The thyrotropin receptor (TSHR) is well known to be not only the primary molecule regulating thyroid function but also one of the autoantigens in autoimmune thyroid disease, especially in Graves’ disease, in which stimulatory antibodies (thyroid-stimulating antibodies, TSAb) against this receptor cause overproduction of thyroid hormones (hyperthyroidism) and thyroid enlargement. Structurally, the TSHR is a member of the G protein coupled receptor superfamily characterized by the exceptionally large ectodomain containing the leucine-rich repeats. Another unique factor of the TSHR is its intramolecular cleavage producing the ligand-binding A-subunit and the transmembrane/cytoplasmic B-subunit. The former is now believed to be the major autoantigen in Graves’ disease .
Unlike Hashimoto's thyroiditis, which is another type of autoimmune thyroid disease but results in the opposite consequence, that is hypothyroidism, a spontaneous animal model of Graves’ disease does not exist. Thus, many groups have attempted to establish the induced type of Graves’ animal model, the details of which can be found in a previous review . Among several models established so far, two methods, one involving repetitive intramuscular (i.m.) injection of recombinant adenovirus expressing the TSHR (Ad-TSHR) and the other involving the combination of repetitive i.m. injections of the eukaryotic expression plasmid coding the TSHR and in-vivo electroporation, are now most widely used. Both methods have good reproducibility and high disease induction rates. In both models, as mentioned above, the A-subunit is more efficient in inducing disease than the full-length holo-receptor. In this review, the recent advances obtained with these two models from 2012 to 2014 are summarized. It is worth noting that it has been 26 years since the cloning of the TSHR cDNA, and 16 years since the first report of the authentic Graves’ mouse model (Shimojo's model) .
ATTEMPTS TO IMPROVE THE MODELS
The models mentioned above originally used the human TSHR. Thus, one should realize that, in these models, immunization induces antihuman TSHR antibodies, which cross-react with and stimulate the mouse TSHR expressed in the mouse thyroid glands. The types of immune cells involved in the disease pathogenesis include a mixture of T helper type 1 (Th1), Th2 and the regulatory T cells (Tregs) (such as CD4+CD25+), but not Th17 cells [3,4]. Thus, Graves’ disease is humoral and antibody-mediated, but cannot be referred to as simply Th2-mediated.
Human and mouse TSHR A-subunits share approximately 87% amino acid homology with each other, suggesting that the human TSHR may not be an authentic autoantigen in mice, and thus the mouse, not human, TSHR should ideally be used as a Graves’ mouse model. Therefore, an adenovirus expressing the mouse TSHR A-subunit has been constructed to evaluate the autoimmune reactions against the mouse TSHR in mice . Interestingly, wild-type BLAB/c mice (a susceptible mouse strain for Graves’ disease) were totally unresponsive to the mouse TSHR, even with the higher doses of adenovirus (1011 particles/mouse instead of 1010) and CD4+CD25+ Treg depletion using anti-CD25 antibody. Thus, Treg plays little role in breaking self-tolerance to the mouse TSHR in wild-type mice . In contrast, the TSHR knockout BALB/c mice responded well to the mouse receptor, producing antimouse TSHR antibodies (TRAb) detected in the flow cytometry and TSAb detected in the bioassay, both using Chinese hamster ovary (CHO) cells expressing the mouse TSHR. Thus, wild-type mice are highly tolerant and the TSHR knockout mice are good responders to the mouse TSHR, indicating that use of the mouse TSHR might be critical for the studies on the mechanisms for immune tolerance to the TSHR and how this tolerance can be broken in mice. Whether there is a difference in relative immunogenicity between the mouse and the human TSHR defined by different numbers of glycosylation sites (five vs. six) remains to be determined .
Of interest, in the TSHR knockout mice immunized with the mouse TSHR, the titres of TRAb gradually declined over 6 months (although the titres were still significantly high at the end of experiments), with a shift from TSAb dominance in the early period to thyroid-blocking antibody (TBAb) dominance in the late stage.
