of July 27, 2018.This information is current as

ResponsesCells Sense RNA and Drive Mucosal IgA Alveolar Macrophages and Lung Dendritic

BachmannPumpens, Philippe Saudan, Pascal Schneider and Martin F. Juliana Bessa, Andrea Jegerlehner, Heather J. Hinton, Paul

http://www.jimmunol.org/content/183/6/3788doi: 10.4049/jimmunol.0804004August 2009;

2009; 183:3788-3799; Prepublished online 26J Immunol 

Referenceshttp://www.jimmunol.org/content/183/6/3788.full#ref-list-1

, 24 of which you can access for free at: cites 55 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2009 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on July 27, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Alveolar Macrophages and Lung Dendritic Cells Sense RNAand Drive Mucosal IgA Responses

Juliana Bessa,* Andrea Jegerlehner,* Heather J. Hinton,* Paul Pumpens,† Philippe Saudan,*Pascal Schneider,‡ and Martin F. Bachmann1*

The mechanisms regulating systemic and mucosal IgA responses in the respiratory tract are incompletely understood. Usingvirus-like particles loaded with single-stranded RNA as a ligand for TLR7, we found that systemic vs mucosal IgA responses inmice were differently regulated. Systemic IgA responses following s.c. immunization were T cell independent and did not requireTACI or TGF�, whereas mucosal IgA production was dependent on Th cells, TACI, and TGF�. Strikingly, both responsesrequired TLR7 signaling, but systemic IgA depended upon TLR7 signaling directly to B cells whereas mucosal IgA required TLR7signaling to lung dendritic cells and alveolar macrophages. Our data show that IgA switching is controlled differently accordingto the cell type receiving TLR signals. This knowledge should facilitate the development of IgA-inducing vaccines. The Journalof Immunology, 2009, 183: 3788–3799.

S uccessful vaccines mediate protection via neutralizing Abs(1, 2). Parenteral routes of immunization are very efficientin inducing neutralizing IgG responses that can be main-

tained as humoral immunological memory over long periods oftime (3, 4). However, these routes usually induce poor IgA re-sponses, especially at mucosal surfaces. For this reason, intranasal(i.n.)2 administration might be an attractive route of vaccination,because in addition to inducing specific IgG responses this routealso efficiently induces IgA responses at mucosal surfaces, whichare the primary sites of pathogen entrance.

The regulation of IgA isotype class switch recombination (CSR)is complex, and both T cell-dependent (TD) and T cell-indepen-dent (TI) mechanisms have been described (5, 6). IgA responsesinduced by TD Ags such as proteins require CD4� T cell helpmediated by CD40 ligand (CD40L) as well as TGF�1 and aremainly mediated by B2 B cells. In contrast, IgA responses inducedby TI type 1 Ag (TI-1) such as LPS or TI type 2 Ag (TI-2) suchas polysaccharides do not require CD40-CD40L interaction. In-stead, they rely on APRIL (a proliferation-inducing ligand; se-creted mainly by activated dendritic cells (DCs) and macrophages)binding to TACI (transmembrane activator, calcium modulator,and cyclophilin ligand interactor) on B cells and are mediatedmainly by B1 B cells (7, 8).

Both TD and TI mechanisms of CSR have been described forintestinal IgA production. Specifically, it has been shown that IgACSR in the gut can occur independently of CD40 signaling andgerminal center formation (9). Retinoic acid has also been impli-cated in TI IgA responses and can stimulate the expression ofgut-homing receptors on B cells. In addition, by synergizing withGALT-DC-derived IL-6 and IL-5, retinoic acid can promote IgAsecretion (10). Recently, NO secreted by DCs has also been im-plicated in the regulation of IgA production in the gut (7).

In contrast to the gut, much less is known about the mechanismscontrolling IgA responses at the respiratory mucosa. One of thefew examples studied is influenza virus infection, where virus-specific IgA responses generated in the respiratory tract occurredindependently of cognate T-B interaction (CD40 expression on Bcells) but did require bystander CD4� T cell help in contrast toIgM and IgG responses, which were dependent on cognate inter-actions (11).

Many reports have investigated the role of TLR signals in B cellresponses, including memory B cell responses (12, 13) as well asCSR and induction of IgG Abs (14–20). The role of TLR signalsin the induction of IgA responses is a subject of investigation. TheMedzhitov group has detected normal levels of serum IgA inMyD88�/� mice, suggesting that TLR signaling is dispensable forIgA responses (16). However, more recently it was shown thatTLR stimulation either on gut DCs (7) or in intestine epithelialcells (21) promotes IgA responses both in sera and intestinal con-tent, indicating a potential role for TLR and MyD88 in regulatingIgA responses.

Virus-like particles (VLPs) are highly effective immunogensthat can induce strong Ab responses. VLPs are classic TI type 1Ags that are able to induce IgM responses in the absence of T helpbecause of multimeric interactions with cognate BCRs, which in-duce a strong activation signal in B cells. In addition, we havepreviously shown that i.n. and s.c. immunization with VLPs de-rived from the bacteriophage Q�, Q�-VLPs (22), are efficient im-munization routes for the induction of IgA responses (23). Q�-VLPs are packaged with RNA derived from Escherichia coliduring its self-assembly process. Thus, Q�-VLPs typically medi-ate TLR3/7 signals that we considered to be important candidatesfor triggering IgA responses.

*Cytos Biotechnology AG, Zurich-Schlieren, Switzerland; †Latvian Biomedical Re-search and Study Center, Riga, Latvia; and ‡Department of Biochemistry, Universityof Lausanne, Epalinges, Switzerland

Received for publication December 1, 2008. Accepted for publication July 21, 2009.

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 Address correspondence and reprint requests to Dr. Martin F. Bachmann, Wagistrasse25, 8952 Zurich-Schlieren, Switzerland. E-mail address: Martin.bachmann@cytos.com2 Abbreviations used in this paper: i.n., intranasal(ly); AFC, Ab-forming cell; APRIL,a proliferation-inducing ligand; BAFF, B cell-activating factor of the TNF family;BAL, bronchoalveolar lavage; BCMA, B cell maturation Ag; BM, bone marrow;CD40L, CD40 ligand; CSR, class switch recombination; DC, dendritic cell; DT,diphtheria toxin; DTR, DT receptor; iNOS, inducible NO synthase; LN, lymph node;MLN, mediastinal lymph node; NP, 4-hydroxy-3-nitrophenylacetate; TACI, trans-membrane activator, calcium modulator, and cyclophilin ligand interactor; TD, T celldependent; TI, T cell independent; VLP, virus-like particle; WT, wild type.

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

The Journal of Immunology

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0804004

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

In the present study, we assessed the role of TLR7 signaling indriving IgA responses to Q�-VLPs both at the systemic level andin the respiratory tract mucosa. We found that depending on thesite of Ag sampling, TLR7 signaling was required either to lungDCs and alveolar macrophages or to B cells and resulted in TD vsTI IgA responses, respectively.

Materials and MethodsMice

C57BL/6 mice were purchased from Harlan. TLR7�/� (24), MyD88�/�

(25), JH�/� (26), MHC II�/� (27), CD40�/� (28), CCR7�/� (29), andCD11c-diphtheria toxin receptor (DTR)/GFP (30) mice on a C57BL/6background have been described earlier. T�RIIfl mice were purchased fromMRC Harwell and crossed with CD19Cre mice (31) on a BALB/c back-ground to delete the floxed target gene (T�RII) in B cells. The T�RII-Bmice used were homozygous for T�RIIfl (T�RIIfl/fl) and heterozygous forCD19Cre (CD19cre/�) and the control group were T�RIIfl/fl and CD19�/�.Depletion of T�RIIfl/fl was �95% on splenic B cells purified from T�RII-Bmice as assessed by PCR (data not shown). TACI�/� and TACI;BCMA�/�

(where BCMA is B cell maturation Ag) were generated at and provided byBiogen Idec (32). All animals were kept under specific pathogen-free con-ditions at BioSupport and were used at 8 to 12 wk of age. Experimentswere conducted in accordance with protocols approved by the Swiss Fed-eral Veterinary Office (Bern, Switzerland).

