airedeﬁciencypromotestrp-1 speciﬁcimmunerejection of melanoma · microenvironment and...

Microenvironment and Immunology

Aire Deficiency Promotes TRP-1–Specific Immune Rejectionof Melanoma

Meng-Lei Zhu1,2, Anil Nagavalli1,2, and Maureen A. Su1,2

AbstractThe thymic transcription factor autoimmune regulator (Aire) prevents autoimmunity in part by promoting

expression of tissue-specific self-antigens, which include many cancer antigens. For example, AIRE-deficientpatients are predisposed to vitiligo, an autoimmune disease of melanocytes that is often triggered by efficaciousimmunotherapies against melanoma. Therefore, we hypothesized that Aire deficiency in mice may elevateimmune responses to cancer and provide insights into how such responses might be triggered. In this study, weshow that Aire deficiency decreases thymic expression of TRP-1 (TYRP1), which is a self-antigen in melanocytesand a cancer antigen in melanomas. Aire deficiency resulted in defective negative selection of TRP-1–specificT cells without affecting thymic numbers of regulatory T cells. Aire-deficient mice displayed elevated T-cellimmune responses that were associatedwith suppression ofmelanoma outgrowth. Furthermore, transplantationof Aire-deficient thymic stroma was sufficient to confer more effective immune rejection of melanoma in anotherwise Aire wild-type host. Together, our work showed how Aire deficiency can enhance immune responsesagainst melanoma and howmanipulating TRP-1–specific T-cell negative selection may offer a logical strategy toenhance immune rejection of melanoma. Cancer Res; 1–13. �2013 AACR.

IntroductionMelanoma is the deadliest form of skin cancer and accounts

for 5% of cancer deaths in the United States (1). Understandingmechanisms that may limit melanoma development is there-fore a significant public health concern. The cellular immuneresponse has the potential to eradicate melanoma cells, andspontaneous regression of melanoma associated with lympho-cytic infiltration of tumors has been reported (2, 3). Despite thispotential, the immune response is often ineffective, with highrates of melanoma growth occurring in immunocompetentindividuals. Furthermore, while immunotherapy for metastat-icmelanoma can result in clinical benefit, a complete responseoccurs in only a small proportion of patients (4). Overcomingimmune tolerance to melanoma in otherwise healthy indivi-duals may therefore be important for preventing and treatingmelanoma.Several antigens targeted by the immune response in mel-

anoma are also antigens in vitiligo, an acquired depigmentingdisorder resulting from autoimmune destruction of melano-cytes (5–7). Consistent with shared antigens in these 2 immune

settings, the development of melanoma is regularly accompa-nied by vitiligo in a number of animal models (6). In addition,development of vitiligo in patients with advanced melanomaundergoing immunotherapy is associated with improved prog-nosis (8). This overlap in immune responses to vitiligo andmelanoma suggests that factors that result in vitiligo may alsoenhance the immune response against melanoma.

The autoimmune regulator (Aire) gene promotes organ-specific immune tolerance and prevents the development ofa number of autoimmune diseases including vitiligo. Patientswith mutations in Aire have an 18- to 36-fold increased risk ofdeveloping vitiligo compared with the general population (9),and genetic polymorphisms in Aire have been associated withvitiligo development in a case–control study (10). Within thethymus, Aire upregulates expression of a large number oftissue-specific self-antigens inmedullary thymic epithelial cells(mTEC; refs. 11–13). Ectopic expression of these self-antigensdrives the clonal deletion of developing T cells recognizingthese antigens with high affinity (14, 15). In addition, Aire mayhave other functions affecting T-cell tolerance including pos-sible roles in antigen processing/presentation and the devel-opment of regulatory T cells (Tregs; refs. 14, 16, 17).

The existence of shared vitiligo/melanoma antigens sug-gests that decreased Aire function in the thymus, in addition topredisposing to vitiligo, may also enhance immune rejection ofmelanoma. In support of this hypothesis, Aire polymorphismshave been associated with melanoma development in a smallcase–control study (18). In addition, Aire-deficient mice mayhave increased memory response to melanoma followingimmunization with irradiated melanoma cells (19). At thesame time, however, recent work also suggests that thymicexpression of a cancer-germline antigen only minimally

Authors' Affiliations: 1Department of Pediatrics and Microbiology/Immu-nology, School of Medicine; and 2Lineberger Comprehensive CancerCenter, University of North Carolina at Chapel Hill, Chapel Hill, NorthCarolina

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Maureen A. Su, University of North Carolina atChapel Hill, 3300 Thurston Building CB7280, Chapel Hill, NC 27599.Phone: 919-966-0259; Fax: 919-843-7588; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-12-3781

�2013 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 1

on June 30, 2021. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 31, 2013; DOI: 10.1158/0008-5472.CAN-12-3781

http://cancerres.aacrjournals.org/

changes T-cell responses to antigen-expressing tumors (20). Inaddition, Aire-regulated thymic expression of tyrosinase, ashared vitiligo/melanoma antigen, does not shape the tyros-inase-specific T-cell repertoire (21). Instead, the tyrosinase-specific T-cell repertoire depends on constitutive presentationof tyrosinase in peripheral lymphoid organs. Thus, althoughthere is some evidence that thymic Aire function regulates theimmune response to melanoma, further studies are needed toclarify this issue. Moreover, the mechanism by which Aire mayregulate the immune response to melanoma has not beendelineated.

We show here that Aire normally induces thymic expressionof the shared vitiligo/melanoma antigen, TRP-1 (tyrosinaserelated protein-1; Tyrp1; gp75), and lack of thymic TRP-1expression in Aire-deficient mice results in the escape ofTRP-1–specific T cells fromclonal deletion. Aire-deficientmicegenerate a more activated primary T-cell response to mela-noma that is associated with delayed growth of melanoma. Inaddition, transplantation of thymic stroma is sufficient toconfer more effective immune rejection of melanoma. Togeth-er, these findings establish that Aire deficiency enhances theprimary T-cell response to melanoma and delineate a novelpathway in which Aire deficiency promotes an antimelanomaT-cell response by decreased TRP-1 expression in the thymus.

Materials and MethodsMice

Tyrp1B-w TRP-1–specific T-cell receptor (TCR) transgenicmouse model (TRP-1 TCR Tg) RAG�/� mice (JAX stock num-ber 008684) were purchased from The Jackson Laboratory. B6.AireGW/þ mice were backcrossed more than 10 generationsfrom a mixed C57BL/6-129 line described in ref. (22). Togenerate TRP-1–specific CD4þ TCR Tg mice in an Aire-defi-cient background, male Tyrp1B-w TRP-1 TCR Tg RAG�/� micewere crossed with female B6.AireGW/þmice. F1male mice withgenotype B6.AireGW/þ Tyrp1B-w/þ TRP-1 TCR Tg RAGþ/�

genotype were further crossed with RAG�/� mice. F2 micewith genotype B6.AireGW/þ Tyrp1þ/þ TRP-1 TCR Tg RAG�/�

and B6.Aireþ/þ Tyrp1þ/þ TRP-1 TCR Tg RAG�/� mice wereused for tumor inoculation studies. B6.Nude mice were pur-chased from Jackson Laboratory. All mice were used in accor-dance with guidelines from the University of North Carolina,Chapel Hill Animal Care and Use Committee. All experimentswere carried out with the approval of the Animal Use and CareCommittees of the University of North Carolina, Chapel Hill(Chapel Hill, NC).

Antibodies and flow cytometryAnti-CD4 (RM4-5), anti-CD3 (145-2c11), anti-CD25 (PC61),

anti-Vb14 (14-2), anti-CD44 (IM7), anti-CD62L (MEL-14), andanti-Foxp3 (FJK-16s) were purchased from eBioscience. Anti-IL-2 (JES6-5H4), anti- IFN-g (XMG1.2), anti–MHC class II (I-A/I-E; M5/114.15.2) were purchased from BD Biosciences. Intra-cellular staining for cytokines was carried out with the Cytofix/Cytoperm Intracellular Staining Kit (BD Biosciences) as permanufacturer's instructions. Cells were stimulated with phor-bol 12-myristate 13-acetate (PMA)/ionomycin for 5 hoursbefore antibody staining. FOXP3 Staining Buffer Set was used

for FOXP3 intracellular staining according to manufacturer'sinstructions (eBioscience). All samples were run on a DakoCyAn (Beckman-Coulter) flow cytometer.

