Modified Adenine (9-Benzyl-2-Butoxy-8-Hydroxyadenine)Redirects Th2-Mediated Murine Lung Inflammation byTriggering TLR71

Alessandra Vultaggio,* Francesca Nencini,* Paul M. Fitch,¶ Lucia Filì,* Laura Maggi,*Paola Fanti,‡ Annick deVries,� Enrico Beccastrini,* Francesca Palandri,* Cinzia Manuelli,§

Daniele Bani,¶ Maria Grazia Giudizi,* Antonio Guarna,‡ Francesco Annunziato,*Sergio Romagnani,* Enrico Maggi,2* Sarah E. M. Howie,� and Paola Parronchi*

Substitute adenine (SA)-2, a synthetic heterocycle chemically related to adenine with substitutions in positions 9-, 2-, and 8- (i.e.,9-benzyl-2-butoxy-8-hydroxyadenine), induces in vitro immunodeviation of Th2 cells to a Th0/Th1 phenotype. In this article, weevaluate the in vivo ability of SA-2 to affect Th2-mediated lung inflammation and its safety. TLR triggering and NF-�B activationby SA-2 were analyzed on TLR-transfected HEK293 cells and on purified bone marrow dendritic cells. The in vivo effect of SA-2on experimental airway inflammation was evaluated in both prepriming and prechallenge protocols by analyzing lung inflam-mation, including tissue eosinophilia and goblet cell hyperplasia, bronchoalveolar lavage fluid cell types, and the functional profileof Ag-specific T cells from draining lymph nodes and spleens. SA-2 induced mRNA expression and production of proinflammatory(IL-6, IL-12, and IL-27) and regulatory (IL-10) cytokines and chemokines (CXCL10) in dendritic cells but down-regulated TGF-�.Prepriming administration of SA-2 inhibited OVA-specific Abs and Th2-driven lung inflammation, including tissue eosinophiliaand goblet cells, with a prevalent Foxp3-independent regulatory mechanism. Prechallenge treatment with SA-2 reduced the lunginflammation through the induction of a prevalent Th1-related mechanism. In this model the activity of SA-2 was route-inde-pendent, but adjuvant- and Ag dose-dependent. SA-2-treated mice did not develop any increase of serum antinuclear autoanti-bodies. In conclusion, critical substitutions in the adenine backbone creates a novel synthetic TLR7 ligand that shows the abilityto ameliorate Th2-mediated airway inflammation by a complex mechanism, involving Th1 redirection and cytokine-mediatedregulation, which prevents autoreactivity. The Journal of Immunology, 2009, 182: 880–889.

I nduction and exacerbation of allergic lung inflammation havebeen associated with respiratory viral infections (1, 2), butthe role of viral Ags in allergic Th2-mediated inflammation

remains to be fully elucidated. The immune response against vi-ruses starts with the recognition of a variety of pathogen-associ-ated molecular patterns by a panel of endosomal TLRs, beingbroadly expressed in cells of the immune system (3). This inter-action activates common signaling pathways leading to dendriticcell (DC)3 maturation and proinflammatory cytokine and chemo-kine production, which are essential in Th1 polarization. Addition-

ally, TLR ligation also triggers a Th1-polarizing program by cell-to-cell contact up-regulating the Notch ligand Delta4 on DC,which induces IFN-� production by naive T cells (4).

The role of Th2 cells specific for common environmental aller-gens characterized by the production of IL-4, IL-5, and IL-13, butlittle or no IFN-�, in allergic inflammation is well documented.However, allergen-specific Th2 responses can be redirected bytriggering endosomal TLRs, at least in vitro, into a Th0 or even aTh1 response, which may be beneficial for the treatment of allergy(5). Indeed, the administration of the Amb a 1 allergen conjugatedto a CpG-containing oligodeoxynucleotide (CpG-ODN) in a spe-cific prechallenge protocol improved clinical symptoms in allergicsubjects, and its effect was related to the redirection of allergen-specific Th2 responses to a less polarized Th0/Th1 phenotype atboth the circulatory and mucosal levels (6, 7). Natural and syn-thetic TLR ligands related to viral structures have been described,including nucleic acid-related sequences triggering endosomalTLR3 (dsRNA), TLR9 (unmethylated CpG-ODNs), and TLR7/8(ssRNA) (3). Heterocycles such as guanosine derivatives and imi-dazoquinolines have been identified as synthetic ligands of TLR7exclusively or of both TLR7 and TLR8, as well as inducers of Th1redirection in vivo and in vitro (8–11).

*Center for Research, Transfer and High Education (DENOThe) and †Department ofInternal Medicine, ‡Department of Organic Chemistry “Ugo Schiff”, §Department ofDermatological Sciences, and ¶Department of Anatomy, Histology and ForensicMedicine, University of Florence, Florence, Italy; and �Immunobiology Group, Med-ical Research Council Centre for Inflammation Research, Queen’s Medical ResearchInstitute, University of Edinburgh, Edinburgh, United Kingdom

Received for publication June 17, 2008. Accepted for publication November 3, 2008.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This research was supported by funds provided by the Italian Ministry of Education,the Italian Ministry of Health, European Union projects (SENS-IT-IV, FP6-LSBH-CT-2006-018861; INNOCHEM, FP6-LSHB-CT-2005-518167), and the NormanSalvesen Emphysema Research Trust (U.K.).2 Address correspondence and reprint requests to Dr. E. Maggi, ImmunoallergologyUnit, Center for Research, Transfer and High Education (DENOThe), University ofFlorence, Policlinico di Careggi, Viale Morgagni, 85, 50134 Florence, Italy. E-mailaddress: e.maggi@dmi.unifi.it3 Abbreviations used in this paper: DC, dendritic cell; ANA, antinuclear autoanti-body; BALF, bronchoalveolar lavage fluid; BMDC, bone marrow-derived DC; CpG-ODN, CpG-containing oligodeoxynucleotide; hd, high dose; i.t., intratracheal; MNC,

mononuclear cell; SA, substitute adenine; SA-1, 2-butoxy adenine; SA-2, 9-benzyl-2-butoxy-8-hydroxyadenine; Treg, regulatory T.

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

The Journal of Immunology

www.jimmunol.org

We recently showed that a synthetic heterocycle chemically re-lated to adenine, substitute adenine (SA)-2 (9-benzyl-2-butoxy-8-hydroxyadenine), which induces TLR7-mediated proinflammatoryand regulatory molecules by human monocytes/macrophages andplasmacytoid DCs, can cause in vitro immunodeviation of aller-gen-specific Th2 cells from allergic patients to a Th0/Th1 pheno-type (11).

In this article, we show that SA-2 triggers murine TLR7 andexerts an immunodeviating effect associated with some regulatorymechanisms not involving Foxp3� regulatory T cells in acutemodels of Th2-mediated allergic lung inflammation. Of note, thein vivo administration of SA-2 did not induce a humoral responseto nuclear autoantigens.

