The Link between Graves’ Disease and Hashimoto’sThyroiditis: A Role for Regulatory T Cells
Sandra M. McLachlan, Yuji Nagayama, Pavel N. Pichurin, Yumiko Mizutori, Chun-Rong Chen,Alexander Misharin, Holly A. Aliesky, and Basil Rapoport
Autoimmune Disease Unit (S.M.M., P.N.P., Y.M., C.-R.C., A.M., H.A.A., B.R.), Cedars-Sinai Research Institute andUniversity of California, Los Angeles, School of Medicine, Los Angeles, California 90048; and Department of Medical GeneTechnology (Y.N.), Molecular Medicine Unit, Atomic Bomb Disease Institute, Graduate School of Biomedical Sciences,Nagasaki University, Nagasaki 852-8523, Japan
Hyperthyroidism in Graves’ disease is caused by thyroid-stim-ulating autoantibodies to the TSH receptor (TSHR), whereashypothyroidism in Hashimoto’s thyroiditis is associated withthyroid peroxidase and thyroglobulin autoantibodies. Insome Graves’ patients, thyroiditis becomes sufficiently exten-sive to cure the hyperthyroidism with resultant hypothyroid-ism. Factors determining the balance between these two dis-eases, the commonest organ-specific autoimmune diseasesaffecting humans, are unknown. Serendipitous findings intransgenic BALB/c mice, with the human TSHR A-subunit tar-geted to the thyroid, shed light on this relationship. Of threetransgenic lines, two expressed high levels and one expressedlow intrathyroidal A-subunit levels (Hi- and Lo-transgenics,respectively). Transgenics and wild-type littermates were de-pleted of T regulatory cells (Treg) using antibodies to CD25(CD4� T cells) or CD122 (CD8� T cells) before TSHR-adenovi-rus immunization. Regardless of Treg depletion, high-expres-sor transgenics remained tolerant to A-subunit-adenovirus
immunization (no TSHR antibodies and no hyperthyroidism).Tolerance was broken in low-transgenics, although TSHR an-tibody levels were lower than in wild-type littermates and nomice became hyperthyroid. Treg depletion before immuniza-tion did not significantly alter the TSHR antibody response.However, Treg depletion (particularly CD25) induced thyroidlymphocytic infiltrates in Lo-transgenics with transient orpermanent hypothyroidism (low T4, elevated TSH). Neitherthyroid lymphocytic infiltration nor hypothyroidism devel-oped in similarly treated wild-type littermates. Remarkably,lymphocytic infiltration was associated with intermolecularspreading of the TSHR antibody response to other self thyroidantigens, murine thyroid peroxidase and thyroglobulin.These data suggest a role for Treg in the natural progressionof hyperthyroid Graves’ disease to Hashimoto’s thyroiditisand hypothyroidism in humans. (Endocrinology 148:5724–5733, 2007)
GRAVES’ DISEASE, ONE of the most common auto-immune diseases affecting humans, is caused by
autoantibodies that induce thyrotoxicosis by mimickingthe action of TSH and activating the TSH receptor (TSHR).Although such thyroid-stimulating autoantibodies are pa-thognomonic of Graves’ disease, autoantibodies to thyroidperoxidase (TPO) and thyroglobulin (Tg) are also present.The latter two autoantibodies are the classical markers ofHashimoto’s thyroiditis, a condition in which thyroid lym-phocytic infiltration and thyrocyte damage may progressto hypothyroidism. Moreover, many Graves’ patients havemild lymphocytic thyroiditis. In some instances, thyroid-itis in Graves’ disease becomes sufficiently extensive as tocure the hyperthyroidism with resultant hypothyroidism.
The relationship between Graves’ disease and Hashimo-to’s thyroiditis has been debated for decades. Although ini-tially considered to be two separate diseases, the presentview is that they represent the opposite sides of the same
coin, or the two ends of a spectrum. On the other hand,whole-genome scanning studies in humans have revealeddistinct differences between loci linked to, or associated with,these two autoimmune thyroid diseases (for example, Ref. 1).Moreover, animal models of Graves’ disease and Hashimo-to’s thyroiditis are studied as two distinct entities. Not sur-prisingly, the pathophysiological relationship betweenTSHR, TPO, and Tg autoantibodies remains an enigma. Forexample, why do TPO and Tg autoantibodies arise in Graves’disease? Do TSHR, TPO, and Tg autoantibodies arise inde-pendently or through intermolecular spreading, and if thelatter, what is the primary antigen? Is the TSHR the autoan-tigen associated with lymphocytic infiltration in Graves’disease?
Thyroid-stimulating antibodies, the proximal cause ofGraves’ hyperthyroidism, arise from the breakdown in self-tolerance to the TSHR, a G protein-coupled receptor withseven transmembrane-spanning domains (reviewed in Ref.2). However, the autoantigen that drives the immune re-sponse in Graves’ disease is not the full-length receptor butthe A-subunit (3, 4), an ectodomain component that is shedafter intramolecular cleavage of the receptor (reviewed inRef. 2). Recently, we generated transgenic mice with thehuman A-subunit targeted to the thyroid gland (5). Thefounder transgenics were crossed with BALB/c mice, a strainthat is susceptible to immunization with adenovirus express-
First Published Online September 6, 2007Abbreviations: A-subunit-Ad, A-subunit-adenovirus; Con-Ad, con-
trol adenovirus; 5H, pentahistidine; m, murine; TBI, TSH binding to theTSH receptor; Tg, thyroglobulin; TPO, thyroid peroxidase; Treg, regu-latory T cells; TSHR, TSH receptor.Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving theendocrine community.
ing either the TSH holoreceptor or its A-subunit (4, 6). Unlikewild-type littermates, the transgenics failed to develop T cellresponses or TSHR antibodies after low-dose A-subunit ad-enovirus (A-subunit-Ad) immunization. However, toleranceto the human A-subunit was partially overcome by immu-nization with high doses of adenovirus expressing the A-subunit or the holoreceptor (5).
Development of tolerance is a complex process thatincludes central and peripheral mechanisms acting inconcert to eliminate self-reactive lymphocytes (7). T celldeletion by central tolerance may not eliminate all self-reactive cells. Another potent mechanism involves regu-latory T cells (Treg), such as naturally occurring CD25�
CD4� T cells or CD8� CD122� cells, that control autore-active effector T cells in the periphery (8, 9). In the presentstudy, we used A-subunit transgenic animals to probe theinfluence of A-subunit transgene expression levels andTreg depletion on the immune response to TSHR-Ad im-munization. We report that Treg are a major factor in theintermolecular spreading of the immune response fromthe TSHR to TPO and Tg as well as in the shift fromhyperthyroidism to full-blown Hashimoto’s thyroiditiswith massive thyroid lymphocytic infiltration and hypo-thyroidism. These findings provide novel insight into theenigmatic balance between hyperfunction and thyroid de-struction in human Graves’ disease.