B16 melanoma modelB16-F10 is a TRP-1þ C57BL/6-derived melanoma cell line

that was kindly provided by Dr. Alan Fong (University of NorthCarolina, Chapel Hill). It was originally from American TypeCulture Collection and maintained in culture media as previ-ously described (23, 24). Mice were injected subcutaneouslywith 1 � 105 B16 melanoma cells in 100 mL volume. Perpen-dicular tumor diameters were measured with calipers. Bodyweight and physiologic status was monitored daily.

Real-time RT-PCRThymic stromal preparations and real-time reverse tran-

scriptase PCR (RT-PCR) were conducted as previouslydescribed (25). TaqMan primer/probe sets for TRP-1, tyrosi-nase, and silver/gp100 were purchased from AppliedBiosystems.

Bone marrow chimeras and thymic transplantsBone marrow chimeras and thymic transplants were con-

ducted as previously described (22). For bone marrow chi-meras, bonemarrowwas harvested from the femurs of 6-week-old Tyrp1B-w TRP-1 Tg RAG�/� mice. T cells were removedfrom the bone marrow by complement depletion using anti-bodies against CD4 and CD8. Recipient mice received 2 dosesof radiation (500 rad each time) at least 4 hours apart. Cells (1�107) were injected retro-orbitally into each recipient mouse.Chimeras were aged for 10 weeks before analysis.

For thymic transplants, thymi from neonates were removedand cultured in Transwell plates for 7 days in 1.35 mmol/L 20-deoxyguanosine (2-dG; Sigma-Aldrich) in complete Dulbecco'smodified Eagle's medium (DMEM) to deplete hematopoieticcells. Thymi were washed in complete DMEM (without 2-dG) 2hours before transplantation. Thymi were transplanted underthe kidney capsule of B6.Nude recipients. Transplanted micewere aged for 10 weeks before analysis. T-cell reconstitutionwas confirmed by the presence of CD3þ, CD4þ, and CD8þ cellsin the peripheral blood.

ELISAB16 melanoma cell extract or TRP-1109–130 peptide

(NCGTCRPGWRGAACNQKILTVR) was incubated overnightat 4�C in 96-well ELISAmicrowell plates (Xenopore) accordingto the manufacturer's instructions. After blocking with 3%bovine serum albumin in PBS, mouse sera were added at a1:200 dilution in PBS and incubated for 2 hours at roomtemperature. Goat anti-mouse immunoglobulin G (IgG;g-chain specific) horseradish peroxidase–conjugate were usedto detect antibodies, respectively. ELISA was developed andread using aMolecular Device ELISA Reader and analyzedwiththe instrument's software. Respective positive controls (anti-TRP-1 antibody) and negative controls (no serum, antihumanIgG alone) were conducted in parallel. Assays were conductedat least twice with wells plated in 5 replicates for verification ofresults.

Zhu et al.

Cancer Res; 2013 Cancer Research2




Immunofluorescence stainingTumors from mice were harvested and fixed overnight in

10% formalin at 4�C, then stored in 70% ethanol. Tumors wereembedded in paraffin for sectioning. Slides were stained over-night with anti-CD3 antibody, washed, and mounted with 40,6-diamidino-2-phenylindole (DAPI). Images were taken on afluorescence microscope (Olympus BX60) and analyzed withImageJ software.

Isolation of tumor-infiltrating lymphocytesTumors were dissected and minced before incubation with

collagenase type I/IV, hyaluronidase, and DNase I. T cells wereisolated using immunomagnetic bead separation using typeMSþ or VSþ columns and anti-CD3–conjugated magneticbeads (Miltenyi) according to the manufacturer's instructions.

Adoptive transferSpleens from Tyrp1B-w TRP-1 TCR Tg RAG�/� male mice

were harvested and made into single-cell suspensions. Redblood cells were eliminated by ammonium-chloride-potassiumlysis. Subsequently, cells were counted and enriched for CD4þ

CD25þ and CD4þCD25� T cells by magnetic bead sorting usinga CD4þCD25þ T-cell enrichment kit from Miltenyi Biotec.CD4þCD25þ and CD4þCD25� T cells were resuspended in PBSand injected retro-orbitally into B6.RAG�/� recipient mice.

Statistical analysisPRISM 5.0 and Microsoft Office Excel software was used to

analyze data (GraphPad Software, Inc.). Unpaired Student ttests were used to compare the differences between 2 groups.Survival curves were compared by Mann–Whitney U test. Pvalues of 0.05 or less were considered significant.

ResultsAire deficiency is associated with decreased growth ofB16 melanoma cells and increased density of tumorinfiltrating lymphocytesMice with Aire deficiency do not spontaneously develop

clinical vitiligo but may have features of the autoimmunevitiligo response. A feature of the autoimmune response invitiligo is the development of autoantibodies against melano-cyte antigens (26). We sought to determine if Aire-deficientmice harbor autoantibodies against melanocyte antigens bymeasuring the frequency of serum autoantibodies against B16melanoma cell extract. As amodel of Aire deficiency, we used aC57/BL6 mouse line harboring a dominant Aire G228Wpoint mutation that results in hypomorphic Aire function[B6.AireGW/þ mice; (22)]. Increased serum autoantibodiesagainst melanoma cell extract were seen in 15- to 25-week-old B6.AireGW/þ mice compared with wild-type (B6.Aireþ/þ)littermate controls (Supplementary Fig. S1A, left). Consistentwith these findings, increased serum autoantibodies againstmelanoma cell extract were also seen in 15- to 25-week-oldAireknockout (B6.Aire�/�) mice (Supplementary Fig. S1A, right).These findings suggest that Aire deficiency promotes an auto-immune response against melanocyte antigens.Because Aire-deficient mice show increased immune re-

sponse to B16 cell extract, we hypothesized that Aire-deficient

mice would eradicate B16 melanoma cells more effectively.To test this hypothesis in vivo, we inoculated B16 melanomacells subcutaneously on the flanks of B6.AireGW/þ mice andfollowed tumor growth. B16 melanoma cells grew more slowlyin B6.AireGW/þ mice compared with B6.Aireþ/þ littermatecontrols (Fig. 1A and B and Supplementary Fig. S1B). Betweenday 16 and 19, the average size of tumors was significantlyreduced in B6.AireGW/þ compared with B6.Aireþ/þ littermatecontrols (Fig. 1B). Consistent with a role for Aire in mediatingB16 melanoma growth, B16 melanoma cells inoculatedinto B6.Aire�/� mice also grew more slowly compared withB6.Aireþ/þ mice (Fig. 1B). In addition, improved survival wasseen in Aire-deficient B6.AireGW/þ and B6.Aire�/� mice inoc-ulated with B16 melanoma cells compared with B6.Aireþ/þ

littermate controls (Fig. 1C).The decreased B16 melanoma growth in Aire-deficient mice

was associated with increased T-cell immune infiltrationwithin the tumors. We used immunofluorescence stainingfor CD3 to quantitate the number of infiltrating T cells withinthe tumor. An increased density of CD3þ T cells per tumorarea were found in the tumors of B6.AireGW/þ mice comparedwith B6.Aireþ/þ mice (Fig. 1D and E). Increased numbers oftumor-infiltrating lymphocytes in the B16melanoma per gramof tumor in Aire-deficient mice was confirmed by flow-cyto-metric analysis of the tumors (data not shown). This findingsuggests that an increased T-cell response directed at mela-noma is associated with decreased tumor growth in Aire-deficient mice.

Aire deficiency is associated with activated immuneresponse against melanoma

We sought to determine whether tumor-infiltrating T cellswere not only more numerous but also more activated in B6.AireGW/þ mice. We pooled together lymphocytes from thetumors of 6 B6.AireGW/þ and 6 wild-type littermate controlsto collect a sufficient number of events to analyze. An increasedfrequency CD4þ and CD8þ T lymphocytes with an activated/memory phenotype (CD62Llow CD44high) were found withinmelanoma of B6.AireGW/þ mice compared with wild-typelittermates (Fig. 2A). An increased frequency of CD62Llow

CD44high CD4þ and CD8þ T cells were also seen in the spleenand tumor-draining lymph node (Supplementary Fig. S2).

