J. Immunol.175, 6303–6310 (2005).

47. Jiang, H.R. et al. IL-18 not required for IRBP peptide-induced EAU: studies in gene-deficient mice. Invest. Ophthalmol. Vis. Sci. 42, 177–182 (2001).

48. Sims, J.E. IL-1 and IL-18 receptors, and their extended family. Curr. Opin. Immunol.14, 117–122 (2002).

49. Veldhoen, M., Hocking, R.J., Atkins, C.J., Locksley, R.M. & Stockinger, B. TGFb in thecontext of an inflammatory cytokine milieu supports de novo differentiation of IL-17-producing T cells. Immunity 24, 179–189 (2006).

50. Bettelli, E. et al. Reciprocal developmental pathways for the generation of pathogeniceffector TH17 and regulatory T cells. Nature 441, 235–238 (2006).

51. Mangan, P.R. et al. Transforming growth factor-b induces development of the TH17lineage. Nature 441, 231–234 (2006).

52. Stockinger, B. & Hausmann, B. Functional recognition of in vivo processed self antigen.Int. Immunol. 6, 247–254 (1994).

NATURE IMMUNOLOGY VOLUME 7 NUMBER 9 SEPTEMBER 2006 953

A R T I C L E S©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y