Transient hyperthyroidism is also observed in the original model using i.m. injections of Ad-human TSHR . Thus, although TSH-binding inhibiting antibody (TBIAb) levels (and also the antibody levels detected by ELISA assay that recognizes nonfunctional antibodies) remained high during the 20-week periods after the first immunization, TSAb values were positive at the 4-week time point, and then gradually declined. No mice were hyperthyroid at the end of the experiment.
The immune tolerance is composed of peripheral and central arms. The significance of peripheral tolerance to the mouse TSHR has been studied using mAbs to deplete CD4+CD25+ Treg, those to antagonize the coinhibitory molecules  and those to stimulate costimulatory molecules . Inhibition of coinhibitory molecules of cytotoxic T lymphocyte antigen-4 (CTLA-4) and programmed cell death ligand-1 (PD-L1) permitted a slight increase in anti-TSHR antibody titres following Ad-mouse TSHR A-subunit immunization without altering thyroid hormone levels, revealing only partial breakdown of immune tolerance to the mouse TSHR . Further studies will be necessary to evaluate the relative importance of central tolerance to the mouse TSHR. Indeed, a recent report emphasized the importance of central tolerance in Graves’ patients by showing that a TSHR gene single nucleotide polymorphism (rs179247) is associated with a decrease in its thymic expression and susceptibility to Graves’ disease .
Knockout mice did not develop Graves’-like hyperthyroidism despite the production of TSAb because of the absence of the TSHR in their thyroid glands, although the monitoring of TRAb and TSAb, not thyroid hormones and thyroid histology, may still be invaluable for the studies on the immune reaction to the mouse TSHR. In order to evaluate thyroid hormone production by antimouse TSHR antibodies elicited in the THSR knockout mice immunized with Ad-mouse TSHR, adoptive transfer of splenocytes from the TSHR knockout mice immunized with the mouse TSHR to immune-deficient athymic nude mice has been attempted [11▪▪]. Like the TSHR knockout mice immunized with the mouse TSHR, TRAb development and gradual decline of its titre with hyperthyroidism/TSAb dominance in the early period and euthyroidism or hypothyroidism/TBAb dominance at the end of 6-month experimental periods were observed. Histologically, the majority of the thyroid glands from hypothyroid mice showed flattened thyroid follicular epithelial cells, a sign of the blockade of TSH action by TBAb. However, some glands were also accompanied by severe intrathyroidal lymphocyte infiltration with a concomitant increase in antithyroglobulin (Tg) antibodies, an indication of a primary TSHR-specific cellular immune response followed by a secondary anti-Tg (humoral and probably also cellular) immune response. Thus, the TSHR can also be a target of the cellular immune response. It should be noted that an attempt to establish a spontaneous Graves’ model by adoptively transferring splenocytes from naive TSHR knockout mice to nude mice resulted in failure. Thus, all of the models mentioned above (the original model with Ad-human TSHR, the TSHR knockout mice immunized with Ad-mouse TSHR and the adoptive transfer model) utilized adenovirus as an immunization vehicle and adjuvant and show transient TSAb production and hyperthyroidism. These data suggest either that the lifespans of TSAb-producing B cells are somehow shorter than those of TBAb-producing B cells or that TBAb-producing B cells may continuously appear by epitope spreading or other mechanisms and dominate over TSAb-producing B cells over an extended time period.
Despite the strong immune tolerance to the full-length of the receptor A-subunit, wild-type mice have been shown to elicit autoimmune responses to the variant mouse TSHR in which the exon 5 (corresponding to one LLR mentioned above) was aberrantly spliced out [12▪▪]. This variant form of the TSHR can bind to TSAb and produces cAMP, but cannot bind to TSH. This report is particularly of interest in terms of a lack of immune tolerance to this variant in wt mice. Importantly, the same variant can be detected in the thyroids of human origin by RT-PCR. Further studies will be necessary to see whether or not this variant is the main autoantigen.