We next assessed the ability of tumor-infiltrating lympho-cytes to secrete proinflammatory cytokines by intracellularcytokine staining. We evaluate the ability of T cells to secreteinterleukin (IL)-2 and IFN-g , as both cytokines have beenimplicated in rejection of tumors by T cells (27–29). Anincreased frequency of IL-2 and IFN-g–secreting CD4þ andCD8þ tumor-infiltrating lymphocytes were detected in B6.AireGW/þ mice compared with wild-type littermate controls(Fig. 2B and C). An increased frequency of CD4þ and CD8þ Tlymphocytes secreting IL-2 and IFN-g was also seen in spleenand tumor-draining lymph nodes of B6.AireGW/þ mice com-pared with wild-type littermates (Fig. 2D and E). Importantly,we did not observe a difference in the frequency of memory/activated or cytokine-secreting T cells in B6.AireGW/þ micebefore tumor inoculation (Supplementary Fig. S3A–S3D). Theincreased immune activation in B6.AireGW/þmice, therefore, is

Aire Deficiency Enhances Immune Response to TRP-1

www.aacrjournals.org Cancer Res; 2013 3




specific to the antitumor response and does not representnonspecific immune activation. Taken together, these datasuggest that melanoma provokes an enhanced T-cell responsein Aire-deficient mice.

Aire deficiency in thymic stroma is sufficient to augmentimmune response against melanoma

Although Aire is most highly expressed in the thymus,peripheral Aire expression in the lymph node and spleen alsopromotes clonal deletion of autoreactive T cells (30). Wesought to determine whether central (thymic) Aire expressionwas sufficient to limit the immune response toward melano-ma. To isolate Aire deficiency to the thymus, we transplantedthymic stroma from B6.AireGW/þmice or wild-type littermatesinto athymic nude (Foxn1�/�; B6.Nude) mice. The reconsti-tuted mice have wild-type Aire expression in the periphery butare either Aire-deficient or Aire-wild-type in the thymus (Fig.3A). Thymic engraftment and immune reconstitution wasdetermined by monitoring for the presence of CD3þ, CD4þ,and CD8þ T cells in the peripheral blood 8 weeks posttrans-plantation. No significant difference in CD4þ and CD8þ T-cellreconstitution was found between recipients of B6.AireGW/þ orB6.Aireþ/þ thymi (Supplementary Fig. S4).

After confirming immune reconstitution in transplantedmice, we inoculated B16melanoma cells andmonitored tumor

growth (Fig. 3A). Between days 9 and 19 postinoculation,melanoma growth was significantly decreased in mice recon-stituted with Aire-deficient thymi compared with wild-typethymi (Fig. 3B). This decreased growth was associated withincreased numbers of CD3þ tumor-infiltrating lymphocytesper area of tumor (mm2; Fig. 3C and D). In addition, anincreased frequency of IL-2 and IFN-g–secreting CD4þ andCD8þ tumor infiltrating lymphocytes were seen inmice recon-stituted with Aire-deficient thymi (Fig. 3E and F). An increasedfrequency of IL-2 and IFN-g–secreting CD4þ and CD8þ sple-nocytes was also observed in mice reconstituted with Aire-deficient thymi (Fig. 3G and H). An enhanced antitumorimmune response therefore tracks with Aire-deficient thymicstroma.

Defective negative selection of TRP-1–specific T cells inAire-deficient thymus

To define the mechanism by which thymic Aire limits theimmune response against melanoma, we sought to determinewhich antigens targeted in both vitiligo and melanoma mightbe Aire-regulated in the thymus. mTECs from B6.AireGW/þ

mice and B6.Aireþ/þ mice were isolated and relative expres-sion levels of 3 shared vitiligo/melanoma antigens (TRP-1,tyrosinase, and gp100) were determined by real-time RT-PCR.While expression of gp100 was independent of Aire, expression

A B C

D

00

102030405060708090

100

WT

Aire

15 17 19 21 23 25

Per

cent

age

surv

ival

GW/+

-/-

E

2,000

4,000

6,000

8,000

10,000

12,000

14,000

# of

CD

3+ c

ells

per

mm

2

2,000

10 11 12 13 14 15 16 17 18 19

Tu

mo

r si

ze (

mm

3 )

Days after tumor inoculation

**

*

* **

*

200 μm200 μm


WT AireGW/+

WT AireGW/+

WT AireGW/+

Aire

WT

AireGW/+

-/-Aire1,500

1,000

500

Figure 1. Slower growth of B16melanoma in Aire-deficient mice is associated with increased density of tumor-infiltrating T cells. A, B16melanoma cells wereinoculated subcutaneously on the flanks of B6.AireGW/þ (AireGW/þ) mice or B6.Aireþ/þ [wild-type (WT)] mice. Representative tumors from AireGW/þ

mice or Aireþ/þmice at day 18 postinoculation are shown. B, tumor growth in either B6.AireGW/þ, B6.Aire�/� (Aire�/�), or B6.Aireþ/þmice. n¼ 7 for all groups.Data shown are representative of 3 independent experiments carried out. C, survival curves of either AireGW/þ, Aire�/�, or Aireþ/þ mice. n¼ 6 for all groups.D, representative immunofluorescence staining for CD3 (red) and DAPI for melanoma cell nuclei (blue) on tumor sections from AireGW/þ mice or Aireþ/þ

mice at day 18 after B16 melanoma inoculation. E, average densities of CD3þ cells per tumor area (mm2) in tumors inoculated into either AireGW/þ or Aireþ/þ

mice. Four sections were evaluated for each group and 6 areas from each section were randomly selected. �, P < 0.05, error bars represent SEM.

Zhu et al.





of TRP-1 and tyrosinase was significantly decreased in B6.AireGW/þ thymi (Fig. 4A). Thus, Aire-mediated thymic expres-sion of either TRP-1 or tyrosinase could potentially mediateT-cell tolerance toward melanoma.We chose to focus on Aire-mediated thymic expression of

TRP-1 rather than tyrosinase, because thymic tyrosinaseexpression has previously been shown to have no effect on

the immune response against melanoma (21). Importantly,TRP-1 is expressed in B16melanoma cells (31), and is thereforea potential antigenic target in immune rejection of B16 mel-anoma cells. Furthermore, both B6.Aire�/� and B6.AireGW/þ

mice have increased serum autoantibodies against TRP-1,suggesting a break in self-tolerance toward this antigen inAire-deficient mice (Fig. 4B).

