Mucosal adjuvant activity of cholera toxin requiresTh17 cells and protects against inhalation anthraxSandip K. Dattaa,1, Mojgan Sabetb,2, Kim Phung L. Nguyenb,2, Patricia A. Valdeza, Jose M. Gonzalez-Navajasb,Shamima Islama, Ivan Mihajlovb, Joshua Fiererb,c, Paul A. Inselb,d, Nicholas J. Websterb, Donald G. Guineyb,and Eyal Razb

aLaboratory of Clinical Infectious Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892;bDepartment of Medicine and dDepartment of Pharmacology, University of California, La Jolla, CA 92093; and cDepartment of Infectious Diseases, VeteransAffairs Healthcare, San Diego, CA 92161

Edited by Rino Rappuoli, Novartis Vaccines, Siena, Italy, and approved April 28, 2010 (received for review February 25, 2010)

Cholera toxin (CT) elicits a mucosal immune response in mice whenused as a vaccine adjuvant. The mechanisms by which CT exertsits adjuvant effects are incompletely understood. We show thatprotection against inhalation anthrax by an irradiated spore vaccinedepends on CT-mediated induction of IL-17-producing CD4 Th17cells. Furthermore, IL-17 is involved in the induction of serum andmucosal antibody responses by CT. Th17 cells induced by CT havea unique cytokine profile compared with those induced by IL-6 andTGF-β, and their inductionbyCT requires cAMP-dependent secretionof IL-1β and β-calcitonin gene-related peptide by dendritic cells.These findings demonstrate that Th17 cells mediate mucosal adju-vant effects of CT and identify previously unexplored pathways in-volved in Th17 induction that could be targeted for developmentof unique mucosal adjuvants.

IL-17 | dendritic cell | T cell | vaccine | cAMP

Cholera toxin (CT), an exotoxin produced by Vibrio cholerae, isresponsible for the diarrhea of cholera and is a potent mu-

cosal-vaccine adjuvant in experimental animals (1–3). After thepentameric B subunit of CT binds the GM1 ganglioside receptor,CT enters the cell and its catalytic-A subunit ADP ribosylatesand activates the heterotrimeric guanine nucleotide binding-protein Gs, which then stimulates cAMP production by adenylylcyclase. In intestinal epithelial cells, cAMP induces chloride se-cretion, leading to the massive fluid loss that characterizescholera (4). The toxicity of CT prevents its use as an adjuvant inhumans, but understanding the mechanisms underlying its ad-juvant activity may lead to development of nontoxic mucosaladjuvants for clinical use.In contrast to what is known about the ability of CT to stim-

ulate intestinal ion transport, the mechanisms by which CT elicitsa mucosal immune response are incompletely understood. CT istypically classified as a “Th2 adjuvant” because it induces Th2-associated cytokines and antibodies, and can either induce orinhibit a Th1 response (2, 5, 6). Th17 cells are a recently iden-tified subset of CD4 T cells that secrete IL-17 (also known as IL-17A), are induced by IL-6 in the context of TGF-β (7–9), andplay an important role in autoimmunity and the control of cer-tain extracellular pathogens, in particular at mucosal sites (10).In light of their important role in mucosal immunity, we hy-pothesized that Th17 cells may contribute to the mucosal adju-vant effects of CT. Indeed, a recent report identified the abilityof CT to induce Th17 cells (11), although the role of these cellsin mediating the adjuvant effects of CT remains unresolved.We show here that Th17 cells are required to mediate the ad-

juvant effects of CT. We developed an intranasal vaccine againstinhalation anthrax using inactivated bacterial spores, and foundthat protection by this vaccine depends on induction of Th17cells by CT. Dissection of the mechanism underlying Th17 in-duction by CT revealed a panel of gene products, in addition toIL-6, which contribute to Th17 differentiation. Furthermore,we found that IL-17 deficiency impairs the antibody response

elicited by CT. These results suggest IL-17 plays a central role inthe mucosal adjuvant activity of CT, and identify previously un-explored pathways involved in the regulation of Th17 responsesthat could be targeted for mucosal adjuvant development.

ResultsIL-17 Contributes to the Mucosal Adjuvant Activity of CT. We firstidentified the IL-17–dependent adjuvant activity of CT whendeveloping a spore-based anthrax vaccine. The currently avail-able anthrax vaccine uses alum as an adjuvant and depends onantibody to the anthrax protective antigen (PA) (12). Based onthe sensitivity of anthrax spores to γ-irradiation (13, 14) andour recent data highlighting the immunogenic properties of a γ-irradiated bacterial vaccine (15), we used irradiated anthraxspores to immunize A/J mice against a model of inhalation an-thrax. Irradiated spore-vaccine preparations derived from bac-teria that either contained or lacked the gene encoding PAprotected against challenge with the toxin-producing Sternestrain (Fig. 1A). Vaccine efficacy required addition of the mu-cosal adjuvant, CT (Fig. 1A).T-cell depletion and serum transfer studies done in immunized

mice at the time of infectious challenge indicated that CD4T cells were critical during the effector phase of protection (Fig.1B and Fig. S1). Mucosal immunization with CT and irradiatedanthrax spores induced spore-specific CD4 T cells that producedIL-17 but not IFN-γ, IL-5, or IL-10 (Fig. 1C). CT alone did notelicit this IL-17 response, indicating an antigen-specific memoryresponse not attributable to prolonged effects of CT on cells ofthe innate immune system (Fig. 1C). Antibody neutralization ofIL-17, but not IFN-γ, at the time of challenge inhibited vaccine-mediated protection against inhalation anthrax (Fig. 1D). Be-cause mice on the C57BL/6 background are not susceptible tothis anthrax model, we could not use IL-17 knockout mice forthese experiments, but antibody neutralization at the time ofchallenge allowed us to evaluate the role of IL-17 specifically invaccine-induced effector mechanisms.C57BL/6 mice immunized either intranasally or by gavage with

CT and ovalbumin (OVA) also showed similar induction ofOVA-specific Th17 cells in addition to the Th2 response de-scribed in the literature (Fig. S2). Immunization with CT plusOVA also induced an OVA-specific IL-22 response but nota detectable IL-4, IL-21, or IL-17F response (Fig. S2).

