Graves' Hyperthyroidism and the Hygiene Hypothesis in a Mouse Model

0013-7227/04/$15.00/0 Printed in U.S.A.

Endocrinology 145(11):5075–5079 Copyright © 2004 by The Endocrine Society doi: 10.1210/en.2004-0683

Graves’ Hyperthyroidism and the Hygiene Hypothesis in a Mouse Model YUJI NAGAYAMA, SANDRA M. MCLACHLAN, BASIL RAPOPORT,

AND

KAZUNORI OISHI

Department of Medical Gene Technology, Atomic Bomb Disease Institute, Graduate School of Biomedical Sciences (Y.N.), and Department of Internal Medicine, Institute of Tropical Medicine (K.O.), Nagasaki University, Nagasaki 852-8523, Japan; and Autoimmune Disease Unit, Cedars-Sinai Research Institute and School of Medicine, University of California (S.M.M., B.R.), Los Angeles, California 90048 Graves’ hyperthyroidism is an organ-specific autoimmune disease mediated by stimulatory autoantibodies against the TSH receptor (TSHR; thyroid-stimulating antibodies), causing thyroid hyperplasia and hyperthyroidism. Development of this ailment is well known to be under polygenic and environmental control. For example, we recently demonstrated that parasite helminth Schistosoma mansoni infection suppressed a T helper cell type 1 (Th1)-type anti-TSHR immune response and prevented disease development in our mouse model of Graves’ disease using adenovirus coding for the TSHR. In the present study we examined the outcome of infection with Mycobacterium bovis bacillus Calmette-Gue´rin (BCG), a Th1-promoting infectious pathogen, on Graves’ disease. Our results show that prior infection with M. bovis BCG

G

RAVES’ DISEASE IS a unique organ-specific autoimmune disease in which autoantibodies directed against the TSH receptor [TSHR; thyroid-stimulating antibodies (TSAb)] activate the thyroid gland and lead to hyperthyroidism and diffuse goiter (reviewed in Ref. 1). The etiology of this ailment appears to be multifactorial, with both genetic and environmental factors being involved in disease development (2). It is well known that pathogenic infectious agents have potent systemic immunomodulatory effects on the host immune system directed to other infectious agents and noninfectious antigens (3). For example, we have recently shown that parasite helminth Schistosoma mansoni infection prevented Graves’ disease induced in a mouse model by im injection of replication-defective recombinant adenovirus coding for the full-length TSHR (Ad-TSHR) (4). In this study S. mansoni egg deposition or injection of soluble egg antigen differentiated the TSHR-specific immune response away from a T helper cell type 1 (Th1) phenotype. Because we have also demonstrated the preventive effect of coinjection of Ad expressing the Th2 cytokine IL-4 and a Th2-inducing agent ␣-galactosylceramide on disease development (4, 5), we concluded that the preventive effect of S. mansoni infection is also attributed to suppression of the Th1-type, TSHR-specific immune response. Thus, Graves’ Abbreviations: Ad, Adenovirus; BCG, bacillus Calmette-Gue´rin; cfu, colony-forming unit; IFN-␥, interferon-␥; TBIAb, TSH-binding inhibiting antibody; Th1, T helper cell type 1; TSAb, thyroid-stimulating antibody; TSHR, TSH receptor. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

differentiates the TSHR-specific immune response toward a Th1 phenotype, as demonstrated by enhanced secretion of a Th1 cytokine interferon-␥ and impaired production of a Th2 cytokine IL-10 from splenocytes stimulated in vitro with TSHR antigen. M. bovis BCG also significantly suppressed disease induction. These data together with our recent report that coinjection of adenovirus expressing the Th1 cytokine IL-12 induced a Th1-polarized, TSHR-specific immune response without affecting disease development support the hygiene hypothesis, rather than Th1-mediated disease suppression. Thus, some infectious pathogens may influence the development of Graves’ disease regardless of their ability to modify the Th1/Th2 balance. (Endocrinology 145: 5075–5079, 2004)

