Supporting InformationOboki et al. 10.1073/pnas.1003059107SI MethodsGeneration of IL-33–Deficient Mice. A BAC clone, RP23-355N22,which contains the translational start site of the IL-33 gene, wasobtained from BACPAC Resources (http://bacpac.chori.org). Thetargeting vector was constructed as described (http://www.cdb.riken.jp/arg/protocol.html). The second exon was replaced witha cassette consisting of the GFP gene and the neomycin resistancegene (neor) flanked by loxP sequences (http://www.cdb.riken.jp/arg/cassette.html). The locations of 5′ and 3′ homologous recombina-tions of the targeting vector are shown in Fig. S1A. The targetingvector was electroporated into ES cells (TT2, derived from CBA ×C57BL/6 mice). Chimera mice were obtained from two distinctclones out of three independently identified targeted ES cells andmated with C57BL/6 female mice (1). For confirmation of correcttargeting, Southern blot hybridization was performed with 5′probes located outside of the regions used in the targeting vector(Fig. S1B). As we reported (2), constitutive expression of IL-33proteins could be detected in homogenates of lungs from IL-33-sufficient, but not IL-33-deficient, mice by Western blot analysis(Fig. S1C). IL-33-deficient mice were obtained at the expectedMendelian ratio from the intercross of heterozygous mice on theC57BL/6 background (N2). IL-33-deficient mice were fertile anddid not show any gross phenotypic abnormalities under specificpathogen-free housing conditions. Genotyping of mice was per-formed by PCR of tail genomic DNAs using the following PCRprimers: ex2-F (5′-cactaagactactcagcctcag), WT-R (5′-cggtgatgctg-tgaagtctg), and KO-R (5′-gtgttctgctggtagtggtcg). The ex2-F andWT-R primers were used for detection of wild-type alleles, and theex2-F and KO-R primers were used for detection of mutant alleles.Detailed information regarding the IL-33-deficient mice (accessionnumber CDB0631K) is available at http://www.cdb.riken.jp/arg/mutant%20mice%20list.html.

Mice. C57BL/6-wild-type mice (SLC Japan), BALB/c-wild-typemice (SLC Japan), C57BL/6-Rag2-deficient mice (Taconic), andBALB/c-4-get mice (The Jackson Laboratory) were used. C57BL/6-IL-1α/β−/− mice and BALB/c-IL-4−/−IL-13−/− mice were kindlyprovided by Yoichiro Iwakura and Andrew McKenzie, respec-tively. All mice were housed under specific pathogen-free con-ditions at the National Research Institute for Child Health andDevelopment, and the animal protocols were approved by theInstitutional Review Boards of the National Research Institutefor Child Health and Development and the Institute of MedicalScience, University of Tokyo.

Contact Hypersensitivity (CHS). FITC-induced CHS and 2,4-dini-trofluorobenzene (DNFB)-induced CHS were examined as de-scribed (3, 4). Briefly, 2 d after shaving the dorsal hair withclippers, mice were sensitized with 200 μL of a 2.0% FITC iso-mer I suspension (Sigma) in a mixture of acetone and dibutylphthalate (1:1) or 25 μL of 0.5% DNFB solution (Wako) ina mixture of acetone and olive oil (1:4). Five days after sensiti-zation, the mice were challenged with 40 μL of 0.5% FITC iso-mer I solution in a mixture of acetone and dibutyl phthalate (1:1)(the left ear, 20 μL on each surface) and 40 μL of the vehiclealone (the right ear, 20 μL on each surface) or with 20 μL ofa 0.2% DNFB solution in a mixture of acetone and olive oil (1:4)(the left ear, 20 μL on the outside skin surface) and 20 μL of thevehicle alone (the right ear, 20 μL on the outside skin surface).Ear thickness was measured before and after FITC or DNFBchallenge by using an engineer’s calipers (Ozaki) by an in-vestigator who was blinded to the mouse genotypes. At 24 h after

FITC challenge, ear tissues were harvested for histologicalanalysis and measurement of the myeloperoxidase (MPO) andeosinophil peroxidase (EPO) activities in ear skin homogenates,as described below. One week after FITC challenge, sera werecollected for measurement of the FITC-specific Ig levels, asdescribed below.

Skin DCMigration.Skin DCmigration was determined as described(5, 6). Mice were treated epicutaneously with 40 μL of a 0.5%(wt/vol) FITC isomer I solution in a mixture of acetone anddibutyl phthalate (1:1) (the left ear, 20 μL on each surface) andthe vehicle alone (the right ear, 20 μL on each surface). Twenty-four hours later, submaxillary lymph nodes (LNs) were collectedseparately from both the FITC-treated left and vehicle-treatedright ears. After incubation with anti-CD16/CD32 mAb (2.4G2;BD Biosciences), LN cells were incubated with PE-anti-mouseCD11c mAb (N418; eBioscience) and APC-anti-mouse I-A/I-EmAb (M5/114.15.2; eBioscience). The proportion of FITC+ cellsamong 7-aminoactinomycin D-negative, MHC class IIhi, CD11c+

cells was determined by using a FACSCalibur (BD Biosciences).