Author contributions: S.K.D., M.S., K.P.L.N., P.A.V., J.M.G.-N., J.F., P.A.I., N.J.W., D.G.G.,and E.R. designed research; S.K.D., M.S., K.P.L.N., P.A.V., J.M.G.-N., S.I., and I.M. performedresearch; S.K.D., M.S., K.P.L.N., P.A.V., J.M.G.-N., S.I., I.M., and N.J.W. analyzed data; andS.K.D. and M.S. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: dattas@niaid.nih.gov.2M.S. and K.P.L.N. contributed equally to this work.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1002348107/-/DCSupplemental.

10638–10643 | PNAS | June 8, 2010 | vol. 107 | no. 23 www.pnas.org/cgi/doi/10.1073/pnas.1002348107

CT Acts on Dendritic Cells to Elicit Th17 Cells. We reasoned that CTacts on antigen-presenting cells, such as dendritic cells (DCs), toinduce Th17 cells in vivo in the studies above. To test the abilityof CT-treated DCs to elicit Th17 cells, we incubated C57BL/6

mouse bone marrow-derived DCs with OVA in the absence orpresence of CT. The next day we added OVA-specific splenicCD4 T cells from OT-II mice. After 6 d, we restimulated the Tcells. Supernatants collected 24 h after restimulation showed thatCT significantly induced IL-17 and the Th17-associated cytokineIL-22, but not IFN-γ, IL-5, IL-10, or IL-21 (Fig. 2A).Because CT was present when T cells were added in the

experiments above, we evaluated whether CT induces IL-17 byacting via DCs or directly on T cells. Treatment of T cells with CTduring stimulation by formalin-fixed, OVA-pulsed DCs did notinduce IL-17 (Fig. 2B, Right bar), similar to CT-untreated cells(Fig. 2B, Left bar). In contrast, OVA-pulsed DCs treated with CTbefore fixation, and then incubated in their original CT-condi-tioned medium with T cells, induced IL-17 (Fig. 2B, Center bar).Thus, CT does not act directly on T cells but instead via DCs toinduce IL-17 production by T cells. Nod2 and Toll-like receptors(TLR) influence IL-17 production (9, 16). DCs from Nod2−/− andMyd88−/− mice retained the ability to elicit IL-17-secreting T cellsafter exposure to CT (Fig. 2C). Intact IL-17 induction by Tlr4−/−

DCs further confirmed TLR-independence (Fig. 2D).

Th17 Induction by CT Involves cAMP. We next investigated whethercAMP is involved in the DC-mediated induction of Th17 cells byCT. We found that either CT or forskolin, a direct activator ofadenylyl cyclase, increased cAMP levels in DCs (Fig. 3A). Treat-ment of DCs with a cell-permeant cAMP analog (8-Br-cAMP),or with agents that increase cellular cAMP [forskolin or 3-isobutyl-1-methylxanthine (IBMX), a phosphodiesterase in-hibitor], mimicked the ability of CT to induce IL-17 secretion byCD4 T cells (Fig. 3B). Treatment of DC with N6-benzoyl-cAMP,an activator of cAMP-dependent protein kinase (PKA), themajor downstream effector of cAMP (17), also induced IL-17(Fig. 3C). However, 8-CPT-2’-O-Me-cAMP, an activator of Epac(exchange protein activated by cAMP), involved in an alterna-tive cAMP signaling pathway (18, 19), failed to generate IL-17

B

C

A

Saline CT CT+spore0

10

20

30 IFN-IL-5IL-10IL-17A

*

Cytokine(ngml-1)

0 5 10 150

25

50

75

100

noneCD4CD8

Depletion:

Days

Percentsurvival

0 5 10 15 200

25

50

75

100

SalineSpore

CT+sporeCT+spore (no PA)

Days

Percentsurvival

D

0 5 10 15 200

25

50

75

100UnimmControl AbAnti-IFN-Anti-IL-17AAnti-IFN- +Anti-IL-17A

Days

Percentsurvival

Fig. 1. Protection by anthrax spore vaccine requires Th17 cell induction byCT. (A) Mice (10 per group) were immunized intranasally, as indicated, andmonitored for survival after live-spore challenge. (B) Mice (10–12 per group)were immunized intranasally with PA-deficient irradiated spores and CT.Indicated depleting antibodies were administered intraperitonially 1 d be-fore challenge. (C) Mice (4 per group) were immunized as indicated andthen antigen challenged with irradiated spores. Three days after challenge,splenic CD4+ cells were isolated and cocultured with antigen-presenting cellspulsed with or without irradiated spores to determine spore-specific cyto-kine production. Data are shown as the mean of each group, and error barsrepresent SD. *, P < 0.05 compared with saline-treated group. (D) Mice (8 pergroup) immunized with irradiated spores and CT were treated with the in-dicated antibodies before infection. Control mice (Unimm) received salinefor the immunization and depletion protocols.

A B

lm

gn(

eni

kot

yC

1 -)

C

0

1

2

3

4*

DC:

CD4+ T cell:No CT +CT No CTNo CT +CT +CT

IL-17 IFN-γ IL-5 IL-10 IL-21 IL-220

2

4

6

8

10 OVACT+OVA*

*

n.d.

