a. marx the autoimmune regulator aire in thymoma biology jto 2010

MALIGNANCIES OF THE THYMUS

The Autoimmune Regulator AIRE in Thymoma BiologyAutoimmunity and Beyond

Alexander Marx, MD,* Peter Hohenberger, MD,† Hans Hoffmann, MD,‡Joachim Pfannschmidt, MD,‡ Philipp Schnabel, MD,‡ Hans-Stefan Hofmann, MD,§Karsten Wiebe, MD,� Berthold Schalke, MD,¶ Wilfred Nix, MD,# Ralf Gold, MD,**

Nick Willcox, MD,†† Part Peterson, PhD,‡‡ and Philipp Strobel, MD*

Abstract: Thymomas are tumors of thymic epithelial cells. Theyassociate more often than any other human tumors with variousautoimmune diseases; myasthenia gravis is the commonest, occur-ring in 10–50% of thymoma patients, depending on the WorldHealth Organization-defined histologic subtype. Most thymomasgenerate many polyclonal maturing T lymphocytes but in disorga-nized microenvironments Failure to induce self-tolerance may be akey factor leading to the export of potentially autoreactive CD4�

progeny, thus predisposing to autoimmune diseases. Normally, themaster Autoimmune Regulator promotes expression of peripheraltissue-restricted antigens such as insulin by medullary thymic epi-thelial cells and induction of tolerance to them. The failure of �95%of thymomas to express autoimmune regulator is another featurepotentially contributing to autoimmunity.

Key Words: Thymoma, AIRE, APECED, APS-1, Myastheniagravis.

(J Thorac Oncol. 2010;5: S266–S272)

The normal thymus generates maturing T lymphocytesfrom bone marrow-derived progenitors, and screens them

for autoreactivity. Those that are potentially autoaggressiveare deleted in the medulla—against major histocompatibilitycomplex (MHC) antigens (including cross-presented peptidesof autoantigens) by dendritic cells (DC) and against tissue-

restricted antigens (TRAgs) by medullary thymic epithelialcells (mTECs) that express them under the control of auto-immune regulator (AIRE) (or perhaps other analogous regu-lators). A minority of the latter group escapes deletion andmatures into the natural regulatory T cell (Treg) subset. Likeother descendants, Tregs then emigrate into the peripheralpool, where they maintain self-tolerance and immunityagainst infection.

ThymomasThymomas are histologically diverse tumors of thymic

epithelial cells (TECs). The new World Health Organization(WHO) classification divides them into types A, AB, B1, B2,and B3 (Figure 1).1,2 By contrast, thymic carcinomas arenamed after their morphologically similar extrathymic coun-terparts and no longer as “type C thymomas.” The localextension of thymomas and their metastatic spread is de-scribed by the staging system first suggested by Masaoka3 orits slightly modified version.2 Stage, in addition to completeresection, have proved to be the best single prognostic markerand an essential indicator for guiding therapy.4–6 Updatearticles dealing with the classification and staging of thymo-mas are published in this monograph (Masaoka A; Kondo K).

In contrast with other cancers (including thymic carci-nomas), �80% of thymomas are thymopoietic and generate“single positive” CD4� and CD8� T cells, often in vastnumber, but in disordered microenvironments. This thymo-poiesis varies according to the thymoma subtype: it is (bydefinition) minimal or even apparently absent in type Athymomas; it is prominent but patchy and relatively ineffi-cient in type AB, generating few mature single CD4� orCD8� T cells; it is abundant and appearing almost normal intypes B1 and B2; and it is sparse but phenotypically quitenormal in type B3.7,8 Thymomas of types A and AB almostnever metastasize; although slow-growing, types B2 and B3can be invasive both locally and by the pleura, althoughblood-borne metastases are very rare.6

With the exception of some type A thymomas, almostall the myasthenia gravis (MG)-associated tumors containstrikingly more mature CD4� CD45RA� cells than those innon-MG cases. Both CD4� and CD8� mature T cells areexported to the blood and peripheral immune system9,10; thisinfiltration by T cells that have not been properly screened for

*Institute of Pathology; †Division of Thoracic Surgery, University MedicalCentre Mannheim, University of Heidelberg, Heidelberg; ‡Departmentsof Thoracic Clinics, University of Heidelberg, Heidelberg; §University ofRegensburg, Regensburg; �University of Munster, Munster; ¶Depart-ments of Neurology, University of Regensburg, Regensburg; #Universityof Mainz, Mainz; **University of Bochum, Bochum, Germany; ††Neu-rosciences Group, Institute of Molecular Medicine, University of Oxford,United Kingdom; and ‡‡Institute of Molecular Pathology, University ofTartu, Estonia.

