Thyrotropin Receptor Autoantibodies Are Associated with Continued ...

0021-972X/03/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 88(9):4135– 4138 Copyright © 2003 by The Endocrine Society doi: 10.1210/jc.2003-030430

Thyrotropin Receptor Autoantibodies Are Associated with Continued Thyrotropin Suppression in Treated Euthyroid Graves’ Disease Patients LEON J. S. BROKKEN, WILMAR M. WIERSINGA,

AND

MARK F. PRUMMEL

Department of Endocrinology and Metabolism, University of Amsterdam, Academic Medical Center, 1105 AZ Amsterdam, The Netherlands Antithyroid treatment effectively restores euthyroidism in patients with Graves’ hyperthyroidism. After a few months of treatment, patients are clinically euthyroid with normal levels of thyroid hormones, but in many patients TSH levels remain suppressed. We postulated that TSH receptor autoantibodies could directly suppress TSH secretion, independently from thyroid hormone levels, via binding to the pituitary TSH receptor. To test this hypothesis, we prospectively followed 45 patients with Graves’ hyperthyroidism who were treated with antithyroid drugs. Three months after reaching euthyroidism, blood was drawn for the analysis of thyroid hormones, TSH, and TSH binding inhibitory Ig (TBII) levels. After 6.7 ⴞ

G

RAVES’ HYPERTHYROIDISM IS caused by stimulating autoantibodies directed against the TSH receptor (TSH-R). Treatment with antithyroid drugs renders most patients euthyroid within 4 – 6 wk as manifested by normal serum free T4 (fT4) and total T3 concentrations. Nevertheless, TSH often remains suppressed for a much longer period (1). This is classically attributed to a delayed recovery of the pituitary-thyroid axis from prolonged thyroid hormone excess (2). Recently we postulated a different mechanism to explain this phenomenon of sometimes long-lasting suppression of TSH values in otherwise euthyroid patients with treated Graves’ hyperthyroidism, involving a pituitary TSH-R (3, 4). We demonstrated that pituitary folliculo-stellate cells express the TSH-R and postulated that these cells could be involved in an ultra-short loop feedback control on TSH secretion (3). In this hypothesis, TSH secreted by the thyrotrophs binds to the TSH-R expressed on neighboring folliculo-stellate cells. These cells then secrete a messenger (e.g. a cytokine), which then suppresses TSH secretion by thyrotrophs in a paracrine way. Such a TSH-R on the folliculo-stellate cells would also be recognizable by circulating TSH-R-stimulating Igs (TSI), because the pituitary resides outside the blood-brain barrier. These stimulating autoantibodies would then—just like TSH itself— down-regulate TSH secretion. This of course will go unnoticed during the hyperthyroid state but will become apparent when the patient is rendered euthyroid but still has circulating TSH-R autoantibodies. This hypothesis was tested first in an animal model. Rats Abbreviations: FT3I, Free T3 index; fT3, free T3 fT4, free T4; TBII, TSH binding inhibitory Ig; TSH-R, TSH receptor; TSI, TSH-R-stimulating Igs.

1.5 months since start of antithyroid treatment, 20 patients still had detectable TBII levels, and 25 had become TBII negative. The two groups had similar levels of free T4 and T3, but TBII-positive patients had lower TSH values than TBIInegative patients: median 0.09 (range < 0.01– 4.30) mU/liter vs. 0.84 (0.01– 4.20; P ⴝ 0.015). In addition, TSH levels correlated only with TBII titers (r ⴝ ⴚ0.424; P ⴝ 0.004), and not with free T4 or T3 values. Our findings suggest that TBII suppress TSH secretion independently of thyroid hormone levels, most likely by binding to the pituitary TSH receptor. (J Clin Endocrinol Metab 88: 4135– 4138, 2003)

