Chronic exposure in vivo to thyrotropin receptor stimulating monoclonal antibodies sustains high thyroxine levels and thyroid hyperplasia in thyroid autoimmunity-prone HLA-DRB1*0301 transgenic mice

Introduction

Some of the most common autoimmune conditions affecting humans belong to thyroid autoimmune disorders, which include Graves' disease and Hashimoto's destructive thyroiditis.¹ In Graves' disease, the hyperthyroidism is directly caused by thyroid-stimulating autoantibodies (TSAbs) to the thyrotropin receptor (TSHR) that mimic the effects of the hormone TSH to hyperstimulate thyroid function.² The thyroid glands in patients with Graves' disease are enlarged as a result of thyroid hyperplasia and some show foci of intrathyroidal inflammatory infiltrates, including well-maintained germinal centres with abundant plasma B cells.³ Another cardinal feature of patients with Graves' disease is the presence of autoantibodies to other thyroid antigens, such as thyroid peroxidase (TPO) and thyroglobulin (Tg), which are commonly associated with Hashimoto's thyroiditis.¹ The reasons for the recruitment of TPO and Tg as autoantigens in patients with Graves' disease are not completely understood, but it is believed that the intense and continuous stimulation of the gland may result in leakage of these potential antigens into the periphery.¹ For example, Tg is a by-product of thyroxine (T₄) release and we have shown a three- to five-fold increase of circulatory Tg after 3 days of TSH infusion via an osmotic pump, coincident with T₄ release that peaked within 24 hr.⁴ There is also strong genetic susceptibility to disease development, contributed predominantly by major histocompatibility complex (MHC) antigens, cytotoxic T-lymphocyte-associated antigen-4 and the protein tyrosine phosphatase 22 (PTPN22) gene.⁵ Amongst the MHC genes, the HLA-DRB1*0301 (DR3) class II genes have been recognized as important susceptibility elements for disease in patients with thyroid autoimmune disease.⁶

One seldom recognized feature of Graves' disease is that there are no animal species other than humans that develop thyroid-stimulating antibodies. Various protocols to induce experimental Graves' disease have been reported over the years and these have led to improved models in terms of potency, disease incidence and severity of hyperthyroidism.⁷ We recently reported that humanized, transgenic mice expressing human leucocyte antigen (HLA) DR3 in the absence of endogenous class II molecules on a non-obese diabetic (NOD) background were more prone to human Tg-induced autoimmune thyroiditis than the wild-type NOD mouse.⁸ Moreover, the DR3 transgenic animals were also susceptible to experimental Graves' hyperthyroidism, which was accompanied by lymphocytic infiltration and antibodies to Tg in some animals.⁹ Interestingly, in other studies using HLA-DRB1*0301 transgenic mice on a C57BL/10 background, no hyperthyroidism or thyroid inflammation developed,¹⁰ confirming the importance of the MHC class II locus together with other autoimmune susceptibility genes for disease.

We recently generated thyroid-stimulating monoclonal antibodies (mAbs) from a mouse model of Graves' disease, which powerfully stimulate the TSHR and act as full agonists, akin to the ligand TSH.¹¹ The two mAbs KSAb1 and KSAb2 are highly potent and act as full agonists to the TSHR at low nanomolar concentrations, but also exhibit important differences in their EC₅₀ values at lower doses.¹¹ In addition, different studies in an acute setting have demonstrated that single injections of low microgram quantities of anti-TSHR mAbs in mice induce rapid elevation of serum T₄ levels, which peak around 24 hr before returning to normal levels by 48–72 hr.^11–13 In the present study, using two stimulatory mAbs, we investigated the effect of chronic stimulation of the TSHR on thyroid function. In addition, we investigated whether chronic stimulation of the thyroid gland over a prolonged period of 9 weeks would provoke autoimmune responses, similar to those observed in patients with Graves' disease. We performed the chronic exposure study in the DR3 transgenic Ab⁰/NOD mice to include the necessary genetic susceptibility elements to provoke potential autoimmunity.