Disclosure: The authors declare no financial interest in this manuscript.Address for correspondence: Alexander Marx, MD, Institute of Pathology,

University Medical Center Mannheim, University of Heidelberg, Theodor-Kutzer-Ufer 1-3, 68135 Mannheim, Germany. E-mail: [email protected]

Copyright © 2010 by the International Association for the Study of LungCancerISSN: 1556-0864/10/0510-0266

Journal of Thoracic Oncology • Volume 5, Number 10, Supplement 4, October 2010S266

autoreactivity appears to be a key factor predisposing tothymoma-associated MG (TAMG) and likely other associ-ated autoimmune diseases in thymoma patients.11,12 To ex-plain the particular focus on muscle targets, others havehypothesized some bias in selection, or even active autoim-munization, of T cells developing within or circulatingthrough thymomas of types AB, B2 to B3 or A, respectively(reviewed in Ref. 12), in which some TEC do express severalrelevant muscle autoepitopes.13–15 Similar autoimmunizationis believed to underlie other paraneoplastic autoimmune syn-dromes.16

Myasthenia GravisMG is a prototypic T cell-dependent autoantibody-

mediated disease, with well-defined patient subgroups (Fig-ure 2). In most MG patients, the characteristic fatigablemuscle weakness is clearly caused by autoantibodies to theacetylcholine receptor (AChR) in its native conformation,and only rarely by autoantibodies to other autoantigens.17,18

In other subgroups without thymomas, MG onset is beforeage 45 in early-onset MG (EOMG), or after that in late-onset

MG (LOMG). In �80% of EOMG patients, the thymusshows “lymphofollicular hyperplasia” with characteristiclymph node-like infiltrates; their germinal centers are focusedon rare muscle-like myoid cells,19 the only cells outsidemuscle that express intact AChR. There is also ongoingproduction of anti-AChR antibodies in situ.20 By contrast, inLOMG, the thymus is apparently normal-for-age (involuted,not inflamed), but there are intriguing serological parallelswith TAMG (see below).

Thymoma-Associated MGTAMG is a distinct MG subgroup: the risk of MG

increases progressively with WHO type from A (�20%)�AB to B1 �B2 to B3 (�60%).11 TAMG patients almostalways have anti-AChR autoantibodies and these are typi-cally not produced within thymomas but rather in the adjacentthymic remnant21 or, hypothetically, lymph nodes and bonemarrow. In addition, already at diagnosis, thymoma patientsshow an almost unique spectrum of additional autoantibodies:to striational muscle autoantigens (Figure 2), especially titin(in �90% of TAMG patients), ryanodine receptors (in 50%);type I interferons, IL-12 (in �50%) and, rarely, severalTh17-related cytokines (see below) (Table 1).22,23

Interestingly, autoantibodies to titin, ryanodine recep-tors, type I interferons, and IL-12 are also found in lowerproportions of LOMG patients (without thymomas) but al-most never in EOMG.24 Further, patients with LOMG andTAMG share several genetic associations, e.g., in the IL10,Fc�2a, TNFB, and TNFA genes,25,26 also, expansions ofnaive CD8� T cells in the blood,9,27 suggest parallel (al-though probably not identical) defects in central toleranceinduction. The similarities are even more striking in theantititin� subgroups of TAMG and LOMG.25–30

The reasons for the serological overlap between TAMGand LOMG, and for its major focus on muscle autoantigens,are unknown. In theory, some LOMG patients might havealready rejected occult thymomas before diagnosis; however,that would predict that these autoantibodies should occurtogether much more than they apparently do. Alternatively, itmight result from the absence or rarity of thymic myoid cellsin thymomas (TAMG) or involuted (LOMG) thymuses, ifthey supply muscle autoantigens to tolerogenic DCs in thenormal thymus.12 That implies continuing thymopoiesis eveninto old age; if that does occur, the autoantigens expressed byTEC might bias T cell selection there, as also suspected inthymomas. In addition, two other features of virtually allMG-associated thymomas might contribute to autoimmuni-zation: first, defective generation of FoxP3� regulatory T-cells within thymomas31 and, second, absence of AIRE�

neoplastic TEC24 and remarkable parallels of TAMG withmonogenic autoimmune diseases.

Autoimmune RegulatorAIRE, the Autoimmune Regulator normally promotes

expression of various TRAgs by occasional mTECs, e.g.,thyroglobulin (Figure 3) and insulin.21 This “promiscuousexpression” of �1000 different TRAgs allows them to deleteautoaggressive T cells.32 To achieve this low-level ectopicexpression, AIRE uses a remarkable variety of molecular

FIGURE 1. Characteristic histology of the main WHO-de-fined thymomas that are commonly associated with myas-thenia gravis (MG) and/or other autoimmune diseases.a,Type A thymoma; b, type AB thymoma; c-e, type B1 thy-moma, H&E: lighter medullary area (med) with a single Has-sall’s corpuscle (HC) sharply separated from darker cortical(cort) area (c); AIRE� medullary thymic epithelial cells(brown nuclei) adjacent to HC (d); desmin-positive myoidcells (brown) in the medullary area of a type B1 thymoma(e). f, Type B2 thymoma; g, Type B3 thymoma. H&E (a-c,f,g); immunoperoxidase (d,e). Reprinted with permissionfrom Autoimmunity.12