were chemically thyroidectomized using methimazole and kept euthyroid by adding T4 to their diet. They were then injected with either TSH-R autoantibody containing IgGs or control IgG. In agreement with our hypothesis, the TSI containing IgGs (and not the control IgGs) suppressed the TSH values without having an effect on serum levels of T4 or T3 (4). Here we test this hypothesis in humans by following 45 patients with Graves’ hyperthyroidism treated with antithyroid drugs. When stable euthyroidism in terms of normal values for fT4 and T3 was reached, TSH values were related to thyroid hormone levels and TSH binding inhibitory Ig (TBII) titers. Patients and Methods Patients We performed a prospective clinical study in 45 consecutive patients with newly diagnosed Graves’ hyperthyroidism. This diagnosis was based on elevated levels of fT4 (⬎1.8 ng/dl [⬎23.0 pmol/liter]) and/or total T3 (⬎179 ng/dl [⬎2.75 nmol/liter]) in the presence of a decreased TSH (⬍0.4 mU/liter), positive TBII titer (⬎12 U/liter), and a diffuse uptake on a technetium scintigram. Excluded were patients with serious concomitant diseases, pregnancy, or on drugs known to influence the pituitary-thyroidal axis. Forty-one patients were treated with 30 mg methimazole, and four with 400 mg propylthiouracil daily, to which was added l-T4 (109 ⫾ 36 ␮g), aiming at normalizing fT4 (0.8 –1.9 ng/dl [10.0 –25.0 pmol/liter]) and total T3 (80 –179 ng/dl [1.20 –2.75 nmol/liter]) but avoiding elevated TSH values (⬎4.0 mU/liter). When the patients were clinically and biochemically euthyroid for at least 3 months, their TBII levels were again determined and related to the levels of thyroid hormones and TSH. Informed consent was obtained from all patients, and the study protocol was approved by the Medical Ethics Committee of the Academic Medical Center in Amsterdam.

4135

4136

J Clin Endocrinol Metab, September 2003, 88(9):4135– 4138

Hormone assays

Brokken et al. • TBII Associated with Long-Time TSH Suppression

Discussion

TSH plasma levels were measured with a highly sensitive chemiluminescent enzyme immunoassay that has a functional sensitivity of 0.01 mU/liter (Immulite Third Generation TSH kit, Diagnostic Products Corp., Los Angeles, CA). fT4 levels were determined with a solid phase time-resolved fluoroimmunoassay (Delfia, Wallac Oy, Turku, Finland). Total T3 plasma levels were determined by in-house RIA (5). To correct for effects of oral contraceptives on total T3 levels, the free T3 index (FT3I) was calculated as the product of total T3 and T3 resin uptake. The latter was determined with a T3 Uptake Kit (Ortho-Clinical Diagnostics, Amersham, Buckinghamshire, UK). TBII titers were measured by TSHRezeptor Antiko¨ rper assay (Brahms Diagnostica, Berlin, Germany).

Statistical analysis Mann-Whitney U test was used to compare patients with negative TBII to patients with positive TBII with respect to fT4, FT3I, TSH, and several demographic parameters. We then calculated the correlation of TBII titers with thyroid hormone and TSH levels using nonparametric two-tailed Spearman’s Rho correlation. TSH values less than 0.01 mU/ liter were substituted by 0.005 mU/liter in the statistical analysis.

Results

Thyroid function tests before treatment and after stable restoration of the euthyroid state are given in Table 1. Mean age ⫾ sd of the total group was 38 ⫾ 12 yr; female/male ratio was 37/8; median (range) duration of the thyroid disease was 6 (1–120) months. After 6.7 ⫾ 1.5 months of treatment with antithyroid drugs and l-T4, stable euthyroidism was restored in all patients. At that moment, 20 patients still had a positive TBII titer, and 25 had become TBII negative. fT4 and FT3I levels did not differ between patients with positive TBII and those with negative values. However, patients in the TBIIpositive group had significantly lower serum TSH levels than those in the TBII-negative group [median (range) 0.09 (⬍0.01– 4.30) mU/liter vs. 0.84 (0.01– 4.20) mU/liter, respectively; P ⫽ 0.015 with Mann-Whitney U; Fig. 1, A–C]. Age, sex ratio, duration of the thyroid disease, baseline hormone levels, and the l-T4 dose used by the patients were similar in both groups. Baseline TBII levels were significantly higher in the group that remained TBII positive after treatment, compared with the group that became TBII negative (Table 2). There was a strong, negative correlation between TBII titers and TSH levels (Spearman correlation coefficient r ⫽ ⫺0.424; P ⫽ 0.004; Fig. 1D), whereas there was no correlation between TBII and fT4, FT3I, duration of the thyroid disease, or the l-T4 dose used by the patients. Similarly, no correlation was found between TSH levels and the latter parameters (Table 3).