Delayed-Type Hypersensitivity (DTH). Methyl-BSA (mBSA)-inducedCHS was examined as described (4, 7). Briefly, 200 μL of 1.25 mg/mL mBSA (Sigma) emulsified with complete Freund’s adjuvant(CFA; Difco) was injected s.c. to the backs of mice. Seven dayslater, the mice were challenged intradermally by injection of 20 μLof 10 mg/mL mBSA in PBS to one footpad and 20 μL of PBS aloneto another footpad as a control. Footpad thickness was measuredbefore and after mBSA or PBS challenge by using an engineer’scalipers (Ozaki) by an investigator who was blinded to the mousegenotypes. One week after the challenge, sera were collected formeasurement of the mBSA-specific Ig levels, as described below.

Airway Inflammation. OVA-induced airway inflammation was ex-amined as described (8, 9). For the IgE-independent protocol (OVAwith alum), mice were immunized intraperitoneally with 200 μL of0.5 mg/mL OVA (grade V; Sigma) emulsified with alum (AlumImmuject; Pierce) (1 mg/mL OVA:alum = 1:1) on days 0 and 14.Themice were then challenged intranasally with 20 μL of 10mg/mLOVA in saline or saline alone on days 28, 29, and 30. For the IgE-dependent protocol (OVA without alum), mice were treated in-traperitoneally with 200 μL of 50 μg/mLOVA in saline on days 0, 2,4, 6, 8, 10, and 12. The mice were then challenged intranasally with20 μL of 10mg/mLOVA in saline or saline alone on days 40, 43, and46. House dust mite (HDM)-induced airway inflammation was es-tablished as described elsewhere (10, 11). Mice were treated in-tranasally with 20 μL of 1.25 mg/mL HDM extract (Greer Lab-oratories) in saline or saline alone on 5 d per week for up to threeconsecutive weeks. For IL-33-induced airway inflammation, micewere treated intranasally with 20 μL of 25 μg/mL recombinanthuman IL-33 (PeproTech) in saline or saline alone once per day for3 d. For papain-induced airway inflammation, mice were treatedintranasally with 20 μL of 5 mg/mL papain (Wako) or heat-inactivated papain (at 100 °C for 15 min) in saline or saline aloneonce per day for 3 d. Twenty-four hours after the last Ag or salineinhalation, bronchoalveolar lavage (BAL) fluids and lungs werecollected for examination of the BAL cell profiles and lung his-tology, respectively. After centrifugation, the BAL cells were re-suspended in 200 μL of Hanks’ buffer, and the total cell numberand leukocyte profile were determined by using a hemocytometer(XT1800iV; Sysmex). Lung function during OVA-induced airway

Oboki et al. www.pnas.org/cgi/content/short/1003059107 1 of 12

inflammation was determined by an invasive approach (Elan Se-ries Mouse RC Site; Buxco Electronics), as described (9, 12).

Experimental Autoimmune Encephalomyelitis (EAE). MOG-inducedEAE was examined as described (13, 14). Briefly, mice were im-munized s.c. with 200 μL of 1.5 mg/mL MOG35-55 peptide(MEVGWYRSPFSRVVHLYRNGK) emulsified with CFA, whichconsisted of incomplete Freund’s adjuvant (Difco) and 5 mg/mLMycobacterium tuberculosis H37RA (Difco), by injection to oneflank on day 0 and the other flank on day 7. The mice were injectedi.v. with 200 μLof 2.5 μg/mLpertussis toxin (Allexis) on days 0 and 2.After the first MOG treatment, the severity of EAE was monitoreddaily and graded on a scale of 0–5 by an investigator whowas blindedto the mouse genotypes. The scale was as follows: 0, no disease; 1,limp tail; 2, hind limb weakness; 3, hind limb paralysis; 4, hind andfore limb paralysis; 5, moribundity and death.

Streptozocin-Induced Diabetes. Streptozocin-induced diabetes wasexamined as described (15), with minor modification. Briefly,mice were injected intraperitoneally with 4 mg/mL streptozocinin 0.05 M citrate buffer (pH 4.5) (50 mg/kg) once per day for 5 d.On days 0, 10, 17, and 24, blood was collected by cutting the tailvein, and the blood glucose levels were measured with an Accu-CHEK Aviva system (Roche Diagnostics).

Con A (ConA)-Induced Hepatitis. ConA-induced hepatitis was ex-amined as described (16). Briefly, mice were i.v. administered 2mg/mL ConA (Sigma) in saline (20 mg/kg). One day later, serawere collected, and the levels of GOT and GPT were measuredby using Transaminase CII-test Wako (Wako).