Journal of Thoracic Oncology • Volume 5, Number 10, Supplement 4, October 2010 Role of AIRE in Thymoma Biology

Copyright © 2010 by the International Association for the Study of Lung Cancer S267

FIGURE 2. Correlation of myas-thenia gravis (MG) subtypes withthymus/thymoma histology, fre-quency of AIRE� medullary thymicepithelial cells (mTECs), and char-acteristic autoantibodies to acetyl-choline receptor (AChR); the stria-tional muscle antigen titin; thesarcoplasmic reticulum calciumchannel ryanodine receptor, RyR;and various cytokines (different in-terferon-alpha subtypes; IFN�s; in-terleukin 12, IL-12). EOMG, early-onset MG (� disease onset �45years of age); LOMG, late-onsetMG (MG without thymoma andonset at 45 years or later); TAMG,thymoma-associated MG. H&E-stained sections: EOMG and LOMG(�100); thymoma (�400). Thethymoma represents a characteris-tic WHO type B2 thymoma (B2).

TABLE 1. Contrasts and Parallels Between Patients With Either APECED (With Germ Line AIRE Mutations) orAIRE-Negative Thymomas

Organ-Specific Autoimmunity

APECED Thymoma

Prevalence of ClinicalManifestations

AutoantigenTargeted Prevalence

AutoantigenTargeted Prevalence

Endocrine

Parathyroid 90% (HP) NALP-5 �45%a

Adrenal cortex 80% (AF) 21-OH �66%b All rare

Gonads 10–20% (HG) SCC �50%b

Pancreatic � cells 5–10% (DM) GAD65 �37%b

Ectodermal

Melanocytes �20% (VI) AADC �50%b All rare

Hair follicles �30% (AP) TH �40%b

Muscle None reported

Motor endplate striations All rare or absent AChR (MG) 100%c

Titin (MG) �90%c

RyR (MG) 50%c

Cytokines

Type I interferons IFN-� �95% �70%c

IFN-� 99% �60%c

IL-12 IL-12 0% 50%c

Th17 series CMC � 90% IL-17A �45% 5–10%c

IL-17F �75% 5–10%c

IL-22 �90% 5–10%c

a Of all APECED patients (50% of those with hypoparathyroidism).67

b Of all APECED patients.68

c Of thymoma patients with associated MG.23,51

AIRE, autoimmune regulator; APECED, autoimmune polyendocrinopathy candidiasis ecodermal dysplasia; HP, hypoparathyroidism; AF, adrenocorticalfailure; HG, hypogonadism; DM, diabetes mellitus; VI, vitiligo; AP, alopecia; CMC, chronic mucocutaneous candidiasis; NALP-5 (NACHT leucine-rich-repeat protein-5); 21-OH, 21-hydroxylase; SCC, side-chain cleavage enzyme; GAD65, glutamic acid decarboxylase 65; AADC, aromatic L-amino aciddecarboxylase; TH, tyrosine hydroxylase; AChR, acetylcholine receptor; RyR, ryanodine receptor; MG, myasthenia gravis.

Marx et al. Journal of Thoracic Oncology • Volume 5, Number 10, Supplement 4, October 2010

Copyright © 2010 by the International Association for the Study of Lung CancerS268

mechanisms.33–35 In addition, it promotes apoptosis of termi-nally mature mTECs36 (see sections Could AIRE deficiencyin thymomas play a role in paraneoplastic MG? and CouldAIRE deficiency have an oncogenic impact on the develop-ment of thymomas?) and thus facilitates tolerogenic “cross-presentation” of mTEC-derived TRAgs by DCs.37 Further-more, it can also promote deletion of T cells reactive againstsome AIRE-independent “bystander” TRAgs,38 possibly by“creating a tolerogenic climate” by alterations in autoantigenprocessing or function of thymic DCs.

The relevance of AIRE for central tolerance in humanswas first highlighted by patients with mostly recessive AIREmutations. 39 The resulting multiorgan autoimmune syndromeis termed “autoimmune polyendocrinopathy candidiasis ec-todermal dystrophy (APECED).” Clinically, it is extremelyvariable in both humans and Aire�/� mice, likely because ofdifferences in genetic background.37 The main symptoms inhumans are chronic muco-cutaneous candidiasis (CMC), usu-ally followed by autoimmunity that most often targets theparathyroid glands and adrenal cortex—the diagnostic triad;also ectodermal tissues, and/or gut, but rarely neuromuscularstructures can be involved.40 In Aire�/� mice, the autoim-mune features result from thymopoiesis in the absence ofAire� mTEC, exactly as we hypothesized in thymomas.However, the autoantibody repertoires show only limitedoverlap between APECED and TAMG, although the verysimilar autoantibodies against cytokines found in manyAPECED and TAMG patients are striking exceptions (Table1), as discussed in section Restricted but significant overlapbetween TAMG and APECED. It is still not clear why AIREdeficiency in APECED mainly affects endocrine and ectoder-mal targets while largely sparing neuromuscular antigens.