Immunization and Ag

To evaluate the Ab response induced by Q�-VLP, C57BL/6, TLR7�/�,MyD88 �/�, MHC II�/�, CD40�/�, T�RII-B, TACI�/� and TACI;BCMA�/� mice were immunized either i.n. or s.c. with 50 �g of Q�-VLPcontaining Escherichia coli-derived ssRNA.

For i.n. immunization with Q�-VLP, mice were anesthetized withisoflurane and vaccine was administered using a 200-�l pipette. The s.c.vaccination was performed by Q�-VLP injection into both sides of theabdomen. For both routes of immunization, Q�-VLP was diluted in PBS toa final administration volume of 100 �l (2 � 50 �l).

Capsids of the RNA phage Q� were cloned into the pQ�10 vector andpurified as described elsewhere (22). The AP205 coat protein (33) wascloned into the pQ�10 vector (34) and expressed and purified similarlyas Q�.

ELISA

ELISA plates (Nunc-Immuno MaxiSorp; Nunc) were coated overnightwith 100 �l of Q�-VLP (1 �g/ml) and ELISA was performed according tostandard protocols using the following HRP-conjugated secondary Abs:goat anti-mouse IgG (Fc� specific; Jackson ImmunoResearch); goat anti-mouse IgA (�-chain specific; Sigma-Aldrich); rat anti-mouse IgG1 (BDPharmingen); and rat anti-mouse IgG2a (BD Pharmingen).

For detection of anti-Q�-VLP IgA in serum, IgG was depleted usingprotein G (GE Healthcare). ELISA controls showed that the elevatedamount of anti-Q�-VLP IgG in serum suppresses the measurement of anti-Q�-VLP IgA. For detection of anti-Q�-VLP IgG and Q�-VLP IgA inbronchoalveolar lavage (BAL), undiluted samples were used. Anti-Q�-VLP IgG titers in serum and in BAL are indicated as dilutions reachinghalf-maximal absorbance at 450 nm. Anti-Q�-VLP IgA titers in serum andin BAL were measured by endpoint titer by calculating the OD (450 nm)cutoff as the average of the OD of the naive sample multiplied by 3-fold theSD of the OD of the naive sample.

ELISPOT assay

Q�-VLP-specific, Ab-forming cell (AFC) frequencies were determined asdescribed (35). Briefly, 24-well plates were coated with 10 �g/ml Q�-VLP.Spleen and mediastinal LN (MLN) were added in DMEM containing 2%FCS and incubated for 5 h at 37°C. Cells were washed off and plates wereincubated with either goat anti-mouse IgG (EY Laboratories) or goat anti-mouse IgA (�-chain specific; Southern Biotech), followed by alkalinephosphatase-conjugated donkey anti-goat IgG Ab (Jackson Immuno-Research Laboratories) before the development of alkaline phosphatasecolor reactions.

Radiation bone marrow chimeras

Radiation BM chimeras in which all B cells were deficient in the expres-sion of TLR7 were generated by i.v. injection of BM mixture containing20% of BM cells isolated from TLR7�/� mice and 80% of BM cells iso-lated from JH�/� mice (1–5 � 107 total injected cells) into JH�/� mice that

had been lethally irradiated (950 rad) 1 day previously. Before cell transfer,BM cell suspension was depleted from T cells. Control chimeras withwild-type (WT) B cells were made using a combination of C57BL/6 andJH�/� donor BM. After 6 wk, BM reconstitution was assessed by stainingperipheral blood lymphocytes. Thereafter, mice were immunized with 100�g of Q�-VLP either i.n. or s.c.

The BM chimeras in which CD11c� cells lack TLR7 expression weregenerated as follows. WT mice were lethally irradiated and reconstitutedwith a BM mixture containing 50% of BM cells isolated from TLR7�/�

mice and 50% of BM cells isolated from CD11c-DTR mice. Control chi-meras were generated using a mixture of BM cells isolated from C57BL/6and CD11c-DTR mice. For CD11c� cells ablation, diphtheria toxin wasadministered i.n. for 5 days. In the first three administrations (every day)the mice received 100 ng of diphtheria toxin (DT). For the last two ad-ministrations (with a 1-day interval), only 50 ng of DT was inoculated.AP205-VLP, similar to Q�-VLP in that they are phage derived, loadedwith E. coli RNA, and exhibit similar sizes, were used for immunization.Mice were immunized i.n. with 100 �g of AP205-VLP 24 h after the firstDT administration.

Lung washes and lung digestion

BAL samples were collected through the trachea by washing the lung threetimes with 300 �l of PBS plus 1% BSA. Samples were stored at �20°Cuntil use.

For the isolation of leukocytes from the lung, mice were perfused with5–10 ml of PBS in the heart ventricle to clear the lungs of blood. Lungswere chopped in small pieces and incubated at 37°C in medium containingcollagenase (2.4 mg/ml). Finally, leukocytes were harvested by using a30% Percoll gradient.

Flow cytometry

For detecting cells in association with Q�-VLP, mice were immunizedeither i.n. or s.c. with Q�-VLP labeled with Alexa Fluor 488. Twenty-four or 4 h after immunization the mice were killed and cells isolatedfrom lung and MLNs were stained with allophycocyanin-conjugated ratanti-mouse CD11b and PE-conjugated hamster anti-mouse CD11c (BDPharmingen). In all cases, Fc receptors were blocked with rat IgG2banti-mouse CD16/32 (2.4G2) and dead cells were excluded by pro-pidium iodide staining.

Quantitative real-time PCR

Synthesis of ss-cDNA was done with total RNA using random nonamers(Microsynth) and SuperScript II reverse transcriptase (Invitrogen) accord-ing to the manufacturer’s protocol. Complementary RNA was digested bytreating the ss-cDNA with 2 U of RNase H (New England Biolabs) at 37°Cfor 20 min. The cDNA was then used as template for real-time quantitativePCR (iCycler instrument; Bio-Rad) with the gene-specific primers �-ac-tin-F (5�-CCCTGAAGTACCCCATTGAAC-3�), �-actin-B (5�-CTTTTCACGGTTGGCCTTAG-3�), APRIL-F (5�-GGGGAAGGAGTGTCAGAGTG-3�), APRIL-B (5�-GCAGGGAGGGTGGGAATAC-3�), BAFF-F(5�-AGGCTGGAAGAAGGAGATGAG-3�), and BAFF-B (5�-CAGAGAAGACGAGGGAAGGG-3�) (where BAFF stands for B cell-activatingfactor of the TNF family, F stands for forward, and B stands for back-wards) using Brilliant SYBR Green QPCR Master Mix (Stratagene) ac-cording to the manufacturer’s protocol.