Materials and MethodsHuman TSHR A-subunit transgenic mice
Mice with the human TSHR A-subunit targeted to the thyroidusing the bovine Tg promoter were described previously (5).Founders were bred to BALB/cJ (Jackson Laboratories, Bar Harbor,ME) to generate five separate lines. Transgene-positive offspringwere repeatedly crossed to BALB/cJ, and three lines (60.6, 50.6, and51.9) maintained as heterozygotes were studied (F6 –F9 generation).Wild-type littermates were used as controls. Because Tg is a recog-nized abbreviation for the thyroid autoantigen thyroglobulin, werefer to our transgenic mice as A-subunit transgenic (or Tg-ic). Thecopy number of the A-subunit transgene was determined by theMurine Genetic Analysis Laboratory Center for Comparative Med-icine (University of California, Davis, Davis, CA). The low-expressortransgenic line was accepted for archiving and cryopreservationby the Mutant Mouse Regional Resource Center (University of Cal-ifornia, Davis) and is available under the following designation:C.Cg-Tg(TG-TSHR)51.9Smcl, no. 014125.
Human A-subunit protein expression in transgenic thyroids
Intrathyroidal expression of human TSHR A-subunit was previ-ously demonstrated by RT-PCR (5). To examine A-subunit proteinexpression, we used a murine anti-pentahistidine (anti-5H) antibody(QIAGEN, Valencia, CA) to detect the C-terminal 6-histidine (6H) tagencoded by the transgene. Immunohistochemistry was performed bythe Research Animal Diagnostic Laboratory (University of Missouri,Columbia, MO) as follows: Paraffin-embedded sections of murinethyroid tissue were dewaxed and rehydrated followed by heat-in-duced epitope retrieval (steam at 97 C for 30 min in 10 mm citratebuffer, pH 6.0). Sections were cooled and treated with 3% H2O2 (toquench endogenous peroxidase) followed by anti-5H (1:100), biotin-ylated rabbit antimouse IgG, and horseradish peroxidase-streptavi-din (Dako North America, Inc., Carpinteria, CA). Color was devel-oped with 3,3� diaminobenzidine (Dako), and H2O2 and the sectionswere counterstained with Mayer’s hematoxylin.
The concentrations of human A-subunit protein were measured inthyroid extracts prepared by homogenization (three glands per trans-
genic line or wild-type littermates) in buffer containing proteaseinhibitors (Roche Applied Science, Indianapolis, IN) and the 14,000 �g supernatant retained. We modified our ELISA for detecting TSHRantibodies in mouse sera (10) to estimate human A-subunit concen-trations in these extracts. To distinguish the mouse and human A-subunits, anti-5H was used to detect the 6H-tagged transgenic pro-tein. Duplicate supernatant aliquots (1:5 to 1:30) were preincubatedwith an equal volume of biotinylated anti-5H (QIAGEN; 1:1000). Inthe absence of thyroid extract, this anti-5H dilution yielded an OD ofabout 1.00 on ELISA wells coated with A-subunit (1 mg/ml) ex-pressed in eukaryotic cells and purified by affinity chromatography(11). Preincubated test samples were transferred to A-subunit-coatedwells and, after further incubation and washing, subsequently ex-posed to horseradish peroxidase-streptavidin (BD Biosciences, SanJose, CA). Color was developed with o-phenylene diamine and H2O2and stopped with H2SO4, and OD values were read at 490 nm.A-subunit concentrations were expressed as milligrams per thyroidgland estimated from a standard curve generated in the same assayusing non-plate-bound A-subunit (0 –20 mg/ml).
Tg concentrations were measured by a similar inhibition assayusing ELISA wells coated with murine (m)Tg (1 mg/ml) and a poly-clonal mouse antihuman Tg that cross-reacts with mTg (generouslyprovided by Dr. Terry F. Davies, Mount-Sinai Medical Center, NewYork, NY; and Dr. Yaron Tomer, University of Cincinnati, Cincinnati,OH). The standard curve, generated using 0 –20 mg/ml mTg and thepolyclonal anti-Tg at 1:1500, was used to calculate mTg concentra-tions. Total protein concentrations were determined by Bradfordassay (12).
Adenovirus immunization
RGD-Ad encoding either the TSH holoreceptor or the A-subunit (5)were used in this study and are referred to as TSHR-Ad. The sameadenovirus stock was always used in an individual experiment. Con-trol immunizations were performed with adenovirus lacking an in-sert [control adenovirus (Con-Ad)] (13). Viruses were propagated inHEK293 cells and purified by CsCl density gradient centrifugation,and viral particle concentration was determined by absorbance at 260nm (14). A-subunit transgenic mice and wild-type littermates (7–10wk old) were immunized three times at three-weekly intervals withhigh doses of TSHR-Ad or Con-Ad (�1010 particles per injection).Blood was drawn 1 wk after two injections, and animals were eu-thanized 1 month after the third injection to harvest blood and thyroidglands. Animal studies were approved by the Institutional AnimalCare and Use Committee at Cedars-Sinai Medical Center and per-formed with the highest standards of care in a pathogen-free facility.
Depletion of Treg
Rat hybridomas PC61 (anti-CD25) and TM�1(15) (anti-CD122; gen-erously provided by Dr. K. Yui, Nagasaki University, Nagasaki, Japan;and Dr. T. Tanaka, Osaka University, Osaka, Japan, respectively) wereinjected into nude mice to induce ascites. The antibodies were purifiedon HiTrap protein G HP columns (Amersham, Piscataway, NJ) and theirefficacy tested in BALB/c mice (Charles River Japan Laboratory Inc.,Tokyo, Japan). Splenocytes from untreated and injected mice were com-pared by FACScan flow cytometry (CellQuest software; BD Biosciences,Mountain View, CA) using the following antibodies: fluorescein iso-thiocyanate (FITC)-anti-CD4 (H129.19) and phycoerythrin-anti-CD25(7D4) (BD Biosciences, San Jose, CA) or FITC-conjugated-anti-CD122(5H4; eBioscience, San Diego, CA) and phycoerythrin-conjugated anti-CD8 (53–6.7, BD Biosciences). Four days after ip injection of anti-CD25(500 �g PC61), CD25� CD24� T cells were reduced from 8 to 2.1%(supplemental Fig. S1, a and b, published as supplemental data on TheEndocrine Society’s Journals Online web site at http://endo.endojournals.org) in accordance with previous observations (16). Afterinjecting anti-CD122 (TM�1, 250 �g), CD122�, CD8� T cells were re-duced from 17.5 to 2.9% (supplemental Fig. S1, c and d). From these data,Treg depletion was performed by ip injections of 500 �g/mouse anti-CD25 or 250 �g/mouse anti-CD122 4 d before adenovirus immuniza-tion. These studies were conducted according to the principles andprocedures in the Guideline for the Care and Use of Laboratory Animals,Nagasaki University.
McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg Endocrinology, December 2007, 148(12):5724–5733 5725
Serum T4, thyroid histology, and TSH
Total T4 was measured in undiluted mouse serum (25 ml) by RIAusing a kit (Diagnostic Products Corp., Los Angeles, CA). Thyroids werefixed in buffered formaldehyde (pH 7.4) and paraffin-embedded, andserial sections were stained with hematoxylin and eosin (Research An-imal Diagnostic Laboratory, University of Missouri, Columbia, MO).Serial thyroid sections were examined without knowing the immuni-zation employed or the origin (transgenic or wild type) of the tissue.Lymphocytic infiltration was assessed as a percentage of the tissueinvolved. TSH levels in some mice were determined (with a fee forservice) by Dr. Roy Weiss (Thyroid Unit, University of Chicago, Chicago,IL) in undiluted serum (50 �l) by RIA (17).