hyperthyroidism is an autoantibody-mediated, but not simplistically Th2-dominant, autoimmune disease. The suppression of other Th1-autoimmune diseases, such as type 1 diabetes, experimental autoimmune encephalitis, and collagen-induced arthritis, by S. mansoni infection has also been reported (6 –9). In contrast, mycobacterial infections are typically inducers of Th1 immune responses (10). Nevertheless, similar to helminth infections, mycobacterial infections or immunization with mycobacterium-containing adjuvants have been demonstrated to suppress Th1-autoimmune diseases, including diabetes (11–14), experimental autoimmune encephalitis (3, 15, 16), and adjuvant arthritis (17). However, these findings contrast with our recent report showing no effect of coinjection of Ad expressing the Th1 cytokine, IL-12, on Graves’ disease development in our mouse model (5). The present study was therefore designed to study the outcome of infection with a Th1-inducing pathogen, Mycobacterium bovis bacillus Calmette-Gue´rin (BCG) on Graves’ disease induction in our mouse model. Materials and Methods Immunization with Ad-TSHR Construction, amplification, purification, and determination of viral particle concentration of Ad-TSHR have previously been described (18). Female BALB/c mice (6 wk old; Charles River Laboratory, Tokyo, Japan) were injected im with 50 ␮l PBS containing Ad-TSHR (1011 particles/ mouse) three times at 3-wk intervals (18). All experiments were conducted in accordance with the principles and procedures outlined in the Guideline for the Care and Use of Laboratory Animals of Nagasaki University. Mice were kept under specific pathogen-free conditions

5075

5076

Endocrinology, November 2004, 145(11):5075–5079

Nagayama et al. • M. bovis BCG Infection and Graves’ Disease

throughout the experiments. Blood was drawn 8 wk after the last immunization.

Infection with M. bovis BCG M. bovis BCG strain was purchased from Nihon BCG Co. (Tokyo, Japan). Titers of M. bovis BCG were confirmed with Middlebrook 7H11 agar plates (Difco, Detroit, MI). Mice were infected with M. bovis BCG by ip injection of the suspension in PBS [1 ⫻ 108 colony-forming units (cfu)/100 ␮l䡠mouse]. Successful infection of M. bovis BCG was monitored by ELISA against mycobacterial antigen (see below).

T4, TSAb and TSH-binding inhibiting antibody (TBIAb) measurements Measurements of serum T4, and TSAb and TBIAb activities were performed as previously described (18).

ELISA for antibodies against TSHR and mycobacterial antigen TSHR-289 protein was purified as previously described (4, 5). Mycobacterial antigen was also prepared as previously shown (19). ELISA wells were coated overnight with 100 ␮l TSHR-289 protein (1 ␮g/ml) or mycobacterial antigen (5 ␮g/ml) and were incubated with mouse serum (diluted 1:100 –10,000). After incubation with horseradish peroxidaseconjugated antimouse IgG (diluted 1:3,000; A3673, Sigma-Aldrich Corp., St. Louis, MO), or subclass-specific antimouse IgG1 and IgG2a (diluted 1:1,000 and 1:1,500; X56 and R19 –15, BD Pharmingen, San Diego, CA), color was developed using orthophenylene diamine and H2O2 as substrate, and OD was read at 492 nm.

Cytokine assays Splenocytes were cultured in triplicate aliquots at 5 ⫻ 105 cells/well in a 96-well, round-bottomed plate in the presence or absence of either mycobacterial antigen (5 ␮g/ml) or TSHR-289 protein (5 ␮g/ml). Five days later, the concentrations of interferon-␥ (IFN-␥), IL-4, and IL-10 in the medium were determined with commercially available ELISA kits (BioSource International, Camarillo, CA). Cytokine production was expressed as picograms per milliliter, using standard curves of recombinant murine cytokines.

Thyroid histology Thyroid tissues were removed and fixed with 10% formalin in PBS. Tissues were embedded in paraffin, and 5-␮m thick sections were prepared and stained with hematoxylin and eosin.

Results

A high dose of M. bovis BCG (108 cfu/mouse), which has previously been shown to suppress type 1 diabetes (11, 13, 14), was used in the present study. All M. bovis BCG-infected mice produced detectable levels of both IgG1 and IgG2a types of antibodies against mycobacterial antigen 4 wk postinfection, and these titers were further increased 15 wk postinfection (Fig. 1A). Because these results indicate mixed Th1 and Th2 antibody responses to mycobacterial antigen, we also studied antigen-specific splenocyte secretions of Th1 and Th2 cytokines. Splenocytes prepared 4 wk postinfection produced significant amounts of IFN-␥ and IL-10, but not IL-4, in response to in vitro stimulation with mycobacterial antigen, and these levels were further elevated 14 wk postinfection (Fig. 1, B–D). Consequently, the immune response against mycobacterial infection observed in this study was of mixed Th1 and Th2 phenotypes, rather than a Th1 phenotype as reported in other studies (10, 19, 20). Graves’ hyperthyroidism was induced by three injections