ResultsSite of Ag exposure governs the requirement for T cell cognatehelp during IgA production

VLPs are classified as TI type 1 (TI-1) Ag because they can effi-ciently induce IgM responses without the need of T cell help. Thisproperty can be accredited to their highly organized structure,which allows them to efficiently crosslink the BCR. At the sametime they can behave as a TD Ag, because they are proteins.Indeed, we have observed that Q�-specific IgM responses areinduced in the absence of T cell help, whereas specific IgGresponses required the cognate interaction between T and Bcells (19, 36). In this study we assessed whether the IgA re-sponse against Q�-VLP would require T cell help or whether,similarly as the IgM response, IgA could also be induced by aTI mechanism. To address this question, we immunized WTand MHC II�/� mice via the i.n. or s.c. routes and compared themucosal (BAL) and systemic (serum) IgA titers, respectively,

3789The Journal of Immunology

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

between the two groups. The mucosal IgA response in MHCII�/� mice immunized i.n. was strongly reduced compared withWT mice. In marked contrast, the levels of systemic IgA weresimilar for MHC II�/� and WT mice immunized s.c. (Fig. 1A).We further assessed the IgA response elicited in CD40�/�

mice. In accordance with our findings in MHC II�/� mice, mu-cosal IgA levels in CD40�/� mice were also significantly re-duced compared with the WT group, whereas the systemic IgAtiter elicited in CD40�/� mice immunized s.c. was comparableto that of the WT group (Fig. 1B). These observations show thatIgA responses against Q�-VLP are differentially regulated de-pending on the site of Ag exposure. Whereas mucosal IgA re-sponses elicited upon mucosal administration of VLPs in thelung requires T cell help and the cognate interaction between Tand B cells, Q�-VLPs administered s.c. induced a TI systemicIgA response. As expected, the IgG response was strongly re-duced in the absence of T cell help, independently of the routeof immunization (data not shown).

It has been previously shown that a high Ag dose can be trappedin the splenic marginal zone and induce TI Ab responses (37). Wetherefore assessed whether the TD IgA observed after i.n. immu-nization was not simply reflecting the low dose of Ag sampledfollowing this immunization route. To address this hypothesis, weadministered high doses of Q�-VLP i.n. and measured IgA titers inBAL and serum. Fig. 2A shows that no Q�-VLP-specific IgAcould be detected in BAL of MHC II�/� mice, even when 1 mg ofQ�-VLP was administered. In serum, we observed that low levelsof IgA could be induced in MHC II�/� mice following a high Agdose (Fig. 2B). This low TI IgA titer in serum might be due tosome Ag leakage in the blood circulation following a high dose ofQ�-VLP i.n. (our unpublished data). Importantly, the TI IgA titer

following 1 mg of Q�-VLP was significantly lower compared withthe TD IgA induced with only 50 �g of Q�-VLP. Thus, the T cellhelp dependency for the induction of IgA following i.n. immuni-zation is not due to the impact of the Ag dose. In contrast to theIgA response, the Q�-VLP-induced IgM response in serum wasalmost completely TI, corroborating the notion that the regulationof IgA Abs is governed by a different mechanism than IgM re-sponses (Fig. 2C).

Differential requirements for TGF� and TACI signaling formucosal and systemic TI IgA production

TGF� is the major cytokine involved in the induction of IgA CSR.In vitro, TGF� has been shown to be involved in the Th-dependentCSR to IgA (7, 38). Evidence for its in vivo role was provided byexperiments showing that mice deficient for the TGF� receptorT�RII selectively on B cells (T�RII-B) were almost completelydevoid of IgA, both in serum and in mucosal washes (39, 40). Inan attempt to address the role of TGF� in the IgA responsesagainst Q�-VLP in vivo, we used these T�RII-B mice. Followings.c. immunization, systemic IgA responses in T�RII-B mice werenot significantly reduced compared with the control group. In con-trast, the mucosal IgA titer in BAL of T�RII-B mice was signif-icantly lower when compared with that of control mice. This dem-onstrates that the systemic TI IgA response to Q�-VLP is notdependent on TGF� signaling to B cells, whereas TGF� plays animportant role for the induction of mucosal IgA response (Fig. 3A).

IgA-specific CSR can be accomplished by TD or TI mecha-nisms. The TD mechanism requires CD40L expressed on acti-vated T cells, whereas the TI mechanism is thought to relylargely on APRIL, a proliferation-inducing ligand, and BAFF, aB lymphocyte stimulator protein (also known as BLyS,

FIGURE 1. Requirement of CD4�

T cell help for induction of Q�-VLP-specific IgA responses. Mice were im-munized either s.c. or i.n. with 50 �gof Q�-VLP, and 20 days later serumand BAL samples were collected fromthe s.c. and i.n. immunized mice, re-spectively, to measure specific IgA lev-els by ELISA. Q�-VLP-IgA titers ofC57BL/6 and MHC II�/� (A) andC57BL/6 and CD40�/� mice (B) areshown. Geometric mean IgA ELISA ti-ters � SEM are indicated; n.d., not de-tectable. The data shown are pooledfrom two independent experiments.Statistical significance was assessed byunpaired Student’s t test (�, p � 0.05;��, p � 0.01).

3790 TLR SIGNALING AND IgA RESPONSES

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

TALL-1, THANK, and zTNF4), produced by DCs that interactwith their cognate receptors, TACI and BCMA, expressed on Bcells (8, 41). To address the role of these molecules in Q�-VLP-specific IgA responses, we compared the IgA levels inserum between the WT and TACI;BCMA double-deficientmice. Surprisingly, the systemic TI IgA levels in TACI;BCMA�/� mice were not significantly reduced compared withthose in the WT group. In contrast, the mucosal IgA response inthe BAL of TACI;BCMA�/� mice was strongly reduced. Thus,in addition to requiring T cell help, local IgA responses alsoneeded signaling through TACI or BCMA whereas TI systemicIgA responses occurred in a TACI/BCMA-independent manner(Fig. 3B). To investigate whether TACI or BCMA signalingwas required for mucosal IgA responses, we also administeredQ�-VLP i.n. to TACI�/� mice. Similarly as the double-defi-

cient TACI;BCMA�/� mice, TACI�/� mice generated a sig-nificantly reduced IgA response in BAL (data not shown). Thisindicates that the TD IgA responses occurring at the airwaymucosa require TACI-mediated signals.

TLR7 signaling is required for induction of optimal mucosal TDand systemic TI IgA responses against Q�-VLP

Our previous data demonstrated that the mucosal IgA productionin response to i.n. immunization required Th cell-derived signals(CD40L) in addition to TACI-mediated signals. In contrast, thesystemic IgA response elicited by s.c. immunization did not re-quire Th cells or TACI signaling. Thus, the IgA responses againstVLPs are differentially regulated depending on the site of Ag ex-posure. We therefore sought to understand which additional factor,other than Th cells and TGF� and APRIL/BAFF production, could

FIGURE 2. High doses of Q�-VLP administered i.n. do not elicit T-independent IgA responses. A and B, Mice were immunized i.n. with 50, 200, 500,and 1000 �g of Q�-VLP, and 20 days later serum and BAL samples were collected to measure specific IgA levels by ELISA. Q�-VLP-IgA titers in BAL(A) and serum (B) are shown. C, Q�-VLP-IgM titers in serum. Geometric mean IgA and IgM ELISA titers � SEM are indicated. Statistical significancewas assessed by unpaired Student’s t test (��, p � 0.01).

FIGURE 3. T�RII and TACI playa critical role for the production ofmucosal IgA responses against Q�-VLP. Mice were immunized eithers.c. or i.n. with 50 �g of Q�-VLP,and 20 days later specific IgA weremeasured in serum and BAL, respec-tively, by ELISA. Q�-VLP-IgA titersof control and T�RII-B (A) andC57BL/6 and TACI;BCMA�/� mice(B) are shown. Geometric mean IgAELISA titers � SEM are indicated.The data shown are pooled from twoindependent experiments. Statisticalsignificance was assessed by unpairedStudent’s t test (��, p � 0.01).

3791The Journal of Immunology

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

3792 TLR SIGNALING AND IgA RESPONSES

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

be involved in the regulation of the TI systemic IgA responseobserved following s.c. immunization.

