Generation of a transgenic animal model of hyperthyroid Graves' disease

Abbreviations:

1 Introduction

Graves' disease (GD) is a common autoimmune disease characterized by the production of antibodies (Ab) against thyrotropin (TSH) receptor (TSHR) 1–4. Agonistic anti-TSHR Ab, or thyroid-stimulating Ab (TSAb), mimic TSH function and induce the overproduction of thyroid hormones. GD patients manifest various symptoms of hyperthyroidism including increased heart rate, excess perspiration and weight loss. TSAb can be detected in the serum of more than 90% of newly diagnosed GD patients with hyperthyroidism 3. The measurement ofTSAb in patients with GD is a useful predictor of relapse and remission 5.

A GD animal model will clarify the mechanisms underlying the development of GD including that of TSAb production. Recently, two attempts were made by immunizing mice using human TSHR: one by the injection of TSHR-transfected cells simultaneously expressing MHC class II 6; the other by genetic immunization using TSHR cDNA 7. Although these immunized mice showed some typical phenotypes associated with hyperthyroidism, the responses were temporary and TSAb were produced against non-self TSHR. The process of autoantibody production in these mice presumably lacks the persistent abrogation of self tolerance.

To circumvent the lack of an appropriate animal model for the study of GD, we undertook a different approach. In previous studies, we established a B cell clone producing a TSAb from a GD patient 8, 9. We subsequently demonstrated that this TSAb bound to TSHR and stimulated the thyroid in vitro 8, 10 and in vivo 11. These observations prompted us to generate a transgenic (Tg) mouse constitutively expressing the human TSAb. Since B cells expressing the TSAb are subjected to the host immune system, we postulated that these TSAb-Tg mice might develop immunological responses associated with autoimmunity, such as clonal deletion, anergy or the breakage of self tolerance. Indeed, we demonstrate in this report that these TSAb-Tg mice manifest many of the clinical features of hyperthyroidism and develop immunological responses against B cells that produce autoantibodies. These TSAb-Tg mice provide a novel and useful model for the study of endocrinological and immunological aspects of GD. In particular, studies concerning the disruption of self tolerance, a process which is responsible for the TSAb production, will be greatly facilitated.

2 Results

2.1 Establishment of transgenic mice expressing a patient TSAb

To express an agonistic TSAb in mice, we designed a transgene construct (Fig. 1A). A 5′ flanking region of the original human Ig (TSAb) gene 8 was used as a provisional promoter and a human Cμ segment was used for the constant region of H chain. We established two founder lines; one carried approximately ten copies of the transgene on chromosome (Ch.) 10 while the other carried ∼20 copies on a Ch. 7–Ch. 9 translocation (data not shown). As no phenotypic differences between these two lines were observed, we show combined results from both lines in the following studies.

2.3 Development of hyperthyroidism in TSAb mice

Thyroid function in TSAb-Tg mice was assessed by measuring serum free thyroxine (FT4) and TSH levels. The FT4 levels of TSAb-Tg mice were significantly higher than those of non-Tg littermates (mean ± SD: Tg: 27.5±8.0 pmol/l, n=74 vs. non-Tg: 14.9±3.9 pmol/l, n=38, p<0.001) (Fig. 2A), while the TSH levels of Tg mice (mean ± SD: 1.3±1.2 ng/ml) were lower than those of non-Tg littermates (mean ± SD: 4.4±2.2 ng/ml) (Fig. 2B). Based on the thyroid function of non-Tg mice, we arbitrarily defined the hyperthyroid status as follows: FT4 >22.7 pmol/l (mean + 2 SD) and TSH <1.4 ng/ml (mean – 2 SD, using normalized logarithmic transformation). Fifty of 74 TSAb-Tg mice (68%) met this criterion (Fig. 2C). Of them, 24 mice had undetectable levels of TSH (<0.1 ng/ml). Thirteen mice that displayed increased FT4 but normal TSH levels. In these mice, the duration or degree of the increased FT4 levels might not be sufficient to suppress TSH levels. In the majority of TSAb-Tg mice examined, we observed positive correlations between serum hIgM and FT4 levels (Fig. 2D). These results indicate that the hIgM (TSAb) is indeed pathogenic, and that its level determines the severity of hyperthyroidism in the Tg mice.