Very recently, the long-lasting hyperthyroid model has been reported by repetitive injection of Ad-human TSHR A-subunit over 9 months [13▪▪]. In this study, the original two or three 3-week interval immunizations were followed by repeated 4-week interval immunizations (in total, nine immunizations). The long-term persistence of hyperthyroidism has originally been described in the model of combined plasmid-human TSHR and in-vivo electroporation , in which immunizations were repeated four times at 3-week intervals, suggesting that long-term provision of antigen by either adenovirus or plasmid may permit the sustained TSAb production.
A spontaneous mouse model has also been reported very recently [15▪▪]. NOD.H2h4 mice, which spontaneously develop iodine-accelerated thyroiditis (positive anti-Tg and thyroid peroxidase (TPO) antibodies with intrathyroidal lymphocyte infiltration), but not anti-TSHR immune response, were used. The authors hypothesized that immunological tolerance to the TSHR might be due to very low levels of the TSHR expression on their thyroid cells, and that forced TSHR overexpression would break the immune tolerance to the TSHR. Thus, their hypothesis was correct, that is NOD.H2h4 mice that overexpressed the human TSHR A-subunit transgene [from the thyroid-specific human TSHR A-subunit transgenic mice of low-expressor (see below)] successfully developed anti-TSHR antibodies with TSAb and TBIAb activities in addition to iodine-accelerated thyroiditis. These spontaneous models will be very useful for studies on not only disease pathogenesis but also development of therapeutic modalities.
Because of the relatively short period of the hyperthyroid state, the adenovirus model is suitable for evaluating the mechanism(s) for disease initiation. The role of CD40, the immuno-stimulatory molecule expressed in the thyroid for disease pathogenesis, has been studied by using two different types of genetically engineered mice, one is transgenic mice specifically expressing CD40 in the thyroid glands under the control of Tg promoter and the other is a chimeric knockout mice lacking its expression in nonbone marrow derived cells including the thyroid follicular epithelial cells . This molecule is of interest because it is a well known susceptible gene for many autoimmune diseases, and it is expressed not only in immune cells but also in nonimmune cells of targets for tissue-specific autoimmune diseases (thyroid follicular epithelial cells in case of Graves’ disease) in humans. Overexpression of CD40 in the thyroid augmented anti-TSHR antibody levels and also the severity of disease and, in contrast, knockdown of its expression in the thyroid decreased disease severity. Following the comparison of thyroid transcriptomes between the acute and chronic phases of disease in wild-type and transgenic mice overexpressing CD40 and also that between thyroid cells cultured in the presence or absence of CD40, and confirmation of production of several proinflammatory cytokines by CD40-stimulated thyroid cells, the authors have focused their attention on interleukin (IL)-6, and shown that anti-IL-6 antibody delayed onset of antibody production and suppressed disease development in transgenic mice. Thus, IL-6 might therapeutically be a suitable target for immune intervention. However, there are two critical points to this article. One is that intrathyroidal lymphocyte infiltration cannot be seen in the adenovirus model, so a question then arises as to how thyroidal CD40 interacts with immune cells. The other is that it has been reported that, although increased CD40 expression was indeed detected by RT-PCR in the thyroids of hyperthyroid mice induced with Ad-human TSHR A-subunit, immunohistochemical analysis showed its expression in ‘multiple cells between thyroid follicular cells’, not in thyroid follicular cells themselves in mice, suggesting a likely distinct role of CD40 in the pathogenesis of human and mouse Graves’ disease .