As mentioned, Q�-VLPs are loaded with E. coli-derived ssRNAand provide efficient TLR7 stimulation. To determine whetherTLR7 signaling played a role in regulating the IgA response, wecompared the IgA and the IgG isotypes in the sera of mice immu-nized either i.n. or s.c. in the presence or absence of TLR7 andMyD88 signaling. To this end, we compared both the Ab responseof mice immunized with Q�-VLPs with or without packaged RNAand the response of WT mice to TLR7- or MyD88-deficient mice.As shown in Fig. 4, A and B, IgA and IgG2a in serum were onlyinduced with Q�-VLP containing RNA. In contrast, the levels ofIgA and IgG2a in WT mice immunized with Q�-VLP devoid ofRNA as well as in TLR7�/� and MyD88�/� mice were verylow. Following s.c. immunization, IgG1 responses were in-duced in the absence of TLR7 signaling and suppressed by thepresence of TLR7 signaling, which confirms previous findingsfor TLR9 signaling (19). In contrast, TLR7 signaling had noinfluence on specific IgG1 titers following i.n. immunization(Fig. 4A). In the BAL, similarly as the systemic Ab response,the levels of IgA and IgG2a were completely abolished in theabsence of TLR7 signaling (Fig. 4C).

The numbers of Q�-VLP specific AFCs correlate with the Ab titer

We have previously seen that MLNs are the major inductive site ofthe Ab response upon i.n. immunization against Q�-VLP (23).Next, the impact of TLR signaling on the number of AFCs secret-ing Q�-VLP-specific IgA in these LNs was investigated. Similarlyas the Ab titer, the number of AFCs secreting IgA was stronglyreduced in mice immunized with VLP devoid of RNA as well asin TLR7�/� and MyD88�/� mice compared with control mice(Fig. 4D). The same was true for the spleen of s.c. immunized mice(Fig. 4E).

In conclusion, in contrast to T cell help, CD40L, TGF�, andTACI, TLR7 signaling was pivotal for the induction of systemicIgA responses after s.c. immunization.

IgA CSR upon s.c. immunization requires TLR7 signalingdirectly on B cells

In the next set of experiments, we investigated how TLR7 signal-ing controls IgA responses to Q�-VLP following i.n. and s.c. im-munization. To this end, we reconstituted lethally irradiated WTC57BL/6 mice with MyD88�/� BM such that hematopoietic cellslacked MyD88 expression, whereas its expression was normal inradiation-resistant cells such as epithelial cells. Irradiated WT micereconstituted with MyD88�/� BM cells failed to produce IgAupon i.n. immunization, suggesting that hematopoietic cells areresponsible for regulating the IgA response to Q�-VLP via TLR7signaling (Fig. 5A). We have previously shown that TLR9 sig-naling in B cells but not in non-B cells is essential for promot-ing IgG2a responses against VLPs loaded with CpGs (19). Wetherefore assessed whether a similar mechanism regulated IgACSR. We addressed this question by generating BM chimerasexhibiting TLR7 expression in all hematopoietic cells except Bcells. To this end, B cell-deficient (JH�/�) recipient mice were

lethally irradiated and reconstituted with a mixture of BM cellsisolated from JH�/� and TLR7�/� mice or from JH�/� and WTmice. This model allowed us to generate chimeric mice con-taining only TLR7 WT or deficient B cells with all other he-matopoietic cells derived from JH�/� mice expressing TLR7.Our results showed that signaling directly to B cells played arole in regulating systemic IgA responses against Q�-VLP fol-lowing s.c. immunization but was less important for mucosalIgA responses following i.n. immunization (Fig. 5, B and C).This shows that TI systemic IgA CSR requires TLR7 signalingdirectly to B cells. In contrast, the impact of TLR7 signaling onmucosal IgA responses was due to an involvement of anotherhematopoietic cell type.

In a marked contrast, the Q�-VLP-specific IgG2a titers, both inserum and BAL, were strongly reduced in the group of mice inwhich the B cells were TLR7 deficient (Figs. 5, B and C). Basedon this finding, we concluded that not only TLR9 and TLR4 sig-naling in B cells (13, 16, 19) but also TLR7 signaling controlsIgG2a CSR independently of the immunization route and requiresTLR7 expression in B cells. Importantly, IgG1 levels in chimericmice exhibiting TLR7-deficient B cells were increased, demon-strating normal responsiveness of the B cells.

Taken together, we conclude that regardless of the immuniza-tion route, IgG2a CSR is governed by TLR signaling directly to Bcells. In contrast, for IgA responses, this was the case only afters.c. immunization. Thus, IgG2a and IgA CSR are differentlyregulated.

Lung DCs and alveolar macrophages are the major cellpopulations transporting Q�-VLP from the lung to the draining LN

TLR7 signaling directly to B cells was not required for mucosalIgA. Therefore, our next attempt was to identify the cell populationrequiring TLR7 signals to promote mucosal IgA responses. Wehave previously shown that upon i.n. administration, Q�-VLP canmainly be found in the lung and that MLNs that drain the lowerairways and are the major sites to which Q�-VLPs are transportedand where the Ab response is initiated (23).

To elucidate which cell population may be involved in TLR7-dependent IgA CSR, we analyzed in detail the cellular populationsinteracting with Q�-VLP within the lung and the MLNs. It hasbeen shown previously that in the lung, CD11c�CD11b� cells arealveolar macrophages whereas CD11c�CD11b� cells are lungDCs (42). By analyzing Q�� cells in the lung, we observed thatmost of the Q�� cells were CD11c�CD11b� cells and a minorproportion were CD11c�CD11b� cells, identifying the major cellpopulation interacting with Q�-VLPs as alveolar macrophages(Fig. 6A). We also found a few Q�-VLPs in association withCD11c�CD11b� cells and B cells in the lung (Fig. 6A and data notshown). In MLNs of mice that received Q�-VLPs i.n., the twomajor cell populations bearing Q�-VLPs were again alveolar mac-rophages and DCs (Fig. 6B). Almost no CD11c�CD11b� cellspositive for Q�-VLPs were found in MLNs, indicating that thiscell population is nonmigratory and remains in the lung. Thus, it ispossible that alveolar macrophages and DCs take up Q�-VLPs inthe lung, sense the endogenous ssRNA, and transport the Ag to

FIGURE 4. Role of TLR7 signaling in regulating IgA responses against Q�-VLP. A and B, Mice were immunized either i.n. (A) or s.c. (B) with 50 �gof Q�-VLP, and 20 days later, Q�-VLP-specific IgA, IgG2a, and IgG1 titers in serum of WT C57BL/6, TLR7�/�, TLR7�/�, and MyD88�/� miceimmunized with Q�-VLP loaded with RNA and WT mice immunized with empty Q�-VLP were measured by ELISA. C, Anti-Q�-VLP IgA and IgG2ain BAL of i.n. immunized mice are shown. Geometric mean IgA, IgG2a, and IgG1 ELISA titers � SEM are indicated. D and E, Number of AFCs secretingQ�-VLP IgA was determined by ELISPOT in MLNs of i.n. (D) and spleens of s.c. (E) immunized groups. ELISPOT data show the mean values � SEM.The data shown are pooled from two independent experiments; n.d., not detectable. Statistical significance was assessed by unpaired Student’s t test (�,p � 0.05; and ��, p � 0.01).

3793The Journal of Immunology

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

MLNs. To address this question, we investigated what happens ina situation where alveolar macrophages and DCs are unable tomigrate. In CCR7�/� mice, these two cell populations were com-

pletely absent in the MLN, and most of Q�-VLP was in associa-tion with CD11b� cells. This finding suggests that in WT micemost Q�-VLPs are carried to MLNs by alveolar macrophages and

FIGURE 5. TLR7 expression di-rectly to B cells is required forsystemic but not mucosal IgA re-sponses against Q�-VLP. A, Q�-VLP IgA levels in BAL of WT andMyD88�/� BM chimeras immu-nized i.n. B and C, JH�/� mice werelethally irradiated and subsequentlyreconstituted with BM cells isolatedfrom JH�/� mice (80%) mixed with20% of BM cells isolated fromeither C57BL/6 (WT B cells) orTLR7�/� (TLR7�/� B cells) mice.After reconstitution, chimeric micewere immunized either i.n. (B) ors.c. (C) with 100 �g of Q�-VLP,and 20 days later specific IgA,IgG1, and IgG2a titers were deter-mined in BAL and serum, respec-tively, by ELISA. Geometric meanIgA ELISA titers � SEM are indi-cated; n.d., not detectable. The datashown are pooled from two inde-pendent experiments. Statistical sig-nificance was assessed by unpairedStudent’s t test (�, p � 0.05; and��, p � 0.01).