*

lm

gn(

A7

1-LI

1-)

D

B6 Myd88-/- Nod2-/-0

1

2

3

4

5 OVA

CT+OVA

* *

*

IL-1

7A(n

gm

l-1)

B6 Tlr4-/-0

1

2

3

4

5OVACT+OVA

p=ns

IL-1

7A

(ng

ml-1

)

Fig. 2. CT-treated DCs elicit Th17 cells. (A) Splenic OT-II CD4 T cells were in-cubated for 6 d with C57BL/6 bonemarrow-derived DCs that had been treatedovernightwithOVA1(0μg/mL) in theabsenceorpresenceofCT (1μg/mL).Afteranti-CD3+anti-CD28 restimulation, supernatantswereanalyzedbyELISA.Datashown are representative of more than six independent experiments. (B) DCswere incubated with OVA or CT+OVA. The DCs were then fixed and resus-pended in their original medium. OT-II CD4 T cells were addedwith or withoutCT. After 6 d, the T cells were restimulated and their supernatants analyzedby ELISA. Data shown are representative of two independent experiments. (Cand D) DCs from C57BL/6 (B6), MyD88−/−, Nod2−/−, or Tlr4−/− mice were in-cubated with CT or OVA overnight before incubation with OT-II CD4+ T cells.After restimulation, supernatants were analyzed by ELISA. Data shown arerepresentativeof three independentexperiments.Allgraphs showthemeanoftriplicate samples and error bars reflect standard deviation. *, P < 0.05 com-pared to CT-untreated group; n.d., not detected.

A B

D

lm l

om

p( P

MA

c1-)

C

lm

gn(

A7

1-LI

1 -)

Media CT Forskolin 0

25

50

75

100

125

*

*

0

1

2

3

4

5

6 - OVA + OVA

*

*

Media CT CPT-O- Me-cAMP

Bnz- cAMP

lm

gn(

A7

1-LI

1-)

0.0

2.5

5.0

7.5

10.0 -OVA +OVA

*

*

* *

Media CT 8-Br- cAMP

IBMX Forsk.

lm

gn(

A7

1-LI

1 -)

Media CT CT+H89 0

1

2

3

4 -OVA

+OVA

p<0.001

Fig. 3. CT elicits Th17 cells via cAMP. (A) cAMP levels in DC lysates weremeasured by ELISA after 3 h incubation with media, CT, or forskolin (10 μM).(B and C) DCs were incubated overnight with media, CT, 8-Br-cAMP (500 μM),IBMX (100 μM), forskolin (10 μM), EPAC agonist (CPT-O-Me-cAMP, 500 μM), orPKAagonist (Bnz-cAMP, 500μM) in theabsenceorpresenceofOVA.OT-II CD4Tcells were then added, restimulated, and their supernatants analyzed by ELISA.(D) DCs were treatedwith or without H89 (10 μM) during a 3 h incubationwithmedia, OVA, or CT. DCs were then fixed and incubated in their original mediawith OT-II CD4 T cells before analysis of restimulated T cell supernatants. Alldata shown are representative of two independent experiments. Data areshown as the mean of triplicate samples, and error bars reflect standard de-viation. *, P < 0.05 compared with media-treated group.

Datta et al. PNAS | June 8, 2010 | vol. 107 | no. 23 | 10639

IMMUNOLO

GY

(Fig. 3C). Treatment of DCs with the PKA-inhibitor H89 beforeincubation with CT blunted the ability of CT to induce IL-17(Fig. 3D). Purified CT B-subunit also had greatly reduced abilityto induce IL-17 (Fig. S3). Taken together, these data imply thata cAMP/PKA-dependent program triggered by CT in DCs isinvolved in inducing IL-17 in T cells.

Th17 Differentiation by CT Depends on a Combination of SecretedFactors. To investigate the mechanisms by which CT-treated DCsinduce T-cell IL-17 production, we first determined whetherinduction of T-cell IL-17 by DCs requires cognate interaction viainduced cell-surface molecules or depends on secreted factors.Compared with unfixed DCs (Fig. 4A, Left bars), DCs that wereformalin-fixed after incubation with CT and OVA failed to in-duce IL-17 secretion in T cells (Fig. 4A, Center bars), indicatingthat cognate interaction was insufficient. Adding back superna-tant from CT-treated DCs restored the ability of fixed DCs toinduce IL-17 (Fig. 4A, Right bars), implicating one or more se-creted factors. Addition of medium from CT-treated DCs (CT-conditioned medium, CT-CM) directly to sorted naive CD4 Tcells (>99% purity) (Fig. S4A) stimulated with anti-CD3 andanti-CD28 in the presence of neutralizing antibodies to IL-4 andIFN-γ, showed that secreted factors were sufficient to induceTh17 cells (Fig. 4B). Consistent with prior reports (7–9), additionof IL-6 and TGF-β to this DC-free system also generated IL-17(Fig. 4B). Although CT-CM generated a lower percentage of IL-17-secreting T cells compared with IL-6 and TGF-β, these cellssecreted comparatively high amounts of IL-17 (Fig. 4B and Fig.S4B), resulting in similar levels in the supernatant (Fig. 4C).Naive T cells differentiated with CT-CM in this DC-free systemproduced IL-17 but not Th1- or Th2-related cytokines (Fig. 4D).CT-CM induced T cells that produced IL-22 (Fig. 4E). However,unlike those generated by IL-6 and TGF-β, T cells differentiatedwith CT-CM did not produce detectable IL-21 (Fig. 4E). Similar

to IL-6 and TGF-β, CT-CM generated T cells with induced levelsof the Th17-defining transcription factor RORγt (Fig. 4F) (20).To investigate the mechanisms by which CT-CM induces Th17