Materials and MethodsAbs and reagents

Anti-murine CD4 FITC, CD11c allophycocyanin, and CD25 PE, as well asanti-I-A�/I-E� FITC Abs were purchased from BD Biosciences. Anti-CCR3-PE and anti-Foxp3-allophycocyanin Abs were purchased fromR&D Systems and eBioscience, respectively. GM-CSF was from Pepro-Tech. Synthetic heterocycle related to adenine SA-2 and its inactive analog(2-butoxy adenine, SA-1) were obtained as described (12). Resiquimod(R848), a TLR7/8 ligand (13), was purchased from InvivoGen. OVA waspurchased from Sigma-Aldrich. Endotoxin levels in modified adenineswere �0.003 EU/ml as measured by the Limulus amebocyte lysate(BioWhittaker).

Mice

Pathogen-free 7-wk-old C57BL/6J female mice were purchased fromCharles River Laboratories and kept under standard housing conditions. Allanimal studies were performed according to institutional national guide-lines and local animal ethics regulations.

Murine models of acute allergic inflammation of the lung andSA-2 treatment

Mice were systemically sensitized by i.p. injection of 10 �g of OVA ad-sorbed in 2.25 �g of aluminum hydroxide (Imject Alum; Pierce) in 100 �lon days 0 and 14. Negative controls were sham sensitized with aluminumhydroxide following the same protocol. All animals were challenged bytwo intratracheal (i.t.) administrations of OVA (10 �g in 50 �l of PBS,sensitized mice) or PBS alone (negative controls) on days 28 and 32, asdescribed (14). Additionally, mice from the sensitized group received SA-2(50 �g in 100 �l of 20% DMSO in saline) or its vehicle only (100 �l of20% DMSO in saline) via i.p. injection 2 days before (day �2) and 4 daysafter (day 4) the first OVA sensitization (protocol A, Fig. 1A). A secondprotocol evaluating the effect of SA-2 on already established Th2 responsein primed mice included a single i.p. dose (50 �g in 100 �l of 20% DMSOin saline) of SA-2 24 h before (day 27) the Ag challenge (protocol B,Fig. 1B).

In some experiments both protocols were performed by immunizinganimals with 300 �g of OVA. Finally, SA-2 alone (and SA-1 as negativecontrol) was i.p. administered to nonsensitized mice, and cytokines (IL-6,IL-10, IL-12, IL-27, IFN-�) and chemokines (CXCL10) in serum wereevaluated before and 2 and 12 h after DMSO and SA-2 administration bycommercial ELISA (R&D Systems). Detection limits in serum were 7pg/ml for IL-6, 16 pg/ml for IL-10, 8 pg/ml for IL-12, 8 pg/ml for IL-27,25 pg/ml for IFN-�, and 31 pg/ml for CXCL10.

Bronchoalveolar lavage

Mice were killed by i.p. injection of pentobarbitone 72 h after the last OVAexposure, and bronchoalveolar lavage was performed as described (14).

Stimulation of TLR-transfected HEK293 cells and reporter assay

HEK293 cells (from American Type Culture Collection/LGC Promochem)were transiently transfected by Nucleofection (Amaxa) with murine TLR7or empty plasmids (InvivoGen) (5 �g of plasmid/1 � 106 cells/transfec-tion) and ELAM-1 promoter NF-�B luciferase or ELAM-1 promoterNF-�B luciferase alone reporter plasmid (pNiFty; InvivoGen) (1/1) andplated in 48-well flat-bottom plates at 1 � 105/ml in Eagle’s MEM (In-vitrogen) supplemented with 5% FCS (HyClone/Thermo Scientific). Eigh-teen hours after transfection, cells were stimulated with or without R848,SA-1, and SA-2 (5, 2.5, and 1.25 �g/ml) for a further 18 h. At the end of

the culture, luciferase activity was determined in cell lysates by a luciferaseassay system (Promega).

Flow cytometry

Spleen cells (106) were stained with anti-CD4 FITC, anti-CD25 PE, andanti-Foxp3 allophycocyanin, and bronchoalveolar lavage cells (105) werestained with anti-CD4 FITC, anti-CCR3 PE, anti-I-��/�-�� FITC, and anti-CD11c allophycocyanin. Stained cells were analyzed on a BD LSRII flowcytometer using the FACSDiva software (BD Biosciences), as described(15).

Histological analysis of lung tissue

Lungs were perfused with PBS and for histological examination were in-flated with and fixed in 4% neutral buffered formalin before paraffin em-bedding. Three-micrometer sections were stained with H&E for assessmentof inflammation. Inflammation was most noticeable in the perivascularcompartment immediately around blood vessels and in the peribronchiolarcompartment. For this reason, severity was evaluated in all experiments at�200 magnification, as described (14), on an increasing score of 1 to 4according to the number of cells surrounding blood vessel walls (1, nodetectable infiltration; 2, �20 cells; 3, �100 cells; and 4, �100 cells) andbronchiolar epithelium (1, no detectable infiltration; 2, �5 cells; 3, �10cells; and 4, �10 cells); in some experiments inflammation in the alveolarwalls (1, normal; 2, focal cellular expansion of the alveolar walls by 2 or3 cells; 3, by 4 or 5 cells; 4, �5 cells) and the bronchiolar epithelium (1,no cells, 2, �5 cells; 3, �10 cells; and 4, �10 cells) was also evaluated.Histological scores were performed by observers blinded to experimentaldetails (S.E.M.H. and A.V.)

Ag-specific T cell proliferation

Ag (OVA)-specific proliferation of T cells was performed by using mono-nuclear cells (MNCs) derived from mediastinal lymph nodes and spleenswith increasing doses of OVA, as described (14). Culture supernatantswere assayed for IL-5, IL-13 and IFN-� content by commercially availableELISAs (R&D Systems). Detection levels were 31 pg/ml for IL-5, 15pg/ml for IL-13, and 15 pg/ml for IFN-�.

Generation of bone marrow-derived DCs (BMDCs)

BMDCs were prepared according to Boonstra et al. (16). Briefly, femursand tibias of C57BL/6J mice were removed, the marrow was flushed outwith a syringe, and erythrocytes were lysed with 0.15 M NH4Cl. Afterwashing, 0.4 � 106 cells/ml were placed in 24-well plates (Corning) in

FIGURE 1. Prepriming and prechallenge protocols of SA-2 treatment.A, C57BL/6J mice were i.p. sensitized and i.t. challenged, as described inMaterials and Methods. OVA-sensitized mice were i.p. injected with 50 �gof SA-2 in 20% DMSO (SA-2/OVA mice) or 100 �l of 20% DMSO insaline (DMSO/OVA mice) at �2 days and at 4 days. At sacrifice (35 days)BALF was taken for cellular analysis, lungs were removed for histologyand mRNA expression, spleens and mediastinal lymph nodes were used forin vitro proliferation and cytokine production, and blood samples wereused for measurement of total IgE and OVA-specific Abs. B, C57BL/6Jmice were sensitized and challenged with OVA, as described in A. Twenty-four hours before the first challenge, OVA-sensitized mice were i.p. in-jected with 50 �g of SA-2 in 20% DMSO in saline or vehicle. At sacrifice(35 days) all the procedures described in A were applied.