3794 TLR SIGNALING AND IgA RESPONSES

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

DCs, whereas in situations where these cells were unable to mi-grate (CCR7�/� mice), Q�-VLPs drained via the lymphatic sys-tem (Fig. 6B).

Following s.c. administration, a large fraction of Q�-VLPs enterthe bloodstream and are distributed throughout the body (our un-published data). By analyzing Q�� cell populations in MLNs afters.c. immunization, we found that Q�-VLP was in association withdistinct cell populations such as macrophages, monocytes, andlymphocytes. Most strikingly, the alveolar macrophage (CD11c�

CD11b�) population was completely absent in the Q�-VLP� gate(Fig. 6C and data not shown).

By comparing the numbers of AFCs secreting Q�-VLP-specificAbs in MLNs of i.n. and s.c. immunized mice, we observed thatboth groups elicited similar numbers of AFCs secreting anti-Q�-VLP IgG; however, the numbers of AFCs secreting anti-Q�-VLPIgA were significantly reduced in MLN of s.c. immunized mice(Fig. 6D). This suggests that CD11c� cells that take up Q�-VLPand migrate to MLNs are the key cells for the induction of mucosalIgA responses.

TLR7 signaling in lung DCs and alveolar macrophages inducesexpression of APRIL and BAFF

Next, we sought to understand why the lack of TLR7 signaling inlung DCs and alveolar macrophages has implications in the mu-cosal IgA response against Q�-VLP. We have seen that TLR7signaling has only a minor impact in lung DC and alveolar mac-rophage migration to the MLNs (data not shown). Therefore, weanticipated that there must be another explanation for the need ofTLR signaling in these cells to induce mucosal IgA responses. Ithas been shown that upon CpG and dsRNA stimulation, mucosalDCs increase BAFF and APRIL expression (7, 43). We weretherefore wondering whether ssRNA signaling via TLR7 ex-pressed in lung DCs and alveolar macrophages has a similar effect.To address this question, we sorted the Q�� cells from MLNs ofWT and TLR7�/� mice that received Q�-VLP i.n. and comparedthe expression level of APRIL and BAFF. Fig. 7A shows that thelevels of APRIL and BAFF were reduced 3- and 2-fold, respec-tively, in the Q�� cells isolated from the TLR7�/� mice whencompared with the WT group. This result indicates that optimalBAFF and APRIL expression by lung DCs and alveolar macro-phages occurs following TLR stimulation in vivo.

Alveolar macrophages and lung DCs sense ssRNA to induceoptimal IgA response

To directly address the hypothesis that the CD11c� cells found inassociation with Q�-VLP in MLN of i.n. immunized mice are thecell population that requires TLR7 signals to promote IgA re-sponses, we generated mixed BM chimeras using BM cells fromCD11c-DTR-transgenic (CD11c-DTR) mice. CD11c-DTR micehave the CD11c promoter driving the expression of the monkeyDTR, which allows the conditional depletion of CD11c� cells fol-lowing DT administration (30). We therefore adopted this strategyto specifically deplete DCs and alveolar macrophages. Specifi-cally, WT C57BL/6 mice were lethally irradiated and reconstitutedwith BM cells isolated from CD11c-DTR mice mixed with BMcells isolated from either TLR7�/� or WT control mice. Upon DTadministration, all members of the CD11c� population in the chi-meric mice completely lacked TLR7 signaling and therefore wecould directly address the role of TLR7 signaling in these cells ininducing IgA responses. Fig. 7B shows that the levels of VLP-specific IgA in serum and BAL were reduced in chimeric mice inwhich DCs lack TLR7 expression, therefore confirming that TLR7signaling in CD11c� DCs and alveolar macrophages is pivotal inregulating mucosal IgA responses.

FIGURE 6. Determination of cells in association with Q�-VLP in thelung and MLNs. A, Mice were immunized i.n. with 50 �g of Q�-VLPlabeled with Alexa Fluor 488 and the lung cell populations in associationwith Q�-VLP were assessed in the lung by flow cytometry 4 and 24 h later.B and C, Cells in association with Q�-VLP in the MLN after 24 h of i.n.(B) and s.c. (C) immunization are shown. Comparison of pattern of Q�-VLP distribution between WT and CCR7�/� mice immunized i.n. is shownin B. FSC, Forward scatter. D, Number of AFCs secreting Q�-VLP IgAand IgG in MLNs of i.n. and s.c. immunized mice were determined byELISPOT on day 20 postimmunization. Mean values � SEM are indicated.The data shown are pooled from two independent experiments. Statisticalsignificance was assessed by the unpaired Student t test (��, p � 0.01).

3795The Journal of Immunology

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

DiscussionMost pathogens invade the body through the mucosa, and somecause local infections at mucosal sites. Secretory IgA plays a crit-ical role in preventing infections at these sites. Furthermore, IgA isthe most abundant Ig isotype in humans and understanding itsregulation is of major importance.

In the present study we dissected the mechanisms involved inthe regulation of IgA responses against Q�-VLP administeredvia mucosal or systemic routes. Our data show that systemicIgA following s.c. immunization does not require Th cells orcognate interaction between T and B cells. In marked contrast,mucosal Q�-VLP-specific IgA response in the lung wasstrongly dependent on T cell help. It has been shown that DCscan induce T cell- and CD40-independent CSR through BAFFand APRIL (44). However, we observed in mice deficient forTACI and BCMA (TACI;BCMA�/�) that BAFF and APRILsignaling were dispensable for the induction of systemic TI IgAtiters in response to Q�-VLP immunization. This finding con-tradicts the literature. Actually, it has been shown that systemicAb responses elicited by 4-hydroxy-3-nitrophenylacetate (NP)-Ficoll (prototype TI-2 Ag) but not by NP-CGG (a TD Ag) arestrongly regulated by TACI (45, 46). This difference can beexplained by differences in Ag nature, i.e., NP-Ficoll is a TI-2Ag whereas VLP is a TI-1 Ag and therefore has a strongercapacity to generate B cell responses on its own. Another pos-sibility is that the RNA present on VLPs overcomes the acti-vation signals required via TACI. It still could be possible thatBAFF-BAFF-R interaction plays a role in the regulation of TIsystemic IgA responses. However, a major role has been attrib-uted to signaling via TACI because TACI�/� mice have lowserum IgA in response to TI type II Ags (45). Furthermore,

APRIL�/� mice also showed reduced systemic IgA in responseto NP-LPS (a TI-1 Ag) (47). Collectively, these data points toan important role of APRIL-TACI interaction in regulating TIIgA CSR and make it unlikely that BAFF-BAFF-R interactionis important to regulate the systemic TI IgA responsesagainst VLPs.