AIRE IN THYMOMASDefective expression of AIRE in nearly all thymomas

(except for �60% of the rare B1s) has implications for theclassification of thymomas, for the establishment of thymo-ma-related biomarkers, for autoimmune predisposition, andhypothetically for the transformation of mTEC.

AIRE Expression is Lacking in �95% ofThymomas

There are controversial proposals to “simplify” theWHO classification of thymomas by “lumping together” typeA, AB, B1 thymomas.41 However, we consider that type B1is entirely different. It is the only one to show any AIRE�

mTEC (Figure 3).24 In addition, type B1 thymomas show themost obvious “organotypic” features among all thymomasubtypes, including (a) the most clear-cut cortico-medullarydifferentiation among all thymoma subtypes, and frequent(although variable) Hassall’s corpuscles; (b) thymopoiesismost similar to that in the normal thymus; (c) myoid cells inmedullary islands in some (but not all) cases, although atlower frequencies than in non-neoplastic thymus; (d) minimalgenomic gains and losses.2,24 To us, these unique featurestaken together argue strongly for retaining the B1 distinction(although in our relatively small series studied so far, AIRE�

and AIRE� B1 thymomas have not shown different autoim-mune susceptibility).

Restricted But Significant Overlap betweenTAMG and APECED

We believe that a key feature common to APECED andthymomas is the absence of AIRE, especially as the T cellsnewly generated in/exported by thymomas clearly do makesubstantial contributions to the peripheral pool.9,10 Hence,one might expect close similarities in the organs and autoan-tigens targeted in these two syndromes. In fact, the similari-ties are very limited.24 Classic APECED manifestations arerare in thymoma patients, although some of the autoantibod-ies/disorders have been noted.12,42 Neither MG and anti-AChR antibodies nor most other neuroimmune conditionshave been reported in APECED; moreover, pure red cellaplasia or Good syndrome (B-cell deficiency and hypogam-maglobulinemia) are virtually unknown in APECED pa-tients.40 Nevertheless, we should remember that �1000APECED patients have been studied in total, so weak asso-ciations cannot yet be excluded.

In striking contrast, almost all APECED patients haveneutralizing antibodies to the IFN�s, just like those inTAMG, and especially IFN-� (Table 1).6,43 Other autoanti-bodies neutralizing Th17-associated cytokines are very prev-alent in APECED, particularly in the patients with CMC(�90%) but also in rare thymoma patients, mostly withCandida infections. Thus, their CMC seems to have anautoimmune basis after all. These autoantibodies apparentlyaffect APC function (anti-IFNs),44 TH1 polarization of naiveT-cells (anti-IL-12 antibodies)45 and defense against CMC (anti-IL-17F and IL-22).23 Nevertheless, life-threatening infectionsare surprisingly uncommon in both APECED and TAMG pa-tients and are largely restricted to a minority of thymomapatients with autoimmune hypogammaglobulinemia.46

The rarity of infections in the other thymoma patients—despite immunosuppressive drug therapy for MG—is likelydue to thymoma development in adulthood, by when TAMGpatients are already well endowed with pathogen-specificmemory T-cells. Similarly, they might, hypothetically, beprotected from APECED-type autoimmune diseases24 by (a)

FIGURE 3. AIRE expression in the normal thymus. a, Ex-pression of AIRE (blue) in the nucleus of one of the pancy-tokeratin-expressing (green; antibodiy AE1/AE3) medullarythymic epithelial cells (mTECs). b, “Promiscuous” expressionof thyroglobulin (green) in one of the AIRE-expressingmTECs (blue nuclei). Immunofluorescence (�400).



robust AIRE-dependent peripheral tolerance established earlyin life47 and (b) AIRE expression in the residual thymus andeven in tolerogenic stromal cells in peripheral lymph nodes.48

Could AIRE Deficiency in Thymomas Play aRole in Paraneoplastic MG?

As described above, the thymopoiesis in the disorga-nized, myoid cell-deficient environment of almost all MG-associated thymomas could explain the preferential autoim-munity against muscle antigens in thymoma patients.12,24 Theabsence of AIRE might contribute to explain this as well.24

The resulting nontolerized T helper cells emigrate fromthymomas as CD4�CD45RA� naive T cells9,10—ready (hy-pothetically) to help autoreactive B cells to respond to AChRand other autoantigens in the periphery. Notably, Strobel etal.7 found a lower—or greatly delayed—risk of MG inpatients whose thymomas generated fewer CD4� T cells, andthey proposed that impaired negative selection plays a keyrole in TAMG.7,49,50 However, thymopoiesis may not beabsolutely necessary as both MG and autoantibodies toAChR, IFN�s and IL-1251 equally occur in rare patients withtype A thymomas that have no apparent thymopoietic activ-ity.8,52 More specific biases in selection, or even activeautoimmunization, may also contribute to the responses tocytokines. Consistently with that, antibody titers increasemarkedly when these tumors recur,53 and there is ongoingproduction of IFN and IL-12 antibodies in the thymomas.21