15.5%

70.6%

25.4%

60.1%

CD

44

CD4

CD8

CD

4

CD

8

CD62L IL-2

18.9%

100

101

102

103

104

100

101

102

103

104

51.9%

100

101

102

103

104

100

101

102

103

104

CD

4

CD

8

IFN-γ

6.75%

100

101

102

103

104

100

101

102

103

104

21.6%

100

101

102

103

104

100

101

102

103

104

WT AireGW/+

WT AireGW/+

WT AireGW/+A

6.72% 30.0%

8.87% 20.6%4.77% 15.6%

BC

D4

35.5%

100

101

102

103

104

100

101

102

103

104

12.6%

100

101

102

103

104

100

101

102

103

104

34.3%

100

101

102

103

104

100

101

102

103

104

16.0%

100

101

102

103

104

100

101

102

103

104

13.8%

100

101

102

103

104

100

101

102

103

104

6.88%

100

101

102

103

104

100

101

102

103

104

9.55%

100

101

102

103

104

100

101

102

103

104

16.3%

100

101

102

103

104

100

101

102

103

104

2.21%

100

101

102

103

104

100

101

102

103

104

1.05%

100

101

102

103

104

100

101

102

103

104

13.5%

100

101

102

103

104

100

101

102

103

104

5.10%

100

101

102

103

104

100

101

102

103

104

1.41%

100

101

102

103

104

100

101

102

103

104

2.29%

100

101

102

103

104

100

101

102

103

104

7.02%

100

101

102

103

104

100

101

102

103

104

3.97%

100

101

102

103

104

100

101

102

103

104

IL-2

WT02468101214

WT0

4

8

12

16

20

% I

L-2+

WT0

10

20

30

40

% I

L-2+

WT05101520253035

%IL

-2+

WT0.00.51.01.52.02.53.03.5

% IF

N-γ

+%

IFN

-γ +

% IF

N-γ

+

WT0

3

6

9

12

15

Aire GW/+

Aire GW/+

Aire GW/+

Aire GW/+

AireGW/+

Aire

Aire

Aire

GW/+

WT0.0

0.5

1.0

1.5

2.0

2.5

GW/+

01234567

WT GW/+

WT AireGW/+ WT AireGW/+

C

D E*

*

*

*

*

*

CD

8C

D4

CD

8

IL-2

CD

4C

D8

IFN-γ

IFN-γ

CD

4C

D8

Spleen

LN

% I

L-2+

Spleen

LN

Figure2. Aire-deficientmice inoculatedwithB16melanomahaveactivatedTcellswith increasedproinflammatory cytokineproduction. A, flowcytometry plotsof CD62L and CD44 expression in tumor-infiltrating lymphocytes from AireGW/þ or wild-type (WT) mice at day 18. B and C, flow cytometry plots ofintracellular expression of IL-2 (B) and IFN-g (C) in tumor-infiltrating lymphocytes. D and E, flow cytometry plots of intracellular expression of IL-2 (D) and IFN-g(E) in either lymphocytes of spleen and tumor-draining lymph node (LN). Cumulative data are shown in the right. n ¼ 4 for both groups. Data shown arerepresentative experiment of 3 experiments. �, P < 0.05, error bars represent SEM.






We used an MHC class II-restricted TRP-1 TCR Tg to trackTRP-1–specific T cells within the thymus (31). The TRP-1TCR Tg mouse line is derived from a CD4þ T cell clonedfrom a TRP-1–deficient mouse (Tyrp1B-w) immunized withTRP-1 protein. TRP-1 TCR Tg mice were crossed onto arecombinase-activating gene (RAG)–deficient (RAG�/�)

background to prevent endogenous TCR rearrangement.Adoptive transfer of na€�ve TRP-1 TCR Tg RAG�/� cells arecapable of eradicating established B16 melanoma tumors ina lymphopenic host (32). Thus, these TRP-1–specific CD4þ

T cells are directly relevant in the immune response againstB16 melanoma.

0

500

1,000

1,500

2,000

2,500

3,000

7 8 9 10 11 12 13 14 15 16 17 18 19

Tum

or s

ize

(mm

3 )

GW/+

*

*

*

*

*

**

****

Aire

A

G

WT GW/+0

10

20

30

40

%IL

-2+

Aire

0.02.55.07.5

10.012.515.017.5

%IL

-2+

WT GW/+Aire

0.0

2.5

5.0

7.5

10.0

%IF

Nγ

+

WT GW/+Aire

0

5

10

15

20

25%

IFN

γ+

WT GW/+Aire

WT

C

D

6.46%

100

101

102

103

104

100

101

102

103

104

11.3%

100

101

102

103

104

100

101

102

103

104

16.4%

100

101

102

103 4

10

4

100

101

102

103

104

21.8%

100

101

102

103

104

100

101

102

103

104

WTWT

13.6%

100

101

102

103

104

100

101

102

103

104

12.8%

100

101

102

103

104

100

101

102

103

104

22.7%

100

101

102

103

10

100

101

102

103

104

21.1%

100

101

102

103

104

100

101

102

103

104

0

10,000

20,000

30,000

40,000

Aire

#C

D3+

T c

ells p

er

mm

2

E

WT GW/+Aire

GW/+Aire

GW/+Aire

GW/+Aire

*

*

*

*

*

B

F

H

Thymi

Aire

WT

GW/+

B16 tumor cells

Recipient: B6.Nude2 mo

Tumor growth


200 μm 200 μm

WTGW/+

WT

WT GW/+Aire Spleen5.44% 8.07%

9.74% 20.8%

19.6% 32.4%

6.62% 15.6%

IL-2

4D

C8

DC

4D

C8

DC

CD

4C

D8

4D

C8

DC

IL-2 IFN-γ

IFN-γ

Spleen

4

Figure 3. Augmented response against melanomamaps to Aire differency in thymic stroma. A, scheme for thymic transplantation experiment. Thymic stromafrom AireGW/þ mice or wild-type (WT) mice were transplanted under the kidney capsule of athymic C57/BL6 Foxn1�/� (B6.Nude) mice. After 2 monthsto allow T-cell reconstitution, mice were injected with B16 melanoma cells subcutaneously and tumor growth was monitored. B, tumor growth inB6.Nude recipientmice reconstitutedwith AireGW/þor wild-type thymic stroma. n¼ 4 for both groups. C, representative immunofluorescence staining of CD3(red) and DAPI for tumor nuclei (blue) on tumor sections from mice reconstituted with AireGW/þ or wild-type thymi. Four tumor sections were evaluatedfor each group and 6 areas from each section were randomly selected. D, average densities of CD3þ T cells per tumor area (mm2) in tumors of micereconstituted with AireGW/þ or wild-type thymic stroma. E and F, tumor-infiltrating lymphocytes were isolated from mice reconstituted with AireGW/þ orwild-type thymi, and intracellular expressionof IL-2 (E) and IFN-g (F) was evaluated by flowcytometry.GandH, intracellular expression of IL-2 (G) and IFN-g (H)from either CD4þ or CD8þ lymphocytes in spleenwas evaluated by flow cytometry. Cumulative data are shown in right. n¼ 4 for both groups. �,P < 0.05, errorbars represent SEM.

Zhu et al.





In thymi deficient for the TRP-1 antigen (Tyrp1B-w), approx-imately 15% of lymphocyte-gated cells were single positive forCD4 (Fig. 4C, left). In thymi in which cognate antigen isexpressed (Tyrp1þ/þ), on the other hand, the frequency ofCD4 single positive cells is significantly decreased (1%–2%; Fig.4C, middle), indicating efficient negative selection of thispopulation. This change was also seen with the frequency ofclonotype-specific Vb14þCD4 single positive cells among totalthymocytes (Fig. 4D). In addition to decreased frequency ofCD4 single positive cells, the absolute number of CD4 singlepositive thymocytes was also significantly decreased [9 � 105cells in Tyrp1B-w thymi compared with 3 � 104 cells inTyrp1þ/þ thymi (Fig. 4E)]. Thus, consistent with a previousreport (31), effective negative selection of TRP-1 TCR TgRAG�/� cells occurs in thymi expressing normal levels ofTRP-1 antigen.Because TRP-1 expression is decreased in Aire-deficient

thymi (Fig. 4A), we reasoned that Aire deficiency would resultin defective negative selection of TRP-1 TCR Tg RAG�/� cells.

We therefore generated TRP-1 TCR Tg RAG�/� mice on anAire-deficient background (B6.AireGW/þ TRP-1 TCR TgRAG�/� mice) and characterized T-cell development withinthe thymi of these mice. The frequency of CD4 single positivecells was significantly greater inAire-deficient thymi comparedwith thymi expressing normal levels of TRP-1 antigen (Fig. 4C,comparemiddle and right). Similar trends were also seenwhenthe frequency of clonotype-specific Vb14þ CD4 single positiveT cells were assessed as a percentage of total thymocytes (Fig.4D). The absolute number of CD4 single positive cells was alsosignificantly greater inAire-deficient thymi (1� 106) comparedwith thymi expressing normal levels of TRP-1 antigen (3� 104cells; Fig. 4E). Splenic CD4þ T cells reflected the findings in thethymus (Fig. 4F–H).