In contrast to the systemic TI IgA responses, the TD IgA titersin BAL were significantly reduced in TACI;BCMA�/� mice aswell as in TACI�/� mice. A dominant role for TACI confirmsprevious findings showing that TACI is important for mediatingisotype switching and IgA production by B cells (48, 49). Surpris-ingly, however, TACI was important for driving a TD IgA re-sponse rather than TI IgA responses. This is not consistent with thedogma that TACI mediates TI IgA CSR (45, 46). Importantly, themucosal IgA levels in BAL were completely abolished in bothMHC II�/� as well as in CD40�/� mice, whereas the reductionwas only partial in TACI;BCMA�/� mice. This possibly indicatesthe synergistic role of Th cells and TACI signaling. Although,CD40-CD40L engagement is crucial for mucosal IgA CSR, TACImay be involved in IgA production and plasma cell survival. Inagreement with this hypothesis, it has been recently shown thatTACI enhances the differentiation of B cells into AFCs in cul-tures containing limiting conditions of CD40 ligation, suggest-ing that TACI may be important also for the Ab response to TDAgs (50).

TGF� has been shown to be an important cytokine for IgACSR in vitro (38, 51, 52). In vivo, it was found that the IgAresponse, both in serum and in mucosal washes, against TD Agswas completely abolished in mice that lack TGF�-RII in B cells(39, 40). More recently, it was found that expression of TGF�-RII on naive B cells is induced by inducible NO synthase

FIGURE 7. After i.n. immuniza-tion, TLR7 signaling in CD11c� cellsis necessary for optimal Q�-VLP IgAresponses. A, Q��-cells were sortedby flow cytometry from MLNs of WTor TLR7�/� mice immunized i.n.,and the expressions of BAFF andAPRIL were measured by quantita-tive real-time PCR. Mean values �SEM are indicated (n � 3). B, WTC57BL/6 mice were lethally irradi-ated and subsequently reconstitutedi.v. with BM cells isolated fromCD11c-DTR mice mixed with BMcells isolated from either C57BL/6(WT DCs) or TLR7�/� (TLR7�/�

DCs) mice. After reconstitution (�6wk), chimeric mice were treated for 5days with DT i.n. After the first DTadministration, mice were immunizedi.n. with 100 �g of AP205-VLP, and10 days later specific IgA titers weredetermined in serum and BAL byELISA. Geometric mean IgA ELISAtiters � SEM are indicated (n � 3).We hypothesized that TLR7 deficiencyin CD11c� cells would impair the IgAresponse after i.n. immunization andtherefore compared the statistical sig-nificance by one-tailed Student’s t test(�, p � 0.05; and ��, p � 0.01).

3796 TLR SIGNALING AND IgA RESPONSES

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

(iNOS) with IgA CSR being impaired in iNOS�/� mice (7).However, the in vivo role of TGF� against viral particles hadnot yet been addressed. In the present study we found a signif-icant reduction of Q�-VLP-specific TD IgA titers in BAL ofmice immunized i.n. that lack TGF�RII in B cells, whereassystemic TI IgA levels were not significantly reduced. Thus,mucosal IgA responses to viral particles seem to depend on theone hand on TGF� and on the other hand on the presence of Thcells in addition to BAFF or APRIL secreted by DCs. In con-trast, the induction of serum IgA upon systemic exposure toVLPs is largely independent of TGF�. Furthermore, systemicIgA was independent of Th cells as well as signaling via TACI.These data indicate that there must be an additional, as yetundetermined, factor important for TI IgA CSR upon systemicexposure to viral particles. The fact that mice deficient inTGF�RII expression in B cells previously have been shown togenerate reduced IgA responses to chicken �-globulin (36) mayreflect differences in Ag size (soluble protein vs VLP) or theabsence of TLR ligands or may even be due to differences in thegenetic background and/or rearing conditions that potentiallyaffect the composition of the mucosal microbiota.

The role of TLR ligands as a third signal for Ab class switch-ing has been previously described in vitro (17). In vivo, severalreports have indicated a role for TLR signaling in B cells as aIgG switching factor (13, 16, 18 –20). However, the role of TLRsignaling in regulating IgA responses has not been well studied.By comparing Q�-VLP IgA titers in serum and BAL betweengroups of mice where TLR7 signaling was present or not, wefound, unexpectedly, that TLR7 signaling had a crucial role inregulating IgA responses against Q�-VLPs independently ofthe route of immunization. The same was true for IgG2a, butnot for IgG1.

In accordance with this notion, other groups have also shownthat microbial Ags are of ultimate importance in activating B cellsand inducing IgA responses at both the serological and the muco-sal level. Tezuka et al. reported that NO plays a crucial role in IgAclass switching and secretion (7). Interestingly, iNOS expressionhas been shown to be dependent on TLR stimulation, as DCs fromgermfree or MyD88�/� mice express reduced levels of iNOS. In-terestingly, it has been shown that BCR-deficient mice could de-velop a germinal center in the GALT (53). However, germinalcenters failed to develop in mice maintained on antibiotics treat-

ment, indicating a major dependence on the intestinal bacterialflora. More recently, it has also been shown that chronic epithelialTLR activation resulted in increased B cell recruitment and, con-sequently, fecal IgA content (21).

Our experiments with BM chimeras demonstrated that TLR7signaling in B cells was required for TI IgA CSR. Together withthe notion that IgA responses under these conditions occur inde-pendently of Th cells, CD40, TGF�, and TACI, we propose thatTLR7 signaling directly to B cells may represent the missing factorrequired for TI IgA CSR. This fits with the idea that, following s.c.immunization, Q�-VLP drains freely to bloodstream and reachesdifferent lymphoid organs, including the spleen (our unpublisheddata). This might facilitate the direct interaction of Q�-VLP withB cells. In contrast, mucosal IgA production requires TLR7 sig-naling by alveolar macrophages and lung DCs rather than directingTLR signaling to B cells. Thus, under these conditions, B cellsmay have little direct contact with VLPs, and IgA is induced in-directly, via CD40-CD40L, TGF�-TGF�RII, and APRIL/BAFF-TACI signaling. Indeed, APRIL and BAFF, the ligands for TACI,were up-regulated in alveolar macrophages and lung DCs uponTLR7 triggering (Fig. 8A). These data are in contrast to a recentreport showing that human upper respiratory mucosa B cells ex-pressing TLR3 can initiate TI IgA response following dsRNAactivation directly to B cells (43). Perhaps soluble ssRNA not inassociation with viral particles also would trigger similar mech-anism, because most likely it would target and be sampled by acell population other than that of alveolar macrophages andlung DCs.

The observation that alveolar macrophages are a key populationfor induction of IgA seems counterintuitive, because macrophagesare usually nonmigratory. Alveolar macrophages, however, seemto be different because it is an intrinsic feature of CD11c� alveolarmacrophages (42) to migrate to MLN and initiate immune re-sponses against particulate Ags (54). Our results have furtherdemonstrated that alveolar macrophages migrate to the MLNsin a CCR7-dependent fashion. It will be interesting to directlydistinguish in vivo the role of lung DCs and alveolar macro-phages in regulating mucosal IgA responses against VLPs. Be-cause different types of DCs and macrophages have distinctroles in the initiation and maintenance of immune responses(55), it might well be that the mechanism described here is not

FIGURE 8. TLR7 regulates ap-propriate IgA responses againstVLPs directly through signaling inB cells and indirectly through activa-tion of CD11c� lung DCs and alveolarmacrophages (M�). A, Mucosal IgA re-sponses require lung DC and alveolarmacrophage activation via TLR7 stim-ulation. Activated DCs/alveolar macro-phages modulate IgA responses on theone hand by activating cognate CD4� Tcells, which in turn provide help viaCD40L-CD40 interaction and cytokinesecretion (i.e., TGF�). Additionally, ac-tivated DCs/alveolar macrophages candirectly modulate IgA responses by se-creting APRIL and BAFF upon TLR7stimulation. B, Optimal TI systemicIgA responses simply require strongBCR cross-linking and TLR7 signalingdirectly in B cells that are provided bydirect interaction with VLPs.

3797The Journal of Immunology

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

applicable for IgA induction in other mucosal sites because al-veolar macrophages and lung DCs are the cell populations re-stricted to the lung tissue.