differentiation, we first tested the role of factors known to in-fluence Th17 differentiation. Blockade of IL-6 using a combina-tion of neutralizing antibodies against IL-6 and IL-6 receptorinhibited the ability of CT-CM to induce IL-17 from naive T cells(Fig. 5A and Fig. S5A). Blockade of TGF-β, which acts in concertwith IL-6 to promote Th17 differentiation, also inhibited Th17differentiation by CT-CM (Fig. 5A and Fig. S5A). An antagonistof aryl hydrocarbon receptor, which is required for Th17 dif-ferentiation by IL-6 and TGF-β (21, 22), similarly inhibited Th17differentiation by CT-CM (Fig. S5B). Furthermore, althoughIL-21 was not detectable in our previous experiments (Fig. 4E),inhibition of IL-21 blocked IL-17 induction by CT-CM (Fig. 5Aand Fig. S5A), consistent with a critical role for low levels of thisautocrine factor in amplifying Th17 cells (23–25). Neutralizationof IL-1β, which enhances Th17 differentiation (26), also inhibi-ted IL-17 induction by CT-CM (Fig. 5A and Fig. S5A). Blockadeof IL-23 did not affect IL-17 induction by CT-CM (Fig. 5B andFig. S5C), consistent with a reportedly more important role forIL-23 in the stabilization and expansion of Th17 cells than intheir initial polarization (27).Our results confirmed the central role of IL-6 and TGF-β in

mediating Th17 differentiation. However, emerging evidencesuggests that other factors may act in concert with IL-6 to in-fluence the development of Th17 cells (27). Indeed, our resultssuggested CT-CM contained additional factors that conferredthe phenotypic differences observed in Th17 cells induced by CT-CM compared with those induced by IL-6 and TGF-β. To identifysuch factors involved in Th17 induction by CT, we performeda microarray-based gene-expression analysis on DCs to detecttranscripts induced after 2 h exposure to CT or forskolin, a timesufficient to induce Th17-differentiating capacity (Fig. S6). Based

lm

gn(

A7

1-LI

1-)

A B

C

0

5

10

15OVACT + OVA

*

*

Fixed DC:DC supe:

- + +

+ - +

D

IL-17A

-N

FIγ Media CT-CM IL-6+TGF-β

E

A71-LI

-NFI

4-LI5-LI

31-LI

01-LI

0.00

0.25

0.5010

20

30MediaCT-CM

*

Cyto

kin

e(n

gm

l-1)

Media CT-CM IL-6+TGF-0

5

10

15 *

*

IL-1

7A

(ng

ml-1

)

IL-22 IL-210.0

0.5

1.0

4

6

8MediaCT-CMIL-6+TGF-

Cyto

kin

e(n

gm

l-1)

n.d. n.d.n.d.

*

**

F

CT-CM IL6+TGF0

2

4

6

8

RO

Rt

(fo

ld-i

nd

ucti

on

)

Fig. 4. Secreted factors from CT-treated DCs are sufficient toinduce Th17 differentiation. (A) DCs were incubated with OVAor CT+OVA and then left unfixed or fixed. OT-II CD4 T cellswere then added with or without the original DC supernatant(DC supe), and IL-17 production assessed after restimulation.(B–E) Sorted naive CD4 T cells were stimulated with anti-CD3,anti-CD28, anti-IL-4, and anti-IFN-γ in the presence of media,10% CT-CM, or IL-6+TGF-β. Cytokine production after restim-ulation with phorbol myristate acetate and ionomycin was de-termined by flow cytometry and ELISA. (F) RORγt expression inCD4 T cells, stimulated as above, was determined by quanti-tative RT-PCR. Fold-induction of cells treated with CT-CM or IL-6+TGF-β compared with cells incubated with media alone isshown. All data are representative of at least two independentexperiments. ELISA data are shown as the mean of duplicatesamples and error bars reflect SD. *, P < 0.05 compared withmedia-treated group; n.d., not detected.

10640 | www.pnas.org/cgi/doi/10.1073/pnas.1002348107 Datta et al.

on our data implicating a cAMP-dependent mechanism, we fo-cused on secreted gene products that were induced by both CTand forskolin (Table S1). These included IL-6 and IL-1β, con-sistent with our results above and confirmed by ELISA (Fig. S7).Of note, TGF-β and IL-21 were not induced in DCs, a resultcompatible with evidence that these molecules are derived fromother sources such as media supplements and T cells (27).Microarray analysis revealed that amphiregulin, an EGF re-

ceptor ligand that activates the Th17-associated transcriptionfactor STAT3 (28, 29), was highly induced by CT and forskolin.However, neutralizing antibodies against amphiregulin failed toinhibit Th17 differentiation by CT-CM (Fig. 5C and Fig. S5D). CTand forskolin also induced prostaglandin endoperoxide synthase-2 (or COX-2). Prostaglandin E2, a product of COX-2 activity thatincreases cAMP levels in target cells, was induced in DC super-natants (Fig. S7). Prostaglandin E2 was recently reported toaugment IL-17 production by differentiated Th17 cells but notinduce Th17 differentiation of naive cells in mice (30, 31). Con-sistent with this, CT-CM from Cox2−/− DCs were capable of in-ducing differentiation of naive T cells into Th17 cells (Fig. 5C andFig. S5D). Thus, amphiregulin and prostaglandins appear not tobe required for CT-induced Th17 differentiation.