Materials and methods

Results

Chronic stimulation of TSHR in vivo and thyroid function

In in vitro studies, KSAb1 and KSAb2 IgG stimulate cAMP production in human TSHR-transfected Chinese hamster ovary cells (JP09 cells) by showing full agonist activity, with EC₅₀ values of 9·4 and 93 ng/ml, respectively.¹¹ For passive transfer studies, we selected a dose of 10 μg for both KSAb1 and KSAb2 mAbs (greater than an EC₅₀ dose of 1000-fold and 100-fold, respectively). In this manner, we ensured an adequate dose of the agonist mAbs in the animals, which exceeded by three and two log folds for KSAb1 and KSAb2, respectively. To maintain the chronicity of TSHR stimulation of the gland, the injections were repeated every week for 9 weeks and serial blood samples were collected 1 day before and 1 day after alternate injections with the mAbs. Serum T₄ levels were measured in all the serial blood samples as an indicator of the hyperthyroid status. After the first intravenous injection of mAb, a robust surge in serum T₄ was observed which peaked on day 1. Serum T₄ levels were elevated by approximately 2·5- and 4-fold in KSAb1- and KSAb2-challenged mice, respectively (shown as means in Fig. 1a,b, respectively). Following the first injection, serum T₄ levels subsequently declined (day 6, Fig. 1a,b), but remained elevated compared with the basal levels for both the mAb-treated animal groups (Fig. 1a,b). Analysis of serum T₄ levels 1 day after the second injection on day 8 revealed a marginal increase in the mean values for both KSAb1- and KSAb2-injected mice (Fig. 1a,b). Subsequent injections of KSAb1 or KSAb2, particularly on day 21 and day 35, when blood samples were also collected before and after the injections, confirmed the continued elevated levels of serum T₄ levels. Interestingly, the overall trend for sustained hyperthyroidism was maintained (Fig. 1a,b). Continued weekly injections of the mAbs, for example, on day 49, showed no dramatic differences in response of the glands to TSHR stimulation where serum T₄ levels stabilized and plateaued in both KSAb1-treated and KSAb2-treated animals (Fig. 1a,b). Finally, analysis of serum samples on day 63 following 9 weeks of sustained stimulation indicated that hyperthyroidism was maintained, compared with the basal serum T₄ levels before the start of the injection schedule (Fig. 1a,b). For the duration of the experiment, the animals did not show any sign of discomfort or weight loss. To summarize, chronic in vivo exposure of mice to continued weekly injections of KSAb1 or KSAb2 led to rapid onset of hyperthyroidism, which was maintained throughout the course of antibody-induced stimulation of the thyroid gland.

Clearance of KSAb1 and KSAb2 IgG following passive transfer

Following the weekly passive transfer injections of KSAb1 and KSAb2, we investigated the residual levels of the mAbs in the animals by assessing the biopotency of the serum in terms of cAMP stimulatory activity in JP09 cells. The results are shown in Fig. 2. After the injection on day 0, a surge in the levels of serum KSAb1 and KSAb2 was apparent on day 1. A small reduction in the potency of the mAbs following the first injection, for example on day 6, indicated clearance (or metabolism) of the mAbs from the circulation (Fig. 2). Subsequent injections for KSAb1 and KSAb2 showed different rates of accumulation and clearance from the circulation. Thus, the rate of clearance of KSAb1 was more pronounced than that observed with KSAb2, such that by day 63 the levels of residual KSAb1 activity appeared substantially different from those for activity of KSAb2. Interestingly, the slow clearance rates of KSAb2 (IgG2a) compared to KSAb1 (IgG2b) correlated with the known half-lives of their subclasses.¹⁷ Nevertheless, despite the accumulation of KSAb2 in the serum of the mice (Fig. 1), there were no apparent differences in the hyperthyroid status of the animals when compared to KSAb1-challenged animals.