TSHR antibodies
TSHR antibody levels were measured by inhibition of TSH bindingto the TSHR (abbreviated TBI) using a commercial kit according to themanufacturer’s protocol (Kronus, Boise, ID). In brief, aliquots (25 �l) ofmouse serum were incubated with detergent-solubilized TSHR; 125I-TSH was added and the TSHR-antibody complexes were precipitatedwith polyethylene glycol. TBI values were calculated from the formula:[1 � (TSH binding in test serum � nonspecific binding)/(TSH bindingin control serum � nonspecific binding)] � 100.
Autoantibodies to mTg
Autoantibody binding to mTg was measured using ELISA wellscoated with mTg (1 �g/ml); the positive control was a cross-reactingpolyclonal mouse antihuman Tg (mTg and antibody from Drs. Daviesand Tomer; see above). Test sera were diluted 1:100, antibody bindingwas detected with horseradish peroxidase-conjugated mouse anti-IgG (Sigma-Aldrich, St. Louis, MO), and the signal was developedwith o-phenylenediamine and H2O2. The data are expressed as OD490 nm.
Autoantibodies to mTPO
The cDNA for mTPO (18) (provided by Dr. S. Ohtaki, Japan) wastransferred to the vector pHMCMV6 (19). Expression of the mTPO-pHMCV6 plasmid was tested by transiently transfecting COS-7 cellsusing Fugene HD (Roche) and antihuman TPO induced in mice usingadenovirus (20) that cross-reacts with mouse TPO. Antibody bindingwas detected using FITC-conjugated affinity-purified goat antimouse
IgG (Caltag Laboratories, Burlingame, CA) and analysis by flowcytometry. Cells stained with propidium iodide (1 �g/ml) were ex-cluded from analysis. Sera (diluted 1:50) from immunized transgenicsand wild-type littermates were tested for mTPO binding in the sameway. Data are expressed as percent positive cells in the gated fraction(M2).
Statistical analyses
The statistical significance of differences between the magnitude ofresponses in multiple groups was determined by ANOVA and testingbetween two groups by Mann Whitney rank sum test or, when normallydistributed, by Student’s t test.
ResultsThyroids from transgenic mice express different levels ofhuman TSHR A-subunit
At the time of previous studies on human TSHR A-sub-unit transgenic mice bred from five founders (5), we hadno information on the level of A-subunit expression. Re-taining three of these transgenic lines, we have now de-termined by immunohistochemistry that thyrocytes fromtransgenic lines 50.6 and 60.6 express high levels of humanTSHR A-subunits, with extensive leakage (or secretion)into the follicular lumens (Fig. 1A, top panels). Althoughthe TSHR holoreceptor is expressed on the basal surface ofthe thyrocyte, the soluble A-subunit is a secreted protein(21). Thyroids from 51.9 transgenics have a much lowerA-subunit expression, similar to that in wild-type mice(Fig. 1A, bottom panels). In thyroid extracts, the highestA-subunit protein level (measured by ELISA inhibition)was in transgenic line 50.6, followed by line 60.6, withminimal or undetectable levels in 51.9 and wild-type mice(Fig. 1B). In contrast, as controls, thyroid concentrations ofmTg and total protein were similar in all three transgeniclines and wild-type littermates. The differences in humanA-subunit expression among transgenic lines were unre-
1.0
1
01
001
0001
µg /
thyr
oid
namuHA-su tinub
torp latoT nieesuoMyht r nilubolgo
50.6
60.6
51.9 WT
50.6
60.6
51.9 WT
50.6
60.6
51.9 WT
B
rosserpxe-oL )9.15(
rosserpxe-iH )6.06(rosserpxe-iH )6.05(
TW
A
FIG. 1. Human A-subunit expression in thyroid tissue from transgenic mice. A, Immunohistochemistry using an anti-5H antibody to detectthe 6H tag on the transgenically expressed human A-subunit. Top, Hi-expressor transgenic lines 50.6 (left) and 60.6 (right); bottom,Lo-expressor line 51.9 (left) and wild-type (WT) littermate (right). Magnification, �100. B, Concentrations of human A-subunit, mouseTg and total protein in thyroid extracts from transgenic lines 50.6, 60.6, and 51.9 and wild-type (WT) littermates. Data are shown asmicrograms per thyroid gland (mean � SD, n � 2). Concentrations of human A-subunit and mouse Tg were estimated by inhibition ELISAand total protein by Bradford assay.
5726 Endocrinology, December 2007, 148(12):5724–5733 McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg
lated to transgene copy number, with two copies presentin 50.6 transgenics and five copies in the 51.9 and 60.6transgenics. In subsequent studies, we focused on humanTSHR A-subunit transgenic lines 50.6 and 51.9, termedhigh (Hi) and low (Lo) expressors, respectively.
Breaking tolerance in A-subunit transgenic mice
In our previous studies on A-subunit transgenic mice be-fore categorization into Hi- and Lo-expressors, tolerance waspartially broken by immunizing with high doses of TSHR-Ad(A-subunit or TSH holoreceptor) (5). We have now ad-dressed the question of breaking tolerance in Hi vs. Lo A-subunit expressor transgenics. In addition, we tested theoutcome of Treg depletion using anti-CD25 or anti-CD122before each immunization with high-dose TSHR-Ad (A-sub-unit or holoreceptor) (Fig. 2A). As expected, wild-type lit-termates had a robust TBI response 1 wk after the secondinjection and 1 month after the third injection (Fig. 2B). Incontrast, Hi-expressor transgenics remained tolerant at both
time intervals, even after Treg depletion with anti-CD25 (Fig.2C). Unlike the Hi-expressor transgenics, TSHR-Ad immu-nization of Lo-expressors did induce TBI activity; Treg de-pletion using anti-CD25 or anti-CD122 did not alter the TSHRantibody response (Fig. 2D). Although some responsestended to be higher in some CD25-treated compared withuntreated transgenics, the differences were not significantlydifferent.
Incidentally, categorization into A-subunit Hi- and Lo-expressor lines permitted reanalysis of previously reporteddata (5). Consistent with the present findings, TBI activitywas induced in Lo- but not Hi-expressor mice with high-doseTSHR-Ad immunization (supplemental Fig. S2).