FIG. 1. Titers of IgG1 and IgG2a antimycobacterial antibodies and in vitro mycobacterial antigen-specific splenocyte secretion of IFN-␥, IL-4, and IL-10 in mice infected with M. bovis BCG. Mice were mockinfected or infected with 108 cfu M. bovis BCG. A, Sera were obtained 4 and 14 wk after infection. IgG1 and IgG2a antibody titers were measured by ELISA (see Materials and Methods). Data are the mean ⫾ SD (n ⫽ 15) obtained with sera diluted 1:10,000. B–D, Splenocytes were prepared 4 and 14 wk after infection and were stimulated with 5 ␮g/ml mycobacterial antigen for 5 d. Cytokine levels in the culture supernatants were measured by ELISA. Data are the mean ⫾ SD of three mice per group and are expressed as picograms per milliliter. 䡺 and f, Splenocytes cultured in the absence or presence of mycobacterial antigen, respectively. *, P ⬍ 0.001, by t test.

of Ad-TSHR as previously described (18). The effect of mycobacterial infection was first evaluated by infecting mice with M. bovis BCG, followed by immunization with AdTSHR 4 wk postinfection. M. bovis BCG infection itself did not impact on T4 levels (Fig. 2A). The incidence of Graves’ hyperthyroidism, based on serum T4 levels measured 8 wk after the last immunization, was significantly lower in mice injected with M. bovis BCG and Ad-TSHR than in those with TSHR alone [13% vs. 60% (P ⫽ 0.004, by ␹2 test); 5.70 ⫾ 4.54 vs. 11.60 ⫾ 6.13 ␮g/dl (P ⫽ 0.011, by t test); Fig. 2A]. Thus, prior infection with M. bovis BCG significantly reduced AdTSHR-induced Graves’ hyperthyroidism. TSAb activity was present in most hyperthyroid mice, and there was a significant difference in TSAb activities between the two groups (113.9 ⫾ 70.5% vs. 201.1 ⫾ 113.8%; P ⫽ 0.026, by t test; Fig. 2B). However, there was no difference between the two groups in terms of TSHR antibody titers determined by TBIAb, IgG ELISA, or IgG subclass ELISAs (Fig. 2, C–E). Antigen-specific splenocyte secretion of the Th1 cytokine IFN-␥ and the Th2 cytokines IL-4 and IL-10 was also studied. As previously reported (4), splenocytes prepared 10 d after immunization with Ad-TSHR produced significant amounts of IFN-␥ and IL-10, but not IL-4, in response to in vitro stimulation with TSHR antigen. However, when M. bovis BCG was given 4 wk earlier, splenocytes prepared 10 d after Ad-TSHR immunization lost their TSHR-specific IL-10 re-


Endocrinology, November 2004, 145(11):5075–5079 5077

FIG. 2. T4, TSAb, TBIAb, and IgG, IgG1, and IgG2a antiTSHR antibody titers in mice infected with M. bovis BCG, followed by Ad-TSHR immunization. Mice were immunized three times with Ad-TSHR 4 wk after mock infection or M. bovis BCG infection. Sera were obtained 8 wk after the last immunization with Ad-TSHR. Composite results of two independent results are shown. A–C, Data are the mean of the duplicate assays. E and 䊐, Euthyroid and hyperthyroid mice, respectively; horizontal lines, the normal upper limits of each assay. *, P ⫽ 0.004 (by ␹2 test); P ⫽ 0.011 (by t test). **, P ⫽ 0.026 (by t test). N.D., Not determined. D and E, Data are the mean ⫾ SD obtained with serum diluted 1:300.

FIG. 4. T4 in mice immunized with Ad-TSHR, followed by M. bovis BCG infection. Sera were obtained 8 wk after the last immunization with Ad-TSHR (7 wk after M. bovis BCG infection). Data are the means of the duplicate assays. E and 䊐, Euthyroid and hyperthyroid mice, respectively; horizontal lines, normal upper limit of assay.