To investigate thymic negative selection using an alternativeapproach, we carried out bonemarrow chimera experiments inwhich Tyrp1B-w TRP-1 TCR Tg RAG�/� bone marrow wastransplanted into lethally irradiated B6.AireGW/þ or B6.Aireþ/þ

recipients (Supplementary Fig. S5). Within these chimeric

CGW/+Aire

F

4.16%

100

101

102

103

10

100

101

102

103

104

0.561%

100

101

102

103

104

100

101

102

103

104

4.04%

100

101

102

103

104

100

101

102

103

104

1.37%

100

101

102

103

104

100

101

102

103

104

15.6%

100

101

102

103

104

100

101

102

103

104

14.6%

100

101

102

103

104

100

101

102

103

104

Tyrp1 B-w Tyrp1

+/+

+/+Aire

+/+Aire

CD8

CD

4C

D4

CD8

TRP-1 TCR Tg Rag -/- TRP-1 TCR Tg Rag -/-

GW/+Aire

Tyrp1 B-w Tyrp1 +/+

+/+Aire

+/+Aire

TRP-1 TCR Tg Rag -/- TRP-1 TCR Tg Rag -/-

Thy

mus

neelpS

V %

βset yco myht l at ot ni PS4DC +41

V %

βset yconel ps l at ot ni +4DC +41

G

Thymus

103

104

105

106

107

# of

CD4

SP

Thymus

0

# of

CD4

+ (x

10^5

)

Spleen Spleen

Tyrp1 +/+Tyrp1 B-w Tyrp1 +/++/+

Aire +/+Aire GW/+AireTyrp1 +/+Tyrp1 B-w Tyrp1 +/+

+/+Aire +/+Aire GW/+Aire

1

2

4

5

3

6

Tyrp1 +/+Tyrp1 B-w Tyrp1 +/++/+Aire +/+Aire GW/+Aire

Tyrp1 +/+Tyrp1 B-w Tyrp1 +/++/+Aire +/+Aire GW/+Aire

D

A B

E

H

** * *

**

*

0

5

10

15

20 *

0

1

2

3

4

5

4

Figure 4. Aire promotes TRP-1 expression and negative selection of TRP-1–specific T cells in the thymus. A, relative mRNA expression of TRP-1, tyrosinase,and silver/gp100 evaluated by real-time RT-PCR in sorted mTECs from AireGW/þ or Aireþ/þ thymi. Data are representative of 4 independent experiments.B, relative amount of TRP-1 autoantibodies in sera fromAireGW/þ, Aire�/�, and Aireþ/þmice determined by ELISA. n¼ 5 in each group. C, representative plotsof lymphocytes expressing CD4 and CD8 in thymus of Tyrp1B-w TRP-1 TCR Tg RAG�/�, AireGW/þ TRP-1 TCR Tg RAG�/�, or Aireþ/þ TRP-1 TCR TgRAG�/�mice. D and E, cumulative data for percentage of Vb14þCD4 single positive T cells in total thymocytes (D) and absolute number of CD4 single positivethymocytes (E). Each shape represents an individual mouse. F, representative plots of CD4 andCD8 expression in spleen of Tyrp1B-w TRP-1 TCR TgRAG�/�,AireGW/þ TRP-1 TCR Tg RAG�/�, or Aireþ/þ TRP-1 TCR Tg RAG�/�mice. G and H, cumulative data for percentage of Vb14þCD4þ T cells in total splenocytes(G) and absolute number of CD4þ splenocytes (H). Error bars represent SEM.






mice, TRP-1–specific T cells would develop in either an Aire-deficient or Aire wild-type thymic stromal environment.Increased frequency of CD4 single positive cells was seen inAire-deficient recipient thymi (�25%) compared with wild-type recipient thymi (�4%). Consistent with this, an increasedfrequency of clonotype-specific T cells was also seen in Aire-deficient thymus. In this bone marrow chimera model, then,Aire-deficient thymic stroma also results in defective negativeselection of TRP-1–specific CD4þ T cells.

Numbers of TRP-1–specific Tregs are unchanged in thethymi of Aire-deficient mice

The TRP-1 TCR Tg mouse line is unusual in that CD4þ

FOXP3þTregs are generated in the thymus even in the RAG�/�

setting (32). To investigate the effect of Aire in the thymicselection of TRP-1–specific Tregs within the thymus, we com-pared the numbers of TRP-1–specific Tregs in thymi of B6.AireGW/þ and B6.Aireþ/þ mice. Although the percentage ofFOXP3þ Tregs among CD4þ single positive thymocytes isdecreased in B6.AireGW/þ TRP-1 TCR Tg RAG�/� thymi com-paredwith B6.Aireþ/þTRP-1 TCRTgRAG�/�mice (Fig. 5A andB), the absolute numbers of these cells was unchanged (Fig.5C). Interestingly, a greater proportion of CD4þ T cells expres-sing an intermediate level of FOXP3 were seen in thymi of Aire-

deficient versus Aire wild-type TRP-1 TCR Tg Rag�/� thymi(Fig. 5A and D). In addition, the frequency of FOXP3þ Tregs isalso decreased in the spleen of Aire-deficient mice but theabsolute numbers are increased (Supplementary Fig. S6A–S6C). Similar to the thymus, a greater proportion of CD4þ Tcells expressing an intermediate level of FOXP3 were seen inthe spleen of Aire-deficient mice (Supplementary Fig. S6A andS6D).

Because Aire deficiency results in defective negative selec-tion of TRP-1–specific CD4þ T cells without affecting CD4þ

Treg numbers, we sought to determine the ratio of CD4þ

effectors (defined as CD4þ FOXP3�) to CD4þ Tregs (CD4þ

FOXP3þ) in both the thymus and the periphery. The effector:Treg ratio was higher in both thymus and spleen of Aire-deficient mice (Fig. 5D). This increased ratio suggests that thebalance of effector:Tregs are tipped toward increased effectorsin Aire-deficient mice.

Decreased B16 melanoma growth and enhanced T-cellresponse in Aire-deficient TRP-1 TCR Tg mice

We tested whether the greater proportion of TRP-1–specificT effectors in Aire-deficient mice might be linked with moreeffective eradication of melanoma in thesemice. To do this, weinoculated B16 melanoma cells into either B6.AireGW/þ TRP-1

A B

4D

C

FOXP3

E

CGW/+Aire

Tyrp1 +/+

+/+Aire

TRP-1 TCR Tg RAG -/-

% F

OX

P3+

in t

ota

l CD

4SP n.s.*

Tyrp1 Tyrp1 +/+Aire GW/+Aire

+/+0

5,000

10,000

15,000#

of

tota

l FO

XP

3+


+/+ +/+

0

10

20

30

100

101

102

103

104

100

101

102

103

104

100

101

102

103

104

100

101

102

103

104

Effector: Treg ratio (Thymus) Effector: Treg ratio (Spleen)

0

1

2

3

050

100150200250300350

Tef

fect

or:

Tre

g

Tef

fect

or:

Tre

g

* *

+/+

2.71% 13.3% 0.450% 0.084%

D FOXP3 intFOXP3 hi

**

% F

OX

P3+

in t

ota

l CD

4SP

0

5

10

15


+/++/+


+/++/+ Tyrp1 Tyrp1 +/+Aire GW/+Aire

+/++/+

Figure 5. Aire does not affect numbers of TRP-1–specific CD4þ FOXP3þ Tregs in the thymus. A, representative flow cytometry plots of CD4 and FOXP3expression in CD4þ thymocytes from AireGW/þ TRP-1 TCR Tg RAG�/� or Aireþ/þ TRP-1 TCR Tg RAG�/� mice. B and C, cumulative data for percentageof CD4þ FOXP3þ cells among total CD4 single positive thymocytes (B) and absolute number of CD4þ FOXP3þ cells (C) in thymi are shown. Eachshape represents an individual mouse. D, cumulative data for percentage of CD4þ FOXP3 intermediate (int) and CD4þ FOXP3 high (hi) cells among totalCD4 single positive thymocytes. E, cumulative data of effector (CD4þ FOXP3�) versus Treg (CD4þ FOXP3þ) ratios in thymocytes (left) and splenocytes (right)from Aireþ/þ or AireGW/þ TRP-1 TCR Tg RAG�/� mice.

Zhu et al.