In summary (Fig. 8), our data demonstrate that for TI systemicIgA responses, direct TLR7 signaling to B cells is crucial, whereasfor TD mucosal IgA responses, regulation occurs indirectlythrough the ability of TLR7 signaling to activate alveolar macro-phages and DCs, leading to increased BAFF and APRIL produc-tion, Th cell activation resulting in TGF� secretion, and deliv-ery of CD40-CD40L cognate signals to B cells. It will beinteresting to further determine whether this is a “unique” fea-ture of TLR7 or whether other TLR agonists have similar prop-erties. We have demonstrated for the first time that direct TLRsignaling to B cells in combination with multivalent Ag resultsin TI IgA CSR.

AcknowledgmentsWe thank Nicola L. Harris, Manfred Kopf (Swiss Federal Institute of Tech-nology, Zurich, Switzerland), and Vania Manolova (Cytos BiotechnologyAG, Zurich-Schlieren, Switzerland) for helpful discussions, Monika Bauerand Simone Muntwiler (Cytos Biotechnology AG) for cell sorting, AnnaFlace, Alexander Titz, Franziska Wagen (Cytos Biotechnology AG), andAubry Tardivel (University of Lausanne, Epalinges, Switzerland) for ex-cellent technical assistance, and Susan Kalled (Biogen Idec, Cambridge,MA) for TACI;BCMA�/� mice.

DisclosuresJ.B., A.J., H.J.H., P.S., and M.F.B. are all employees of Cytos Biotech-nology AG and may hold shares or share options in the company.

References1. Bachmann, M. F., and M. Kopf. 1999. The role of B cells in acute and chronic

infections. Curr. Opin. Immunol. 11: 332–339.2. Zinkernagel, R. M. 2001. Maternal antibodies, childhood infections, and auto-

immune diseases. N. Engl. J. Med. 345: 1331–1335.3. Scheibel, I., M. W. Bentzon, P. E. Christensen, and A. Biering. 1966. Duration

of immunity to diphtheria and tetanus after active immunization. Acta Pathol.Microbiol. Scand. 67: 380–392.

4. Ahmed, R., and D. Gray. 1996. Immunological memory and protective immunity:understanding their relation. Science 272: 54–60.

5. Fagarasan, S., and T. Honjo. 2003. Intestinal IgA synthesis: regulation of front-line body defences. Nat. Rev. Immunol. 3: 63–72.

6. Cerutti, A. 2008. The regulation of IgA class switching. Nat. Rev. Immunol. 8:421–434.

7. Tezuka, H., Y. Abe, M. Iwata, H. Takeuchi, H. Ishikawa, M. Matsushita,T. Shiohara, S. Akira, and T. Ohteki. 2007. Regulation of IgA production bynaturally occurring TNF/iNOS-producing dendritic cells. Nature 448: 929–933.

8. Macpherson, A. J., K. D. McCoy, F. E. Johansen, and P. Brandtzaeg. 2008. Theimmune geography of IgA induction and function. Mucosal Immunol. 1: 11–22.

9. Bergqvist, P., E. Gardby, A. Stensson, M. Bemark, and N. Y. Lycke. 2006. GutIgA class switch recombination in the absence of CD40 does not occur in thelamina propria and is independent of germinal centers. J. Immunol. 177:7772–7783.

10. Mora, J. R., M. Iwata, B. Eksteen, S. Y. Song, T. Junt, B. Senman, K. L. Otipoby,A. Yokota, H. Takeuchi, P. Ricciardi-Castagnoli, et al. 2006. Generation of gut-homing IgA-secreting B cells by intestinal dendritic cells. Science 314:1157–1160.

11. Sangster, M. Y., J. M. Riberdy, M. Gonzalez, D. J. Topham, N. Baumgarth, andP. C. Doherty. 2003. An early CD4� T cell-dependent immunoglobulin A re-sponse to influenza infection in the absence of key cognate T-B interactions.J. Exp. Med. 198: 1011–1021.

12. Bernasconi, N. L., E. Traggiai, and A. Lanzavecchia. 2002. Maintenance of se-rological memory by polyclonal activation of human memory B cells. Science298: 2199–2202.

13. Meyer-Bahlburg, A., S. Khim, and D. J. Rawlings. 2007. B cell intrinsic TLRsignals amplify but are not required for humoral immunity. J. Exp. Med. 204:3095–3101.

14. Lin, L., A. J. Gerth, and S. L. Peng. 2004. CpG DNA redirects class-switchingtowards “Th1-like” Ig isotype production via TLR9 and MyD88. Eur. J. Immu-nol. 34: 1483–1487.

15. He, B., X. Qiao, and A. Cerutti. 2004. CpG DNA induces IgG class switch DNArecombination by activating human B cells through an innate pathway that re-quires TLR9 and cooperates with IL-10. J. Immunol. 173: 4479–4491.

16. Pasare, C., and R. Medzhitov. 2005. Control of B-cell responses by Toll-likereceptors. Nature 438: 364–368.

17. Ruprecht, C. R., and A. Lanzavecchia. 2006. Toll-like receptor stimulation as athird signal required for activation of human naive B cells. Eur. J. Immunol. 36:810–816.

18. Gavin, A. L., K. Hoebe, B. Duong, T. Ota, C. Martin, B. Beutler, andD. Nemazee. 2006. Adjuvant-enhanced antibody responses in the absence ofToll-like receptor signaling. Science 314: 1936–1938.

19. Jegerlehner, A., P. Maurer, J. Bessa, H. J. Hinton, M. Kopf, and M. F. Bachmann.2007. TLR9 signaling in B cells determines class switch recombination to IgG2a.J. Immunol. 178: 2415–2420.

20. Heer, A. K., A. Shamshiev, A. Donda, S. Uematsu, S. Akira, M. Kopf, andB. J. Marsland. 2007. TLR signaling fine-tunes anti-influenza B cell responseswithout regulating effector T cell responses. J. Immunol. 178: 2182–2191.

21. Shang, L., M. Fukata, N. Thirunarayanan, A. P. Martin, P. Arnaboldi,D. Maussang, C. Berin, J. C. Unkeless, L. Mayer, M. T. Abreu, and S. A. Lira.2008. Toll-like receptor signaling in small intestinal epithelium promotes B-cellrecruitment and IgA production in lamina propria. Gastroenterology 135:529–538.

22. Cielens, I., V. Ose, I. Petrovskis, A. Strelnikova, R. Renhofa, T. Kozlovska, andP. Pumpens. 2000. Mutilation of RNA phage Q� virus-like particles: from ico-sahedrons to rods. FEBS Lett. 482: 261–264.

23. Bessa, J., N. Schmitz, H. J. Hinton, K. Schwarz, A. Jegerlehner, andM. F. Bachmann. 2008. Efficient induction of mucosal and systemic immuneresponses by virus-like particles administered intranasally: implications for vac-cine design. Eur. J. Immunol. 38: 114–126.

24. Hemmi, H., T. Kaisho, O. Takeuchi, S. Sato, H. Sanjo, K. Hoshino, T. Horiuchi,H. Tomizawa, K. Takeda, and S. Akira. 2002. Small anti-viral compounds acti-vate immune cells via the TLR7 MyD88-dependent signaling pathway. Nat Im-munol. 3: 196–200.

25. Adachi, O., T. Kawai, K. Takeda, M. Matsumoto, H. Tsutsui, M. Sakagami,K. Nakanishi, and S. Akira. 1998. Targeted disruption of the MyD88 gene resultsin loss of IL-1- and IL-18-mediated function. Immunity. 9: 143–150.

26. Chen, J., M. Trounstine, C. Kurahara, F. Young, C. C. Kuo, Y. Xu, J. F. Loring,F. W. Alt, and D. Huszar. 1993. B cell development in mice that lack one or bothimmunoglobulin � light chain genes. EMBO J. 12: 821–830.