Microarray analysis also revealed induction of calcitonin-related polypeptide beta (or β-calcitonin gene-related peptide,β-CGRP), a neuropeptide with described immunological prop-erties that triggers cAMP-mediated pathways (32). We con-firmed induction of β-CGRP in DCs by CT (Fig. S7). Addition ofCGRP8–37, a competitive antagonist of full-length β-CGRP,inhibited the ability of CT-CM to induce Th17 cells (Fig. 5D andFig. S5E) but had no effect on generation of Th17 cells by IL-6and TGF-β (Fig. 5E). Addition of full-length β-CGRP to T cellsdid not elicit Th17 differentiation, indicating that β-CGRP isnecessary but not sufficient to induce Th17 differentiation by CT(Fig. 5D). β-CGRP did not enhance Th17 differentiation by IL-6and TGF-β (Fig. 5F), but the combination of β-CGRP, IL-1β,and TGF-β in the presence of IL-23 generated Th17 cells (Fig.5F). Because naive T cells do not express IL-23 receptor (27),these results are consistent with the ability of cAMP to induceIL-23 receptor (30) and suggest that β-CGRP and IL-1β cancoordinate with TGF-β in the absence of IL-6 to induce a Th17differentiation program that is then reinforced by IL-23 once itsreceptor is up-regulated (Fig. S8).

CT-Induced Antibody Response Requires IL-17. A hallmark of CT isits ability to induce mucosal and systemic antibody responses. Totest whether IL-17 influenced antibody production after immu-nization with CT, we immunized wild-type and IL-17-deficientmice either intranasally or orally with OVA or CT+OVA. In-terestingly, IL-17-deficient mice showed a statistically significantimpairment in mucosal IgA and systemic IgA and IgG1 pro-duction in response to CT+OVA only after oral immunization,although intranasally immunized IL-17–deficient mice did showa trend toward impaired IgA production (Fig. 6).

DiscussionIn two experimental systems involving in vivo immunization, and intwo in vitro systems that used either APC- or antibody-mediatedTCR activation, we have identified IL-17 as a central mediator ofthe mucosal adjuvant activities of CT. Not only did IL-17 con-tribute to cell-mediated immunity induced after immunizationwith CT, but also IL-17–deficient mice had diminished antibodyresponses to oral immunization. This is consistent with reports thatIL-17 influences B-cell activation (33), germinal center formation(34), and up-regulation of polymeric Ig receptor (35).Our data identifies a role for Th17 cells in mediating protection

against infection with anthrax spores. Furthermore, our resultssuggest a strategy for development of an anthrax vaccine thattargets multiple anthrax spore antigens and does not solely de-pend on immunity against PA. Mechanisms required for effectiveacquired immunity against spores have not been previously de-fined, but may have important implications for vaccine de-velopment against anthrax and other spore-borne diseases. Thereported activities of Th17 cells suggests they may play an im-portant role in maintaining the epithelial barrier and recruitinglocal effectors, such as neutrophils and antimicrobial peptides, tocontain spore-transmitted pathogens at the mucosal surface.The current Th17 differentiation paradigm, involving IL-6,

TGF-β, aryl hydrocarbon receptor, and IL-21, plays a centralrole in the generation of Th17 cells by CT. Indeed, a recentreport observed that CT is able to generate Th17 cells in an IL-6–dependent manner (11), and vaccines containing CT orEscherichia coli heat-labile toxin also induced IL-17 (36–39), butthe mechanisms involved and the role of IL-17 in mediatingadjuvant effects (i.e., antibody responses and protection againstinfection) were not explored. In addition, our data show thata natural biological molecule such as CT induces DCs to secretea complex mixture of products that are required to act in com-bination to determine T-cell fate (Fig. S8). The Th17 phenotypeinduced by this mixture of products differs from that induced byIL-6 and TGF-β, suggesting that the specific combination of

A B

C D

E

Media B6 Cox2-/- - Amph0

10

20

CT-CM CT-CM

p=ns

p=ns

IL-1

7A

(ng

ml-1

)

eM

ida -

-

PR

GC

73-8

-

PR

GC

0

5

10

15 p<0.001

CT-CM

IL-1

7A

(ng

ml-1

)

Media - IL-23p190

1

2

3

4

CT-CM

p=ns

IL-1

7A

(ng

ml-1

)

0.00

0.25

0.50

0.75

1.00

Media

CT-CM

-IL

-6

TGF-

IL-2

1

* **

*

IL-1

*IL-1

7A

(ng

ml-1

)

0

1

2

3

4

5

TGF- - + + + + + + + +

IL-6 - + + - - - - - -

IL-1 - - - + - + + - +

-CGRP - - + - + + - + +

IL-23 - - - - - - + + +

*

*

*IL-1

7A

(ng

ml-1

)

*

Media - CGRP8-37

0

1

2

3

4

5 p=ns

IL-1

7A

(ng

ml-1

)

IL-6+TGF-

F

Fig. 5. A combination of secreted factors mediates Th17 differentiation byCT. (A and B) Cytokine production was determined after stimulation of naiveCD4 T cells with 10% CT-CM and the indicated neutralizing antibodies. (C)Ten-percent CT-CM from wild-type (B6) or Cox2−/− DC, or 10% CT-CM with orwithout neutralizing amphiregulin antibody was used to stimulate naive CD4T cells. (D and E) Cytokine production was determined after stimulation ofnaive CD4 T cells with 10% CT-CM, IL-6+TGF-β, β-CGRP inhibitor (β-CGRP8–37),or β-CGRP. (F) Cytokine production was determined after stimulation of naiveCD4 T cells with the indicated factors. All data are representative of at leasttwo independent experiments. ELISA data are shown as the mean of dupli-cate samples, and error bars reflect SD. *, P< 0.05 comparedwith CT-CM (A) ormedia-treated (F) group.