Ag-Specific LN or Spleen Cell Responses. The FITC-, mBSA-, OVA-and MOG-specific LN cell proliferative responses were examinedas described (4, 6, 14). Briefly, for FITC-specific LN cell responses,mice were treated epicutaneously with 2.0% FITC on both the leftand right ears (20 μL on one surface of each ear). Five days later,submaxillary LNs were collected. For mBSA-specific LN cell re-sponses, the backs ofmice were injected s.c. with 200 μL of 1.25mg/mLmBSA emulsified with CFA. Five days later, inguinal LNs wereharvested. For OVA-specific LN and spleen cell responses, micewere immunized intraperitoneally with 200 μL of 0.5 mg/mL OVAemulsifiedwith alum (1mg/mLOVA:alum=1:1) on days 0 and 14.Seven days later, mesenteric LNs and the spleen were collected.For MOG-specific LN cell responses, the backs of mice were s.c.injected with 200 μL of 1.5 mg/mL MOG35-55 peptide emulsifiedwith CFA. Seven days later, inguinal LNs were harvested.LN or spleen cells (4 × 105 cells per well in 96-well flat-bottom

plates) were cultured in the presence and absence of 40 μg/mLFITC, mBSA, OVA, or MOG at 37 °C for 72 h. Cell proliferativeresponses were determined by pulsing with 0.25 μCi/mL [3H]-labeled thymidine for 6 h.The intracellular cytokine profiles during mBSA- and MOG-

specific LN cell responses were determined as described (14) withminor modification. LN cells (5 × 106 cells per well in a 24 well-plate) were cultured in the presence of 40 μg/mL mBSA or MOGat 37 °C for 72 h, and then stimulated with 1 μg of ionomycin(Sigma) and 0.1 μg/mL PMA in the presence of 1 μM monensin(Sigma) for 5 h. After washing, the cells were incubated withanti-CD16/CD32 mAb (2.4G2; BD Biosciences) in FACS buffer(Hanks’ buffer containing 2% FCS) for FcR blocking on ice for15 min, and then incubated with APC-conjugated anti-mouseCD4 mAb (GK1.5; eBioscience) on ice for 30 min. After wash-ing, the cells were treated with Fix Buffer I (BD Biosciences) atroom temperature for 15 min. Then the cells were washed with0.1% saponin (Sigma) in FACS buffer and incubated with FITCanti-mouse Foxp3 mAb (FJK-16s: eBiosciences) and PE anti-mouse IL-17 mAb (TC11-18H10, BD Biosciences) at 4 °C for 30min. IL-17 and Foxp3 expressions in CD4+ T cells were analyzed

on a FACSCalibur (Becton Dickinson) by using CellQuestsoftware (Becton Dickinson).

Measurement of Cytokines. Cytokine levels in the culture super-natantsofAg-specificLNcells andBALfluidsduringOVA-inducedairway inflammationwere determinedwithmouse IFN-γ, IL-1α, IL-1β, IL-4, IL-5, IL-13, and IL-17 ELISA kits obtained from BDBiosciences or eBioscience. The level of IL-33 in colon homoge-nates (prepared as described below) was measured with an ELISAkit (BioLegend).

Measurement of MPO and EPO Activities. The levels of MPO andEPO activities in tissues were examined as described (9, 12).Briefly, tissues (ear and colon) were homogenized in a 0.5% ce-tyltrimethylammonium chloride solution (Sigma). After centri-fugation, the supernatants of the homogenates were collected.Total protein levels in the tissue homogenates were measuredwith a Bio-Rad DC protein assay kit. Recombinant human MPOandEPO (Calbiochem) were used as standard proteins. TheMPOand EPO activities per milligram of total protein in tissue ho-mogenates were calculated.

Measurement of Ag-Specific Ig Levels. Ninety-six-well ELISA plates(Nunc; 442404) were coated with 2 μg/mL FITC-OVA (6), 10 μg/mL mBSA, 10 μg/mL OVA, or 10 μg/mL MOG at 4 °C overnight.After the wells were blocked with PBS containing 10% FCS, op-timally diluted serum samples (IgG1 = 1:10,000, IgG2a = 1:100,and IgE = undiluted for FITC-specific Igs; IgG1 = 1:50, IgG2a =1:10, IgG2b = 1:50, and IgG3 = 1:2 for mBSA-specific Igs; IgE =undiluted for OVA-specific IgE; and IgG1 = 1:2 and IgG2a = 1:2for MOG-specific Igs) were applied, and the plates were incubatedat room temperature for 1 h. After washing, biotinylated anti-mouse IgG1 (A85-1; BD Biosciences), IgG2a (R19-15; BD Bio-sciences), IgG2b (R12-3; BD Biosciences), IgG3 (R40-82; BDBiosciences), or IgE (R35-118; BD Biosciences) mAb was added,followed by incubation at room temperature for 1 h. Then, afterwashing, HRP-conjugated streptavidin (BD Biosciences) wasadded, followed by incubation at room temperature for 1 h. Forenzymatic reaction, TMB substrate (KPL) was used as the sub-strate. The reaction was stopped by addition of 1 M H2SO4, andthen the absorbance at 450 nm was measured by using a platereader. Data show the absorbance value at 450 nm. The levels ofOVA-specific serum IgE were normalized by using an anti-OVAIgE mAb (TOS-2; kindly provided by Mamoru Kiniwa, TaihoPharmaceutical, Saitama, Japan) as the standard antibody (12).

Quantitative PCR.Total RNA in the colon and lung specimens wasisolated by using ISOGEN (Nippon Gene) and RNeasy Mini Kit(Qiagen). Using the isolated RNA, cDNA was obtained by RT-PCR with an iScript cDNA Synthesis Kit (Bio-Rad). Quantitativereal-time PCR was performed with THUNDERBIRD SYBRqPCR Mix (Toyobo) and an Applied Biosystems 7300 real-timePCR system. The relative gene expression was normalized againstGAPDH gene expression. PCR primers were designed as shownin Table S1.