FIG. 3. In vitro TSHR-specific splenocyte secretion of IFN-␥, IL-4, and IL-10 in mice infected with M. bovis BCG, followed by Ad-TSHR immunization. Splenocytes were prepared 10 d or 14 wk after immunization with Ad-TSHR in mice mock-infected or infected with M. bovis BCG and were stimulated with 5 ␮g/ml TSHR-289 protein for 5 d. Cytokine levels in the culture supernatants were measured by ELISA. Data are the mean ⫾ SD of three mice per group and are expressed as picograms per milliliter. 䡺 and f, Splenocytes cultured in the absence or presence of TSHR-289 protein, respectively. *, P ⬍ 0.001 (by t test).

sponse. Furthermore, an augmented antigen-specific IFN-␥ response was observed when splenocytes were prepared 14 wk after Ad-TSHR immunization (Fig. 3). These data indicate immune deviation toward Th1 of the anti-TSHR immune response after M. bovis BCG infection. In the second set of experiments, mice were first immunized three times with Ad-TSHR, followed by infection with M. bovis BCG 1 wk after the last immunization, at a time when the anti-TSHR immune response was already fully activated. As shown in Fig. 4, M. bovis BCG infection after Ad-TSHR immunization had no effect on disease development. The hyperthyroid mice had diffuse goiters with hypertrophy and hypercellularity of thyroid epithelial cells as previously reported (18), findings consistent with those in over-

5078

Endocrinology, November 2004, 145(11):5075–5079

stimulated thyroid glands. No lymphocytic infiltration was observed (data not shown). Discussion

We investigated the effect of M. bovis BCG infection on Graves’ disease development using a mouse model we have recently established (18). Although M. bovis BCG infection has previously been reported to induce a Th1-dominant immune response (10, 19, 20), the antimycobacterial immune response observed in our study had characteristics of both Th1 and Th2 phenotypes. Thus, antibodies against mycobacterial antigen were of IgG1 and IgG2a subclasses, and splenocytes secreted both Th1 and Th2 cytokines, IFN-␥ and IL-10, in response to in vitro antigen stimulation. Nevertheless, although the reason(s) for this discrepancy is not clear, prior infection with M. bovis BCG differentiated the antiTSHR immune response toward a Th1 phenotype, as demonstrated by the loss and enhancement of antigen-specific splenocyte secretion of IL-10 and IFN-␥, respectively, in mice immunized with Ad-TSHR after M. bovis BCG infection, although the titers of IgG1 and IgG2a anti-TSHR antibodies remained unchanged. Importantly, this Th1 deviation of the anti-TSHR immune response by M. bovis BCG infection was accompanied by suppression of TSAb and Graves’ disease, but not of antiTSHR antibody titers, as determined with TBIAb assay and ELISA. Thus, prior infection with M. bovis BCG did not prevent the initiation of autoimmunity against TSHR, but did halt the progression to overt Graves’ disease. These data appear inconsistent with our recent study with Ad expressing IL-12 (4), in which overexpression of IL-12 skewed the anti-TSHR immune response toward Th1 without affecting disease incidence. Consequently, Th1 polarization of the TSHR-specific immune response cannot solely explain the preventive effect of M. bovis BCG infection on the induction of Graves’ disease. As mentioned above, mycobacterial infections or immunization with mycobacterium-containing adjuvants are reported to suppress other Th1-autoimmune diseases (3, 11– 17). Although the mechanisms for the suppressive effect of mycobacterial infection remains to be elucidated, Martins et al. (14) have demonstrated that protection of nonobese diabetic mice against diabetes by Mycobacterium avium is due to Th1-type immune responses that increase Fas/Fas ligand expression and cytotoxic effects on T cells, leading to deletion of autoreactive T lymphocytes. In contrast, Shedaheh et al. (12) previously reported Th2 deviation of the autoimmune response by complete Freund’s adjuvant in nonobese diabetic mice. Our present and recent (4) studies together with those from other laboratories (3, 6 –9, 11–17) clearly demonstrate that both Th1- and Th2-promoting infectious pathogens can prevent Th1-autoimmune diseases. These apparently discrepant results are consistent with the hygiene hypothesis or counterregulatory model (21, 22). Infectious agents may play a role in the induction of autoimmunity (23). However, the hygiene hypothesis proposes that decreased exposure to both Th1- and Th2-type infectious pathogens during childhood impairs the development of an appropriately educated