Datta et al. PNAS | June 8, 2010 | vol. 107 | no. 23 | 10641

IMMUNOLO

GY

factors influences the stability and characteristics of Th17 cellsthat are induced. The CT-CM system described, where IL-6levels generated by CT are low (Fig. S7) and the use of 10% CT-CM further lowers cytokine concentrations to limiting levels,likely enhanced our ability to detect the contribution of mole-cules such as IL-1β and β-CGRP to Th17 differentiation by CT.Our data suggest that CT activates cAMP-dependent pathways

in DCs to drive Th17 differentiation. Further studies will beneeded to address whether cAMP activation in vivo is sufficientto mimic the adjuvant effects of CT. Interestingly, ATP canstimulate purinergic receptors to induce both cAMP (40) and IL-1β (41), suggesting involvement of these molecules in the abilityof bacterially-derived ATP to drive Th17 responses in the gut(42). Furthermore, the involvement of neuropeptides, such asβ-CGRP and vasoactive intestinal peptide (43) in Th17 differ-entiation implies a role for these molecules in T-cell physiologyat richly innervated locations, such as the gut mucosa.Our findings with CT highlight the ability of this bacterial toxin

to influence host immune responses beyond its contribution tobacterial virulence. Our results suggest that Th17 cells are cen-tral to the mucosal adjuvant activities of CT, likely acting toimprove epithelial barrier function, recruit neutrophils, induceantimicrobial peptides, and enhance antibody production andsecretion (33–35, 44). Understanding the mechanisms by whichCT exerts its mucosal adjuvant activities may identify pathwaysthat can be targeted for development of nontoxic mucosal adju-

vants and enhance our understanding of the regulation of mu-cosal T-cell differentiation.

MethodsMethods related to animals, reagents, ELISA, microarray analysis, PCR, andstatistics are in SI Methods.

Bacterial Strains and Preparation of Spores. Bacillus anthracis Sterne (pXO1+,pXO2−) has been described (45). The PA-deficient Sterne MS1 strain (pXO1−,pXO2−) was derived by curing the pXO1 plasmid from the Sterne strain (46).Elimination of pXO1 was confirmed by PCR and PA immunoblot. B. anthracisspores were purified and CFUs after heat activation were determined asdescribed (45).

Anthrax Immunization. Purified spores (109 CFU/mL)were γ-irradiatedwith 2.27Mrads (137Cs). Loss of viability was confirmed before each vaccination by cul-turing0.1mLof the irradiated suspension for 72h inBrainHeart Infusionbroth.Anesthetized A/J mice were immunized intranasally three times at 2-weekintervals with 5 × 107 irradiated spores and CT (1 μg). Mice were challengedintranasally 4 weeks after the last immunization with 4 × 106 CFU of live, heat-activated spores. Animal survival was monitored daily for up to 21 d. Unlessotherwise indicated, PA-deficient spores (SterneMS1 strain, pXO1−) were usedfor immunization and PA-containing spores (Sterne strain, pXO1+) were usedfor challenge.

In some experiments, CD4+ or CD8+ cells were depleted by i.p. injection ofthe indicated antibodies (1 mg) the day before intranasal challenge (15).Cytokines were depleted by i.p. injection of the appropriate neutralizingantibody (0.2 mg anti-IL-17, 1 mg anti-IFN-γ) on the day prior and the dayof challenge.

To detect cytokine production after immunization, immunized mice werechallenged intranasally with 107 irradiated spores. Three days after challenge,we incubated isolated splenic CD4 T cells (Miltenyi Biotec) with naive A/Jsplenocytes that had been pulsed overnight with 107 irradiated spores ormedia alone and then treated with mitomycin C (Sigma). We collected super-natants after 3 d andmeasured cytokine levels by ELISA. To determine antigen-specific responses, we subtracted cytokine levels from samples that were in-cubated with splenocytes that had not been pulsed with irradiated spores.

Immunization with OVA. Mice were immunized intranasally with PBS, OVA(20 μg), or CT (2 μg) plus OVA (20 μg) three times at 2-week intervals. Twoweeks after the last immunization, mice were killed for collection of bron-choalveolar lavage, serum, and organs. By gavage, mice were immunized onday 0 and day 14 with PBS, OVA (200 μg), or CT (20 μg) plus OVA (200 μg) in0.1 M sodium bicarbonate. On day 39, the mice were challenged with OVA(200 μg intragastrically plus 50 μg intravenously) and then killed on day 42.Single-cell suspensions from spleens and lymph nodes were incubated for 3 dwith media or OVA (200 μg/mL), and cytokine content in the supernatantswas analyzed by ELISA.

In Vitro T-Cell Differentiation.Mouse bone-marrow cells were cultured for 6 dwith GM-CSF (BD Biosciences) to obtain myeloid DCs (15). The DCs were in-cubated with OVA and indicated stimuli. In some instances, DCs were thenfixed with 2% buffered formalin without methanol (Sigma-Aldrich). IsolatedCD4+ T cells (Miltenyi Biotec) from OT-II mouse spleens were then incubatedwith the DCs. After 6 d, the T cells were restimulated with anti-CD3 (10 μg/mL plate-bound) and anti-CD28 (1 μg/mL soluble). For intracellular cytokineanalysis, GolgiStop (BD Biosciences) was added after 2 h. After 4 additionalhours, the cells were stained for the indicated cytokines before data ac-quisition on an LSRII or FACSCalibur flow cytometer (BD Biosciences). Datawere analyzed with FlowJo software (Treestar). For ELISA analysis, the cellswere restimulated for 24 h before supernatant collection.