2.5 Increased thyroid technetium uptake by hyperthyroid Tg mice

Technetium pertechnetate (TPT) scintigraphy was performed to confirm that the hyperthyroidism in the TSAb-Tg mouse was caused by hyperactivity of the thyroid gland, not by destructive thyroiditis. The pertechnetate ion is transported into thyroid tissue by an iodide-concentrating mechanism but has a shorter half-life (6 h) than other radioiodines (¹³¹I: 8 days; ¹²³I: 13 h). We injected mice with TPT via the tail vein and assessed the time course of TPT transport into the thyroid. In non-Tg mice, TPT was predominantly found in the stomach though trace amounts were present in the thyroid (Fig. 3A). Thyroidal TPT accumulation slightly increased up to 9% (mean ± SD: 7.8±0.64%) and saturated at 90–120 min (Fig. 3B) post-injection. In contrast, hyperthyroid TSAb-Tg mice accumulated TPT up to significantly higher levels (mean ± SD: 16.5±3.1%) and sustained this for at least 150 min. These results demonstrate that iodide transport is substantially increased in the hyperthyroid TSAb-Tg mice.

2.7 Clonal deletion of TSAb-positive B cells in peripheral lymphoid organs

As described earlier, B cell numbers in the TSAb-Tg mice were much lower than expected, suggesting that autoreactive TSAb-positive B cells might have been partially eliminated by clonal deletion. To test this hypothesis, B cells from bone marrow (BM), spleen (SP) and peritoneal cavity (PerC) of TSAb-Tg mice were analyzed. The percentages of hIgM⁺/hIgκ⁺ B cells in the whole lymphoid cells from the SP and BM of Tg mice were similar to those in the peripheral blood lymphocytes (PBL) (Fig. 4A). Total B cell numbers decreased in the SP and BM as well as in PBL (Fig. 4B). Statistical analysis confirmed that the total B cell numbers in TSAb-Tg PBL were significantly decreased compared with that of non-Tg littermates. However, percentages of hIgM⁺ B cells in the PBL and PerC of TSAb-Tg mice varied among mice. These results suggest that clonal deletion of TSAb-bearing B cells may be occurring in the TSAb-Tg mice.

3 Discussion

In this study, we generated hyperthyroidism in mice by expressing a patient-derived anti-TSHR Ab as a transgene. We were particularly interested in determining whether this strategy would evoke in mice similar autoimmune responses to those that are commonly seen in humans with GD. The human TSAb was successfully expressed in our Tg mice, and stimulated thyroid cells to produce excessive amounts of thyroid hormones (FT4). Consequently, these TSAb-Tg mice developed a series of hyperthyroidism symptoms typically seen in GD, which include increased serum FT4 levels concomitant with suppressed or decreased TSH levels, hyperthermia, hyperactivity, and an increase in body temperature. In addition, thyroid cells in these mice showed increased technetium uptake and displayed hyperplastic changes, all of which are typically seen in GD.

We took a new approach in generating a GD animal model. In the previous models, mice were immunized with human TSHR, which resulted in a transient autoimmune response. The non-persistent nature of the autoimmune response in the immunized mice is, supposedly, a consequence of the transient expression of the nominal Ag. However, the continuous presence of the Ag could induce tolerance, rather than leading to a persistent autoimmune response. To circumvent this problem, we persistently expressed the autoantibody, derived from a GD patient, in our Tg mouse. Indeed, many of the pathological and clinical changes associated with GD appeared in our TSAb-Tg mice. This is the first mouse model in which the constitutive expression of a human autoantibody has resulted in the successful development of an autoimmune disease.

Using this mouse model, we investigated the immunological factors which influence the development of hyperthyroidism in mice. The serum FT4 levels, a direct indicator of hyperthyroidism, strongly correlated with the serum hIgM levels. Nevertheless, the numbers of total B cells in PBL, SP and BM were significantly lower in the Tg mice compared to those of non-Tg littermates. Moreover, the number of Tg B cells variably decreased in the PBL, BM and SP in Tg mice and did not show correlation with serum hIgM or FT4 levels. These pieces of evidence suggest that a deletion process against the self-reactive B cells took place but the deletion of Tg B cells in the periphery did not abrogate the production of autoreactive TSAb in the Tg mice. Interestingly, we observed that the number of total B cells dramatically increased in the PerC of the Tg mice and that many of those B cells expressed hIgM (TSAb). Collectively, these findings suggest that B cells in PerC, rather than those in the periphery, are implicated in the development of GD in the Tg mice.