immune system and increases the likelihood of developing Th1-autoimmune as well as Th2-allergic diseases in adult life. In general, the prevalence of allergic and autoimmune diseases is higher in developed countries (21, 22, 24). In particular, Graves’ disease appears to be relatively uncommon in developing countries (25, 26). Moreover, there is a significant association between Th1-autoimmune and Th2allergic diseases in individual patients (27, 28), suggesting the existence of a common etiology, e.g. the hygiene hypothesis. Based on this background, we interpret our present data as supporting the hygiene hypothesis, rather than as Th1mediated disease suppression. Alternatively, but not mutually exclusively, the Th1 vs. Th2 paradigm may be too simplistic to explain the highly complex nature of the pathogenesis of autoimmune diseases. It should be noted that M. bovis BCG infection after immunization with Ad-TSHR has no effect on disease development. Of interest, Esaguy and Aguas (17) have shown that BCG injection during, but not after, the neonatal period prevented adjuvant arthritis in Lewis rats. Consequently, the timing of infection may be crucial to influence disease development. In conclusion, our present and recent (4) studies indicate that the induction of Graves’ disease is affected by certain infectious pathogens regardless of their ability to modify Th1/Th2 balance. It should be emphasized that nonpathogenic environmental microbial agents (conventional housing) and microorganism components (Escherichia coli lipopolysaccharide and yeast zymosan A) had little effect on the induction of Graves’ disease (29). Additional studies clarifying the mechanism(s) of these suppressive effects of infectious pathogens may ultimately facilitate the development of novel therapies for Graves’ disease. Acknowledgments Received May 28, 2004. Accepted August 2, 2004. Address all correspondence and requests for reprints to: Dr. Yuji Nagayama, Department of Medical Gene Technology, Atomic Bomb Disease Institute, Graduate School of Biomedical Sciences, Nagasaki University, 1-12-4 Sakamoto, Nagasaki 852-8523, Japan. E-mail: [email protected].

References 1. Rapoport B, Chazenbalk GD, Jaume JC, McLachlan SM 1998 The thyrotropin (TSH) receptor: interaction with TSH and autoantibodies. Endocr Rev 19:673– 716 2. McLachlan SM, Rapoport B 2000 Genetic factors in thyroid disease. In: Braverman LE, Utiger RD, eds. Werner, Ingbar’s the thyroid. 8th ed. Philadelphia: Lippincott Williams & Wilkins; 474 – 487 3. Sewell DL, Reinke EK, Hogan LH, Sandor M, Fabry Z 2002 Immunoregulation of CNS autoimmunity by helminth and mycobacterial infections. Immunol Lett 82:101–110 4. Nagayama Y, Watanabe K, Niwa M, McLachlan SM, Rapoport B 2004 Schistosoma mansoni and ␣-galactosylceramide: prophylactic effect of Th1 suppression in a mouse model of Graves’ hyperthyroidism. J Immunol 173:2167–2173 5. Nagayama Y, Mizuguchi H, Hayakawa T, Niwa M, McLachlan SM, Rapoport B 2003 Prevention of autoantibody-mediated Graves’-like hyperthyroidism in mice by IL-4, a Th2 cytokine. J Immunol 170:3522–3527 6. Cooke A, Tonks P, Jones FM, O’Shea H, Hutchings P, Fulford AJ, Dunne DW 1999 Infection with Schistosoma mansoni prevents insulin dependent diabetes mellitus in non-obese diabetic mice. Parasitol Immunol 21:169 –176 7. Mattsson L, Larsson P, Erlandsson-Harris H, Klareskog L, Harris RA 2000 Parasite-mediated down-regulation of collagen-induced arthritis (CIA) in DA rats. Clin Exp Immunol 122:477– 483 8. Zaccone P, Fehervari Z, Jones FM, Sidobre S, Kronenberg M, Dunne DW,


9. 10. 11. 12. 13. 14. 15. 16.

17. 18.