To induce T-cell differentiation in a DC-free system, naive (CD62Lhigh,CD44low) CD4+ T cells sorted by flow cytometry (>99% purity) (Fig. S4A) fromspleens of C57BL/6 mice were stimulated for 3 d with anti-CD3, anti-CD28,anti-IL-4 (10 μg/mL), and anti-IFN-γ (10 μg/mL) in the absence or presence ofIL-6 (20 ng/mL), TGF-β1 (3 ng/mL), or 10% CT-CM. Other neutralizing anti-bodies were used at 5 to 10 μg/mL based on neutralizing activity. The T cellswere expanded in fresh media containing IL-2 (5 ng/mL) for 3 additionaldays. Expansion with IL-23 (20 ng/mL) instead of IL-2 in select experimentsgave largely similar results. The cells were then stimulated with phorbolmyristate acetate (50 ng/mL) and ionomycin (1 μM) to assess cytokine pro-duction, as described above.

B6 Il17-/-0.0

0.5

1.0

1.5

2.0

2.5p=ns

p=ns

p=ns

OVAOVA+CT

BA

LIg

A(μ

gm

l-1)

B6 Il17-/-0

1000

2000

3000

4000

p=ns

p<0.001p<0.001

Seru

mIg

G1

(μg

ml-1

)

B6 Il17-/-0

2

4

6p<0.01

p=ns

p=ns

Seru

mIg

A(μ

gm

l-1)

B6 Il17-/-0

100

200

300

p=ns

p=ns p=ns

Seru

mIg

G2c

(μg

ml-1

)

B6 Il17-/-0.0

0.5

1.0

1.5

p<0.001

p=ns

p<0.001

FecalIg

A(μ

gm

l-1)

B6 Il17-/-0.0

2.5

5.0250500750

1000

p=ns

p<0.05 p<0.05

Seru

mIg

G1

(μg

ml-1

)

B6 Il17-/-0

10

20

30

40

p=ns

p<0.01 p<0.01

Seru

mIg

A(μ

gm

l-1)

B6 Il17-/-0.0

0.5

1.0

1.5

p=ns

p=ns p=ns

Seru

mIg

G2c

(μg

ml-1

)

A B

C D

E F

G H

Fig. 6. IL-17 is required for CT-induced antibody production after gavage.OVA-specific bronchoalveolar lavage (BAL), fecal, and serum antibody levelswere determined in wild-type (B6) and IL-17-deficient mice immunized withOVA or CT+OVA intranasally (A–D, 4–6 mice per group) or by gavage (E–H,10–12 mice per group). Intranasal experiment is representative of two in-dependent experiments and gavage experiment is representative of threeindependent experiments.

10642 | www.pnas.org/cgi/doi/10.1073/pnas.1002348107 Datta et al.

ACKNOWLEDGMENTS. This research was supported by the IntramuralResearch Program of the National Institutes of Health/National Institute ofAllergy and Infectious Diseases, the Ellison Medical Foundation, the

Leukemia and Lymphoma Society, the Crohn and Colitis Foundation ofAmerica, and by extramural National Institutes of Health Grants AI077989,AI083328, AI058743, AI68685, and DK35108.

1. Northrup RS, Fauci AS (1972) Adjuvant effect of cholera enterotoxin on the immuneresponse of the mouse to sheep red blood cells. J Infect Dis 125:672–673.

2. Elson CO, Ealding W (1984) Generalized systemic and mucosal immunity in mice aftermucosal stimulation with cholera toxin. J Immunol 132:2736–2741.

3. Freytag LC, Clements JD (2005) Mucosal adjuvants. Vaccine 23:1804–1813.4. Vanden Broeck D, Horvath C, De Wolf MJ (2007) Vibrio cholerae: Cholera toxin. Int J

Biochem Cell Biol 39:1771–1775.5. Xu-Amano J, et al. (1993) Helper T cell subsets for immunoglobulin A responses: Oral

immunization with tetanus toxoid and cholera toxin as adjuvant selectively inducesTh2 cells in mucosa associated tissues. J Exp Med 178:1309–1320.

6. Braun MC, He J, Wu CY, Kelsall BL (1999) Cholera toxin suppresses interleukin (IL)-12production and IL-12 receptor beta1 and beta2 chain expression. J Exp Med 189:541–552.

7. Bettelli E, et al. (2006) Reciprocal developmental pathways for the generation ofpathogenic effector TH17 and regulatory T cells. Nature 441:235–238.

8. Mangan PR, et al. (2006) Transforming growth factor-beta induces development ofthe T(H)17 lineage. Nature 441:231–234.

9. Veldhoen M, Hocking RJ, Atkins CJ, Locksley RM, Stockinger B (2006) TGFbeta in thecontext of an inflammatory cytokine milieu supports de novo differentiation of IL-17-producing T cells. Immunity 24:179–189.

10. Ouyang W, Kolls JK, Zheng Y (2008) The biological functions of T helper 17 celleffector cytokines in inflammation. Immunity 28:454–467.

11. Lee JB, Jang JE, Song MK, Chang J (2009) Intranasal delivery of cholera toxin in-duces th17-dominated T-cell response to bystander antigens. PLoS ONE 4:e5190.

12. Friedlander AM, Pittman PR, Parker GW (1999) Anthrax vaccine: Evidence for safetyand efficacy against inhalational anthrax. JAMA 282:2104–2106.

13. Horne T, Turner GC, Willis AT (1959) Inactivation of spores of Bacillus anthracis bygamma-radiation. Nature 183:475–476.

14. Spotts Whitney EA, et al. (2003) Inactivation of Bacillus anthracis spores. Emerg InfectDis 9:623–627.

15. Datta SK, et al. (2006) Vaccination with irradiated Listeria induces protective T cellimmunity. Immunity 25:143–152.

16. van Beelen AJ, et al. (2007) Stimulation of the intracellular bacterial sensor NOD2programs dendritic cells to promote interleukin-17 production in human memoryT cells. Immunity 27:660–669.

17. Walsh DA, Perkins JP, Krebs EG (1968) An adenosine 3′,5′-monophosphate-dependantprotein kinase from rabbit skeletal muscle. J Biol Chem 243:3763–3765.