Thyroid function and pathology in Lo and Hi A-subunitexpressor transgenics
TSHR antibodies detected in the TBI assay may also havethyroid stimulatory activity with resultant thyrotoxicosis.We, therefore, assessed thyroid function by determining se-
oC n T RHS oC n RHST- - α 52DC α 221DC - - α 52DC α 221DC
nedA urivo s:An ydobit :
TBI (
% In
hibi
tion)
0
02
04
06
08
001
jnI 2x 3xyt-dliW peBA
DC
TBI (
% In
hibi
tion)
oC n- -
RHSTα 52DC α 221DC
oC n- -
RHSTα 52DC α 221DC
nedA urivo s:
0
02
04
06
08
001jnI 2x 3xrosserpxe-oL
oC n T RHSoC n T RHSAn ydobit : α 52DC-- α 52DC --
TBI (
% In
hibi
tion)
0
02
04
06
08
001jnI 2x 3xrosserpxe-iH
deelB liaT tuE h aisana
dA-RHSTdA-RHST
ro enoNα ro 52DCα 221DC
ro enoNα ro 52DCα 221DC
dA-RHST
skeeW64 5321 10 0987
ro enoNα ro 52DCα 221DC
nedA urivo s:An ydobit :
FIG. 2. A, Overview of antibody administration to deplete Treg before immunization with TSHR-Ad, obtaining blood 1 wk after the secondimmunization and euthanasia 1 month after the third immunization. B–D, Induction of TSHR antibodies in A-subunit Hi- or Lo-expressortransgenic mice depleted of Treg before immunization with high-dose TSHR-Ad. Mice were untreated (�) or injected with anti-CD25 (�CD25)or �CD122 to deplete Treg, and 4 d later immunized with TSHR-Ad (1010 particles per injection). Some mice received Con-Ad (Con). Sera weretested 1 wk after two adenovirus injections (Inj 2x) and 1 month after the third immunization (3x). TSHR antibodies were measured as percentageinhibition of TSH binding to the TSHR (TBI). Values for individual transgenics are shown: E, wild-type mice (B); F, Hi-expressor transgenics(C); speckled circles, Lo-expressor transgenics (D). The shaded area represents the mean � 2 SD TBI levels for Con-Ad-immunized wild-typemice.
McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg Endocrinology, December 2007, 148(12):5724–5733 5727
rum T4 levels in the mice immunized with TSHR-Ad. Asexpected in wild-type littermates, high-dose TSHR-Ad in-duced elevated serum T4 in about one third of mice (Fig. 3A).Consistent with the absence of TSHR antibodies (Fig. 2), allHi-expressor transgenics remained euthyroid, even afterTreg depletion with anti-CD25 (Fig. 3B). Despite detectableTBI activity, no Lo-expressor transgenics developed hyper-thyroidism. Surprisingly, 1 wk after the second immuniza-tion, serum T4 levels were subnormal in all Lo-expressorspretreated with anti-CD25 (Fig. 3C). Some recovery in T4levels was evident at euthanasia, 4 wk after the third im-munization. However, four of seven mice pretreated withanti-CD25 had extremely low or undetectable serum T4 levels(Fig. 3C). Because of these unusual findings, we measuredserum TSH levels in the Lo-expressors and wild-type litter-mates for which sufficient serum was available at euthanasia.TSH was markedly elevated in both mice with very low T4levels (Fig. 3D, arrows), confirming their hypothyroid status.
Thyroid lymphocytic infiltration in A-subunit transgenicswith Treg depletion
Thyroid pathology provided the explanation for thehypothyroidism described above. In scores of wild-type
mice of different strains (including BALB/c) immunizedwith TSH holoreceptor or A-subunit-Ad, neither we norothers ever observed significant thyroid lymphocytic in-filtration (4, 6, 22, 23). Consistent with these observations,wild-type littermates and euthyroid Lo-expressor trans-genics immunized with TSHR-Ad had normal thyroid his-tology (Fig. 4, A and B). Hyperthyroid wild-type micedeveloped thyroid hyperplasia without infiltrating lym-phocytes (Fig. 4C), as observed by ourselves and others (4,6, 22, 23). Some Lo-expressor transgenics pretreated withanti-CD122 before TSHR-Ad immunization had modestlymphocytic infiltrates (Fig. 4D). Most striking, however,thyroids of hypothyroid Lo-expressor transgenics Tregdepleted with anti-CD25 had massive lymphocyte infil-trates, encompassing much of the thyroid (Fig. 4, E and F).These findings are consistent with a shift from Graves’disease to Hashimoto’s thyroiditis.
The extent of thyroid lymphocytic infiltration in both ex-periments was quantified as the percentage of the thyroidarea invaded by lymphocytes. All Lo-expressor transgenicsdepleted of Treg with anti-CD25 before TSHR-Ad immuni-zation had thyroid lymphocytic infiltration. In this group,extensive infiltrates, encompassing more than 50% of the
RHST
oC n oC n RHSTnedA urivo s:An ydobit :
RHSTα 52DC
0
2
4
6
8
01
21
41
- α 52DC- --
B
T4 (µ
g/dL
)jnI 2x 3x
D
1
01
001
0001
TSH
(mU
/L)
jnI 3x
oC n
- - α 52DC α 221DC
dliW yt- pecigT oL-
nedA urivo s:An ydobit :
A
-- α 52DC α 221DC - - α 52DC α 221DCoC n T RHS oC n RHSTnedA urivo s:
An ydobit :
T4 (µ
g/dL
)
0
2
4
6
8
01
21
41jnI 2x 3x
-
C
- --oC n T RHS Con RHSTnedA urivo s:
An ydobit : α 52DC α 221DC α 52DC α 221DC0
2
4
6
8
01
21
41jnI 2x 3x
T4 (µ
g/dL
)
yt-dliW pe
oL-cigT
iH-cigT
FIG. 3. Serum T4 and TSH in A-subunit transgenics depleted of Treg before immunization with high doses of TSHR-Ad. As shown in Fig. 2A,mice were untreated (�) or injected with anti-CD25 (�CD25) or anti-CD122 (�CD122) 4 d before each TSHR-Ad immunization. Sera were tested1 wk after two adenovirus injections (Inj 2x) and 1 month after the third immunization (3x). Some mice were injected with Con-Ad (Con). Dataare shown for individual animals. A–C, T4 values (�g/dl) in wild-type littermates (A), Hi-expressor transgenics (Tgic-Hi) (B), and Lo-expressortransgenics (Tgic-Lo) (C); D, TSH values (mU/liter) in wild-type and Lo-expressor transgenics. Serum TSH levels were determined whensufficient serum was available (at euthanasia). Arrows indicate elevated TSH values in hypothyroid mice. For each panel, the shaded arearepresents the mean � 2 SD T4 levels for Con-Ad-immunized wild-type littermates.
5728 Endocrinology, December 2007, 148(12):5724–5733 McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg
thyroid, were observed in four mice (Fig. 5, right), all ofwhom were hypothyroid at the time of euthanasia (Fig. 3C).Smaller infiltrates (less than 20% of the gland) were presentin euthyroid animals, including some transgenics pretreatedwith anti-CD122, as well as in one wild-type mouse depletedof Treg using anti-CD25.
Intermolecular spreading of autoantibodies to otherthyroid autoantigens
The appearance of extensive lymphocytic infiltrationand hypothyroidism suggested a shift in this animal modelof Graves’ disease toward Hashimoto’s thyroiditis. Im-munological hallmarks of Hashimoto’s thyroiditis are au-toantibodies to TPO and Tg. Indeed, in association withthyroid lymphocytic infiltration, autoantibodies to murineTg were present in all anti-CD25-treated Lo-expressortransgenics but in no other groups of transgenics or wild-type mice (Fig. 6A). Similarly, autoantibodies to mTPOwere positive in four of six anti-CD25-treated transgenics(Fig. 6B).