27. Grusby, M. J., R. S. Johnson, V. E. Papaioannou, and L. H. Glimcher. 1991.Depletion of CD4� T cells in major histocompatibility complex class II-deficientmice. Science 253: 1417–1420.

28. Kawabe, T., T. Naka, K. Yoshida, T. Tanaka, H. Fujiwara, S. Suematsu,N. Yoshida, T. Kishimoto, and H. Kikutani. 1994. The immune responses inCD40-deficient mice: impaired immunoglobulin class switching and germinalcenter formation. Immunity 1: 167–178.

29. Forster, R., A. Schubel, D. Breitfeld, E. Kremmer, I. Renner-Muller, E. Wolf, andM. Lipp. 1999. CCR7 coordinates the primary immune response by establishingfunctional microenvironments in secondary lymphoid organs. Cell 99: 23–33.

30. Jung, S., D. Unutmaz, P. Wong, G. Sano, K. De los Santos, T. Sparwasser, S. Wu,S. Vuthoori, K. Ko, F. Zavala, et al. 2002. In vivo depletion of CD11c� dendriticcells abrogates priming of CD8� T cells by exogenous cell-associated antigens.Immunity 17: 211–220.

31. Rickert, R. C., K. Rajewsky, and J. Roes. 1995. Impairment of T-cell-dependentB-cell responses and B-1 cell development in CD19-deficient mice. Nature 376:352–355.

32. Shulga-Morskaya, S., M. Dobles, M. E. Walsh, L. G. Ng, F. MacKay, S. P. Rao,S. L. Kalled, and M. L. Scott. 2004. B cell-activating factor belonging to the TNFfamily acts through separate receptors to support B cell survival and T cell-independent antibody formation. J. Immunol. 173: 2331–2341.

33. Klovins, J., G. P. Overbeek, S. H. van den Worm, H. W. Ackermann, andJ. van Duin. 2002. Nucleotide sequence of a ssRNA phage from Acinetobacter:kinship to coliphages. J. Gen. Virol. 83: 1523–1533.

34. Kozlovska, T. M., I. Cielens, D. Dreilinna, A. Dislers, V. Baumanis, V. Ose, andP. Pumpens. 1993. Recombinant RNA phage Q� capsid particles synthesized andself-assembled in Escherichia coli. Gene 137: 133–137.

35. Gatto, D., T. Pfister, A. Jegerlehner, S. W. Martin, M. Kopf, andM. F. Bachmann. 2005. Complement receptors regulate differentiation of bonemarrow plasma cell precursors expressing transcription factors Blimp-1 andXBP-1. J. Exp. Med. 201: 993–1005.

36. Gatto, D., C. Ruedl, B. Odermatt, and M. F. Bachmann. 2004. Rapid response ofmarginal zone B cells to viral particles. J. Immunol. 173: 4308–4316.

37. Ochsenbein, A. F., D. D. Pinschewer, B. Odermatt, A. Ciurea, H. Hengartner, andR. M. Zinkernagel. 2000. Correlation of T cell independence of antibody re-sponses with antigen dose reaching secondary lymphoid organs: implications forsplenectomized patients and vaccine design. J. Immunol. 164: 6296–6302.

38. Sonoda, E., R. Matsumoto, Y. Hitoshi, T. Ishii, M. Sugimoto, S. Araki,A. Tominaga, N. Yamaguchi, and K. Takatsu. 1989. Transforming growth factor� induces IgA production and acts additively with interleukin 5 for IgA produc-tion. J. Exp. Med. 170: 1415–1420.

39. Cazac, B. B., and J. Roes. 2000. TGF-� receptor controls B cell responsivenessand induction of IgA in vivo. Immunity 13: 443–451.

40. Borsutzky, S., B. B. Cazac, J. Roes, and C. A. Guzman. 2004. TGF-� receptorsignaling is critical for mucosal IgA responses. J. Immunol. 173: 3305–3309.

41. Massacand, J. C., P. Kaiser, B. Ernst, A. Tardivel, K. Burki, P. Schneider, andN. L. Harris. 2008. Intestinal bacteria condition dendritic cells to promote IgAproduction. PLoS ONE. 3: e2588.

42. Jakubzick, C., F. Tacke, J. Llodra, N. van Rooijen, and G. J. Randolph. 2006.Modulation of dendritic cell trafficking to and from the arways. J. Immunol. 176:3578–3584.

43. Xu, W., P. A. Santini, A. J. Matthews, A. Chiu, A. Plebani, B. He, K. Chen, andA. Cerutti. 2008. Viral double-stranded RNA triggers Ig class switching by ac-tivating upper respiratory mucosa B cells through an innate TLR3 pathway in-volving BAFF. J. Immunol. 181: 276–287.

3798 TLR SIGNALING AND IgA RESPONSES

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

44. Litinskiy, M. B., B. Nardelli, D. M. Hilbert, B. He, A. Schaffer, P. Casali, andA. Cerutti. 2002. DCs induce CD40-independent immunoglobulin class switch-ing through BLyS and APRIL. Nat. Immunol. 3: 822–829.

45. Yan, M., H. Wang, B. Chan, M. Roose-Girma, S. Erickson, T. Baker, D. Tumas,I. S. Grewal, and V. M. Dixit. 2001. Activation and accumulation of B cells inTACI-deficient mice. Nat Immunol. 2: 638–643.

46. von Bulow, G. U., J. M. van Deursen, and R. J. Bram. 2001. Regulation of theT-independent humoral response by TACI. Immunity 14: 573–582.

47. Castigli, E., S. Scott, F. Dedeoglu, P. Bryce, H. Jabara, A. K. Bhan,E. Mizoguchi, and R. S. Geha. 2004. Impaired IgA class switching in APRIL-deficient mice. Proc. Natl. Acad. Sci. USA 101: 3903–3908.

48. Castigli, E., S. A. Wilson, S. Scott, F. Dedeoglu, S. Xu, K. P. Lam, R. J. Bram,H. Jabara, and R. S. Geha. 2005. TACI and BAFF-R mediate isotype switchingin B cells. J. Exp. Med. 201: 35–39.

49. Sakurai, D., H. Hase, Y. Kanno, H. Kojima, K. Okumura, and T. Kobata. 2007.TACI regulates IgA production by APRIL in collaboration with HSPG. Blood109: 2961–2967.

50. Castigli, E., S. A. Wilson, A. Elkhal, E. Ozcan, L. Garibyan, and R. S. Geha.2007. Transmembrane activator and calcium modulator and cyclophilin ligandinteractor enhances CD40-driven plasma cell differentiation. J. Allergy Clin. Im-munol. 120: 885–891.

51. Stavnezer, J. 1996. Antibody class switching. Adv. Immunol. 61: 79–146.52. Fayette, J., B. Dubois, S. Vandenabeele, J. M. Bridon, B. Vanbervliet, I. Durand,

J. Banchereau, C. Caux, and F. Briere. 1997. Human dendritic cells skew isotypeswitching of CD40-activated naive B cells towards IgA1 and IgA2. J. Exp. Med.185: 1909–1918.

53. Casola, S., K. L. Otipoby, M. Alimzhanov, S. Humme, N. Uyttersprot,J. L. Kutok, M. C. Carroll, and K. Rajewsky. 2004. B cell receptor signal strengthdetermines B cell fate. Nat. Immunol. 5: 317–327.

54. Harmsen, A. G., B. A. Muggenburg, M. B. Snipes, and D. E. Bice. 1985. The roleof macrophages in particle translocation from lungs to lymph nodes. Science 230:1277–1280.

55. Iwasaki, A. 2007. Mucosal dendritic cells. Annu. Rev. Immunol. 25: 381–418.

3799The Journal of Immunology

by guest on July 27, 2018http://w

ww

.jimm

unol.org/D

ownloaded from