Histology. Tissues were fixed in Carnoy’s fluid and embeddedin paraffin. Then sections were prepared and stained withhematoxylin-eosin.

Score During DSS-Induced Colitis. The severity of diseases wasscored as described (17).

Statistical Analyses. Data show the mean ± SE. Differences wereevaluated by the Kaplan–Meier test (survival), two-way ANOVAfollowed by the Holm–Sidak post hoc test (airway hypersensi-tivity), the Mann–Whitney u test (score in EAE), or the two-tailed Student’s t test (other studies, unless otherwise specified).

Oboki et al. www.pnas.org/cgi/content/short/1003059107 2 of 12

1. Murata T, et al. (2004) ang is a novel gene expressed in early neuroectoderm, but itsnull mutant exhibits no obvious phenotype. Gene Expr Patterns 5:171–178.

2. Ohno T, et al. (2009) Caspase-1, caspase-8, and calpain are dispensable for IL-33release by macrophages. J Immunol 183:7890–7897.

3. Zabel BA, et al. (2008) Mast cell-expressed orphan receptor CCRL2 binds chemerin andis required for optimal induction of IgE-mediated passive cutaneous anaphylaxis.J Exp Med 205:2207–2220.

4. Nakae S, et al. (2002) Antigen-specific T cell sensitization is impaired in IL-17-deficientmice, causing suppression of allergic cellular and humoral responses. Immunity 17:375–387.

5. Nakae S, et al. (2001) IL-1 α, but not IL-1 β, is required for contact-allergen-specific Tcell activation during the sensitization phase in contact hypersensitivity. Int Immunol13:1471–1478.

6. Suto H, et al. (2006) Mast cell-associated TNF promotes dendritic cell migration.J Immunol 176:4102–4112.

7. Nambu A, Nakae S, Iwakura Y (2006) IL-1β, but not IL-1α, is required for antigen-specific T cell activation and the induction of local inflammation in the delayed-typehypersensitivity responses. Int Immunol 18:701–712.

8. Williams CM, Galli SJ (2000) Mast cells can amplify airway reactivity and features ofchronic inflammation in an asthma model in mice. J Exp Med 192:455–462.

9. Nakae S, et al. (2007) TNF can contribute to multiple features of ovalbumin-inducedallergic inflammation of the airways in mice. J Allergy Clin Immunol 119:680–686.

10. Johnson JR, et al. (2004) Continuous exposure to house dust mite elicits chronicairway inflammation and structural remodeling. Am J Respir Crit Care Med 169:378–385.

11. Rydell-Törmänen K, Johnson JR, Fattouh R, Jordana M, Erjefält JS (2008) Induction ofvascular remodeling in the lung by chronic house dust mite exposure. Am J Respir CellMol Biol 39:61–67.

12. Nakae S, et al. (2007) Mast cell-derived TNF contributes to airway hyperreactivity,inflammation, and TH2 cytokine production in an asthmamodel in mice. J Allergy ClinImmunol 120:48–55.

13. Matsuki T, Nakae S, SudoK, Horai R, Iwakura Y (2006) Abnormal T cell activation causedby the imbalance of the IL-1/IL-1R antagonist system is responsible for the developmentof experimental autoimmune encephalomyelitis. Int Immunol 18:399–407.

14. Komiyama Y, et al. (2006) IL-17 plays an important role in the development ofexperimental autoimmune encephalomyelitis. J Immunol 177:566–573.

15. Lukic ML, Mensah-Brown E, Wei X, Shahin A, Liew FY (2003) Lack of the mediators ofinnate immunity attenuate the development of autoimmune diabetes in mice.J Autoimmun 21:239–246.

16. Tagawa Y, Sekikawa K, Iwakura Y (1997) Suppression of concanavalin A-inducedhepatitis in IFN-γ(-/-) mice, but not in TNF-α(-/-) mice: Role for IFN-γ in activatingapoptosis of hepatocytes. J Immunol 159:1418–1428.

17. Siegmund B, Lehr HA, Fantuzzi G (2002) Leptin: A pivotal mediator of intestinalinflammation in mice. Gastroenterology 122:2011–2025.

A

B

13.3 kb

16.5 kb

KpnI KpnI

2 kb 13.3 kb

DT-A KpnI

loxP loxP GFP

KpnI

neor

KpnI

16.5 kb

WT allele

Targeting vector

MT allele

Probe

Probe GFP neor

loxP loxP

MT

WT

IL-33+/+ IL-33+/-

+/+ +/+ +/+ -/- -/- -/- kDa 37

25

C

IL-33

β-actin

Fig. S1. Generation of IL-33-deficient mice. (A) IL-33 gene targeting strategy. The second exon containing a translational start codon was replaced witha tandemly arrayed promoter-less GFP gene and a floxed neomycin resistance gene (neor) (targeted allele). (B) Southern blot analysis of genomic DNA ob-tained from wild-type or mutant ES cells. The DNA probes used for Southern blot analysis are shown in A. By digestion of genomic DNA with KpnI, the probesdetected endogenous wild-type (WT; 13.3 kb) and/or targeted (MT; 16.5 kb) fragments. (C) Western blot analysis of whole-lung homogenates.