This study once again shows that TSH levels can remain suppressed in many patients treated for Graves’ hyperthyroidism, despite normal levels of fT4 and T3. This has always been attributed to a delayed recovery of the pituitary-thyroid axis after prolonged hyperthyroidism (2). Our results suggest an alternative explanation because we found that there was a significant negative correlation between TSH levels and TBII titers after approximately 7 months of antithyroid drug therapy (Fig. 1D). It thus seems that TBII are capable of suppressing TSH secretion, independently from thyroid hormone levels, most likely by binding to the TSH-R expressed on folliculo-stellate cells in the pituitary (3). In physiological circumstances, this pituitary TSH-R is probably involved in ultra-short loop negative feed back control on TSH secretion, but in Graves’ disease binding of TBII would also lead to suppression of TSH secretion, as suggested by our results. This explanation for the observed low levels of TSH in patients otherwise euthyroid during antithyroid treatment seems more plausible than to attribute it only to a delayed recovery of the pituitary-thyroid axis, for a number of reasons. First, patients with Graves’ hyperthyroidism do not in general have a long history of thyrotoxicosis. The delay between onset of symptoms and start of treatment in this study was only 4 – 6 months, and the duration of disease was not correlated to the observed TSH levels after 7 months of therapy. Secondly, in other cases of thyrotoxicosis, most notably T4 suppressive therapy in patients with papillary or follicular thyroid cancer, TSH becomes clearly elevated in a number of weeks after discontinuation of T4 treatment. Thirdly, our hypothesis explains why some patients with treated Graves’ hyperthyroidism attain normal TSH values in a couple of months and others do not. For, in approximately half of the patients TBII levels will become negative after a number of months, whereas in the other half of the patients they remain positive. The patients who remain positive for TBII have a higher relapse rate after discontinuation of antithyroid drug treatment, presumably because of the persistence of TBII. Other independent risk factors for a relapse of hyperthyroidism after a course of antithyroid drugs for Graves’ hyperthyroidism are persistent goiter, and the presence of low TSH values, even in the absence of detectable TBII. We therefore propose that the persistence of low TSH values can be seen as evidence for still circulating TSH-R-stimulating autoantibodies. Our results also support the functional relevance of the

TABLE 1. Thyroid function tests of 45 consecutive patients with Graves’ hyperthyroidism before treatment and after 6.7 ⫾ 1.5 months of antithyroid drug treatment, when all had reached a stable euthyroid state (values as median with range)a

TSH (mU/liter) TBII (U/liter) Total T4 (␮g/dl) Total T3 (ng/dl) fT4 (ng/dl) FT3I

Before treatment

After treatment

Ref. range

⬍0.01 (⬍0.01– 0.19) 29 (6 – 400) 19.4 (6.6 –29.5) 370 (172–740) 4.4 (0.9 –5.8) 428 (215–977)

0.37 (⬍0.01– 4.30) 9 (3–278) 9.6 (3.5–17.9) 133 (81–237) 1.1 (0.4 –2.0) 127 (78 –185)

0.4 – 4.0 ⱕ12 5.4 –11.7 80 –179 0.8 –1.9 80 –179

a To convert to Syste`me Internationale units, multiply total T4 and fT4 values by 12.87 (to nmol/liter and pmol/liter, respectively), and total T3 and FT3I by 0.0154 (to nmol/liter and Index, respectively).