Discussion
Because of human diversity as well as for obvious ethicallimitations, syngeneic animal models of autoimmune dis-
eases are invaluable investigative tools. Previously, we de-rived transgenic mice with the human TSHR A-subunit tar-geted to the thyroid gland. We studied animals derived fromfive transgenic founders, all bred to BALB/c, a genetic back-ground susceptible to Graves’ hyperthyroidism induced byTSH holoreceptor (6) or A-subunit-Ad (4, 22, 23). Unlike theirwild-type littermates, A-subunit transgenic mice were resis-tant to immunization with low-dose A-subunit-Ad, althoughhigh-dose A-subunit or holoreceptor adenovirus immuniza-tion elicited low-level immune responses (5). In the presentreport, using immunohistochemistry and analysis of thyroidextracts, we categorized the extent of intrathyroidal A-sub-unit expression in progeny from the five founders and havestudied one Lo-expressor and two Hi-expressor transgeniclines. The data obtained provide novel insight into toleranceand the role of Treg in the pathogenesis of thyroid autoim-mune disease.
Immunization of the transgenic lines with TSHR-Adindicated that Hi-expressor mice had a suppressed or ab-sent immune response compared with Lo-expressors.These data are consistent with the Hi-expressors having agreater degree of central tolerance, a process in whichself-reactive T cells, which bind with high affinity to pep-tides from self-antigens expressed in the thymus, aredeleted. Intrathymic expression of the TSHR has beenreported for humans (24, 25) and rats (26), but A-subunitmRNA was undetectable in thymic tissue from Hi-expres-sor mice (not shown). Although some self-antigens, forexample Tg (27), can be studied in thymic tissue (predom-inantly nonexpressing thymocytes), others are studied inthymic medullary epithelial cells (28). We cannot, there-fore, exclude a role for peripheral tolerance in the A-sub-unit transgenics. However, as in mice transgenic for hen
0
02
04
06
08
Thyr
oid
Infil
trat
ion
(%)
oitcejnI 3 ns
yt-dliW pe oL-cigT
nedA urivo s:
An ydobit :
oC n
- -RHST
α 52DC α 221DC -RHSTA
α 52DC α 221DC-oC n
FIG. 5. Thyroid lymphocytic infiltration. Data are shown for Lo-ex-pressor transgenics (Tgic-Lo) and wild-type mice whose T4 levels areshown in Fig. 3. Mice were untreated (�) or pretreated with anti-CD25 (�CD25) or anti-CD122 (�CD122) before immunization withTSHR-Ad or Con-Ad (Con) (Fig. 2A). Thyroid tissue was obtained ateuthanasia 4 wk after the third immunization. The extent (percent)of the thyroid infiltrated with lymphocytes was estimated withoutknowledge of the type of mouse or its treatment regimen. Values forthe four transgenic mice with hypothyroidism at the time of eutha-nasia (Fig. 3) are circled.
dA-noC :TW dA-RHST :cigT-oL
:cigT-oL αα T ,52DC S dA-RH :cigT-oL α T ,52DC S dA-RH
:cigT-oL α dA-RHST ,221DCdA-RHST :TW
A
FE
DC
B
FIG. 4. Thyroid pathology in wild-type (WT) and A-subunit trans-genics after immunization with TSHR-Ad or Con-Ad. A and B, Normalthyroid histology in Con-Ad-immunized WT mice (A) and TSHR-Ad-immunized Lo-expressor transgenic (Lo-Tgic) with intact Treg (B); C,thyroid hyperplasia in a wild-type mouse that developed hyperthy-roidism after TSHR-Ad immunization; D–F, thyroid inflammation inLo-expressor transgenics pretreated with �CD122 (D) or �CD25 (Eand F) before TSHR-Ad immunization. Magnification, �100 (A–E)and �40 (F).
McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg Endocrinology, December 2007, 148(12):5724–5733 5729
egg lysozyme (29), it is likely that high peripheral expres-sion of the human A-subunit correlates with increasedcentral tolerance (and low peripheral expression with de-creased tolerance) to the TSHR. It has recently been rec-ognized that thymic expression of several tissue-restrictedantigens varies considerably between individuals (28).Therefore, in humans, variability in intrathymic expres-sion of TSHR could play a role in the breakdown of tol-erance in Graves’ disease.
Because Treg, such as naturally occurring CD25� CD4�
T cells and CD122� CD8� T cells could also contribute totolerance to TSHR immunization in the A-subunit trans-genic animals, we probed the influence of Treg depletionon the immune response to TSHR-Ad immunization. Im-munization with a high dose of TSHR-Ad alone or pre-ceded by anti-CD25 Treg depletion failed to induce TSHRantibodies in Hi-expressor transgenics. In contrast, Lo-expressors developed TSHR antibodies regardless of Tregdepletion. Therefore, breaking tolerance is inversely re-lated to the extent of A-subunit transgene expressed in the
thyroid, and Treg do not appear to play an important rolein this process.
The above studies on tolerance led to serendipitousfindings providing novel insight into the role of Treg in theprogression of Graves’ hyperthyroidism to severe auto-immune thyroiditis and hypothyroidism. A clinical rela-tionship between these two diseases has long been recog-nized. For example, without surgical or radioiodinethyroid ablation, the long-term natural course of hyper-thyroid Graves’ disease is not uncommonly hypothyroid-ism (30). As in Hashimoto’s thyroiditis, diffuse lympho-cytic infiltration is present in Graves’ thyroids, althoughtypically much less extensive and without thyroid follicledestruction. Although considerable progress has beenmade in understanding the pathogenesis of Graves’ dis-ease and Hashimoto’s thyroiditis, the mechanism under-lying a shift in the balance between these two phenotypesremains enigmatic.
In experimental animals, Graves’ disease and Hashimo-to’s thyroiditis are generally studied as separate diseases,the former induced by immunization with the TSHR, thelatter with Tg or, less commonly, TPO. Unlike in humandisease, Graves’ hyperthyroidism induced in mice haspreviously not been associated with significant thyroidlymphocytic infiltration (4, 6, 22, 23) or the appearance ofTPO and Tg autoantibodies. With minimal or no lympho-cytic infiltration, it is self-evident that there are no reportsof progression of hyperthyroidism to hypothyroidism inthese mice. Lymphocytic thyroiditis is readily inducedusing adjuvant combined with self protein, murine Tg (31)or mTPO (32). Despite extensive lymphocytic infiltration,in most studies, the animals remain euthyroid. Indeed,transgenic mice with thyroid-restricted expression of thechemokine CCL21 develop massive lymphocytic (B and Tcell) thyroid infiltration without thyroid autoantibodies,and thyroid function remains unaffected (33). TargetingTPO by immunizing with a particular mTPO peptide (34)or by expressing a pathogenic TPO-specific T cell receptorin transgenic mice lacking normal T cells or B cells (Ragknockout) (35) can cause thyroiditis and hypothyroidism.However, until the present report, no animal model hasincluded all the pathological features of human autoim-mune thyroid disease, namely hyperthyroidism, massivelymphocytic infiltration leading to hypothyroidism, andautoantibody spreading from the TSHR to TPO and Tg, theother major thyroid-specific autoantigens.