Oboki et al. www.pnas.org/cgi/content/short/1003059107 3 of 12

A

0

5

10

15

20

25

[ 3 n o i t a r o p r o c n i

R

d T ] H

0 1

x ( 3

) m

p c

OVA Med

s l e v e l e n i k o t y

C

s t n a t a n r e p u s e r u t l u c

n i 0

1

2

3

0

100

200

300

0

10

20

30

OVA Med OVA Med OVA Med

IL-17 (pg/ml) IL-4 (pg/ml) IFN- g (ng/ml)

B

s l e v e l e n i k o t y

C

) l m

/ g p (

s d i u l f L

A

B

n i 0

500

1000

0

80

160

0

20

40

0

50

100

150

Saline OVA Saline OVA Saline OVA Saline OVA

IFN- g IL-4 IL-5 IL-17

C

0

20

40

60 ) l m

/ g p (

E g I c i f i c e p s -

A

V O

Naive Saline OVA

IL-33 +/+ (n = 8) IL-33 -/- (n = 10)

IL-33 +/+ (Saline, n = 4; OVA, n = 11) IL-33 -/- (Saline, n = 5; OVA, n = 11)

IL-33 +/+ (Saline, n = 4; OVA, n = 11) IL-33 -/- (Saline, n = 5; OVA, n = 11)

Naive (n = 3)

Fig. S2. IL-33 is not required for induction of antigen-specific T and B cell function. The levels of OVA-specific IgE in sera (A) and the levels of IFN-γ, IL-4, IL-5,and IL-17 in BAL fluids (B) from the mice shown in Fig. 1 or naïve mice. (C) OVA-specific proliferative responses and cytokine secretion of spleen cells from micesensitized twice with OVA emulsified in alum. Data show the mean ± SE. No significant differences were found between IL-33+/+ and IL-33−/− groups.

0

25

50

75

10 0

0 1 x ( s l l e c

L A

B

4 )

0

20

40

60

80

0

5

10

15

20

0

2

4

6

8

0

2

4

6

8

Saline IL-33 Saline IL-33 Saline IL-33 Saline IL-33 Saline IL-33

Total Eosinophils Macrophages Neutrophils Lymphocytes

C57BL/6-Wild-type (Saline, n = 6; IL-33, n = 16) C57BL/6-Rag-2 -/- (Saline, n = 4; IL-33, n = 5)

* * * *

*

*

* *

*

*

Fig. S3. IL-33 induces airway eosinophilia independently of T and B cells. At 24 h after the last IL-33 or saline inhalation, the number of inflammatory cellsin BAL fluids from wild-type and Rag-2-deficient mice on the C57BL/6 background was determined. Data show the mean ± SE. *P < 0.05 vs. correspondingvalues for saline-treated mice.

Oboki et al. www.pnas.org/cgi/content/short/1003059107 4 of 12

0

20

40

60

0

1

2

3

4

0

1

2

3

4

0

5

10

15

20

0

10

20

30

Saline Papain Saline Papain Saline Papain Saline Papain Saline Papain

Total Eosinophils Macrophages Neutrophils Lymphocytes

0 1 x ( s l l e c

L A

B

4 )

BALB/c (Saline, n = 3; Papain, n = 5) C57BL/6 (Saline, n = 3; Papain, n = 5)

*

*†

*

*†

*

* *

*

0

20

40

60

Saline IL-33 0

10

20

30

40

Saline IL-33 0

5

10

Saline IL-33 0

2

4

6

8

Saline IL-33 0

1

2

3

4

Saline IL-33

Total Eosinophils Macrophages Neutrophils Lymphocytes

0 1 x ( s l l e c

L A

B

4 )

BALB/c (Saline, n = 5; IL-33, n = 6) C57BL/6 (Saline, n = 5; IL-33, n = 6)

*

*†

*

*†

*†

† *†

* * *

) s t i n u y r a r t i b r

A

( s l e v e l

3 3 - L I

Saline (n = 5) Papain (n = 5)

A

B

C

WT Rag2 -/- 0

0.5

1.0

1.5 *

Fig. S4. The susceptibility to papain and IL-33 differs between BALB/c and C57BL/6 mice. (A) Expression of IL-33 mRNA in the lungs after papain inhalation. (B)The number of inflammatory cells in BAL fluids from C57BL/6- and BALB/c-wild-type mice at 24 h after the last papain or saline inhalation. (C) The number ofinflammatory cells in BAL fluids from C57BL/6- and BALB/c-wild-type mice at 24 h after the last IL-33 or saline inhalation. Data show the mean ± SE. *P < 0.05 vs.corresponding values for saline-treated mice, and †P < 0.05 vs. papain- or IL-33-treated C57BL/6-wild-type mice.