We therefore believe that in these (and probably other)thymomas: (a) IFN-producing macrophages and DC couldprime recirculating T-cells; indeed, we have noted activatedCD25� mature T helper cells in an MG� type A thymoma8;(b) AIRE-deficiency creates a milieu in AIRE� thymomas—and APECED thymuses—where autoantigen expression is“dangerously” autoimmunogenic; for example, AIRE� DCswere found (by some but not all authors) to have augmentedantigen-presenting capacity37,54; (c) the locally abundantIFN-�s might autoimmunize within both thymomas andAIRE-mutant APECED thymi; (d) the autoantigens availabledetermine the responses evoked; abundantly expressed intactIFNs, which are small single-chain molecules, might immu-nize both local T helper cells and recirculating B-cells inthymomas, eliciting the autoantibody production observedthere against IFN� and IL-12 but not AChR21; (e) theisolated AChR subunits expressed by thymoma TEC can

prime T helper cells only against linear epitopes in situ,and these T cells must subsequently induce B-cell re-sponses in sites where complete, native AChR is available,e.g., in myoid cells in the adjacent non-neoplastic thymus(where lymphoid follicles are common in TAMG) or inmuscle-draining lymph nodes.12,55

We also predict (f) endocrine and ectodermal biases inthe repertoire of TRAgs expressed by AIRE-mutant mTEC inAPECED; (g) that early thymectomy in young patientsshould help to prevent development of later autoimmunemanifestations.

Autoantibodies as Biomarkers for ThymomaAntititin autoantibodies are almost specific and highly

sensitive biomarkers for both MG� and MG� thymomas(sensitivity, 70–90%) and also for LOMG, particularly withonset after 60 years of age.53 By contrast, the anti-IFN�2autoantibodies that are almost diagnostic of APECED aremore prevalent in TAMG (�70%) and non-MG thymomapatients (50%) than in LOMG (20–30%). Notably, the anti-cytokine autoantibodies do not cosegregate with any partic-ular thymoma subtype.51 Therefore, titin antibodies havelimited use in predicting the presence of a thymoma, unlessthe patient is younger than 60, but careful titration of IFN�2and IL-12 antibodies may help in diagnosing thymoma re-currences.53

Defective AIRE Expression—A Hint of EpithelialImmaturity?

The molecular basis underlying the defective AIREexpression in most thymomas is enigmatic. We tested for arange of signaling molecules (related to TNF) that are knownto be essential for inducing Aire expression in mice, includ-ing TRAF6, RelB, RANK, RANKL, LT-�, LT-�R (reviewedin Refs. 56,57). However, none of them was significantlyaltered in the many AIRE� thymomas we tested (Figure 4).Furthermore, even in AIRE-deficient thymomas, T cell mat-uration appeared broadly normal,7,9 unlike in patients withsuch T cell-intrinsic defects as severe combined immunode-ficiencies or Omenn syndrome.58,59 Recently, complete T cellmaturation was shown to depend on NF-kappa B signalinginduced by interactions of CD40 and RANK with theirrespective ligands and that these signals are delivered byauto-antigen-specific CD4� thymocytes to mTECs display-

FIGURE 4. a, Normal expression of AIRE andlymphotoxin-� (LT-�) in the normal thymic me-dulla. b, Apparently normal expression (mRNAand/or protein) of TRAF6, RANK, RANKL, LT-�,and LT-�-receptor in various thymomas comparedwith normal thymus. Beta-actin and GAPDHserved as loading controls. The indicated factorsare known to be required for AIRE expression inthe thymus.56–57



ing cognate autoantigen-MHC-class-II complexes.57 Thus,the observed defects in MHC class II expression/inducibilityin the TEC in most thymomas both in situ and in vitro11,60

could also explain their defective AIRE expression, whichmight alternatively reflect some maturational block earlier inTEC development.

Could AIRE Deficiency Have An OncogenicImpact on the Development of Thymomas?

Aire deficiency predisposes to marginal zone B-celllymphoma in old Aire�/� mice, likely because of overstimu-lation of B-cells by peripheral Aire-deficient DC.61 By con-trast, thymoma development has not been reported either inAire�/� mice or in APECED patients, even though AIREexpression in mTECs,36 and its forced expression in otherepithelial cells,62 supports apoptosis. Although AIRE couldthus formally exert tumor suppressor-like functions in thethymic medulla, there appears to be redundancy in the main-tenance of mTEC homeostasis in the normal adult thymus.Indeed, overexpression of antiapoptotic factors such as bcl-2,63,64 survivin,65 and XIAP66 is characteristic of thymomasand thymic carcinomas. Nevertheless, we speculate that ab-sence of AIRE in the setting of thymoma could contribute tothe combined antiapoptotic mechanisms operating there.