The adenovirus model is also useful to study the short-term effect of potential therapeutic candidates. Cai et al. have recently used diosgenin, a naturally occurring steroid saponin, for the purpose of treating experimental Graves’ hyperthyroidism, which, when given after induction of hyperthyroidism, successfully normalized serum thyroxine (T4) concentrations. They interpret their data that this compound inhibited thyroid cell proliferation because there was no alteration in anti-TSHR antibody levels or splenic CD4+/CD8 T lymphocyte ratios. However, no alteration in the antibody titres does not necessarily indicate the nonimmunological effect of this compound. Evaluation of either alteration in TSAb/TBAb ratios or a shift away from functional TSAb production towards production of nonstimulatory antibodies may help clarify its therapeutic mechanisms.
THE THYROTROPIN RECEPTOR TRANSGENIC MICE
Two lines of the thyroid-specific human TSHR A-subunit transgenic mice have previously been generated; one is high-expressors showing higher expression levels of the TSHR protein in the thyroid and thymus and strong immune tolerance to the human TSHR, and the other low-expressors showing weak tolerance. A difference in the levels of self-tolerance is thought to be due to different expression levels of the TSHR protein in the thymus. The induction of extensive intrathyroidal lymphocyte infiltration and antibody spreading to other thyroid antigens [Tg and TPO] in low-expressor depleted of Treg by anti-CD25 antibody and then immunized with Ad-human TSHR A-subunit has previously been shown. An attempt, made to break tolerance in high-expressor by depleting Treg before immunization with complete Freund's adjuvant (CFA) along with the human TSHR A-subunit protein and two boosts with Ad-human TSHR A-subunit, worked well in terms of antibody induction, but not thyroiditis induction. Thus, high-expressor immunized developed anti-TSHR antibody, levels comparable with those in low-expressor and wild-type mice, but intrathyroidal lymphocyte infiltration was not observed .
GRAVES’ ORBITOPATHY MODEL
The patients with Graves’ disease often develop extrathyroidal manifestations such as Graves’ orbitopathy and dermopathy. Although the precise underlying causal mechanism(s) remained unknown, several lines of circumstantial evidence suggest that the TSHR is the potential candidate autoantigen for these complications . A few articles have recently described the orbital lesions in the models mentioned above. First, immunization with the human TSHR by either adenovirus or plasmid (without in-vivo electroporation) induced no orbital lesion despite significant expression of the TSHR protein in the orbital fat . Second, adoptive transfer of splenocytes from Ad-mouse TSHR-immunized TSHR knockout mice to nude mice induced macrophage infiltration into the retrobulbar muscles and adipose tissues at low frequency at the end of the experiment when mice are either euthyroid or hypothyroid with positive TBAb (see above) [11▪▪]. Third, Banga's group has reported that in-vivo electroporation of plasmid-coding human TSHR in wild-type BLAB/c mice elicited the sustained hyperthyroidism and fibrosis in the retro-orbital regions in their first article [22▪] but, in their second article [23▪▪], that induction of the stronger immune response by ‘deeper injection of the plasmid over a larger muscle area’ led to hypothyroidism, and orbital lesions of retro-orbital lymphocyte infiltration and eye protrusion. These data, suggesting a linkage between hypothyroidism and development of Graves’ orbitopathy-like lesions, are difficult to reconcile, as clinical data show a strong association between TSAb and orbitopathy . Also, the lymphocyte infiltration mentioned above indicates the elicitation of a cellular immune response to the mouse TSHR, which does not fit with the data showing the lack of cross-reactivity of human reactive T cells with the mouse receptor .
A genetic basis for induced Graves’ disease by adenovirus has been extensively studied using large sets of recombinant inbred mice by McLachlan et al.  showing a linkage of TBIAb to major immunohistocompatibility complex (MHC) region genes on chromosome 17 and immunoglobulin G (IgG) heavy chain variable (V) region genes (IGVH) on chromosome 12, and that of the increase in T4 levels to IGVH. These findings could be expected when considering the importance of both of MHC and IgG in the development of antibody-mediated autoimmune disease, because T lymphocytes are activated by the presentation of the antigenic peptides complexed with MHC class II on antigen-presenting cells, and then stimulate B cells to produce antibodies.