Oboki et al. www.pnas.org/cgi/content/short/1003059107 5 of 12

10 0 10 1 10 2 10 3 10 4 0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

F - c e l g i S

10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

1 r G

10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

b 1 1 D

C

c 1 1

D

C

10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

t i K

- c

cεF α

R

5 X D

IL-4-EGFP IL-4-EGFP

10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

F - c e l g i S

A B BALB/c-4-get BALB/c-wild-type

Fig. S5. Eosinophils are a producer of IL-4 in BAL cells after papain inhalation. BAL cells were harvested from BALB/c-4-get mice (A), which express EGFP withIL-4, and BALB/c-wild-type (B) at 24 h after the last papain inhalation. EGFP-IL-4-expressing cells in 7-aminoactinomycin D-, CD3- and B220-negative cells weredetected by flow cytometry. Data show a representative result from two independent experiments.

Oboki et al. www.pnas.org/cgi/content/short/1003059107 6 of 12

0

40

80

120

Time after challenge (h) 0 24 48 72 In

crea

se in

ear

thic

knes

s(μ

m)

0

40

80

120 *

*

0 24 48 72

A

0

50

100

Vehicle DNFB 0

40

80

*†

Vehicle DNFB

IL-33-/-

(n = 18) IL-33+/+, +/-

(n = 19)

Vehicle FITC

IL-1α/β-/-

(n = 6)WT

(n = 6)

Vehicle FITC

IL-33+/+ (n = 8) IL-33-/- (n = 10)

WT (n = 5) IL-1α/β-/- (n = 5)

0

50

10 0

15 0

20 0

25 0

Incr

ease

in e

ar th

ickn

ess

(μm

)

Vehicle

WT (n = 10) Rag-2-/- (n = 7)

*

FITC

B

C

Incr

ease

in e

ar th

ickn

ess

(μm

)*

*†

* *

* * *

*

*†

*† *†

* * *

Fig. S6. The role of IL-33 in the development of CHS. Mice were sensitized with FITC or DNFB and then challenged with the same hapten. Ear thickness wasmeasured before and after hapten challenge. (A) FITC-induced CHS in wild-type and Rag-2−/− mice on the C57BL/6 background. (B) FITC-induced CHS in IL-33+/+

and IL-33−/− mice (Left) and in wild-type and IL-1α/β−/− mice on the C57BL/6 background (Right). (C) DNFB-induced CHS in IL-33+/+ and IL-33−/− mice (Left) and inwild-type and IL-1α/β−/− mice on the C57BL/6 background (Right) at 24 h after the challenge. Data show the mean ± SE *P < 0.05 vs. corresponding values forvehicle-treated mice, and †P < 0.05 vs. hapten-challenged wild-type mice.

Oboki et al. www.pnas.org/cgi/content/short/1003059107 7 of 12

C

B

A

MPO

(μg/

mg)

0

1

2

3

Vehicle FITC

EPO

(U/m

g)

0

1

2

3

4

Vehicle FITC

FITC

spe

cific

Ig le

vels

(OD

450)

D

Naive CHS 0

0.5

1.5

1.0

E

Vehicle FITC

[3 H] T

dR u

ptak

e (x

103 c

pm)

IL-4

(pg/

ml)

0

20

40

60

Medium FITC

0

20

40

60

Vehicle FITC

FITC

+ cel

ls

in M

HC

IIhi C

D11

c+ cel

ls

(%)

Cel

l cou

nt

FITC+ cells in MHCIIhi CD11c+ cells (%) 10 0 10 1 10 2 10 3 10 4

0 10

20

30

40

50

10 0 10 1 10 2 10 3 10 4

0 10

20

30

40

50

45.0 0.4

40.9 0.5

IL-33+/+ IL-33-/-

IL-33+/+

IL-33-/-

IL-33+/- (n = 6) IL-33-/- (n = 5)

Naive (n = 3) IL-33+/- (n = 9) IL-33-/- (n = 9)

IL-33+/+ (n = 9) IL-33-/- (n = 10)

0 2.5 5.0 7.5

10.0 12.5

Medium FITC

Naive CHS 0

0.5

1.5

1.0

IL-33+/- (n = 9) IL-33-/- (n = 9)

IgG2a

IgG1

Vehicle FITC

Fig. S7. IL-33 is not essential for hapten-specific T cell induction and activation. (A) Skin DC migration. The proportion of FITC+ cells among 7-amino-actinomycin D-negative, MHC class IIhi, CD11c+ cells was determined by flow cytometry. Shaded areas, LN cells from the vehicle-treated side; bold lines, LN cellsfrom the FITC-treated side. Representative FACS results are shown. (B) Mice were treated epicutaneously with 2.0% FITC on both the left and right ears. Fivedays later, submaxillary LNs were collected, and LN cells were cultured in the presence and absence of 40 μg/mL FITC for 72 h. FITC-specific proliferative re-sponses and IL-4 secretion of LN cells from mice sensitized with FITC are shown. (C) Histology (hematoxylin-eosin staining) of ear skin at 24 h after FITCchallenge (200×; representative data from three to five mice are shown). (D) The levels of MPO and EPO activities in ear tissue homogenates at 24 h after FITCchallenge. (E) Sera were collected 1 wk after FITC challenge. The levels of FITC-specific IgG1, IgG2a, and IgE in sera from naïve IL-33+/+ mice and FITC-challengedIL-33+/+ and IL-33−/− mice are shown. Data show the mean ± SE. No significant differences were found between the IL-33+/+ and IL-33−/− groups.