TCRTgRAG�/� or B6.Aireþ/þTRP-1TCRTgRAG�/�mice andmonitored tumor growth. Growth of B16melanoma tumorwassignificantly decreased in Aire-deficient mice compared withAire-wild-type controls between days 19 and 26 (Fig. 6A and B).In addition, Aire-deficientmice inoculatedwith B16melanomashowed significantly increased survival compared with Aire-wild-type controls (Fig. 6C).Decreased B16 melanoma growth was associated with

increased density of tumor-infiltrating T cells in B6.AireGW/þ

TRP-1 TCR Tg RAG�/� mice. On immunofluorescent stain-ing of tumor sections, the density of CD3þ T cells was higher intumors fromAire-deficientmice compared with Aire wild-typecontrols (Fig. 6D and E). This increased density of tumor-infiltrating lymphocytes was associated with increased abso-lute numbers of activated/memory (CD62Llow CD44high) TRP-1–specific CD4þ T cells in the spleen and tumor-draininglymph nodes (Fig. 6F). In addition, increased absolute numbersof IL-2 and IFN-g secreting TRP-1–specific CD4þ T cells in thespleen and tumor-draining lymph nodes were also seen (Fig.6G). Thus, enhanced melanoma rejection in Aire-deficientTRP-1 TCR Tg mice is associated with both increased densityof tumor-infiltrating lymphocytes and increased numbers ofactivated cells producing proinflammatory cytokines.As shown in Fig. 5E, the ratio of effector:Tregs is increased in

Aire-deficient TRP-1 TCR Tg RAG�/� mice. We thereforesought to determine whether this increased ratio in Aire-deficient mice might contribute to increased melanomagrowth suppression. To test this, we conducted adoptivetransfers of T cells into RAG-deficient recipients at the 2effector to Treg ratios seen in wild-type (effector:regulatory1:3) and Aire-deficient (effector:regulatory 2:1) mice. Adoptivetransfer of T cells at the effector to Treg ratio in spleen of Aire-deficient mice protected against melanoma outgrowth, where-as adoptive transfer of T cells at the effector to Treg ratio inspleen of wild-type mice did not (Fig. 6H). These findingssuggest that the increased proportion of Tregs may preventefficient antimelanoma immunity in Aire wild-type mice.

DiscussionThis study defines a novel pathway by which Aire deficiency

enhances T-cell–mediated immune rejection of melanoma. Inwild-type mice, Aire-mediated expression of the shared vitili-go/melanoma antigen TRP-1 in thymic mTECs promotesclonal deletion of TRP-1–specific CD4þ T cells and immunetolerance toward melanoma (Fig. 7A). In Aire-deficient mice,on the other hand, lack of Aire-driven TRP-1 expression inthymicmTECs results in defective negative selection of TRP-1–specific T cells. Escape of TRP-1–specific T cells from thethymus then enhances immune rejection of melanoma (Fig.7B). These findings lend support to the proposition thatimmune-mediated control of tumor growth is modified bythymic expression of self-antigens that are also expressed intumors (33, 34).A recent study reported that Aire-deficient mice, if primed

with irradiated melanoma cells, have increased immuneresponse to melanoma antigens and have increased likelihoodof tumor-free survival (19). We report here that Aire-deficientmice have an enhanced immune response tomelanoma result-

ing in slower melanoma growth, even without priming withirradiated melanoma cells. Why the requirement for primingdiffers between the 2 studies is not clear, but may reflectdifferences in housing conditions or in the number of B16cells inoculated permouse (2� 105 vs. 1� 105). In both studies,an elevated antibody response tomelanoma antigens was seenin Aire deficiency. Whether these antibodies participate in theimmune rejection of melanoma, however, is unknown. Elevat-ed autoantibodies against noncancer self-antigens have pre-viously been described in Aire-deficient mice but do not seemto play an active role in autoimmune disease pathogenesis (35,36).

Loss of Aire function has multiple effects on developingthymocytes. In addition to decreasing expression of a largenumber of tissue-specific self-antigens (37), Aire deficiency hasalso been proposed to downregulate the thymic chemokinesCCR4 and CCR7 (38), to decrease thymic medullary accumu-lation of dendritic cells via decreased XCL1 expression (16),and to impede antigen processing andpresentation (14). In thisstudy, Aire's role in upregulating TRP-1 expression in thethymus seems to be sufficient for mediating changes inTRP-1–specific CD4þ T-cell development in the thymus. Com-pared with Aire-deficient thymi, TRP-1 antigen-deficient(Tyrp1B-w) thymi exhibited similar defects in the negativeselection of TRP-1–specific CD4þ T cells.

In this study, Aire deficiency significantly delayedmelanomagrowth in TRP-1 TCR Tg mice. This finding differs from aprevious report that melanoma growth was only marginallydelayed in TRP-1 TCR Tgmice with genetic mutation in TRP-1(31). A potential explanation for this discrepancy is that TRP-1is lacking not only in the thymus but also in the periphery (i.e.,melanocytes) of TRP-1 TCR transgenic mice with a geneticmutation in TRP-1. Thus T cells are na€�ve to the TRP-1 antigenbefore introduction of B16 melanoma cells. In TCR transgenicmice deficient in Aire, on the other hand, TRP-1 is deficient inthe thymus but not in the periphery. Thus, TRP-1 antigen-availability in melanocytes may prime TRP-1–specific T cellsand allow for more effective rejection of melanoma cells.

In addition to promoting T-cell tolerance within the thymus,Aire also enforces self-tolerance within secondary lymphoidorgans (30). We use thymic transplantation studies to showthat immune tolerance toward B16 melanoma is conferred byAire expression in the thymus and does not require extrathy-mic Aire expression. In contrast to our results, a previous studyhas reported that CD8þ T-cell tolerance toward tyrosinase,another shared vitiligo/melanoma antigen, is not mediated bytyrosinase expression in the thymus (21). Furthermore, extra-thymic induction of T-cell tolerance toward tyrosinase hasbeen shown to be Aire-independent (39). Thus, distinct toler-ance mechanisms govern CD4þ T-cell responses to TRP-1compared with CD8þ T-cell responses to tyrosinase.

To date, studies investigating the role of Aire in thymicnegative selection have used mouse models involving forcedexpression of a "neo"-self-antigen [e.g., ovalbumin (OVA) andhen egg lysomzyme (HEL)] under tissue-specific promoterswithin the thymus (13, 14, 40, 41). Because these antigens areoverexpressed foreign proteins, their expression pattern andlevels may not mirror physiologic self-antigens (42). We show






Days after tumor inoculationTu

mor

siz

e (m

m3 ) TRP-1 TCR Tg Aire

GW/+TRP-1 TCR Tg Aire

0

.5

1.0

1.5

#IL

-2 e

xpre

ssin

g (

x105

)

0

1

2

3

4

5

6

B

D

F

* **

**

*

**

200 μΜ

TRP-1 TCR Tg: +/+

GW/+AireAire+/+

GW/+AireAire+/+

+/+

GW/+

#IL

-2 e

xpre

ssin

g (

x103

)

CD

3+ T

cel

ls p

er m

m2

E

+/+

0 10 20 300

25

50

75

100

Per

cen

t su

rviv

al

C

*

0

.4

.8

1.2#

I

0

.4

.8

1.2

#IF

N-γ

exp

ress

ing

(x1

03)

FN

-γ e

xpre

ssin

g (

x103

)

GW/+AireAire

+/+

GW/+AireAire+/+

G

0

.5

1.0

1.5

2.0

2.5

# C

D62L

CD

44+

(x10

5)

0123456

# C

D62L

CD

44+

(x10

3)

GW/+AireAireTRP-1 TCR Tg :

+/+

GW/+AireAire

+/+

Spl

een

Days after tumor inoculation14 15 16 17 18 19 20 21 22 23 24 25 26

0

200

400

600

800

1,000

1,200

TRP-1 TCR Tg Aire GW/+TRP-1 TCR Tg Aire

+/+

TRP-1 TCR Tg Aire GW/+TRP-1 TCR Tg Aire+/+

1,000

2,000

7,000

6,000

5,000

4,000

3,000

*

*

*

*

* *

*

A

200 μΜTRP-1 TCR Tg Aire Aire

GW/+

Aire

Aire

LN

neel pS

LN

TRP-1 TCR Tg : TRP-1 TCR Tg :

TRP-1 TCR Tg : TRP-1 TCR Tg : TRP-1 TCR Tg :

8 9 10 11 12 13 14 15 16 170

500

1,000

1,500

2,000Teff:Treg 2:1Teff:Treg 1:3


Tu

mo

r siz

e (

mm

3)

H

Figure 6. Slower growth of B16 melanoma in Aire-deficient TRP-1 TCR Tg mice is associated with an increased density of TRP-1–specific tumor-infiltratingT cells. A, B16 melanoma cells were inoculated subcutaneously into AireGW/þ or Aireþ/þ TRP-1 TCR Tg RAG�/� mice. Representative tumors at day 26 aftertumor inoculation are shown.B, tumorgrowth in eitherAireGW/þorAireþ/þTRP-1TCRTgRAG�/�mice.n¼6 for bothgroups.Data shownare representativeof2 experiments carried out. C, survival curves of either AireGW/þ or Aireþ/þ TRP-1 TCR Tg RAG�/� mice. n ¼ 6 for both groups. D, immunofluorescence