Similar observation was obtained from studies using a Tg model of autoimmune hemolytic anemia (AIHA), in which the mouse anti-erythrocyte autoantibody was expressed as a transgene 13, 14. Using this model, Honjo and co-workers demonstrated the importance of a subset of B cells, namely B1 cells 15, 16, in the development of AIHA. These investigators further demonstrated that B1 cells which had survived clonal deletion accumulated in the PerC, and that these B1 cells could produce autoantibodies following LPS, IL-5 or IL-10 administration 17, 18. In contrast, conventional B cells, even in the PerC as well as in the periphery, were deleted in this model. These findings, combined with our own observation, lead us to speculate that B cells in PerC can produce the TSAb, causing the development of hyperthyroidism in our Tg mice.

Administration of LPS into the Tg mice resulted in the elevated serum FT4 levels. Consistently, clinical studies demonstrated that the increase of CD5⁺ B cells, though in the PBL compartment, correlates well with the disease activity in GD patients 19, 20. Interestingly, this fact correlates with the data that the number of CD5⁺ B cells in the PerC tends to decrease; on the other hand, those in PB tend to increase when hyperthyroidism was induced by LPS administration. However, we did not determine the number of plasma cells, a population presumably secreting the TSAb, in these compartments (i.e. periphery and PerC). Moreover, a role of T cells has not yet been studied in the present model. In addition to B cell dysregulation, intrinsic T cell dysfunction may be involved in chronic autoimmunity 21, 22. Therefore, it is not clear how the present model reflects T cell help and dependency of the chronic Ab production on innate stimuli in the autoimmune response. For example, T cell-driven B2 differentiation into memory plasma cells has not been ruled out as a possible pathological mechanism in GD. Thus future investigation is needed to address these issues.

We noticed one difference between the AIHA and TSAb-Tg mice in addition to the similar accumulation of B1 cells in PerC and the deletion of peripheral B cells. Both B1 and B2 cells in TSAb-Tg mice increased in number in the PerC (data not shown), whereas only B1 cells but not B2 cells increased in AIHA mice. This difference may be associated with the distribution of the nominal Ag in each system. TSHR is widely expressed by the adipose tissue, thymus, kidney, heart and brain, in addition to the thyroid 23, 24, while erythrocytes are seen only in the blood. Such difference in Ag distribution could supposedly lead to different consequences in the development of autoimmunity in each animal model because autoreactive B cells might encounter their Ag at different locations and/or stages in their development. However, both animal models manifested the development of autoimmune responses, suggesting that the distribution of Ag did not meaningfully affect the onset of the immune responses in these animal models.

Most TSAb detected in patients with GD belong to the IgG class 1, 25. In fact, clone B6B7, which was isolated from PBL of a GD patient, is originally the IgG clone and has a significant number of somatic mutations in V region genes of H and L chains, indicating the involvement of somatic mutations for the TSAb specificity 8, 9. Although the isotype was changed from IgG to IgM in this TSAb-Tg mouse, this would not influence the binding affinity to TSHR or TSAb activity per se. Furthermore, we found IgM TSAb from PBL of GD patients and indicated the affinity maturation of TSAb driven by Ag in IgM-producing lymphocytes 8, 26. The isotype alteration, however, mightaffect on effector functions, i.e. complement fixation. This issue should be further investigated in the future study.

Clinical studies implicated the involvement of environmental factors in the development of GD in humans. We observed, as noted above, that an oral administration of LPS into the Tg mice causedan elevation of serum FT4 levels suggesting that the activation of B cells in PerC may augment the TSAb production in the Tg mice. Furthermore, this observation implies that certain bacterial infection in humans, through the activation of B1 cells in PerC, may episodically trigger or augment autoantibody production in human GD patients.

In summary, we developed a Tg mouse model of GD. These mice should provide us with unique opportunities not only to study the pathogenesis of GD but also to develop a new treatment strategy for this autoimmune disease.

4 Materials and methods