Induction of autoimmune response and thyroid gland histology

Sustained stimulation of the thyroid gland by TSHR-stimulating antibodies may lead to the induction of an autoimmune response to thyroid antigens such as Tg and TPO, a cardinal feature of patients with active Graves' disease. We tested the day 65 serum samples for antibodies to mouse Tg by ELISA. As shown in Table 1, compared to sera from control, age-matched DR3⁺ Ab⁰/NOD mice, only one mouse (KSAb2 #3) was positive for mTg antibodies. Examination of the thyroid glands from KSAb1-challenged and KSAb2-challenged animals revealed grossly enlarged glands with vascular congestion compared to control animals (Fig. 3a,b). Histological analysis of all the glands from mAb-treated mice revealed enlarged follicles with hyperplastic papillary infolding of the epithelium (Fig. 3c,d). However, none of the thyroid sections demonstrated any significant inflammatory cell infiltrate above background.

Table 1. Serum anti-mTg titres and thyroid pathology in HLA-DR3 transgenic mice after repeated injections of anti-TSHR mAbs

Anti-TSHR mAb	Mouse no.	Anti-mTg titres			Thyroiditis % thyroid infiltration
		Serum dilution (O.D.)
		1 : 5	1 : 10	1 : 20
KSAb1	1	0·26	0·22	0·17	>0–10
	2	0·23	0·18	0·13	>0–10
	3	0·19	0·16	0·15	0
KSAb2	1	0·28	0·22	0·17	0
	2	0·26	0·22	0·17	0
	3	0·70	0·47	0·34	0
Control	1	0·21	0·17	0·13	0
	2	0·26	0·24	0·20	>0–10
	3	0·15	0·12	0·09	0
	4	0·14	0·11	0·08	>0–10
	5	0·16	0·13	0·10	0
	6	0·23	0·18	0·13	0

• Male (KSAb1) and female (KSAb2) HLA-DR3 transgenic Ab⁰/NOD mice were injected intravenously with 10 μg anti-TSHR mAb every week for 8 weeks. On day 65, mice were killed, and serum antibody titres to mTg and thyroid pathology were determined. Control mice were unmanipulated male and female DR3⁺ Ab⁰/NOD mice.

Discussion

The development of mAbs with strong thyroid-stimulating properties has led to studies evaluating their pathogenic mechanisms in vivo by passive transfer to naive mice.^11–13 Using two mAbs with powerful agonistic properties to the TSHR, our data in animals subject to chronic stimulation of the TSHR show that the thyroid gland continues to respond by secreting thyroid hormone, similar to the clinical situation in Graves' disease patients. Measurement of the residual bioactivity of the mAbs in the serum of the injected animals showed the presence of the stimulating antibodies throughout the period of chronic exposure, confirming sustained stimulation of the gland during the course of the study.

The isotypes of KSAb1 (IgG2b) and KSAb2 (IgG2a) are different.¹¹ Analysis of the decay of biopotency of KSAb1 and KSAb2 shows that the residual levels of KSAb2 were generally maintained at higher levels than KSAb1. This was particularly pronounced after several weeks of repeated injections, where the levels of KSAb2 appeared to plateau compared to KSAb1, where the levels continued to fluctuate. These findings are in accordance with the known half-lives of mouse IgG subclasses in serum, which have been determined to be 4–6 days (IgG2b) and 6–8 days (IgG2a).¹⁷ Thus, one explanation for the observed heterogeneity of the induced experimental Graves' disease models in inbred strains of mice⁷ may be related to the specific subclasses of induced antibodies with TSAb activity, where their half-lives determine the degree of hyperthyroidism present in the different animals. In patients with Graves' disease, such a scenario may not be relevant because almost all TSAb activity is known to reside in the IgG1 subclass.¹⁸ On the other hand, in Graves' patients, the clearance rate would be offset by continuing production of TSAb to maintain a persistent level of stimulation. It is not inconceivable that continuous stimulation of the TSHR could help promote the autoimmune process leading to thyroid infiltration as seen in some patients, although our attempt to provide chronic stimulation did not result in autoimmune thyroiditis in these mice.