Various studies have improved mouse Graves’ models, which have made it possible to perform numerous experiments to study disease pathogenesis and the development of therapeutic modalities that cannot be done in humans. However, at the same time, it is also true that the different models have their own advantages and disadvantages, and that their clinical features do not always reflect those of human disease exactly. Therefore, improvement of the models and studies on disease pathogenesis should be continued in parallel to accomplish the final goal of establishing a means for the prevention and treatment of Graves’ disease and of Graves’ orbitopathy.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
REFERENCES AND RECOMMENDED READING
Papers of particular interest, published within the annual period of review, have been highlighted as:
- ▪ of special interest
- ▪▪ of outstanding interest
1. McLachlan SM, Rapoport B. Breaking tolerance to thyroid antigens: changing concepts in thyroid autoimmunity. Endocr Rev 2014; 35:59–105.
2. McLachlan SM, Nagayama Y, Rapoport B. Insight into Graves’ hyperthyroidism
from animal models. Endocr Rev 2005; 26:800–832.
3. Rapoport B, McLachlan SM. Graves’ hyperthyroidism
is antibody-mediated but is predominantly a Th1-type cytokine disease. J Clin Endocrinol Metab 2014; 99:4060–4061.
4. Zhou J, Bi M, Fan C, et al. Regulatory T cells but not T helper 17 cells are modulated in an animal model of Graves’ hyperthyroidism
. Clin Exp Med 2012; 12:39–46.
5. Nakahara M, Mitsutake N, Sakamoto H, et al. Enhanced response to mouse thyroid-stimulating hormone (TSH) receptor immunization in TSH receptor-knockout mice. Endocrinology 2010; 151:4047–4054.
6. McLachlan SM, Aliesky HA, Chen CR, et al. Role of self-tolerance and chronic stimulation in the long-term persistence of adenovirus
-induced thyrotropin receptor
antibodies in wild-type and transgenic mice. Thyroid 2012; 22:931–937.
7. Ott PA, Hodi FS, Robert C. CTLA-4 and PD-1/PD-L1 blockade: new immunotherapeutic modalities with durable clinical benefit in melanoma patients. Clin Cancer Res 2013; 19:5300–5309.
8. Verbrugge I, Hagekyriakou J, Sharp LL, et al. Radiotherapy increases the permissiveness of established mammary tumors to rejection by immunomodulatory antibodies. Cancer Res 2012; 72:3163–3174.
9. Yasui J, Nakahra M, Shiammura M, et al. Minor contribution of cytotoxic T lymphocyte antigen 4 and programmed cell death ligand 1 in immune tolerance against mouse thyrotropin receptor
in mice. Acta Med Nagasaki 2014; 59:13–17.
10. Colobran R, Armengol Mdel P, Faner R, et al. Association of an SNP with intrathymic transcription of TSHR and GD: a role for defective thymic tolerance. Hum Mol Genet 2011; 20:3415–3423.
11▪▪. Nakahara M, Johnson K, Eckstein A, et al. Adoptive transfer of antithyrotropin receptor (TSHR) autoimmunity from TSHR knockout mice to athymic nude mice. Endocrinology 2012; 153:2034–2042.
Given the good responsiveness of TSHR knockout mice to the mouse TSHR in the ref. , adoptive transfer of splenocytes from immunized TSHR knockout mice to immunodeficient nude mice successfully induced TSAb and hyperthyroidism, which however is transient followed by development of TBAb and hypothyroidism. Macrophage infiltration was also observed in the retrobulbar muscles and adipose tissues of a fraction of mice.
12▪▪. Endo T, Kobayashi T. Immunization of mice with a newly identified thyroid-stimulating hormone receptor splice variant induces Graves’-like disease. J Autoimmun 2013; 43:18–25.
The authors cloned the variant mouse TSHR cDNA in which the exon 5 was aberrantly spliced out. Unlike wild-type TSHR, this variant form of the receptor was shown to be immunogenic in wild-type mice, inducing TSAb and hyperhtyroidism upon immunization using adenovirus.