Cooke A 2003 Schistosoma mansoni antigens modulate the activity of the innate immune response and prevent onset of type 1 diabetes. Eur J Immunol 33: 1439 –1449 Sewell D, Qing Z, Reinke E, Elliot D, Weinstock J, Sandor M, Fabry Z 2003 Immunomodulation of experimental autoimmune encephalitis by helminth ova immunization. Int Immunol 15:59 – 69 Mutis T, Cornelisse YE, Ottenhoff TH 1993 Mycobacteria induce CD4⫹ T cells that are cytotoxic and display Th1-like cytokine secretion profile: heterogeneity in cytotoxic activity and cytokine secretion levels. Eur J Immunol 23:2189 –2195 Harada M, Kishimoto Y, Makino S 1990 Prevention of overt diabetes and insulitis in NOD mice by a single injection of BCG vaccination. Diabetes Res Clin Pract 8:85– 89 Shehadeh NN, LaRosa F, Lafferty KJ 1993 Altered cytokine activity in adjuvant inhibition of autoimmune diabetes. J Autoimmunol 6:291–300 Bras A, Aguas AP 1996 Diabetes-prone NOD mice are resistant to Mycobacterium avium and the infection prevents autoimmune disease. Immunology 89:20 –25 Martins TC, Aguas AP 1999 Mechanisms of Mycobacterium avium-induced resistance against insulin-dependent diabetes mellitus (IDDM) in non-obese diabetic (NOD) mice: role of Fas and Th1 cells. Clin Exp Immunol 115:248 –254 Meyer TJ, Azuma I, Ribi EE 1975 Biologically active components from mycobacterial cell walls. III. Production of experimental allergic encephalomyelitis in guinea-pigs. Immunology 28:219 –229 Lehmann D, Ben-Nun A 1992 Bacterial agents protect against autoimmune disease. I. Mice pre-exposed to Bordetella pertussis or Mycobacterium tuberculosis are highly refractory to induction of experimental autoimmune encephalitis. J Autoimmunol 5:675– 690 Esaguy N, Aguas AP 1996 Prevention of adjuvant arthritis in Lewis rats by neonatal bacille Calmette-Guerin (BCG) infection. Clin Exp Immunol 104:103– 107 Nagayama Y, Kita-Furuyama M, Ando T, Nakao K, Mizuguchi H, Hayakawa T, Eguchi K, Niwa M 2002 A novel murine model of Graves’ hyperthyroidism

Endocrinology, November 2004, 145(11):5075–5079 5079

19. 20. 21. 22. 23. 24. 25. 26. 27.

28.

29.

with intramuscular injection of adenovirus expressing the thyrotropin receptor. J Immunol 168:2789 –2794 Orme IM, Roberts AD, Griffin JP, Abrams JS 1993 Cytokine secretion by CD4 T lymphocytes acquired in response to Mycobacterium tuberculosis infection. J Immunol 151:518 –525 Rhodes SG, Palmer N, Graham SP, Bianco AE, Hewinson RG, Vordermeier HM 2000 Distinct response kinetics of ␥-interferon and interleukin-4 in bovine tuberculosis. Infect Immun 68:5393–5400 Bach J-F 2002 The effect of infections on susceptibility to autoimmune and allergic diseases. N Engl J Med 347:911–920 Wills-Karp M, Santeliz J, Karp CL 2001 The germless theory of allergic disease: revisiting the hygiene hypothesis. Nature Rev 1:69 –75 Wucherpfennig KWW 2001 Mechanisms for the induction of autoimmunity by infectious agents. J Clin Invest 108:1097–1104 Stene LC, Nafstad P 2001 Relation between occurrence of type 1 diabetes and asthma. Lancet 357:607– 608 Kalk WJ, Kalk J 1989 Incidence and causes of hyperthyroidism in blacks. S Afr Med J 75:114 –117 Famuyiwa OO, Bella AF 1990 Thyrotoxicosis in Nigeria. Analysis of a five year experience. Trop Geogr Med 42:248 –254 Kero J, Gissler M, Hemminki E, Isolauri E 2001 Could TH1 and TH2 diseases coexist? Evaluation of asthma incidence in children with coeliac disease, type 1 diabetes, or rheumatoid arthritis: a register study. J Allergy Clin Immunol 108:781–783 Simpson CR, Anderson WJA, Helms PJ, Taylor MW, Watson L, Prescott GJ, Godden DJ, Barker RN 2002 Coincidence of immune-mediated diseases driven by Th1 and Th2 subsets suggests a common aetiology; a populationbased study using computerized general practice data. Clin Exp Allergy 32: 37– 42 Nagayama Y, McLachlan SM, Rapoport B, Niwa M 2003 A major role for non-major histocompatibility complex genes but not for microorganisms in a novel murine model of Graves’ hyperthyroidism. Thyroid 13:233–238

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.