There are a few studies that have examined the in vivo effects of chronic stimulation of TSHR with TSH or antibodies with contradictory results. Our earlier study⁴ and those of others^19,20 have examined the chronic stimulation of mice (and rats) with osmotic pumps or regular injections of TSH, respectively. Thus, a rapid decline in serum T₄ levels in response to further stimulation with TSH, suggesting desensitization of the gland, was noted in some studies^4,19 but not in others.²⁰ Similarly, stimulation of the thyroid gland in vivo with antibodies in both acute and chronic situations has also given mixed results.^11–13,21 In a recent chronic exposure study¹³ immunocompromised mice treated with pristane to grow a hamster-derived hybridoma as ascites secreting TSAb over a 2-week period resulted in refractoriness of the gland, with serum T₄ levels in the treated mice comparable to those in the preimmune controls. This in vivo protocol is significantly different from the procedure described in this report, which used a measured bolus of purified mAb for the stimulating protocol. Other studies using thyroid-stimulating mAbs in passive transfer by single injections to mimic acute stimulatory conditions have also given contradictory results.^11–13 In two different studies, including our study with KSAb1 and KSAb2 mAbs, passive transfer of microgram quantities of the stimulatory mAbs into BALB/c mice led to rapid onset of hyperthyroidism.^11,12 The elevated serum T₄ levels were maintained following stimulation of the receptor by the TSAbs for a number of days in our study at a dose that was two to three log fold greater than the EC₅₀ value of the stimulatory concentration of the mAbs.^11,12 It can be argued that single serum T₄ measurements may not accurately reflect the degree of hyperthyroidism in vivo as a variety of factors, such as T₄ half-life and frequent blood drawings may affect the readings. It is for this reason that the animals were bled every alternate week following administration of the stimulating mAbs. Other studies have examined the effect of stimulating antibodies at an acute level. For example, the acute passive transfer study of the hamster TSAb mAb injected into CBA/J mice showed that serum T₄ levels were elevated within 24 hr, followed subsequently by significant decrease in T₄ levels.¹³ The results from this study indicated antibody-mediated receptor desensitization as a potential mechanism, which was also supported in part by in vitro studies. Nevertheless, the possibility remains that the differences in these acute study models may be the result of the clearance rates of the murine-derived and hamster-derived mAbs in mice or perhaps the two log difference in the K_d of the thyroid-stimulating mAbs.^11–13

Acute stimulation studies by passive transfer in vivo of thyroid-stimulating mAbs have also given contrasting results on thyroid inflammation. In our recent study with single injections of KSAb1 and KSAb2 into BALB/c mice, we failed to trigger an inflammatory response to the thyroid gland¹¹ but contrasting results were reported in another study.¹² In this report, sustained stimulation of the TSHR in autoimmunity-prone, DR3 transgenic animals did not provoke an inflammatory response because the thyroid glands did not show any signs of mononuclear cell infiltration over the background level that we had observed in DR3⁺ Ab⁰/NOD mice.⁹ Nor did we observe antibodies to Tg except in one of the six mice. It is clear, however, that regulatory influences to suppress autoimmune responses persist despite prolonged stimulation of the thyroid. The possibility remains that more aggressive and sustained exposure to the stimulating mAbs may trigger higher levels of anti-Tg antibodies and perhaps thyroid inflammation, as is present in patients with Graves' disease.

In summary, we provide evidence for sustained in vivo stimulation of the TSHR by passive transfer of potent mAbs with thyroid-stimulating activity. The possibility exists that there may be desensitization of the gland following repeated stimulation of the receptor with the TSAbs, although we observed stimulation. We did not examine in vivo desensitization of the thyroid gland, but the overall results show that, if desensitization was present, the phenotype observed was dominated by stimulation to maintain the serum T₄ levels. Since only stimulating antibodies were given, there were no blocking antibodies to provide competition, as would be present in experimentally induced models or patients with Graves' disease. Thus, thyrotoxicosis as maintained in the animals during the period of chronic exposure could help explain continuing hyperthyroidism in patients. Understanding the mechanisms of the dominance of stimulation by chronic exposure in vivo with these TSAbs can provide a greater insight into the maintenance of thyrotoxicosis in patients with Graves' disease.

Acknowledgments

We are grateful to Julie Hanson and her staff for breeding the mice and to Ann Mazurco for excellent histological preparations. This work was supported in part by DK45960 from NIDDK and St John Hospital (to Y.M.K.) and by a Training grant (to J.A.G.).