Oboki et al. www.pnas.org/cgi/content/short/1003059107 8 of 12

0 24 48 72 0 24 48 72 0

500

1000

1500

( s s e n k c i h t

d a p t o o f n i

e s a e r c n I )

μm

Time after challenge (h)

0

0.1

0.2

0.3

0.4

0.5

) l m

/ g n (

n o i t a r t n e c n o C

IL-4 IL-5

0

1

2

3 IL-17

0

2

4

6 IFN- g

Medium mBSA Medium mBSA

Medium mBSA Medium mBSA

0

5

10

15 n o i t a r o p r o c n i R

d T ]

H

3 [ 0 1

x ( 4

) m

p c

Medium mBSA

3.67

4.67 10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

3.02

3.69 10 0 10 1 10 2 10 3 10 4

0 1 0

0 1 1

0 1 2

0 1 3

0 1 4

7 1 - L I

Foxp3

0

0.1

0.2

0.3

0.4

0.5

Naive (n = 4)

0

0.5

1.0

1.5 s l e v e l g I

c i f i c e p s A

S

B

m

) 0 5 4 D

O

(

IgG1

0

0.8

1.6 IgG2a

0

1

2 IgG2b

0

0.4

0.8 IgG3

Naive DTH Naive DTH Naive DTH Naive DTH

IL-33 +/+, +/- (n = 15) IL-33 -/- (n = 14)

A

B

C

* * *

D

IL-33 +/+ (n = 5) IL-33 -/- (n = 4)

IL-33 +/+

IL-33 -/-

IL-33 -/- (n = 15)

IL-33 +/+, +/-

(n = 14)

PBS mBSA

IL-1 a/b -/- (n = 5)

WT (n = 5)

PBS mBSA

0.51

0.28

Fig. S8. The role of IL-33 in the pathogenesis of mBSA-induced DTH. (A) Footpad thickness was measured before and after mBSA or PBS challenge. (B) mBSA-specific proliferative responses and cytokine secretion of LN cells from mice sensitized with mBSA emulsified in CFA. (C) Profiles of IL-17- and Foxp3-expressingCD4+ T cells in the culture with mBSA in B. Representative FACS results are shown. (D) mBSA-specific Ig levels in sera from mice 1 wk after mBSA challenge in Aor naïve IL-33+/+ mice. Data show the mean ± SE: *P < 0.05 vs. corresponding values for WT mice.

Oboki et al. www.pnas.org/cgi/content/short/1003059107 9 of 12

Days after MOG immunization

0

1

2

3

4

10 15 20 25 30 35 0 5

IL-1α/β-/- (n = 5)WT (n = 5)

Clin

ical

sco

re

0

1

2

3

4

10 15 20 25 30 35 0 5

IL-33-/- (n = 11) IL-33+/+ (n = 10)

0

2.5

5.0

7.5

10.0

12.5

[TdR

] Inc

orpo

rtio

n (x

104 c

pm) IL-33+/+ (n = 4) IL-33-/- (n = 5)

0

0.1

0.2

0.3

0.4

0.5

Con

cent

ratio

n (n

g/m

l)

0

1

2

3

0

2

4

6

8

Med MOG 0

0.1

0.2

0.3

0.4

0.5

IL-1

7

Foxp3 10 0 10 1 10 2 10 3 10 4

10 0

10 1

10 2

10 3

10 4

3.82

8.85

10 0 10 1 10 2 10 3 10 4

10 0

10 1

10 2

10 3

10 4

4.62

11.1

0

0.5

1.5

MO

G-s

peci

fic Ig

leve

l (O

D45

0)

0

0.5

1.5

2.5

Naive EAE

Naive (n = 4)

Naive EAE

1.0 2.0

1.0

IL-4 IL-5 IL-17 IFN-γ

Med MOG Med MOG Med MOG Med MOG

IL-33+/+ IL-33-/-

IgG2a IgG1

*

A

B

C

D IL-33-/- (n = 11) IL-33+/+ (n = 10)

1.33 1.60

Fig. S9. The role of IL-33 in the development of EAE. (A) The severity of MOG-induced EAE. (B) MOG-specific proliferative responses and cytokine secretion ofLN cells from mice sensitized with MOG emulsified in CFA. (C) Profiles of IL-17- and Foxp3-expressing CD4+ T cells in the culture with MOG in B. RepresentativeFACS results are shown. (D) MOG-specific Ig levels in sera from mice 33 d after the first MOG immunization in A or naïve IL-33+/+ mice. The levels of MOG-specific IgG1 and IgG2a were measured by ELISA. Data show the mean ± SE: *P < 0.01 (Mann–Whitney u test). No significant differences were found betweenthe IL-33+/+ and IL-33−/− groups.

levelesoculg

doolB

)Ld/gm(

0100

200

300

400

0 16 24Time (day)

levelT

OG 01

x(3

)lm/

Une

mraK

Naive Con A Con ANaive

levelTP

G 01x(

3)l

m/U

nemra

K

8

A

B

0

5

10

15

0

1

2

3

4

5

Vehicle Streptozocin

IL-33-/-IL-33+/+, +/- (n = 5)

(n = 5)(n = 12)(n = 15)

Naive (n = 4)IL-33+/+ (n = 18)IL-33-/- (n = 18)

Fig. S10. The role of IL-33 in the development of T/NKT cell-dependent streptozocin-induced diabetes or Con A-induced hepatitis. (A) The levels of bloodglucose in mice during streptozocin-induced diabetes. (B) GOT and GPT activity levels in sera from mice during Con A-induced hepatitis. Data show the mean ±SE. No significant differences were found between the IL-33+/+ and IL-33−/− groups.