The present study suggests that although Treg do notappear to be responsible for tolerance in our A-subunittransgenic mice, they (particularly CD25� Treg) are thepathogenetic key to the shift in the balance from Graves’hyperthyroidism to Hashimoto’s hypothyroidism. Previ-ous studies have demonstrated a role for CD4�CD25�
Treg depletion in experimentally induced thyroiditis. Tgimmunization of BALB/c mice, normally resistant to in-duction of thyroiditis, develop thyroiditis in conjunctionwith CD25 Treg depletion (36, 37). Mild thyroiditis wasinduced by CD25� T cell depletion before immunizationwith TSHR-expressing adenovirus in thyroiditis-suscep-tible C57BL/6 (but not BALB/c) mice (16). Part of thisstrain difference may involve the genetically controlled
yt-dliW pe oL-cigTA
mTg
ELIS
A (O
D49
0 nm
)
0
4.0
8.0
2.1
6.1 *oitcejnI 3 ns
- -- α 52DC
eniruMyhT nilubolgor
an eidobit s
nedA urivo s:An ydobit :
oC n /bus-A RHSTα 52DC α 221DC
/bus-A RHSTα 221DC-
oC n
0
01
02
03
04
mTP
OPo
sitiv
e C
ells
(%) *05 oitcejnI 3 ns
yt-dliW pe oL-cigTB
eniruMOPT
an eidobit s
nedA urivo s:An ydobit : α 52DC α 221DC
oC n /bus-A RHST /bus-A RHSTα 221DC-
oC nα 52DC- --
FIG. 6. Intermolecular autoantibody spreading from the TSHR toother thyroid autoantigens. Wild-type littermates and Lo-expressortransgenics (Tgic-Lo) were pretreated with anti-CD25 (�CD25) oranti-CD122 (�CD122) before immunization with TSHR-Ad or Con-Ad(Con) (Fig. 2A). Autoantibody spreading to mTg (A) and mTPO (B)occurs in Tgic-Lo animals Treg depleted with �CD25 before immu-nizations with A-subunit-Ad (A-sub) or TSHR-Ad. These groups ofanimals also had the greatest degree of thyroid lymphocytic infiltra-tion (Fig. 5). Autoantibodies to mTPO were assayed by flow cytometryusing mTPO-expressing COS-7 cells. Data are expressed as the per-cent positive gated cells. Murine Tg autoantibodies were detected byELISA. The dashed horizontal line represents the mean � 2 SD forcontrol-Ad-immunized mice. *, Values significantly greater than forother transgenic groups (ANOVA, P � 0.05).
5730 Endocrinology, December 2007, 148(12):5724–5733 McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg
thymic development of Treg (CD4�CD25�FoxP3�), withhigher percentages in BALB/c than C57BL/6 mice (38). Inthyroiditis-susceptible NOD mice expressing a humanHLA-DR3 transgene, anti-CD25 treatment enhances io-dide-induced thyroiditis (39). However, it should be em-phasized that none of these studies involving Treg deple-tion led to severe thyroiditis or to hypothyroidism. Incontrast, after CD25 Treg depletion, all Lo-expressor micestudied developed hypothyroidism after the second im-munization (Fig. 3C). Moreover, all four mice whose hy-pothyroidism persisted 1 month after the final immuni-zation had extensive thyroiditis encompassing 60 – 80% ofthe gland.
Besides inducing extensive thyroid lymphocytic infil-tration and hypothyroidism in Lo-expressor A-subunittransgenics, immunization had another unexpected resultthat mimics human thyroid autoimmune disease, namelyspreading of the humoral autoantibody response from theTSHR to other thyroid-specific autoantigens. We providea hypothesis to explain these intriguing findings (Fig. 7).Unimmunized wild-type and transgenic mice are tolerantto the TSHR, TPO, and Tg (all self murine proteins) andtransgenics are tolerant to the human A-subunit. Afterimmunization with human TSHR-Ad, wild-type andtransgenic mice develop T cells that recognize humanTSHR peptides. Presumably, cross-reactivity to mouseTSHR peptides is limited or absent. The immune responseis also restrained by Treg. Therefore, lymphocytes rarelyinfiltrate the thyroid in mice of this genetic background(BALB/c). However, after Treg depletion (particularlyCD25� T cells), restraint on the immune response is di-minished. In transgenics (but not wild-type mice), lym-phocytes home to the thyroid, the source of their targetpeptides, the human A-subunit. The infiltrating lympho-cytes (likely including cytotoxic cells and generating cy-tokines) cause thyrocyte damage that can lead to overthypothyroidism. Release of thyroid antigens in this in-flammatory milieu breaks tolerance to other thyroid an-
tigens. Thus, in the Lo-expressor A-subunit transgenics,CD25 (but not CD122) Treg depletion was accompanied bystrong autoantibody responses to mTg and mTPO. Previ-ously, very low-level mTg autoantibodies were observedin a single human TSHR-DNA-vaccinated DR3� NODmouse (40), a background susceptible to developing spon-taneous thyroiditis and Tg autoantibodies. Consequently,our findings provide the first unequivocal evidencefor intermolecular autoantibody spreading in thyroidautoimmunity.
Incidentally, it should be noted that hyperthyroidismdeveloped only in wild-type littermates immunized withTSHR-Ad and not in Lo-expressor transgenics. The rela-tively low proportion of hyperthyroid wild-type mice isanticipated; to break tolerance in A-subunit transgenicanimals, immunizations required high-dose adenovirusimmunizations, which generate fewer hyperthyroid ani-mals than low-dose immunizations (41). Extensive thy-roiditis explains why Lo-expressor transgenics depleted ofCD25 Treg became hypothyroid. However, it is less clearwhy Lo-expressor transgenics that were not Treg depletedremained euthyroid. It is possible that sufficient A-subunitprotein is generated in the thyroid to neutralize thyroid-stimulating antibodies.
What evidence is available of a role for Treg in humanthyroid autoimmunity? In humans, the number and func-tion of Treg are still unclear, depending on the Treg mark-ers and assays employed as well as the disease. AbundantTreg were found infiltrating the thyroid gland of Graves’patients in one study, but the suppressor function of pe-ripheral Treg was decreased (42). In another study, intra-thyroidal Treg were reduced compared with those in pe-ripheral blood, possibly because of increased apoptosis(43). Despite the limited number of studies (and in somecases the limited number of patients investigated), thesedata are consistent with our findings for the associationbetween thyroiditis and Treg in mice and (incidentally)with the early studies of Volpe and Iitaka (44) concerninga suppressor T cell defect. Future studies on the evolutionof Graves’ disease into Hashimoto’s thyroiditis with hy-pothyroidism will be of interest to determine whether thisshift is accompanied by an alteration in Treg number orfunction.