Zhu et al.





here that Aire deficiency results in reduced thymic expressionof the endogenously expressed self-antigen, TRP-1. This loss ofTRP-1 expression in the thymus directly results in the defectivenegative selection of TRP-1–specific CD4þ thymocytes. Thisfinding is significant because it shows that Aire-mediatedexpression of endogenous thymic self-antigens is sufficient todrive efficient negative selection of self-reactive T cells.The role of Aire deficiency in the generation of CD4þ Treg

development in the thymus is controversial. Most studies havereported that Aire deficiency does not affect the absolutenumbers of thymic CD4þ Tregs (13, 14, 43). Lei and colleagues,on the other hand, reported that thymic Tregs are decreased by50% in Aire-deficient mice (16). Similar to the first group ofstudies, our study did not detect any difference in the absolutenumber of TRP-1–specific CD4þ Tregs developing in Aire-deficient thymi. A possible explanation for the differing con-clusions is the methodologies used to quantify Tregs. Unlikethe other studies, Lei and colleagues used immunofluores-cence analysis of thymus sections to quantify Tregs within themedulla. It is also possible that Aire deficiency may alter Tregfunction in a qualitative manner. In our study, there seems tobe an increased frequency of TRP-1TCRTg cells expressing lowlevels of FoxP3 in the periphery of Aire-deficient mice. Theimplications for this observation will require further investi-

gation. However, it is noteworthy that CD4þ T cells expressinglow levels of FoxP3 may play a role in autoimmunity (44).Notably, we show here that the ratio of effector:Tregs differdramatically between Aire-deficient and wild-type TRP-1 TCRTg mice, and an increased proportion of Tregs is inverselyassociated with antimelanoma immunity. This finding sug-gests an inhibitory role for TRP-1–specific Tregs in antimela-noma immunity.

Our finding that Aire-driven TRP-1 expression in the thymusresults in clonal deletion of TRP-1–specific T cells has potentialimplications for designing immunotherapies in the treatmentof metastatic melanoma. In metastatic melanoma, immuno-therapeutic approaches result in complete responses in only asmall subset of patients (4), and normal Aire-mediated self-tolerancemechanismsmay contribute to the limited efficacy ofthese approaches (45). Pursuing strategies that modulateeither Aire expression or TRP-1 expression in the thymus, forexample, may allow increased escape of TRP-1–specific T-cellclones important for effective melanoma eradication. Alter-natively, strategies that increase the number of TRP-1–reactiveT cells in the periphery, such as adoptive transfer of TRP-1–reactive T cells, may help overcome Aire-mediated clonaldeletion of these T cells in the thymus. Future studies to testthese strategies will be of direct clinical relevance.

staining for CD3 (red) andDAPI staining for tumor nuclei (blue) onB16melanoma sections fromAireGW/þor Aireþ/þ TRP-1 TCRTgRAG�/�mice at day 26 aftertumor inoculation. Four tumor sections were evaluated for each group and 6 areas from each section were randomly selected. E, average densitiesof CD3þ cells per tumor area (mm2) in either AireGW/þ or Aireþ/þ TRP-1 TCR Tg RAG�/� mice. F, absolute numbers of activated (CD62Llow CD44high)CD4þ T cells from spleen and draining lymph nodes (LN) as evaluated by flow cytometry. Each shape represents individual mouse. G, absolute numbersof CD4þ T cells expressing IL-2 and IFN-g from spleen and draining lymph nodes as evaluated by flow cytometry. Each shape represents individualmouse. H, melanoma growth in RAG-deficient mice receiving 2 � 105 CD4þCD25� effector T cells (Teff) and variable CD4þCD25þ Tregs at the indicatedratios of effector to Tregs. �, P < 0.05, error bars represent SEM. �, P < 0.05, error bars represent SEM.

Aire Aire

TRP-1-specific T cell TRP-1-specific

T cell

Wildtype thymus Aire-deficient thymus

mTEC mTEC

Apoptosis

Tumor growth Tumor rejection

XX

TRP-1-specific T cells TRP-1-specific

T cells

TRP-1 TRP-1

A B

Figure 7. Model of howAire deficiency elicits amore robust immune response againstmelanoma. A, Aire normally induces TRP-1 expression in thymicmTECs.Presentation of TRP-1 to developing T cells results in deletion of TRP-1–specific T cells via apoptosis. B, Aire deficiency results in decreased TRP-1expression in mTECs. Lack of TRP-1 presentation to TRP-1–specific T cells allows escape of these T cells from negative selection and enhances the immuneresponse against melanoma.






Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: M.-L. Zhu, M.A. SuDevelopment of methodology: M.-L. Zhu, M.A. SuAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): M.-L. Zhu, A. NagavalliAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): M.-L. ZhuWriting, review, and/or revision of the manuscript: M.-L. Zhu, M.A. SuAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): A. NagavalliStudy supervision: M.A. Su

AcknowledgmentsThe authors thank Jenny Ting, Yisong Wan, and Fatima Larry for critical

review of the article.

Grant SupportThis study was supported by NIH grant (K08 AI076429).The costs of publication of this article were defrayed in part by the

payment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact.

Received October 1, 2012; revised January 16, 2013; accepted January 20, 2013;published OnlineFirst January 31, 2013.

References1. Siegel R, Ward E, Brawley O, Jemal A. Cancer statistics, 2011: the

impact of eliminating socioeconomic and racial disparities on prema-ture cancer deaths. CA Cancer J Clin 2011;61:212–36.

2. Woods GM, Malley RC, Muller HK. The skin immune system andthe challenge of tumour immunosurveillance. Eur J Dermatol 2005;15:63–9.

3. McGovern VJ. Spontaneous regression of melanoma. Pathology1975;7:91–9.

4. Sosman JA. Ipilimumab (anti-CTLA-4) immunotherapy in advancedmelanoma. UpToDate. 2012 [cited 2011 Dec 13]. Available from:http://www.uptodate.com.

5. Ongenae K, Van Geel N, Naeyaert JM. Evidence for an autoimmunepathogenesis of vitiligo. Pigment Cell Res 2003;16:90–100.

6. Ram M, Shoenfeld Y. Harnessing autoimmunity (vitiligo) to treatmelanoma: a myth or reality? Ann N Y Acad Sci 2007;1110:410–25.

7. Hodi FS. Well-defined melanoma antigens as progression markers formelanoma: insights into differential expression and host responsebased on stage. Clin Cancer Res 2006;12:673–8.

8. Gogas H, Ioannovich J, Dafni U, Stavropoulou-Giokas C, Frangia K,Tsoutsos D, et al. Prognostic significance of autoimmunity duringtreatment of melanoma with interferon. N Engl J Med 2006;354:709–18.

9. Ahonen P, Myllarniemi S, Sipila I, Perheentupa J. Clinical variation ofautoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy(APECED) in a series of 68 patients. N Engl J Med 1990;322:1829–36.

10. Tazi-Ahnini R, McDonagh AJ, Wengraf DA, Lovewell TR, VasilopoulosY, Messenger AG, et al. The autoimmune regulator gene (AIRE) isstrongly associated with vitiligo. Br J Dermatol 2008;159:591–6.

11. Derbinski J, Schulte A, Kyewski B, Klein L. Promiscuous gene expres-sion in medullary thymic epithelial cells mirrors the peripheral self. NatImmunol 2001;2:1032–9.

12. AndersonMS, Venanzi ES, Klein L, Chen Z, Berzins SP, Turley SJ, et al.Projection of an immunological self shadow within the thymus by theaire protein. Science 2002;298:1395–401.

13. Liston A, Lesage S, Wilson J, Peltonen L, Goodnow CC. Aire regulatesnegative selection of organ-specific T cells. Nat Immunol 2003;4:350–4.

14. Anderson MS, Venanzi ES, Chen Z, Berzins SP, Benoist C, Mathis D.The cellular mechanism of Aire control of T cell tolerance. Immunity2005;23:227–39.

15. Koble C, Kyewski B. The thymic medulla: a unique microenvironmentfor intercellular self-antigen transfer. J Exp Med 2009;206:1505–13.

16. Lei Y, Ripen AM, Ishimaru N, Ohigashi I, Nagasawa T, Jeker LT, et al.Aire-dependent production of XCL1mediatesmedullary accumulationof thymic dendritic cells and contributes to regulatory T cell develop-ment. J Exp Med 2011;208:383–94.