881The Journal of Immunology

FIGURE 2. SA-2 triggers murine TLR7 and up-regulates proinflammatory and regulatory cytokines and chemokines by DCs. A, NF-�B activation ofHEK293 cells cotransfected with murine TLR7 (gray columns) or empty plasmid (hatched columns) and NF-�B luciferase reporter plasmid (expressed asfold increase vs that of HEK293 cells transfected with NF-�B luciferase reporter plasmid) upon stimulation with R848 or increasing doses of SA-2, asdescribed in Materials and Methods. The results of one representative of three consecutive experiments are reported. B, mRNA expression of Jagged1 andDelta4 ratios in BMDCs cultured in vitro for 3 h with 2.5 �g/ml of SA-2 or SA-1 or 6 �M R848, as described in Materials and Methods. The mean values(SEM) of three experiments are reported. �, p � 0.05; ��, p � 0.001. C, mRNA expression (left panels) and supernatant protein levels (right panels)of IL-6, IL-10, IL-12, IL-27, and CXCL10 of BMDCs cultured in vitro with medium, SA-2, SA-1, or R848, as described in Materials and Methods. Dataare expressed as mean values (SEM) of gene/ubiquitin ratios or of proteins (above the levels of cultures with medium alone) in culture supernatants ofseven separate experiments. Statistical significance between SA-2- and SA-1-treated BMDCs is indicated as follows: �, p � 0.05; ��, p � 0.02; ���, p �0.01; ∧, p � 0.001.

882 ADJUVANT EFFECTS OF MODIFIED ADENINES IN VIVO

complete medium plus 10 ng/ml recombinant murine GM-CSF (Pepro-Tech). Cultures were fed with fresh medium plus GM-CSF every 3 days.At day 8, �75% of cells consistently expressed CD11c, with the recoverybeing 50% of the cultured cells. After washings, BMDCs were culturedin vitro for 3 h (mRNA detection) and 3, 12, and 72 h (protein detection)with 2.5 �g/ml SA-2 or SA-1 (as negative control), or 6 �M R848 (aspositive control), or medium alone in the absence or presence of 50 �g/mlOVA. At the end of the cultures, cells were collected for total RNA ex-traction and supernatants were assayed for their IL-6, IL-10, IL-12, IL-27,and CXCL10 content by commercially available ELISA (R&D Systems).Detection limits of the kits were 15, 7.5, 19, 7.5, and 15 pg/ml,respectively.

Quantitative mRNA analysis

Total lung RNA from snap-frozen mouse lungs and BMDCs were ex-tracted using TRIzol reagent (RNAwiz; Invitrogen). Real-time quantitative

PCR was performed on an ABI PRISM 7700 sequence detector (AppliedBiosystems) with Applied Biosystems predesigned TaqMan gene expres-sion assays and reagents, according to the manufacturer’s instructions.

OVA-specific Abs

Detection of total serum IgE and OVA-specific IgE, IgG1, and IgG2a wasperformed by ELISA as described (14).

Antinuclear autoantibodies (ANAs)

Both homogeneous and speckled patterns of ANAs were detected by in-direct immunofluorescence using Hep-2 substrate slides (Scimedx) accord-ing to the manufacturer’s instructions, except for the secondary Ab(FITC-conjugated F(ab�)2 goat anti-mouse IgG, Santa Cruz Biotechnology)at 1/20 serum as starting dilution. Slides were scored by an observerblinded to the experimental conditions as described (17).

FIGURE 3. Prepriming SA-2 treatment inhibits lung inflammation. A, Left panels, Representative H&E-stained lung tissue sections (�40) are shown;right panel, results are reported as mean values (SEM) of lung inflammation scores of 10 animals/group. �, p � 0.005; ��, p � 0.02. B, Left panel, BALFcell analysis of SA-2/OVA, DMSO/OVA, and PBS/PBS mice is reported. Right panel, Morphological cytometric parameters and CCR3 and CD4expression in BALF cells from the same animals are shown. A representative experiment out of 10 is shown. �, p � 0.0025; ��, p � 0.005; ���, p � 0.001.C, mRNA expression of Muc5b and GOB-5 in lung from the same animals as in A and B is shown. �, p � 0.05; ��, p � 0.005.

Table I. Cytokine and chemokine serum levels upon SA-2 in vivo administrationa

Treatment andSample Timing

Serum Level (pg/ml)

IFN-� CXCL10 IL-10 IL-27

DMSOBefore �25 �31 �16�2 h 71 15 47 23 �16 �8�12 h �25 �31 �16 �8

SA-2Before 82 70 �31 �16 �8�2 h 5,804 315*** 10,401 2,266** 653 151* 337 117**�12 h �25 1,449 159*** �16 �8

a Nonimmunized mice (n � 5) were administered with SA-2 or DMSO by i.p. injection as described in Materials andMethods. Data are presented as the mean values (SEM). Significant difference between serum levels from SA-2- and DMSO-treated mice are indicated: �, p � 0.05; ��, p � 0.02; ���, p � 0.001.

883The Journal of Immunology

Statistical analyses

Results are presented as means SEM. Statistical analysis was performedusing Student’s t test. Values of p �0.05 were considered to be significant.

ResultsSA-2 triggers TLR7 and up-regulates proinflammatory cytokinesand chemokines on murine DCs

We first determined whether SA-2 was recognized by murineTLR7 by using TLR7-transfected HEK293 cells with or without areporter gene, as described (11). SA-2 and R848, but not SA-1,directly activate TLR7-transfected HEK293 cells as shown by theup-regulation of NF-�B-dependent luciferase expression, thus in-dicating that SA-2 acts as a TLR7L on murine cells (Fig. 2A).SA-1 (not shown), SA-2, and R848 failed to induce any responseon empty vector-transfected HEK293 cells (Fig. 2A).

To evaluate the effect of SA-2 on cells of innate immunity,purified BMDCs were cultured in the absence or presence of SA-2(or SA-1) or R848. After 3 h of culture, SA-2, but not SA-1,decreased Jagged1/Delta4 mRNA ratios in a comparable way toR848-cultured cells (Fig. 2B). Similarily, SA-2 and R848, but notSA-1, up-regulated on BMDCs the mRNA expression of proin-flammatory (IL-6 and IL-27), regulatory (IL-10), and Th1-associ-ated (IL-12, CXCL10) molecules (Fig. 2C). Of note, under theseconditions, mRNA expression of the Th2-associated chemokinesCCL17 and CCL22 was not affected, whereas TGF-� mRNA was

inhibited (data not shown). When analyzed at the protein level,SA-2 induced a time-related production of cytokines or chemo-kines by BMDCs. While IL-6 increased early (3 h) and peaked at72 h, the other molecules were barely detectable at 3 h and peakedat 12 h (IL-27) or 72 h (IL-10, IL-12, CXCL10) (Fig. 2C). The invitro synthesis of these molecules (except for IL-6) increasedmarkedly when OVA (50 �g/ml) was added to TLR7 ligands, thusindicating a synergistic effect (data not shown).