CONCLUSIONS

1. The restriction of AIRE expression to some of the rare“organotypic” type B1 thymomas is a strong biologicargument for their separation from other WHO sub-types.

2. AIRE deficiency in thymomas very rarely leads to thetypical manifestations caused by germ-line AIRE muta-tions in APECED, except for (a) highly prevalent anddiagnostically useful anti-IFN� autoantibodies, and (b)rare APECED-type anti-TH17 cytokine autoantibodiesthat apparently predispose to paraneoplastic candidiasisin rare thymoma patients.

3. Although AIRE deficiency alone is insufficient to causeMG in thymoma patients, it might help to create a“dangerous” milieu in which pre-emigrant thymo-cytes—in the absence of tolerogenic myoid cells—areactively autoimmunized against autoantigens (such asthe AChR �-subunit) that are aberrantly expressed byneoplastic thymic epithelial cells.

4. Deficiency of AIRE—which is normally pro-apoptoticin medullary epithelial cells—might contribute to thespectrum of tumor-promoting antiapoptotic features ofthe neoplastic epithelial cells in most thymomas.

ACKNOWLEDGMENTSSupported by Grant 10–1740 of the Deutsche Kreb-

shilfe (to A.M. and P. St.). The authors gratefully acknowl-edge assistance by Dr. Guiseppe Giaccone in the editorialprocessing of this manuscript.

REFERENCES1. Rosai J. Histological typing of tumours of the thymus, 2nd ed. Berlin and

Heidelberg: Springer-Verlag, 1999.2. Muller-Hermelink HK, Engel PJ, Harris NL, et al. Tumours of the

thymus. In WD Travis, E Brambilla, HK Muller-Hermelink, CC Harris,eds. Tumours of the Lung, Thymus, and Heart Pathology and Genetics.Lyon: IARC Press, 2004.

3. Masaoka A, Monden Y, Nakahara K, et al. Follow-up study of thymo-mas with special reference to their clinical stages. Cancer 1981;48:2485–2492.

4. Margaritora S, Cesario A, Cusumano G, et al. Thirty-five-year follow-upanalysis of clinical and pathologic outcomes of thymoma surgery. AnnThorac Surg 2010;89:245–252; discussion 252.

5. Casey EM, Kiel PJ, Loehrer PJ Sr. Clinical management of thymomapatients. Hematol Oncol Clin North Am 2008;22:457–473.

6. Strobel P, Bauer A, Puppe B, et al. Tumor recurrence and survival inpatients treated for thymomas and thymic squamous cell carcinomas: aretrospective analysis. J Clin Oncol 2004;22:1501–1509.

7. Strobel P, Helmreich M, Menioudakis G, et al. Paraneoplastic myasthe-nia gravis correlates with generation of mature naive CD4� T cells inthymomas. Blood 2002;100:159–166.

8. Nenninger R, Schultz A, Hoffacker V, et al. Abnormal thymocytedevelopment and generation of autoreactive T cells in mixed and corticalthymomas. Lab Invest 1998;78:743–753.

9. Hoffacker V, Schultz A, Tiesinga JJ, et al. Thymomas alter the T-cellsubset composition in the blood: a potential mechanism for thymoma-associated autoimmune disease. Blood 2000;96:3872–3879.

10. Buckley C, Douek D, Newsom-Davis J, et al. Mature, long-lived CD4�and CD8� T cells are generated by the thymoma in myasthenia gravis.Ann Neurol 2001;50:64–72.

11. Strobel P, Chuang WY, Chuvpilo S, et al. Common cellular and diversegenetic basis of thymoma-associated myasthenia gravis: role of MHCclass II and AIRE genes and genetic polymorphisms. Ann N Y Acad Sci2008;1132:143–156.

12. Marx A, Willcox N, Leite MI, et al. Thymoma and paraneoplasticmyasthenia gravis. Autoimmunity. In press.

13. Maclennan CA, Vincent A, Marx A, et al. Preferential expression ofAChR epsilon-subunit in thymomas from patients with myastheniagravis. J Neuroimmunol 2008;201–202:28–32.

14. Romi F, Bo L, Skeie GO, et al. Titin and ryanodine receptor epitopes areexpressed in cortical thymoma along with costimulatory molecules.J Neuroimmunol 2002;128:82–89.

15. Skeie GO, Freiburg A, Kolmerer B, et al. Titin transcripts in thymomas.Ann N Y Acad Sci 1998;841:422–426.

16. Chalk CH, Murray NM, Newsom-Davis J, et al. Response of theLambert-Eaton myasthenic syndrome to treatment of associated small-cell lung carcinoma. Neurology 1990;40:1552–1556.

17. Drachman DB. Myasthenia gravis. N Engl J Med 1994;330:1797–1810.18. Lang B, Vincent A. Autoimmune disorders of the neuromuscular junc-

tion. Curr Opin Pharmacol 2009;9:336–340.19. Roxanis I, Micklem K, McConville J, et al. Thymic myoid cells and

germinal center formation in myasthenia gravis; possible roles in patho-genesis. J Neuroimmunol 2002;125:185–197.