18. de Rooij J, et al. (1998) Epac is a Rap1 guanine-nucleotide-exchange factor directlyactivated by cyclic AMP. Nature 396:474–477.

19. Kawasaki H, et al. (1998) A family of cAMP-binding proteins that directly activateRap1. Science 282:2275–2279.

20. Ivanov II, et al. (2006) The orphan nuclear receptor RORgammat directs thedifferentiation program of proinflammatory IL-17+ T helper cells. Cell 126:1121–1133.

21. Veldhoen M, et al. (2008) The aryl hydrocarbon receptor links TH17-cell-mediatedautoimmunity to environmental toxins. Nature 453:106–109.

22. Quintana FJ, et al. (2008) Control of T(reg) and T(H)17 cell differentiation by the arylhydrocarbon receptor. Nature 453:65–71.

23. Nurieva R, et al. (2007) Essential autocrine regulation by IL-21 in the generation ofinflammatory T cells. Nature 448:480–483.

24. Korn T, et al. (2007) IL-21 initiates an alternative pathway to induce proinflammatoryT(H)17 cells. Nature 448:484–487.

25. Zhou L, et al. (2007) IL-6 programs T(H)-17 cell differentiation by promotingsequential engagement of the IL-21 and IL-23 pathways. Nat Immunol 8:967–974.

26. Chung Y, et al. (2009) Critical regulation of early Th17 cell differentiation byinterleukin-1 signaling. Immunity 30:576–587.

27. Korn T, Bettelli E, Oukka M, Kuchroo VK (2009) IL-17 and Th17 Cells. Annu RevImmunol 27:485–517.

28. David M, et al. (1996) STAT activation by epidermal growth factor (EGF) andamphiregulin. Requirement for the EGF receptor kinase but not for tyrosinephosphorylation sites or JAK1. J Biol Chem 271:9185–9188.

29. Yang XO, et al. (2007) STAT3 regulates cytokine-mediated generation of inflamma-tory helper T cells. J Biol Chem 282:9358–9363.

30. Boniface K, et al. (2009) Prostaglandin E2 regulates Th17 cell differentiation and func-tion through cyclic AMP and EP2/EP4 receptor signaling. J Exp Med 206:535–548.

31. Yao C, et al. (2009) Prostaglandin E2-EP4 signaling promotes immune inflammationthrough Th1 cell differentiation and Th17 cell expansion. Nat Med 15:633–640.

32. Brain SD, Grant AD (2004) Vascular actions of calcitonin gene-related peptide andadrenomedullin. Physiol Rev 84:903–934.

33. Doreau A, et al. (2009) Interleukin 17 acts in synergy with B cell-activating factor toinfluence B cell biology and the pathophysiology of systemic lupus erythematosus.Nat Immunol 10:778–785.

34. Hsu HC, et al. (2008) Interleukin 17-producing T helper cells and interleukin 17orchestrate autoreactive germinal center development in autoimmune BXD2 mice.Nat Immunol 9:166–175.

35. Jaffar Z, Ferrini ME, Herritt LA, Roberts K (2009) Cutting edge: Lung mucosal Th17-mediated responses induce polymeric Ig receptor expression by the airway epithe-lium and elevate secretory IgA levels. J Immunol 182:4507–4511.

36. Lu YJ, Forte S, Thompson CM, Anderson PW, Malley R (2009) Protection againstPneumococcal colonization and fatal pneumonia by a trivalent conjugate of a fusionprotein with the cell wall polysaccharide. Infect Immun 77:2076–2083.

37. Velin D, et al. (2009) Interleukin-17 is a critical mediator of vaccine-induced reductionof Helicobacter infection in the mouse model. Gastroenterology 136:2237–2246.e1.

38. McNeal MM, et al. (2007) Intrarectal immunization of mice with VP6 and either LT(R192G) or CTA1-DD as adjuvant protects against fecal rotavirus shedding afterEDIM challenge. Vaccine 25:6224–6231.

39. Smiley KL, et al. (2007) Association of gamma interferon and interleukin-17production in intestinal CD4+ T cells with protection against rotavirus shedding inmice intranasally immunized with VP6 and the adjuvant LT(R192G). J Virol 81:3740–3748.

40. Post SR, Jacobson JP, Insel PA (1996) P2 purinergic receptor agonists enhance cAMPproduction in Madin-Darby canine kidney epithelial cells via an autocrine/paracrinemechanism. J Biol Chem 271:2029–2032.

41. Piccini A, et al. (2008) ATP is released by monocytes stimulated with pathogen-sensingreceptor ligands and induces IL-1beta and IL-18 secretion in an autocrine way. ProcNatl Acad Sci USA 105:8067–8072.

42. Atarashi K, et al. (2008) ATP drives lamina propria T(H)17 cell differentiation. Nature455:808–812.

43. Yadav M, Rosenbaum J, Goetzl EJ (2008) Cutting edge: Vasoactive intestinal peptide(VIP) induces differentiation of Th17 cells with a distinctive cytokine profile. JImmunol 180:2772–2776.

44. Aujla SJ, Dubin PJ, Kolls JK (2007) Th17 cells and mucosal host defense. SeminImmunol 19:377–382.

45. Sabet M, Cottam HB, Guiney DG (2006) Modulation of cytokine production andenhancement of cell viability by TLR7 and TLR9 ligands during anthrax infection ofmacrophages. FEMS Immunol Med Microbiol 47:369–379.

46. Mikesell P, Ivins BE, Ristroph JD, Dreier TM (1983) Evidence for plasmid-mediatedtoxin production in Bacillus anthracis. Infect Immun 39:371–376.

Datta et al. PNAS | June 8, 2010 | vol. 107 | no. 23 | 10643

IMMUNOLO

GY