Finally, when SA-2, but not SA-1, were i.p. administered tonaive mice, a transient peak of serum IL-6, IL-10, and IL-27 wasseen at 2 h after treatment, which disappeared in the subsequent10 h (Table I and data not shown). Under the same conditions,IL-12 was never detectable in serum of SA-2-treated mice, whileIFN-� and CXCL10 reached extraordinary high levels at 2 h aftertreatment (Table I). Interestingly, CXCL10 was the only moleculestill detectable at 12 h after drug administration (Table I).

SA-2 treatment inhibits the development of Th2-mediated lunginflammation

To determine whether SA-2 treatment could modulate Th2-medi-ated lung inflammation, C57BL/6J mice were sensitized and i.t.challenged with OVA and i.p. treated with SA-2 at �2 days and at4 days (SA-2/OVA mice) according to protocol A (Fig. 1A). In thepositive control group SA-2 was replaced by SA-2 vehicle(DMSO/OVA mice), whereas in the negative control group OVA

FIGURE 4. Prepriming SA-2 treatment inhibits the development of Th2 responses. A, mRNA expression of cytokines (IL-4, IL-5, IL-13, IFN-�) andchemokines (CCL11, CCL17, CCL22, CXCL10) in lung tissue of mice treated with protocol A are shown. mRNAs were detected by quantitative RT-PCR,as described in Materials and Methods. Data are expressed as gene/ubiquitin ratio (mean values SEM, five animals/group). �, p � 0.01; ��, p � 0.02;���, p � 0.005. B, IL-5 and IL-13 production by MNCs from mediastinal lymph nodes of SA-2/OVA and DMSO/OVA mice stimulated with increasingdoses of Ag (OVA 25, 50, 100, 200 �g/ml) is reported. Data are expressed as the mean values (SEM) of protein levels in culture supernatants from fivemice/group. �, p � 0.05; ��, p � 0.02; ���, p � 0.001. C, Serum OVA-specific IgE (left panel) and IgG1 (right panel) Abs in SA-2/OVA, DMSO/OVA,or in control (PBS/PBS) mice (five mice/group) are shown. Statistical significance between the DMSO- and SA-2-treated groups is indicated as follows:�, p � 0.05; ��, p � 0.01; ���, p � 0.001. The level of significance between the PBS and the SA-2-treated groups is indicated as follows: �, p � 0.05;��, p � 0.01; ���, p � 0.001.

884 ADJUVANT EFFECTS OF MODIFIED ADENINES IN VIVO

was replaced by PBS (PBS/PBS mice). Both peribronchiolar andperivascular infiltration scores evaluated on H&E-stained tissuesections were significantly ( p � 0.02 and p � 0.005, respectively)decreased in lungs of SA-2/OVA mice compared with DMSO/OVA mice (Fig. 3A). Moreover, SA-2/OVA mice showed a sig-nificant reduction in the proportions and absolute values of totalcells ( p � 0.0025), lymphocytes ( p � 0.005), and eosinophils( p � 0.001) in BALF after two OVA challenges compared withDMSO/OVA mice (Fig. 3B and data not shown). This effect wasaccompanied by a reduction of lung mRNA expression of GOB-5(CLCA2) and Muc5b (mucin 5B), typically associated with mucusproduction and goblet cell hyperplasia (Fig. 3C). After the secondchallenge, mRNA expression of both Th2-associated (IL-4, IL-5,IL-13, CCL11, CCL17, CCL22) molecules and IL-10 and TGF-�was significantly reduced in lungs of SA-2/OVA compared withDMSO/OVA mice, whereas IFN-� and CXCL10 mRNAs wereunaffected (Fig. 4A and data not shown). Even though MNCs frommediastinal lymph nodes from SA-2/OVA and DMSO/OVA miceexhibited comparable proliferative responses to OVA, 3 day-cul-ture supernatants of these cells from SA-2/OVA mice showed asignificant reduction of IL-5 and IL-13 compared with those ofcontrol mice, without any change in IFN-� secretion (Fig. 4B anddata not shown). At the Ab level, Th2-dependent total IgE (1116 284 vs 411 77 ng/ml, p � 0.05), OVA-specific IgE, and IgG1serum levels (Fig. 4C) were significantly decreased, while Th1-dependent OVA-specific IgG2a levels remained unaffected (datanot shown).

Finally, we sought to investigate whether regulatory T (Treg)cells were involved in SA-2-induced prevention of Th2-mediatedlung inflammation. No differences in mRNA expression of Foxp3,EBI3 (Epstein-Barr virus-induced gene 3), and IL-12p35 in thelungs from SA-2- vs DMSO-treated mice were found (Fig. 5A).Moreover, when an earlier time point was analyzed (10 days afterSA-2 treatment), IL-10 and TGF-� mRNA expressions (data notshown), as well as the proportions of CD4�CD25highFoxp3�

spleen T cells, were comparable in SA-2/OVA and DMSO/OVAmice (Fig. 5B).

SA-2 down-regulates the lung inflammation in primed mice byimproving tissue and systemic Th1-oriented responses

To determine whether SA-2 treatment could modulate establishedTh2 sensitization, we evaluated the previously described protocolB (Fig. 1B). A significant ( p � 0.05) reduction of peribronchiolar(but not of perivascular, bronchiolar, or alveolar wall) inflamma-tion score was found in the lung of SA-2 (OVA/SA-2) mice com-pared with control-treated (OVA/DMSO) mice (Fig. 6A, leftpanel). The proportions of inflammatory cells in BALF were notreduced in OVA/SA-2 mice (Fig. 6A, right panel). At 6 h afterOVA challenge, a significant ( p � 0.05) increase of IFN-� andCXCL10 mRNAs and a parallel decrease (not significant) of Th2-associated cytokines (IL-4, IL-5, and IL-13) and chemokines(CCL11, CCL17, and CCL22) were seen in lungs from OVA/SA-2mice compared with controls (Fig. 6B, left panel, and data notshown). IFN-�, but not IL-5 and IL-13, was significantly ( p �0.005) increased in 3-day-culture supernatants of OVA-stimulatedMNC from mediastinal lymph nodes from OVA/SA-2 mice com-pared with controls (Fig. 6B, right panel, and data not shown). Asignificant ( p � 0.05) reduction of total serum IgE levels with nochange of OVA-specific IgE and IgG1 levels was detected inOVA/SA-2 mice compared with controls (data not shown). OVA-specific IgG2a Ab levels, even though not significant, resulted inhigher levels in the serum of OVA/SA-2 mice compared with con-trols (data not shown). Finally, comparable levels of mRNA ex-pression of Foxp3, IL-12p35, and EBI3 in the lung of OVA/SA-2and DMSO/SA-2 mice were found (data not shown).