17. Pomie C, Vicente R, Vuddamalay Y, Lundgren BA, van der Hoek M,Enault G, et al. Autoimmune regulator (AIRE)-deficient CD8þCD28low

regulatory T lymphocytes fail to control experimental colitis. Proc NatlAcad Sci U S A 2011;108:12437–42.

18. Conteduca G, Ferrera F, Pastorino L, Fenoglio D, Negrini S, SormaniMP, et al. The role of AIRE polymorphisms in melanoma. Clin Immunol2010;136:96–104.

19. Trager U, Sierro S,DjordjevicG, BouzoB, KhandwalaS,Meloni A, et al.The immune response to melanoma is limited by thymic selection ofself-antigens. PLoS ONE 2012;7:e35005.

20. Huijbers IJ, Soudja SM,UyttenhoveC,BuferneM, Inderberg-SusoEM,Colau D, et al. Minimal tolerance to a tumor antigen encoded by acancer-germline gene. J Immunol 2012;188:111–21.

21. Nichols LA, Chen Y, Colella TA, Bennett CL, Clausen BE, EngelhardVH. Deletional self-tolerance to a melanocyte/melanoma antigenderived from tyrosinase is mediated by a radio-resistant cell in periph-eral and mesenteric lymph nodes. J Immunol 2007;179:993–1003.

22. SuMA, Giang K, Zumer K, Jiang H, Oven I, Rinn JL, et al. Mechanismsof an autoimmunity syndrome in mice caused by a dominant mutationin Aire. J Clin Invest 2008;118:1712–26.

23. Kruczynski A, Hill BT. Classic in vivo cancermodels: three examples ofmouse models used in experimental therapeutics. Curr Protoc Phar-macol 2002;Chapter 5:Unit5.24.

24. Yu YR, Fong AM, Combadiere C, Gao JL, Murphy PM, Patel DD.Defective antitumor responses in CX3CR1-deficientmice. Int JCancer2007;121:316–22.

25. DeVossJ,HouY, JohannesK, LuW,LiouGI,Rinn J, et al. Spontaneousautoimmunity prevented by thymic expression of a single self-antigen.J Exp Med 2006;203:2727–35.

26. Alikhan A, Felsten LM, Daly M, Petronic-Rosic V. Vitiligo: a compre-hensive overview Part I. Introduction, epidemiology, quality of life,diagnosis, differential diagnosis, associations, histopathology, etiol-ogy, and work-up. J Am Acad Dermatol 2011;65:473–91.

27. Corthay A, Skovseth DK, Lundin KU, Rosjo E, Omholt H, Hofgaard PO,et al. Primary antitumor immune response mediated by CD4þ T cells.Immunity 2005;22:371–83.

28. Mumm JB, Emmerich J, Zhang X, Chan I, Wu L, Mauze S, et al. IL-10elicits IFNgamma-dependent tumor immune surveillance. Cancer Cell2011;20:781–96.

29. Chapuis AG, Thompson JA,Margolin KA, RodmyreR, Lai IP, DowdyK,et al. Transferred melanoma-specific CD8þ T cells persist, mediatetumor regression, and acquire central memory phenotype. Proc NatlAcad Sci U S A 2012;109:4592–7.

30. Gardner JM, Devoss JJ, FriedmanRS,Wong DJ, Tan YX, Zhou X, et al.Deletional tolerance mediated by extrathymic Aire-expressing cells.Science 2008;321:843–7.

31. Muranski P, Boni A, Antony PA, Cassard L, Irvine KR, Kaiser A, et al.Tumor-specific Th17-polarized cells eradicate large established mel-anoma. Blood 2008;112:362–73.

32. Xie Y, Akpinarli A, Maris C, Hipkiss EL, Lane M, Kwon EK, et al. Naivetumor-specific CD4(þ) T cells differentiated in vivo eradicate estab-lished melanoma. J Exp Med 2010;207:651–67.

33. Cloosen S, Arnold J, Thio M, Bos GM, Kyewski B, Germeraad WT.Expression of tumor-associated differentiation antigens,MUC1 glyco-forms and CEA, in human thymic epithelial cells: implications for self-tolerance and tumor therapy. Cancer Res 2007;67:3919–26.

34. Kyewski B, Klein L. A central role for central tolerance. Annu RevImmunol 2006;24:571–606.

35. Devoss JJ, Shum AK, Johannes KP, Lu W, Krawisz AK, Wang P, et al.Effector mechanisms of the autoimmune syndrome in the murine

Zhu et al.





model of autoimmune polyglandular syndrome type 1. J Immunol2008;181:4072–9.

36. Gavanescu I, Benoist C,MathisD.Bcells are required for Aire-deficientmice todevelopmulti-organautoinflammation: a therapeutic approachfor APECED patients. Proc Natl Acad Sci U S A 2008;105:13009–14.

37. Oliveira EH, Macedo C, Donate PB, Almeida RS, Pezzi N, Nguyen C,et al. Expression profile of peripheral tissue antigen genes inmedullarythymic epithelial cells (mTECs) is dependent on mRNA levels ofautoimmune regulator (Aire). Immunobiology 2013;218:96–104.

38. LaanM,KisandK,Kont V,Moll K, Tserel L, Scott HS, et al. Autoimmuneregulator deficiency results in decreased expression of CCR4 andCCR7 ligands and in delayed migration of CD4þ thymocytes. J Immu-nol 2009;183:7682–91.

39. Cohen JN, Guidi CJ, Tewalt EF, Qiao H, Rouhani SJ, Ruddell A, et al.Lymph node-resident lymphatic endothelial cells mediate peripheraltolerance via Aire-independent direct antigen presentation. J ExpMed2010;207:681–8.

40. Liston A, Gray DH, Lesage S, Fletcher AL, Wilson J, Webster KE,et al. Gene dosage–limiting role of Aire in thymic expression, clonal

deletion, and organ-specific autoimmunity. J Exp Med 2004;200:1015–26.

41. Schuler P, Contassot E, Irla M, Hugues S, Preynat-Seauve O, Beer-mann F, et al. Direct presentation of a melanocyte-associated antigenin peripheral lymph nodes induces cytotoxic CD8þ T cells. Cancer Res2008;68:8410–8.

42. Hogquist KA, Baldwin TA, Jameson SC. Central tolerance: learningself-control in the thymus. Nat Rev Immunol 2005;5:772–82.

43. Kuroda N, Mitani T, Takeda N, Ishimaru N, Arakaki R, Hayashi Y, et al.Development of autoimmunity against transcriptionally unrepressedtarget antigen in the thymus of Aire-deficient mice. J Immunol2005;174:1862–70.

44. Wan YY, Flavell RA. Regulatory T-cell functions are subverted andconverted owing to attenuated Foxp3 expression. Nature 2007;445:766–70.

45. Engelhard VH, Bullock TN, Colella TA, Sheasley SL, Mullins DW.Antigens derived from melanocyte differentiation proteins: self-toler-ance, autoimmunity, and use for cancer immunotherapy. Immunol Rev2002;188:136–46.






2013;73:2104-2116. Published OnlineFirst January 31, 2013.Cancer Res Meng-Lei Zhu, Anil Nagavalli and Maureen A. Su Melanoma

Specific Immune Rejection of−Aire Deficiency Promotes TRP-1

Updated version

10.1158/0008-5472.CAN-12-3781doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2013/01/31/0008-5472.CAN-12-3781.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/73/7/2104.full#ref-list-1

This article cites 43 articles, 20 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/73/7/2104.full#related-urls

This article has been cited by 7 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/73/7/2104To request permission to re-use all or part of this article, use this link


http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-12-3781http://cancerres.aacrjournals.org/content/suppl/2013/01/31/0008-5472.CAN-12-3781.DC1http://cancerres.aacrjournals.org/content/73/7/2104.full#ref-list-1http://cancerres.aacrjournals.org/content/73/7/2104.full#related-urlshttp://cancerres.aacrjournals.org/cgi/alertsmailto:[email protected]://cancerres.aacrjournals.org/content/73/7/2104http://cancerres.aacrjournals.org/

airedeﬁciencypromotestrp-1 speciﬁcimmunerejection of melanoma · microenvironment and...

Documents