J Clin Endocrinol Metab, September 2003, 88(9):4135– 4138 4137

FIG. 1. A–C, Plasma TSH, fT4, and FT3I levels in patients with Graves’ hyperthyroidism after 6.7 ⫾ 1.5 months of antithyroid drug treatment, grouped according to the presence or absence of circulating TBII. Horizontal continuous lines indicate medians. Horizontal dashed lines indicate the normal reference range. TSH concentrations below the detection limit (0.01 mU/liter) are represented as 0.005 mU/liter. To convert to metric units, divide the fT4 and FT3I values by 12.87 (to ng/dl) and 0.0154, respectively. D, Correlation between serum TSH values and TBII titers in all 45 patients after 6.7 ⫾ 1.5 months of antithyroid drug treatment. TABLE 2. Baseline data and treatment characteristics (median and range) in patients with Graves’ hyperthyroidism who were TBII negative or TBII positive after 6.7 ⫾ 1.5 months of antithyroid drug treatmenta TBII negative

TBII positive

P value

Baseline data n Age (yr) Sex (F/M) Duration of disease (months) TSH (mU/liter) TBII (U/liter) Total T4 (␮g/dl) Total T3 (ng/dl) fT4 (ng/dl) FT3I

25 31 (21–58) 19/6 4 (2–24) ⬍0.01 (⬍0.01– 0.04) 14 (6 –114) 18.3 (6.6 –23.7) 341 (172– 633) 4.0 (1.2–5.8) 401 (215– 868)

20 41 (21–71) 18/2 6 (1–120) ⬍0.01 (⬍0.01– 0.19) 52 (19 – 400) 20.2 (8.2–29.5) 396 (208 –740) 5.3 (0.9 –5.8) 455 (218 –977)

0.073 0.227 0.437 0.279 ⬍0.001 0.209 0.322 0.310 0.217

Treatment characteristics Methimazole ⫹ L-T4 (n) Propylthiouracil ⫹ L-T4 (n) L-T4 dose (␮g/d) TBII after treatment (U/liter) Duration of treatment (months)

23 2 100 (0 –200) 5 (3–12) 6 (4 –9)

18 2 100 (75–175) 30 (17–278) 7 (5–10)

0.315 0.000 0.139

a To convert to Syste`me Internationale units, multiply total T4 and fT4 by 12.87 (to nmol/liter and pmol/liter, respectively), and total T3 and FT3I by 0.0154 (to nmol/liter and Index, respectively).

pituitary TSH-R. We and others (3, 6) found the TSH-R expressed on folliculo-stellate cells, which have always been thought to be involved in the paracrine regulation of pituitary hormone secretion (7, 8). The presence of the TSH-R on these cells makes such a role more plausible, and we suggest

that they might also be involved in the pulsatility of TSH secretion. Such an ultra-short loop control is probably not limited to TSH secretion, because other receptors for pituitary hormones, like the GH receptor and the prolactin receptor, have also been found in the pituitary (9 –11). In fact,

4138

J Clin Endocrinol Metab, September 2003, 88(9):4135– 4138

TABLE 3. Nonparametric two-tailed Spearmann’s Rho correlation between individual serum TSH levels and TBII titers (from 45 patients with Graves’ hyperthyroidism after 6.7 ⫾ 1.5 months of antithyroid drug treatment) with TBII, fT4, FT3I, duration of treatment, and the L-T4 dose that the patients received TSH Correlation coefficient (r)

TBII fT4 FT3I Duration L-T4 dose a

⫺0.424 ⫺0.086 ⫺0.132 ⫺0.082 0.242

0.004a 0.576 0.399 0.591 0.114

Correlation coefficient (r)

Significance (P)

⫺0.008 0.081 0.203 0.087

0.957 0.606 0.180 0.575

Significance at P ⬍ 0.01.