Due to the milder effects exerted by SA-2 in protocol B com-pared with protocol A, we asked whether SA-2 activity could bepotentiated by changing the route of administration and doses ofadjuvant or Ag. I.t. or i.p. SA-2 administration similarly increasedTh1-related (IFN-� and T-bet) and impaired Th2-related mole-cules (IL-5 and IL-13) in the lung (Fig. 6C and data not shown).Additionally, IL-12 and CXCL10 mRNAs were both increased inthe lung after SA-2 administration, although IL-12 and CXCL10mRNA levels were higher in i.t.-treated and in i.p.-treated mice,

FIGURE 5. SA-2 effects are not mediated by expansion of CD4�CD25high Treg cells. A, mRNA expression of Foxp3�, IL-12p35, and EBI3 in the lung3 days after the last OVA challenge is reported. Quantitative RT-PCR was used to detect mRNA, as described in Materials and Methods. Data obtainedfrom 12 mice/group are expressed as the mean values (SEM) of gene/ubiquitin ratios. B, The intracellular expression of Foxp3 on gated splenicCD4�CD25high T cells 10 days after i.p. SA-2 injection was analyzed, as described in Materials and Methods. The proportions of CD4�CD25�Foxp3�

T cells in spleens of SA-2- (or DMSO-) treated mice are reported (upper panel). Coexpression of Foxp3 and CD25 on CD4� splenic T cells from onerepresentative experiment is shown (lower panel).

885The Journal of Immunology

respectively (data not shown). Similarly, the i.p. treatment with250 �g of SA-2 induced increased (not significant) lung mRNAexpression of Th1-related molecules compared with those obtainedwith a low dose of SA-2 (50 �g) (Fig. 6C). Finally, we found thatwhen a high immunization dose (300 �g instead of 10 �g) of OVAwas used, prechallenge SA-2 administration induced improvementof several lung inflammatory parameters. Indeed, the high dose(hd) OVA model resulted in a significant decrease of all lung(perivascular, peribronchiolar, bronchiolar, and alveolar cell infil-trates) inflammation scores of OVA/SA-2 mice compared withOVA/DMSO mice (Fig. 7A). Of note, similar results were obtainedin hdOVA-sensitized mice treated with protocol A (Fig. 7A).Moreover, the hdOVA model in protocol B, but not in protocol A,induced a significant ( p � 0.05) decrease of lung eosinophilia(Fig. 7B, left panel, and data not shown). Finally, in a prechallengeprotocol, hdOVA/SA-2 mice exhibited a reduction of the percent-

ages and absolute values of eosinophils ( p � 0.01) and lympho-cytes (not significant) and the parallel increase of macrophages( p � 0.001) in BALF (Fig. 7B, right panel, and data not shown).When we measured cytokines in the supernatants of lymph nodeOVA-stimulated T cells from five hdOVA/SA-2 and five hdOVA/DMSO mice, comparable levels of IFN-� (48.6 31.6 vs 43.3 19.8 pg/ml) were found, whereas IL-10 was significantly ( p �0.05) reduced (17.8 17.8 vs 222.6 98.5 pg/ml) in culturesupernatants from hdOVA/SA-2 mice compared with those fromhdOVA/DMSO mice. Additionally, hdOVA/SA-2 mice showedlower ( p � 0.05) serum IgE levels and OVA-specific IgE andIgG1 Abs compared with those of hdOVA/DMSO mice (Fig. 7C,right panel, and data not shown).

To evaluate the ability of SA-2 treatment to induce autoreactiv-ity, different patterns of ANAs were finally determined in the serafrom mice treated with both A and B protocols. After 35 days from

FIGURE 6. Effects of prechallenge SA-2 administration on the development of Th2-mediated lung inflammation. A, Effect of SA-2 on lung peribronchiolarand perivascular inflammation scores (left panel) and the absolute number of BALF cells (right panel) in PBS/PBS control (open columns) in OVA/DMSO (blackcolumns) and in OVA/SA-2 (gray columns) mice (10/group) is reported. �, p � 0.05. B, The IFN-� mRNA expression in the lung (left panel) and IFN-� proteinproduction by MNCs from mediastinal lymph nodes cultured in the presence or absence of 200 �g/ml OVA (right panel) of the same groups of mice as A arereported. �, p � 0.05; ��, p � 0.005. C, Effects of different routes (i.p. and i.t.) and doses (50 and 250 �g) of SA-2 administration in protocol B on mRNAexpression of Th1- or Th2-associated molecules in lung tissue are reported. Data are expressed as gene/ubiquitin ratios (five animals/group).

886 ADJUVANT EFFECTS OF MODIFIED ADENINES IN VIVO

SA-2 administration in a prepriming protocol, 10 SA-2/OVA miceexhibited significantly ( p � 0.02) lower levels of ANAs than did10 DMSO/OVA mice (Table II). Furthermore, in the prechallengeprotocol B (with 50 �g and 250 �g of SA-2), ANA serum levelsmeasured at 30, 60, and 90 days from SA-2 administration in sevenOVA/SA-2 mice were significantly lower than those found inseven controls (Table II). Moreover, when the sera from the sameanimals were analyzed for speckled nuclear pattern, typical forSm/RNP-reactive (anti-RNA) autoantibodies, comparable titerswere observed at 30 (166.6 74 vs 54.1 4.1), 60 (125 75 vs100 15), and 90 (186.6 70.6 vs 30 10) days after SA-2treatment.

DiscussionThe most promising strategies to potentiate allergen vaccines con-sist of novel adjuvants able to induce Th1 redirection through TLR

triggering, such as CpG-ODNs acting as TLR9L (18, 19). Re-cently, a series of synthetic 8-hydroxyadenines as novel in vitroand in vivo IFN-�-inducing agents in experimental animals hasbeen reported (20). We showed that one of these compounds, 2-bu-toxy-8-hydroxy-9-benzyladenine, called SA-2, stimulates highproduction of proinflammatory and regulatory cytokines and che-mokines by human BDCA-4� plasmacytoid DCs and CD14�

cells. By triggering TLR7, SA-2 redirects the in vitro differentia-tion of allergen-specific human Th2 cells toward the Th1/Th0 phe-notype (11).

The present study addressed the in vivo role of SA-2 as a syn-thetic adjuvant in regulating Th2-mediated airway inflammationby using protocols A and B in OVA-immunized and i.t.-challengedmice. Previous data indicated that SA-2 was able to induce IFN-�production both in vitro and in vivo (20, 21), but the direct evi-dence of its interaction with any of murine TLRs was lacking. Our

FIGURE 7. Effects of SA-2 administration inmice sensitized with high dose of Ag. A, Lung in-flammation scores in mice immunized with highdose of OVA and treated according to the protocolsA and B are shown. The inflammation scores wereevaluated as described in Materials and Methods.The mean values (SEM) of scores obtained in 10mice/group are reported: �, p � 0.05; ��, p � 0.01;���, p � 0.005. B, Percentages of lung-infiltratingeosinophils (left panel) and BALF cell analysis(right panel) in OVA/DMSO and OVA/SA-2 micewere evaluated, as described in Materials andMethods. The mean values (SEM) of 10 mice/group are reported. �, p � 0.05; ��, p � 0.01. C,Serum levels of OVA-specific IgE Abs in OVA/DMSO and OVA/SA-2 groups of mice described inA. OVA-specific IgE was evaluated as described inMaterials and Methods. The mean values (SEM)of OD obtained in 10 mice/group are reported:�, p � 0.05.