20. Hill ME, Shiono H, Newsom-Davis J, et al. The myasthenia gravisthymus: a rare source of human autoantibody-secreting plasma cells fortesting potential therapeutics. J Neuroimmunol 2008;201–202:50–56.

21. Shiono H, Roxanis I, Zhang W, et al. Scenarios for autoimmunization ofT and B cells in myasthenia gravis. Ann N Y Acad Sci 2003;998:237–256.

22. Willcox N, Leite MI, Kadota Y, et al. Autoimmunizing mechanisms inthymoma and thymus. Ann N Y Acad Sci 2008;1132:163–173.

23. Kisand K, Boe Wolff AS, Podkrajsek KT, et al. Chronic mucocutaneouscandidiasis in APECED or thymoma patients correlates with autoimmu-nity to Th17-associated cytokines. J Exp Med 2010;207:299–308.

24. Strobel P, Murumagi A, Klein R, et al. Deficiency of the autoimmuneregulator AIRE in thymomas is insufficient to elicit autoimmune poly-endocrinopathy syndrome type 1 (APS-1). J Pathol 2007;211:563–571.

25. Alseth EH, Nakkestad HL, Aarseth J, et al. Interleukin-10 promoterpolymorphisms in myasthenia gravis. J Neuroimmunol 2009;210:63–66.

26. Amdahl C, Alseth EH, Gilhus NE, et al. Polygenic disease associationsin thymomatous myasthenia gravis. Arch Neurol 2007;64:1729–1733.

27. Tackenberg B, Schlegel K, Happel M, et al. Expanded TCR Vbetasubsets of CD8(�) T-cells in late-onset myasthenia gravis: novel par-allels with thymoma patients. J Neuroimmunol 2009;216:85–91.

28. Giraud M, Beaurain G, Yamamoto AM, et al. Linkage of HLA to



myasthenia gravis and genetic heterogeneity depending on anti-titinantibodies. Neurology 2001;57:1555–1560.

29. Aarli JA, Romi F, Skeie GO, et al. Myasthenia gravis in individuals over40. Ann N Y Acad Sci 2003;998:424–431.

30. Giraud M, Vandiedonck C, Garchon HJ. Genetic factors in autoimmunemyasthenia gravis. Ann N Y Acad Sci 2008;1132:180–192.

31. Strobel P, Rosenwald A, Beyersdorf N, et al. Selective loss of regulatoryT cells in thymomas. Ann Neurol 2004;56:901–904.

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

33. Abramson J, Giraud M, Benoist C, et al. Aire’s partners in the molecularcontrol of immunological tolerance. Cell 2010;140:123–135.

34. Kyewski B, Peterson P. Aire, master of many trades. Cell 2010;140:24 –26.

35. Peterson P, Org T, Rebane A. Transcriptional regulation by AIRE:molecular mechanisms of central tolerance. Nat Rev Immunol 2008;8:948–957.

36. Gray D, Abramson J, Benoist C, et al. Proliferative arrest and rapidturnover of thymic epithelial cells expressing Aire. J Exp Med 2007;204:2521–2528.

37. Mathis D, Benoist C. Aire. Annu Rev Immunol 2009;27:287–312.38. Kuroda N, Mitani T, Takeda N, et al. Development of autoimmunity

against transcriptionally unrepressed target antigen in the thymus ofAire-deficient mice. J Immunol 2005;174:1862–1870.

39. Nagamine K, Peterson P, Scott HS, et al. Positional cloning of theAPECED gene. Nat Genet 1997;17:393–398.

40. Perheentupa J. Autoimmune polyendocrinopathy-candidiasis-ectoder-mal dystrophy. J Clin Endocrinol Metab 2006;91:2843–2850.

41. Marchevsky AM, Gupta R, McKenna RJ, et al. Evidence-based pathol-ogy and the pathologic evaluation of thymomas: the World HealthOrganization classification can be simplified into only 3 categories otherthan thymic carcinoma. Cancer 2008;112:2780–2788.

42. Cheng MH, Fan U, Grewal N, et al. Acquired autoimmune polyglandu-lar syndrome, thymoma, and an AIRE defect. N Engl J Med 2010;362:764–766.

43. Meager A, Visvalingam K, Peterson P, et al. Anti-interferon autoanti-bodies in autoimmune polyendocrinopathy syndrome type 1. PLoS Med2006;3:e289.

44. Kisand K, Link M, Wolff AS, et al. Interferon autoantibodies associatedwith AIRE-deficiency decrease the expression of IFN-stimulated genes.Blood 2008;112:2657–2666.

45. Zhang W, Liu JL, Meager A, et al. Autoantibodies to IL-12 in myas-thenia gravis patients with thymoma; effects on the IFN-gamma re-sponses of healthy CD4� T cells. J Neuroimmunol 2003;139:102–108.