In summary, the present study provides the first de-scription of a complete animal model of autoimmune thy-roid disease mimicking all the clinical and pathologicalfeatures of human disease. Moreover, it describes an im-munological mechanism whereby induction of thyroid-stimulating antibodies and Graves’ hyperthyroidism canbe diverted to spreading of the immune response to en-dogenous thyroid autoantigens (mTPO and mTg) withextensive lymphocytic infiltration and hypothyroidism.Our data, taken together with other studies on experi-mental thyroiditis, support the clinical observation thatHashimoto’s thyroiditis in humans is only rarely associ-ated with TSHR autoantibodies, the antithesis of Graves’disease in which autoantibodies to Tg and TPO are com-monly present. This finding has two important clinicalimplications. First, in Graves’ disease, TPO and Tg auto-antibodies are secondary to the immune response to the
amuh nA-su tinub
Wild t- ype Tra sn cineg ( L )o
gerT tneserp
yhT r noitartlifni dio
esnopser llec T RHSTh + RHSTh +
ger T d detelpe
enoN enoN
yhT r noitartlifni dioan /d ro ad am eg enoN evisnetxE
tnA i eg r sn e ael des enoN gTm a OPTm dn
nareloT ot tyht 'fles' r dio
na git ens
OPTm
gTm
RHSTm
OPTm
gTm
RHSTm
esnopser llec T RHSTh +++ RHSTh +++
dezinumminU
-RHSThsurivonedAnoitazinummi
FIG. 7. Schematic representation of the hypothesis whereby Tregdepletion in TSHR-immunized Lo-expressor transgenics leads to hy-pothyroidism and spreading of the immune response to other thyroidself antigens. Additional details of this hypothesis are described in thetext (Discussion). h, Human.
McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg Endocrinology, December 2007, 148(12):5724–5733 5731
TSHR. Second, lymphocytic infiltration in Graves’ diseaseis likely to reflect spreading of the immune response toTPO and to Tg.
Acknowledgments
We thank our colleagues for generously providing us with re-agents: Dr. K. Yui (Nagasaki University, Nagasaki, Japan) for anti-CD25 hybridoma PC61, Dr. T. Tanaka (Osaka University, Osaka,Japan) for anti-CD122 hybridoma TM�1, Dr. T. F. Davies (Mount-Sinai Medical Center, New York, NY) and Dr. Y. Tomer (Universityof Cincinnati, Cincinnati, OH) for mouse Tg and mouse anti-Tg, andDr. S. Ohtakhi (Miyazaki Medical College, Miyazaki, Japan) formouse TPO cDNA. We are also grateful for contributions by Dr. BorisCatz, Los Angeles.
Received July 26, 2007. Accepted August 24, 2007.Address all correspondence and requests for reprints to: Sandra M.
McLachlan, Cedars-Sinai Medical Center, 8700 Beverly Boulevard,Suite B-131, Los Angeles, California 90048. E-mail: [email protected].
This work was supported by National Institutes of Health GrantsDK54684 (S.M.M.) and DK19289 (B.R.) and a Winnick Family ClinicalResearch Scholar Award (S.M.M).
Disclosure Summary: The authors have nothing to disclose.
References
1. Tomer Y, Ban Y, Concepcion E, Barbesino G, Villanueva R, Greenberg DA,Davies TF 2003 Common and unique susceptibility loci in Graves and Hashi-moto diseases: results of whole-genome screening in a data set of 102 multiplexfamilies. Am J Hum Genet 73:736–747
2. Rapoport B, Chazenbalk GD, Jaume JC, McLachlan SM 1998 The thyrotropinreceptor: Interaction with thyrotropin and autoantibodies. Endocr Rev 19:673–716
3. Chazenbalk GD, Pichurin P, Chen CR, Latrofa F, Johnstone AP, McLachlanSM, Rapoport B 2002 Thyroid-stimulating autoantibodies in Graves diseasepreferentially recognize the free A subunit, not the thyrotropin holoreceptor.J Clin Invest 110:209–217
4. Chen C-R, Pichurin P, Nagayama Y, Latrofa F, Rapoport B, McLachlan SM2003 The thyrotropin receptor autoantigen in Graves’ disease is the culprit aswell as the victim. J Clin Invest 111:1897–1904
5. Pichurin PN, Chen C-R, Chazenbalk GD, Aliesky H, Pham N, Rapoport B,McLachlan SM 2006 Targeted expression of the human thyrotropin receptorA-subunit to the mouse thyroid: insight into overcoming the lack of responseto A-subunit adenovirus immunization. J Immunol 176:668–676
6. Nagayama Y, Kita-Furuyama M, Ando T, Nakao K, Mizuguchi H, HayakawaT, Eguchi K, Niwa M 2002 A novel murine model of Graves’ hyperthyroidismwith intramuscular injection of adenovirus expressing the thyrotropin recep-tor. J Immunol 168:2789–2794
7. Kappler JW, Roehm N, Marrack P 1987 T cell tolerance by clonal eliminationin the thymus. Cell 49:273–280
8. Miyara M, Sakaguchi S 2007 Natural regulatory T cells: mechanisms of sup-pression. Trends Mol Med 13:108–116
9. Rifa’I M, Kawamoto Y, Nakashima I, Suzuki H 2004 Essential roles ofCD8�CD122� regulatory T cells in the maintenance of T cell homeostasis. J ExpMed 200:1123–1134
10. Pichurin P, Pichurina O, Chazenbalk GD, Paras C, Chen CR, Rapoport B,McLachlan SM 2002 Immune deviation away from Th1 in interferon-� knock-out mice does not enhance TSH receptor antibody production after naked DNAvaccination. Endocrinology 143:1182–1189
11. Chazenbalk GD, Wang Y, Guo J, Hutchison JS, Segal D, Jaume JC,McLachlan SM, Rapoport B 1999 A mouse monoclonal antibody to a thyro-tropin receptor ectodomain variant provides insight into the exquisite anti-genic conformational requirement, epitopes and in vivo concentration of hu-man autoantibodies. J Clin Endocrinol Metab 84:702–710
12. Bradford MM 1976 A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dye binding.Anal Biochem 72:248–254
13. Chen CR, Aliesky HA, Guo J, Rapoport B, McLachlan SM 2006 Blockade ofcostimulation between T cells and antigen-presenting cells: an approach tosuppress murine Graves’ disease induced using thyrotropin receptor-express-ing adenovirus. Thyroid 16:427–434
14. Mittereder N, March KL, Trapnell BC 1996 Evaluation of the concentrationand bioactivity of adenovirus vectors for gene therapy. J Virol 70:7498–7509
15. Tanaka T, Tsudo M, Karasuyama H, Kitamura F, Kono T, Hatakeyama M,Taniguchi T, Miyasaka M 1991 A novel monoclonal antibody against murineIL-2 receptor �-chain. Characterization of receptor expression in normal lym-phoid cells and EL-4 cells. J Immunol 147:2222–2228