Table II. Titers of ANAs in SA-2-treated mice

Treatment Protocol(SA-2 dose)

Days afterTreatment

Groups of Mice

PBS/PBS DMSO treateda SA-2 treateda

Protocol A (50 �g) 35 23 20 235 60°° 87 17*Protocol B (50 �g) 35 25 24 387 78°°° 30 10*

90 20 18 480 160° 373 141Protocol B (250 �g) 35 21 13 346 162 20 7**

60 22 14 280 114° 21 390 20 8 315 178 30 7

a OVA-immunized mice received i.p. injection of DMSO or SA-2 according to different protocols as described in Materialsand Methods. PBS/PBS mice were included as control mice. ANA titration in the serum was assessed as described in Materialsand Methods at different points after treatment. Results are reported as the mean values (SEM) of the reciprocal ANA titers.Significant differences between SA-2- and DMSO-treated mice are indicated as �, p � 0.02 and ��, p � 0.005, whereas thosebetween DMSO-treated mice and the PBS control group are indicated as �, p � 0.05; ��, p � 0.01; ���, p � 0.005.

887The Journal of Immunology

data with a TLR-transfected cell line provide the first demonstra-tion that SA-2 triggers the activation of murine TLR7, as shown byenhanced NF-�B-dependent luciferase activity. Furthermore, thisTLR7 agonist was able to promote the activation of BMDCs andthe production of high amounts of proinflammatory (IL-6, IL-12,and IL-27) and regulatory (IL-10) cytokines and chemokines(CXCL10). Th2-associated chemokines (CCL17 and CCL22)were not affected, whereas TGF-� was down-regulated in BMDCs.

Our study shows that systemic SA-2 administration during Agimmunization (protocol A) almost completely protected C57BL/6Jmice from the development of Th2-mediated airway inflammation.These animals showed a significant reduction of eosinophils andlymphocytes in BALF as well as of eosinophils and perivascularand peribronchiolar inflammatory infiltration in the lung. More-over, mRNAs for goblet cell-related genes (GOB-5 and Muc5b)were impaired in the lung of SA2-treated mice. Accordingly,mRNAs for Th2-associated cytokines or chemokines werestrongly reduced in the lung, and OVA-specific memory T cellsfrom mediastinal lymph nodes produced lower levels of IL-5 andIL-13. Such Th2 imbalance was in agreement with the peak ofserum IFN-� and CXCL10 seen a few hours after TLR7 agonistadministration (22) and with the early decreased levels of total IgEand OVA-specific IgE and IgG1 Abs found in the serum of SA-2-treated animals.

Some reports suggest that the major role in prevention of lunginflammation driven by endosomal TLR agonists is played by Tregcells through the inhibition of both Th2 and Th1 responses (23).This finding is at odds with the present data (despite showing in-creased Foxp3 expression relative to nonallergic controls) in sen-sitized mice, as no significant changes were shown in the propor-tions of splenic CD4�CD25highFoxp3� Treg cells after 10 days ofSA-2 administration. Moreover, Foxp3, as well as EBI3 and IL-12p35 (which both identify the Treg-specific cytokine IL-35) (24)mRNA expression was unmodified in the lung of SA2-treatedmice, and a significant reduction of IL-10 and TGF-� mRNA werefound in the lung of SA-2-treated animals. On the whole, thesedata strongly suggest that the inhibition of inflammatory airwaydisease in the prepriming protocol was possibly related to regula-tory mechanisms other than Foxp3� cells.

It has been also reported that TLR9 agonists (CpG-ODNs) in-hibit Th2-driven experimental asthma through the up-regulation oflung IDO through the hyperproduction of ODN-induced IL-12p40and/or IFN-� (25). In the murine models reported herein, however,SA-2 did not up-regulate IDO activity on BMDCs, and IDOmRNA was not expressed in the lung of SA-2-treated mice (A.Vultaggio, unpublished data). In agreement with studies usingother TLR (including TLR7) ligands (23, 26), our data indicatethat SA-2 inhibits lung inflammation through a prevalent long-lasting non-Foxp3-dependent regulatory mechanism, in associa-tion with systemic imbalance of Th1/Th2 responses, likely actingearly (27). Indeed, immunoregulation is suggested in vitro by theincreased production of IL-10 and IL-27 (the most relevant inducerof IL-10 by T cells) by SA-2-stimulated BMDCs and in vivo bythe high serum levels of the same cytokines 2 h after SA-2 treat-ment. Moreover, under the same in vitro conditions, SA-2 inducedBMDC expression of Delta4 and the secretion of type IFN-� andCXCL10, all related to Th1 redirection.

A further intriguing finding was that SA-2 administration beforechallenge (protocol B) slightly affected the established lung in-flammation. To improve the effects of SA-2 on airway flogosis, weassessed different protocols. Intratracheal as well as i.p. adminis-trations of SA-2 up-regulated Th1-related cytokine and chemokinemRNA expression in a comparable manner. Moreover, we foundthat SA-2 dose was crucial to prevent allergic Th2 responses in

lungs. Of note, high SA-2 dose (250 �g) induced higher IFN-�,CXCL10, and T-bet mRNA expression than did a low dose (50�g). Finally, the most pronounced effects were found when a highimmunizing dose (300 �g instead of 10 �g) of OVA was used,with a striking reduction of eosinophils in BALF in comparisonwith that obtained with the original protocol. These mice alsoshowed a decreased infiltration in all lung compartments and areduction of lung eosinophilia. How SA-2 exhibits a higher effi-ciency in controlling allergic inflammation in mice sensitized withhdOVA is unclear. Indeed, no evidence of IFN-� up-regulation inculture supernatants of OVA-stimulated lymph node T cells fromhdOVA/SA-2 mice was realized, whereas IL-10 production wasvirtually abrogated. An expansion of memory T cells toward aphenotype more susceptible to be redirected to a Th1 response ora “high dose immunotolerance” (or both) may be envisaged (28).

The last relevant finding was that SA-2 does not favor the onsetof a humoral autoimmune response. Indeed, it has been recentlyshown that TLR7-mediated signals can be involved in the patho-genesis of systemic lupus erythematosus and generation of auto-antibodies to RNAs in systemic lupus erythematosus-prone mice(29, 30). This activity, which has been related to the TLR7/8 li-gand-mediated inhibition of Treg functions (23), could affect thesafety of novel anti-allergic strategies based on vaccination withallergen-conjugated TLR7/8 agonists. In contrast and surprisingly,our data indicate that systemic administration of SA-2 in both Aand B protocols not only did not up-regulate, but, conversely, sig-nificantly decreased the ANA (all staining patterns) serum levelsfor 3 mo after SA-2 administration. Of note, the serum titers ofRNA-related autoantibodies (speckled fluorescence pattern) in thesame condition remained unchanged (29). This finding is at oddswith a relevant role of SA-2 in the onset of autoreactivity andsuggests that a long-lasting inhibitory mechanism is induced andmaintained in SA-2-treated mice. In contrast, these data are inkeeping with other in vivo studies on redirection of Th2 inflam-mation by TLR7/9 agonists that did not refer any induction ofautoimmunity (28, 30). Of note, increase of ANA level has beenobserved in several strains of mice after OVA sensitization (31),and this is likely due to polyclonal B cell activation, involvingbystander autoreactive B cells with ANA specificity, following Agplus adjuvant administration.