13▪▪. Holthoff HP, Goebel S, Li Z, et al. Prolonged TSH receptor A subunit immunization of female mice leads to a long-term model of GD, tachycardia, and cardiac hypertrophy. Endocrinology 2015; 156:1577–1589.
This article described a long-lasting mouse model of hyperthyroidism, which was established by the original two to three times 3-week interval immunizations followed by repeated 4-week boots (in total, nine immunizations) with adenovirus expressing the human TSHR A-subunit.
14. Kaneda T, Honda A, Hakozaki A, et al. An improved GD model established by using in vivo electroporation exhibited long-term immunity to hyperthyroidism
in BALB/c mice. Endocrinology 2007; 148:2335–2344.
15▪▪. Rapoport B, Aliesky HA, Banuelos B, et al. A unique mouse strain that develops spontaneous, iodine-accelerated, pathogenic antibodies to the human thyrotrophin receptor. Endocrinology 2015; 194:4154–4161.
This article for the first time described a spontaneous mouse model of Graves’ like disease, in which NOD.H2h4 mice, a spontaneous model of iodine-accelerated thyroiditis, were crossed with thyroid-specific human TSHR A-subunit TG mice, and the resultant mice developed anti-TSHR antibodies with TSAb and TBIAb activities in addition to iodine-accelerated thyroiditis (anti-Tg and TPO antibodies, and intrathyroidal lymphocyte infiltration).
16. Huber AK, Finkelman FD, Li CW, et al. Genetically driven target tissue overexpression of CD40: a novel mechanism in autoimmune disease. J Immunol 2012; 189:3043–3053.
17. Ye F, Hou P, Wu X, et al. The significance of immune-related molecule expression profiles in an animal model of GD. Autoimmunity 2012; 45:143–152.
18. Cai H, Wang Z, Zhang HQ, et al. Diosgenin relieves goiter via the inhibition of thyrocyte proliferation in a mouse model of GD. Acta Pharmacologica Sinica 2014; 35:65–73.
19. McLachlan SM, Aliesky HA, Chen CR, et al. Breaking tolerance in transgenic mice expressing the human TSH receptor A-subunit: thyroiditis, epitope spreading and adjuvant as a ’double edged sword’. PLoS One 2012; 7:e43517.
20. Bahn RS. News and views: at long last, an animal model of Graves’ orbitopathy
. Endocrinology 2013; 154:2989–2991.
21. Johnson KT, Wiesweg B, Schott M, et al. Examination of orbital tissues in murine models of GD reveals expression of UCP-1 and the TSHR in retrobulbar adipose tissues. Horm Metab Res 2013; 45:401–407.
22▪. Zhao SX, Tsui S, Cheung A, et al. Orbital fibrosis in a mouse model of GD induced by genetic immunization of thyrotropin receptor
cDNA. J Endocrinol 2011; 210:369–377.
In-vivo electroporation and the plasmid encoding the human TSHR A-subunit induced the long-lasing hyperthyroidism and fibrosis in the retro-orbital regions.
23▪▪. Moshkelgosha S, So PW, Deasy N, et al. Retrobulbar inflammation, adipogenesis, and acute orbital congestion in a preclinical female mouse model of Graves’ orbitopathy
induced by thyrotropin receptor
plasmid-in vivo electroporation. Endocrinology 2013; 154:3008–3015.
By modifying the immunization method shown in the ref. , induction of the stronger immune response by deeper injection of the plasmid over a larger muscle area and in-vivo electroporation led to hypothyroidism and lymphocyte infiltration in the retro-orbital regions and eye protrusion.
24. McLachlan SM, Aliesky H, Banuelos B, et al. Immunoglobulin heavy chain variable region and major histocompatibility region genes are linked to induced GD in females from two very large families of recombinant inbred mice. Endocrinology 2014; 155:4094–4103.