Oboki et al. www.pnas.org/cgi/content/short/1003059107 10 of 12

0

5

10

15

0

10 0

20 0

30 0

40 0

50 0 0

500

1000

1500

0

1000

2000

3000 0

20

40

60

80

0

5

10

15

20 0

20

40

60

80

0

50

100

150

200

Naive

IL-33+/+ IL-33-/-

*

*

*

*

* * *

*

*

* *

*

*

*

*†

*†

Day 15 Day 8

0

0.5

1.0

IL-3

3 (n

g/m

g)

*

* † †

IL-33+/+

IL-33-/-

Naive Day 8 Day 15

A B

n = 7 8 10 7 7 6

Expr

essi

on le

vel (

Arb

itrar

y un

its)

IL-1β

TNF

MIP2

KC

IL-1β

TNF

MIP2

KC

Fig. S11. IL-33 is involved in the expression of neutrophil chemoattractant factors during DSS-induced colitis. The colon was collected from mice on day 8 or 15during 3.0% DSS-induced colitis or from naïve mice. (A) The levels of IL-1β, TNF, MIP2, and KC expression in the colon were determined by quantitative PCR. (B)The levels of IL-33 in the colon homogenates were determined by ELISA. Data show the mean ± SE: *P < 0.05 vs. corresponding values for naive mice, and †P <0.05 vs. DSS-treated IL-33+/+ mice.

0

5

10

15

) l m

/ g n (

6 - L I

Med IL-33 LPS LPS IL-33

* A

0

0.5

1.0

1.5 ST2

0

0.5

1.0

1.5

IL-33 +/+ IL-33 -/-

TLR4

) s t i n u y r a r t i b r

A

( n o i s s e r p x E

B

IL-33 -/- (n = 10) IL-33 +/+ (n=10)

0

25

50

75

100

0 2 4 6 8 1 0 0

25

50

75

100

) %

( l a v i v r u S

0 2 4 6 8 Day

: 2.5 mg/kg

: 15 mg/kg

Re-challenge IL-33 -/- (n = 10) IL-33 +/+ (n=9)

: 0.5 mg/kg

: 20 mg/kg

Tolerance protocol

Day

) %

( l a v i v r u S

C D

protocol

Fig. S12. IL-33 is not essential for secondary responses or tolerance to LPS. (A) TGC-induced peritoneal macrophages from C57BL/6 mice were cultured withplain medium or medium containing IL-33, LPS, or IL-33 + LPS for 48 h. IL-6 levels in the culture supernatants were determined by ELISA. Data show the mean +SE (n = 3). *P < 0.05 vs. medium alone, IL-33, and LPS stimulation. (B) The expressions of ST2 and TLR4 in peritoneal macrophages from IL-33+/+ and IL-33−/− micewere determined by quantitative PCR. Data show the mean ± SE (n = 3). (C) The survival ratios in IL-33+/+ and IL-33−/− mice, which were first sensitized witha low dose of LPS (2.5 mg/kg; black arrow), after treatment with a lethal dose of LPS (15 mg/kg; red arrow). (D) The survival ratios in IL-33+/+ and IL-33−/− mice,which were first made tolerant to LPS by repeated injection of a low dose of LPS (0.5 mg/kg; black arrowhead), after treatment with a lethal dose of LPS (20mg/kg; green arrow). No significant differences were found between the IL-33+/+ and IL-33−/− groups (B–D).

Oboki et al. www.pnas.org/cgi/content/short/1003059107 11 of 12

Table S1. Primer design

Genes Forward (5′ - 3′) Reverse (5′ - 3′)

GAPDH CCCACTCTTCCACCTTCGATG AGGTCCACCACCCTGTTGCTIL-1β CAACCAACAAGTGATATTCTCCATG GATCCACACTCTCCAGCTGCAIL-33 CAGGCCTTCTTCGTCCTTCAC TCTCCTCCACTAGAGCCAGCTGTNF GCCTCCCTCTCATCAGTTCT CACTTGGTGGTTTGCTACGAKC CACGTGTTGACGCTTCCCTT TGAACGTCTCTGTCCCGAGCMIP-2 AACTGACCTGGAAAGGAGGAGC ACTCTCAGACAGCGAGGCACATST2 TCAACCGCCTAGTGAACACACC CAAAGCCCAAAGTCCCATTCTCTLR4 CGCTTTCACCTCTGCCTTCACTACAG ACACTACCACAATAACCTTCCGGCTC

Oboki et al. www.pnas.org/cgi/content/short/1003059107 12 of 12