Other synthetic TLR7 ligands have been described to redirectTh2-oriented airway inflammation by multiple mechanisms, in-cluding regulatory plasmacytoid DC activation, inhibition of genescoding Th2-polarizing or chemotactic molecules in plasmacytoidDCs, hyperproduction of molecules (CXCR3 ligands) impairinglung angiogenesis, stimulation of IFN-� production by memoryCD4� T cells, IgE production by B cells, reduced secretion ofpreformed Th2 cytokines by mast cells, and direct or IDO-medi-ated inhibition of Treg cells (32). Although more than one of thesemechanisms can be suggested, in vivo and in vitro data from ourand other reports support the notion that the ligation by TLR7agonists induces both a Th1 redirection of T cell responses andsome immunoregulatory mechanisms dominated by IL-10 and IL-10-inducing IL-27 (4, 27).

Although a single mechanism is unlikely to fully explain our invivo data, reciprocal interactions among molecules associated withTh1 and regulatory activity cannot be excluded. There are prece-dents for such interactions. It has been recently shown that IL-27up-regulates local production of the Th1-associated chemokineCXCL10 by endothelial cells (33), and type I and II IFNs caninduce IL-27 in DCs and subsequently IL-10 by T cells (34).Moreover, a new Th1-like regulatory cell subset producing bothIFN-� and IL-10 has been described (35) that can suppress exper-imental allergic lung inflammation, and airway hyperreactivity

888 ADJUVANT EFFECTS OF MODIFIED ADENINES IN VIVO

(36) and Th1 cells are induced to produce IL-10 by the Notch1/Delta4 interaction in the presence of IL-12 and IL-27 (37). Ourdata show that SA-2 imbalances Jagged1/Delta4 surface expres-sion and secretion of both IL-12 and IL-27 by BMDCs.

Further investigations might determine whether the redirectingmechanism occurs early after SA-2 administration and a non-Foxp3-mediated immunoregulatory mechanism predominateslater. This view is in agreement with the observation of a rapidproinflammatory response (TNF-�-mediated) and a delayed inhib-itory cytokine production (IL-10) induced in murine macrophagesupon in vitro exposure to different TLR ligands (38).

If additional in vivo results confirm the safety of these com-pounds, especially if bound to purified allergens, they could beconsidered as the pharmacological substitutes for environmentalmicrobes to protect against allergy as has been described in animaland human models with TLR9 ligand (CpG-ODN)-allergen con-jugates (6, 7).

DisclosuresThe authors have no financial conflicts of interest.

References1. Pelaia, G., A. Vatrella, L. Gallelli, T. Renda, M. Cazzola, R. Maselli, and

S. A. Marsico. 2006. Respiratory infections and asthma. Resp. Med. 100:775–784.

2. Johnstone, S. L. 2007. Innate immunity in the pathogenesis of virus-inducedasthma exacerbations. Proc. Am. Thorac. Soc. 4: 267–270.

3. Liew, F. Y., D. Xu, E. K. Brint, and L. A. O’Neill. 2005. Negative regulation ofToll-like receptor-mediated immune responses. Nat. Rev. Immunol. 5: 446–458.

4. Liotta, F., F. Frosali, V. Querci, L. Maggi, B. Mazzinghi, R. Angeli, E. Ronconi,V. Santarlasci, T. Biagioli, C. Ballerini, et al. 2008. Human immature myeloiddendritic cells trigger a TH2-polarizing program via Jagged-1/Notch interaction.J. Allergy Clin. Immunol. 121: 1000–1005.

5. Parronchi, P., E. Maggi, and S. Romagnani. 1999. Redirecting TH2 responses inallergy. Curr. Top. Microbiol. Immunol. 238: 27–56.

6. Simons, F. E., Y. Shikishima, G. Van Nest, J. J. Eiden, and K. T. HayGlass. 2004.Selective immune redirection in humans with ragweed allergy by injecting Amba 1 linked to immunostimulatory DNA. J. Allergy Clin. Immunol. 113:1144–1151.

7. Tulic, M. K., P. O. Fiset, P. Christodoulopoulos, P. Vaillancourt, M. Desrosiers,F. Lavigne, J. Eiden, and Q. Hamid. 2004. Amb a 1-immunostimulatory oligode-oxynucleotide conjugate immunotherapy decreases the nasal inflammatory re-sponse. J. Allergy Clin. Immunol. 113: 235–241.

8. Moisan, J., P. Camateros, T. Thuraisingam, D. Marion, H. Koohsari, P. Martin,M. L. Boghdady, A. Ding, M. Gaestel, M. C. Guiot, et al. 2006. TLR7 ligandprevents allergen-induced airway hyperresponsiveness and eosinophilia in aller-gic asthma by a MYD88-dependent and MK2-independent pathway.Am. J. Physiol. 290: L987–L995.

9. Diebold, S. S., T. Kaisho, H. Hemmi, S. Akira, and C. Reis e Sousa. 2004. Innateantiviral responses by means of TLR7-mediated recognition of single-strandedRNA. Science 303: 1529–1531.

10. Brugnolo, F., S. Sampognaro, F. Liotta, L. Cosmi, F. Annunziato, C. Manuelli,P. Campi, E. Maggi, S. Romagnani, and P. Parronchi. 2003. The novel syntheticimmune response modifier R-848 (resiquimod) shifts human allergen-specificCD4� TH2 lymphocytes into IFN-�-producing cells. J. Allergy Clin. Immunol.111: 380–388.

11. Filì, L., S. Ferri, F. Guarna, S. Sampognaro, C. Manuelli, F. Liotta, L. Cosmi,A. Matucci, A. Vultaggio, F. Annunziato, et al. 2006. Redirection of allergen-specific TH2 responses by a modified adenine through Toll-like receptor 7 inter-action and IL-12/IFN release. J. Allergy Clin. Immunol. 118: 511–517.

12. Hirota, K., K. Kazaoka, I. Niimoto, H. Kumihara, H. Sajiki, Y. Isobe, H. Takaku,M. Tobe, H. Ogita, T. Ogino, et al. 2002. Discovery of 8-hydroxyadenines as anovel type of interferon inducer. J. Med. Chem. 45: 5419–5422.

13. Jurk, M. F. Heil, J. Vollmer, C. Schetter, A. M. Krieg, H. Wagner, G. Lipford,and S. Bauer. 2002. Human TLR7 or TLR8 independently confer responsivenessto the antiviral compound R-848. Nat. Immunol. 3: 196–200.

14. Leech, M. D., R. A. Benson, A. De Vries, P. M. Fitch, and S. E. Howie. 2007.Resolution of Derp1-induced allergic airway inflammation is dependent onCD4�CD25�Foxp3� regulatory cells. J. Immunol. 179: 7050–7058.