46. Kelesidis T, Yang O. Good’s syndrome remains a mystery after 55years: a systematic review of the scientific evidence. Clin Immunol2010;135:347–363.

47. Guerau-de-Arellano M, Martinic M, Benoist C, et al. Neonatal tolerancerevisited: a perinatal window for Aire control of autoimmunity. J ExpMed 2009;206:1245–1252.

48. Gardner JM, Devoss JJ, Friedman RS, et al. Deletional tolerance medi-ated by extrathymic Aire-expressing cells. Science 2008;321:843–847.

49. Chuang WY, Strobel P, Gold R, et al. A CTLA4high genotype isassociated with myasthenia gravis in thymoma patients. Ann Neurol2005;58:644–648.

50. Chuang WY, Strobel P, Belharazem D, et al. The PTPN22gain-of-function�1858T(�) genotypes correlate with low IL-2 expression in

thymomas and predispose to myasthenia gravis. Genes Immun 2009;10:667–672.

51. Meager A, Wadhwa M, Dilger P, et al. Anti-cytokine autoantibodies inautoimmunity: preponderance of neutralizing autoantibodies againstinterferon-alpha, interferon-omega and interleukin-12 in patients withthymoma and/or myasthenia gravis. Clin Exp Immunol 2003;132:128–136.

52. Nenninger R, Schultz A, Muller-Hermelink HK, et al. Abnormal T cellmaturation in myasthenia gravis associated thymomas. Verh Dtsch GesPathol 1996;80:256–260.

53. Buckley C, Newsom-Davis J, Willcox N, et al. Do titin and cytokineantibodies in MG patients predict thymoma or thymoma recurrence?Neurology 2001;57:1579–1582.

54. Anderson MS, Venanzi ES, Chen Z, et al. The cellular mechanism ofAire control of T cell tolerance. Immunity 2005;23:227–239.

55. Meager A, Peterson P, Willcox N. Hypothetical review: thymic aberra-tions and type-I interferons; attempts to deduce autoimmunizing mech-anisms from unexpected clues in monogenic and paraneoplastic syn-dromes. Clin Exp Immunol 2008;154:141–151.

56. Derbinski J, Kyewski B. Linking signalling pathways, thymic stromaintegrity and autoimmunity. Trends Immunol 2005;26:503–506.

57. Irla M, Hollander G, Reith W. Control of central self-tolerance inductionby autoreactive CD4� thymocytes. Trends Immunol 2010;31:71–79.

58. Poliani PL, Facchetti F, Ravanini M, et al. Early defects in human T-celldevelopment severely affect distribution and maturation of thymic stro-mal cells: possible implications for the pathophysiology of Omennsyndrome. Blood 2009;114:105–108.

59. Cavadini P, Vermi W, Facchetti F, et al. AIRE deficiency in thymus of2 patients with Omenn syndrome. J Clin Invest 2005;115:728–732.

60. Kadota Y, Okumura M, Miyoshi S, et al. Altered T cell development inhuman thymoma is related to impairment of MHC class II transactivatorexpression induced by interferon-gamma (IFN-gamma). Clin Exp Im-munol 2000;121:59–68.

61. Hassler S, Ramsey C, Karlsson MC, et al. Aire-deficient mice develophematopoetic irregularities and marginal zone B-cell lymphoma. Blood2006;108:1941–1948.

62. Colome N, Collado J, Bech-Serra JJ, et al. Increased apoptosis afterautoimmune regulator expression in epithelial cells revealed by a com-bined quantitative proteomics approach. J Proteome Res 2010;9:2600–2609.

63. Tateyama H, Eimoto T, Tada T, et al. Apoptosis, bcl-2 protein, and Fasantigen in thymic epithelial tumors. Mod Pathol 1997;10:983–991.

64. Pan CC, Chen PC, Wang LS, et al. Expression of apoptosis-relatedmarkers and HER-2/neu in thymic epithelial tumours. Histopathology2003;43:165–172.

65. Hiroshima K, Iyoda A, Toyozaki T, et al. Proliferative activity andapoptosis in thymic epithelial neoplasms. Mod Pathol 2002;15:1326–1332.

66. Wu M, Sun K, Gil J, et al. Immunohistochemical detection of p63 andXIAP in thymic hyperplasia and thymomas. Am J Clin Pathol 2009;131:689–693.

67. Alimohammadi M, Bjorklund P, Hallgren A, et al. Autoimmune poly-endocrine syndrome type 1 and NALP5, a parathyroid autoantigen.N Engl J Med 2008;358:1018–1028.

68. Soderbergh A, Myhre AG, Ekwall O, et al. Prevalence and clinicalassociations of 10 defined autoantibodies in autoimmune polyendocrinesyndrome type I. J Clin Endocrinol Metab 2004;89:557–562.



a. marx the autoimmune regulator aire in thymoma biology jto 2010

Documents

cd8 t cells

thymic carcinomas

themaster autoimmune

thymoma subtype

localextension of thymomas

type c thymomas

thymoma biologyautoimmunity

thymoma patients