16. Saitoh O, Nagayama Y 2006 Regulation of Graves’ hyperthyroidism withnaturally occurring CD4�CD25� regulatory T cells in a mouse model. Endo-crinology 147:2417–2422
17. Pohlenz J, Maqueem A, Cua K, Weiss RE, Van Sande J, Refetoff S 1999Improved radioimmunoassay for measurement of mouse thyrotropin in se-rum: strain differences in thyrotropin concentration and thyrotroph sensitivityto thyroid hormone. Thyroid 9:1265–1271
18. Kotani T, Umeki K, Yamamoto I, Takeuchi M, Takechi S, Nakayama T,Ohtaki S 1993 Nucleotide sequence of the cDNA encoding mouse thyroidperoxidase. Gene 123:289–290
19. Mizuguchi H, Kay MA 1998 Efficient construction of a recombinant adeno-virus vector by an improved in vitro ligation method. Hum Gene Ther 9:2577–2583
20. Guo J, Pichurin P, Nagayama Y, Rapoport B, McLachlan SM 2003 Insight intoantibody responses induced by plasmid or adenoviral vectors encodingthyroid peroxidase, a major thyroid autoantigen. Clin Exp Immunol 132:408–415
21. Chazenbalk GD, Jaume JC, McLachlan SM, Rapoport B 1997 Engineering thehuman thyrotropin receptor ectodomain from a non-secreted form to a se-creted, highly immunoreactive glycoprotein that neutralizes autoantibodies inGraves’ patients’ sera. J Biol Chem 272:18959–18965
23. Land KJ, Gudapati P, Kaplan MH, Seetharamaiah GS 2006 Differential re-quirement of Stat4 and Stat6 in a thyrotropin receptor-289-adenovirus inducedmodel of Graves’ hyperthyroidism. Endocrinology 147:111–119
24. Murakami M, Hosoi Y, Negishi T, Kamiya Y, Miyashita K, Yamada M,Iriuchijima T, Yokoo H, Yoshida I, Tsushima Y, Mori M 1996 Thymic hy-perplasia in patients with Graves’ disease. Identification of thyrotropin re-ceptors in human thymus. J Clin Invest 98:2228–2234
25. Dutton CM, Joba W, Spitzweg C, Heufelder AE, Bahn RS 1997 Thyrotropinreceptor expression in adrenal, kidney, and thymus. Thyroid 7:879–884
26. Murakami M, Hosoi Y, Araki O, Morimura T, Imamura M, Ogiwara T,Mizuma H, Mori M 2001 Expression of thyrotropin receptors in rat thymus.Life Sci 68:2781–2787
27. Li HS, Carayanniotis G 2005 Detection of thyroglobulin mRNA as truncatedisoform(s) in mouse thymus. Immunology 115:85–89
28. Taubert R, Schwendemann J, Kyewski B 2007 Highly variable expression oftissue-restricted self-antigens in human thymus: implications for self-toleranceand autoimmunity. Eur J Immunol 37:838–848
29. Akkaraju S, Ho WY, Leong D, Canaan K, Davis MM, Goodnow CC 1997 Arange of CD4 T cell tolerance: partial inactivation to organ-specific antigenallows nondestructive thyroiditis or insulitis. Immunity 7:255–271
30. Wood LC, Ingbar SH 1979 Hypothyroidism as a late sequela in patientwith Graves’ disease treated with antithyroid agents. J Clin Invest 64:1429–1436
31. Kong YM, David CS, Lomo LC, Fuller BE, Motte RW, Giraldo AA 1997Role of mouse and human class II transgenes in susceptibility to andprotection against mouse autoimmune thyroiditis. Immunogenetics 46:312–317
32. Ng HP, Banga JP, Kung AW 2004 Development of a murine model of auto-immune thyroiditis induced with homologous mouse thyroid peroxidase.Endocrinology 145:809–816
33. Martin AP, Coronel EC, Sano G, Chen SC, Vassileva G, Canasto-ChibuqueC, Sedgwick JD, Frenette PS, Lipp M, Furtado GC, Lira SA 2004 A novelmodel for lymphocytic infiltration of the thyroid gland generated bytransgenic expression of the CC chemokine CCL21. J Immunol 173:4791–4798
34. Ng HP, Kung AWC 2006 Induction of autoimmune thyroiditis and hypothy-roidism by immunization of immunoactive T cell epitope of thyroid peroxi-dase. Endocrinology 147:3085–3092
35. Quaratino S, Badami E, Pang YY, Bartok I, Dyson J, Kioussis D, Londei M,Maiuri L 2004 Degenerate self-reactive human T-cell receptor causes sponta-neous autoimmune disease in mice. Nat Med 10:920–926
36. Morris GP, Chen L, Kong YC 2003 CD137 signaling interferes with activationand function of CD4�CD25� regulatory T cells in induced tolerance to ex-perimental autoimmune thyroiditis. Cell Immunol 226:20–29
37. Wei WZ, Jacob JB, Zielinski JF, Flynn JC, Shim KD, Alsharabi G, GiraldoAA, Kong YC 2005 Concurrent induction of antitumor immunity and auto-immune thyroiditis in CD4� CD25� regulatory T cell-depleted mice. CancerRes 65:8471–8478
38. Romagnoli P, Tellier J, van Meerwijk JP 2005 Genetic control of thymicdevelopment of CD4�CD25�FoxP3� regulatory T lymphocytes. Eur J Immu-nol 35:3525–3532
39. Flynn JC, Meroueh C, Snower DP, David CS, Kong YM 2007 Depletion of
5732 Endocrinology, December 2007, 148(12):5724–5733 McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg
CD4CD25 regulatory T cells exacerbates sodium iodide-induced experimentalautoimmune thyroiditis in human leucocyte antigen DR3 (DRB1*0301) trans-genic class II-knock-out non-obese diabetic mice. Clin Exp Immunol 147:547–554
40. Flynn JC, Rao PV, Gora M, Alsharabi G, Wei W, Giraldo AA, David CS,Banga JP, Kong YM 2004 Graves’ hyperthyroidism and thyroiditis in HLA-DRB1*0301 (DR3) transgenic mice after immunization with thyrotropin re-ceptor DNA. Clin Exp Immunol 135:35–40
41. Chen C-R, Pichurin P, Chazenbalk GD, Aliesky H, Nagayama Y, McLachlanSM, Rapoport B 2004 Low-dose immunization with adenovirus expressing thethyroid-stimulating hormone receptor A-subunit deviates the antibody
response toward that of autoantibodies in human Graves’ disease. Endocri-nology 145:228–233
42. Marazuela M, Garcia-Lopez MA, Figueroa-Vega N, de la FH, Alvarado-Sanchez B, Monsivais-Urenda A, Sanchez-Madrid F, Gonzalez-Amaro R2006 Regulatory T cells in human autoimmune thyroid disease. J Clin Endo-crinol Metab 91:3639–3646
43. Nakano A, Watanabe M, Iida T, Kuroda S, Matsuzuka F, Miyauchi A,Iwatani Y 2007 Apoptosis-induced decrease of intrathyroidal CD4�CD25�
regulatory T cells in autoimmune thyroid diseases. Thyroid 17:25–3144. Volpe R, Iitaka M 1991 Evidence for an antigen-specific defect in suppressor
T-lymphocytes in autoimmune thyroid disease. Exp Clin Endocrinol 97:133–138
Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving theendocrine community.
McLachlan et al. • Graves’ Disease, Hashimoto’s Thyroiditis, and Treg Endocrinology, December 2007, 148(12):5724–5733 5733