15. Van Rijt, L. S., H. Kuipers, N. Vos, D. Hijdra, H. C. Hoogsteden, andB. N. Lambrecht. 2004. A rapid flow cytometric method for determining the

cellular composition of bronchoalveolar lavage fluid cells in mouse models ofasthma. J. Immunol. Methods 288: 111–121.

16. Boonstra, A., C. Asselin-Paturel, M. Gilliet, C. Crain, G. Trinchieri, Y. J. Liu, andA. O’Garra. 2003. Flexibility of mouse classical and plasmacytoid-derived den-dritic cells in directing T helper type1 and 2 cell development: dependency onantigen dose and differential Toll-like receptor ligation. J. Exp. Med. 197:101–109.

17. Christensen, S. R., J. Shupe, K. Nickerson, M. Kashgarian, R. A. Flavell, andM. J. Shlomchik. 2006. Toll-like receptor 7 and TLR9 dictate autoantibody spec-ificity and have opposing inflammatory and regulatory roles in a murine model oflupus. Immunity 25: 417–428.

18. Till, S. J., J. N. Francis, K. Nouri-Aria, and S. R. Durham. 2004. Mechanisms ofimmunotherapy. J. Allergy Clin. Immunol. 113: 1025–1034.

19. P. Parronchi, F. Brugnolo, F. Annunziato, C. Manuelli, S. Sampognaro,C. Mavilia, S. Romagnani, and E. Maggi. 1999. Phosphorothioate oligode-oxynucleotides promote the in vitro development of human allergen-specificCD4� T cells into TH1 effectors. J. Immunol. 163: 5946–5953.

20. A. Kurimoto, T. Ogino, S. Ichii, Y. Isobe, M. Tobe, H. Ogita, H. Takaku,H. Sajiki, K. Hirota, and H. Kawakami. 2004. Synthesis and structure-activityrelationships of 2-amino-8-hydroxyadenines as orally active interferon inducingagents. Bioorg. Med. Chem. 12: 1091–1099.

21. Lee, J., C. C. Wu, K. J. Lee, T. H. Chuang, K. Katakura, Y. T. Liu, M. Chan,R. Tawatao, M. Chung, C. Shen, et al. 2006. Activation of anti-hepatitis C virusresponses via Toll-like receptor 7. Proc. Natl. Acad. Sci. USA 103: 1828–1833.

22. Asselin-Paturel, C., G. Brizard, K. Chemin, A. Boonstra, A. O’Garra, A. Vicari,and G. Trinchieri. 2005. Type I interferon dependence of plasmocytoid dendriticcell activation and migration. J. Exp. Med. 201: 1157–1167.

23. Peng, G., Z. Guo, Y. Kiniwa, K. S. Voo, W. Peng, T. Fu, D. Y. Wang, Y. Li,H. Y. Wang, and R. F. Wang. 2005. Toll-like receptor 8-mediated reversal ofCD4� regulatory T cell function. Science 309: 1380–1384.

24. Collison, L. W., C. J. Workman, T. Kuo Timothy, K. Boyd, Y. Wang,K. M. Vignali, R. Cross, D. Sehy, R. S. Blumberg, and D. A. A. Vignali. 2007.The inhibitory cytokine IL-35 contributes to regulatory T-cell function. Nature450: 566–569.

25. Hayashi, T., L. Beck, C. Rossetto, C. Gong, O. Takikawa, K. Takabayashi,D. H. Broide, D. A. Carson, and E. Raz. 2004. Inhibition of experimental asthmaby indoleamine 2,3-dioxygenase. J. Clin. Invest. 114: 270–279.

26. Quarcoo, D., S. Weixler, R. A. Joachim, P. Stock, T. Kallinich, B. Ahrens, andE. Hamelmann. 2004. Resiquimod, a new immune response modifier from thefamily of imidazoquinolinamines, inhibits allergen-induced Th2 responses, air-way inflammation and airway hyper-reactivity in mice. Clin. Exp. Allergy 34:1314–1320.

27. Sel, S., M. Wegmann, S. Sel, S. Bauer, H. Garn, G. Alber, and H. Renz. 2007.Immunomodulatory effects of viral TLR ligands on experimental asthma dependon the additive effects of IL-12 and IL-10. J. Immunol. 178: 7805–7813.

28. Herrick, C. A., and K. Bottomly. 2003. To respond or not to respond: T cells inallergic asthma. Nat. Rev. Immunol. 3: 405–412.

29. Pisitkun, P., J. A. Deane, M. J. Difilippantonio, T. Tarasenko, A. B. Satterthwaite,and S. Bolland. 2006. Autoreactive B cell responses to RNA-related antigens dueto TLR7 gene duplication. Science 312: 1669–1672.

30. Ehlers, M., and J. V. Ravetch. 2007. Opposing effects of Toll-like receptor stim-ulation induce autoimmunity or tolerance. Trends Immunol. 28: 74–79.

31. Kool, M., T. Soullie, M. van Nimwegen, M. A. Willart, F. Muskens, S. Jung,H. C. Hoogsteden, H. Hammad, and B. N. Lambrecht. 2008. Alum adjuvantboosts adaptive immunity by inducing uric acid and activating inflammatory den-dritic cells. J. Exp. Med. 205: 869–882.

32. M. Goldman. 2007. Toll-like receptor ligands as novel pharmaceuticals for al-lergic disorders. Clin. Exp. Immunol. 147: 208–216.

33. Feng, X. M., X. L. Chen, N. Liu, Z. Chen, Y. L. Zhou, Z. B. Han, L. Zhang, andZ. C. Han. 2007. Interleukin-27 upregulates major histocompatibility complexclass II expression in primary human endothelial cells through induction of majorhistocompatibility complex class II transactivator. Hum. Immunol. 68: 965–972.

34. Pirhonen, J., J. Siren, I. Julkunen, and S. Matikainen, S. 2007. IFN-� regulatesToll-like receptor-mediated IL-27 gene expression in human macrophages.J. Leukocyte Biol. 82: 1185–1192.

35. Jankovic, D., and G. Trinchieri. 2007. IL-10 or not IL-10: that is the question.Nat. Immunol. 8: 1281–1283.

36. Stock, P., O. Akbari, G. Berry, G. J. Freeman, R. H. Dekruyff, and D. T. Umetsu.2004. Induction of T helper type 1-like regulatory cells that express Foxp3 andprotect against airway hyper-reactivity. Nat. Immunol. 5: 1149–1156.

37. Benson, R. A., K. Adamson, M. Corsin-Jimenez, J. V. Marley, K. A. Wahl,J. R. Lamb, and S. E. Howie. 2005. Notch1 co-localizes with CD4 on activatedT cells and Notch signaling is required for IL-10 production. Eur. J. Immunol. 35:859–869.

38. Crabtree, T. D., L. Jin, D. P. Raymond, S. J. Pelletier, C. W. Houlgrave,T. G. Gleason, T. L. Pruett, and R. G. Sawyer. 2001. Preexposure of murinemacrophages to CpG oligonucleotide results in a biphasic tumor necrosis factoralpha response to subsequent lipopolysaccharide challenge. Infect. Immun. 69:2123–2129.

889The Journal of Immunology