evidence is accumulating that both GH and prolactin can down-regulate their own secretion (12, 13). We therefore think, that the postulated ultra-short loop feedback is operative in the secretion of all pituitary hormones. However, because of the specific characteristics of Graves’ disease (TBII), this will only have clinical consequences in this TSH-R-mediated disease. It should be noted that such an ultra-short loop feedback would only serve for fine-regulation of TSH secretion. In other words, if the patient truly becomes hypothyroid, this will override the fine regulation through the pituitary TSH-R, and TSH values may nonetheless become elevated. In conclusion, this study shows that elevated TBII levels are associated with continued suppression of TSH secretion in patients with Graves’ hyperthyroidism rendered euthyroid, most likely via interaction with a pituitary TSH-R. Acknowledgments We thank Iris M. M. J. Wakelkamp for excellent statistical assistance. Received March 11, 2003. Accepted May 22, 2003.

Address all correspondence and requests for reprints to: Leon J. S. Brokken, Department of Physiology, Institute of Biomedicine, University of Turku, Kiinamyllynkatu 10, FIN-20520 Turku, Finland. E-mail: [email protected].

References

TBII

Significance (P)


1. Franklyn JA 1994 The management of hyperthyroidism N Engl J Med 330: 1731–1738 2. Ross DS, Daniels GH, Gouveia D 1990 The use and limitations of a chemiluminescent thyrotropin assay as a single thyroid function test in an out-patient endocrine clinic. J Clin Endocrinol Metab 71:764 –769 3. Prummel MF, Brokken LJS, Meduri G, Misrahi M, Bakker O, Wiersinga WM 2000 Expression of the thyroid stimulating hormone-receptor in the folliculostellate cells of the human anterior pituitary gland. J Clin Endocrinol Metab 85:4347– 4353 4. Brokken LJS, Scheenhart JW, Wiersinga WM, Prummel MF 2001 Suppression of serum TSH by Graves’ Ig: evidence for a functional pituitary TSH receptor. J Clin Endocrinol Metab 86:4814 – 4817 5. Wiersinga WM, Chopra IJ 1982 Radioimmunoassay of thyroxine (T4), 3, 5, 3⬘-triiodothyronine (T3), 3, 3⬘, 5⬘-triiodothyronine (reverse T3, rT3), and 3, 3⬘-diiodothyronine (T2). Methods Enzymol 84:272–303 6. Theodoropoulou M, Arzberger T, Gruebler Y, Korali Z, Mortini P, Joba W, Heufelder AE, Stalla GK, Schaaf L 2000 Thyrotrophin receptor protein expression in normal and adenomatous human pituitary. J Endocrinol 167:7–13 7. Arzt E, Paez Pereda M, Costas M, Sauer J, Renner U, Holsboer F, Stalla GK 1998 Cytokine expression and molecular mechanisms of their auto/paracrine regulation of anterior pituitary function and growth. Ann NY Acad Sci 840: 525–531 8. Allaerts W, Carmeliet P, Denef C 1990 New perspectives in the function of pituitary folliculo-stellate cells. Mol Cell Endocrinol 71:73– 81 9. Mertani HC, Waters MJ, Jambou R, Gossard F, Morel G 1994 Growth hormone receptor binding protein in rat anterior pituitary. Neuroendocrinology 59:483– 484 10. Mertani HC, Pechoux C, Garcia-Caballero T, Waters MJ, Morel G 1995 Cellular localization of the growth hormone receptor/binding protein in the human anterior pituitary gland. J Clin Endocrinol Metab 80:3361–3367 11. Jin L, Qian X, Kulig E, Sanno N, Scheithauer BW, Kovacs K, Young WF, Lloyd RV 1997 Prolactin receptor messenger ribonucleic acid in normal and neoplastic pituitary tissues. J Clin Endocrinol Metab 82:963–968 12. Agustsson T, Bjornsson BT 2000 Growth hormone inhibits growth hormone secretion from the rainbow trout pituitary in vitro. Comp Biochem Physiol C Toxicol Pharmacol 126:299 –303 13. Bentley AM, Wallis M 1987 In vitro evidence for the autoregulation of prolactin secretion at the level of the pituitary gland in the rat. Endocrinology 115:13–18