A New Small-Molecule Antagonist Inhibits Graves' Disease Antibody Activation of the TSH Receptor

A novel TSH receptor inverse agonist was developed and shown to inhibit stimulation of TSH receptor by all 30 sera from patients tested who had Graves' disease.

Abstract

Context:

Graves' disease (GD) is caused by persistent, unregulated stimulation of thyrocytes by thyroid-stimulating antibodies (TSAbs) that activate the TSH receptor (TSHR). We previously reported the first small-molecule antagonist of human TSHR and showed that it inhibited receptor signaling stimulated by sera from four patients with GD.

Objective:

Our objective was to develop a better TSHR antagonist and use it to determine whether inhibition of TSAb activation of TSHR is a general phenomenon.

Design:

We aimed to chemically modify a previously reported small-molecule TSHR ligand to develop a better antagonist and determine whether it inhibits TSHR signaling by 30 GD sera. TSHR signaling was measured in two in vitro systems: model HEK-EM293 cells stably overexpressing human TSHRs and primary cultures of human thyrocytes. TSHR signaling was measured as cAMP production and by effects on thyroid peroxidase mRNA.

Results:

We tested analogs of a previously reported small-molecule TSHR inverse agonist and selected the best NCGC00229600 for further study. In the model system, NCGC00229600 inhibited basal and TSH-stimulated cAMP production. NCGC00229600 inhibition of TSH signaling was competitive even though it did not compete for TSH binding; that is, NCGC00229600 is an allosteric inverse agonist. NCGC00229600 inhibited cAMP production by 39 ± 2.6% by all 30 GD sera tested. In primary cultures of human thyrocytes, NCGC00229600 inhibited TSHR-mediated basal and GD sera up-regulation of thyroperoxidase mRNA levels by 65 ± 2.0%.

Conclusion:

NCGC00229600, a small-molecule allosteric inverse agonist of TSHR, is a general antagonist of TSH receptor activation by TSAbs in GD patient sera.

Graves' disease (GD) is caused by persistent, unregulated stimulation of thyroid cells by thyroid-stimulating antibodies (TSAbs) that activate the TSH receptor (TSHR) (1–3). TSAbs, like TSH, bind primarily to the large amino-terminal ectodomain of TSHR. For the majority of antibodies tested, TSAbs and TSH compete for binding to TSHR. We previously reported the first small-molecule TSHR antagonist (NIDDK/CEB-52) (4), which inhibited TSH-stimulated signaling, and the first TSHR inverse agonist (NCGC00161856) (5), which is an antagonist that inhibits basal (or constitutive or agonist-independent) TSHR signaling in addition to TSH-stimulated signaling. TSHR is one of a minority of G protein-coupled receptors that exhibit easily measurable basal signaling activity in vitro (6). In our previous reports, we provided compelling evidence that these drug-like compounds bind to TSHR in its serpentine region at what is termed an allosteric site and do not compete for binding with TSH. We also reported that one of these antagonists inhibited activation of TSHR by sera from four patients with GD (4). Because we measured the effect of the antagonist on a small number of GD sera, it was possible that allosteric ligands would not inhibit TSHR activation by some TSAbs.

We have continued in our attempts to optimize small-molecule TSHR ligands and have developed a better inverse agonist (NCGC00229600) (1), which is an analog of NCGC00161856. Herein we show that 1 is a TSHR inverse agonist in a model system overexpressing TSHRs and in human thyrocytes in primary culture and show that 1 inhibits TSHR activation by all 30 sera from patients with GD tested.

Subjects and Methods

The clinical features of the patients are summarized in Table 1.

Table 1.

Patient TSH, free T₄, TPO-Ab, and TSAb levels and treatment at the time of serum acquisition

Sample	TSH (IU/liter)	Free T4 (ng/dl)	TPO-Ab (IU/ml)	TSAb (AU)	TR-Ab (U/liter)	Thyroid volume (ml)	RX
GD10	<0.01	4.2	225	23.5	4.7	22.2	At diagnosis
GD13	2.6	0.7	<15	24.1	5.7	20.2	MMI for 3 months
GD15	0.01	2.8	NA	23.6	NA	21.4	MMI for 30 d
GD16	5.2	0.6	166	19.7	49.8	46.8	MMI for 21 months
GD19	<0.01	2.5	NA	55.9	3.2	36.3	At diagnosis
GD20	0.015	1.3	<15	53.5	4.5	15.2	At diagnosis
GD23	0.01	1.5	<15	63.1	1.6	15.5	MMI for 2 months
GD24	0.9	1.3	NA	276	NA	0.3	Post-TX/RAI on l-T4 for 10 yr; ophthalmopathy
GD25	0.01	2.3	<15	56.2	7.4	27	MMI for 9 months
GD26	0.01	1.4	>2000	96.6	2.9	50.9	MMI for 4 months
GD27	5.9	0.7	<15	34.7	1.4	11.3	MMI for 3 months
GD28	0.66	1.1	849	26.4	2.1	18.4	MMI for 4 months
GD31	0.02	1.6	101	9.5	3.5	12.5	MMI for 30 d
GD32	0.01	1.7	>2000	7.6	5.1	11.7	At diagnosis
GD33	0.06	1.2	>2000	10.6	1.3	12.5	At diagnosis
GD38	0.09	0.8	438	33	7.8	10.9	MMI for 20 d
GD42	0.04	2.2	NA	133	2.7	17	At diagnosis
GD43	0.06	1.4	NA	28.6	NA	18.7	MMI for 2 months
GD44	0.01	1.9	1058	5.9	3.4	18.1	MMI for 15 d
GD45	24.8	0.5	>2000	5.39	5.6 (at diagnosis)	12.6	MMI for 4 months
GD54	18.9	0.7	>2000	387	11.4	32.5	MMI for 4 months
GD55	13.6	0.6	101	86	6.2	13.3	MMI for 24 months
GD56	0.7	1.2	790	263	4.1	2.4	RAI on l-T4 for 3 months
GD57	1.2	1.1	<15	1.1	0.57	10.6	Remission of GD
GD66	0.01	4.4	>2000	45.9	4.4	16.8	At diagnosis
GD67	13.2	0.6	99	1.34	3.4	16.1	PTU for 6 months
GD68	0.01	3.6	931	5.18	9.7	10.8	At diagnosis
GD69	3.8	0.6	<15	11.2	0.8	30.1	MMI for 7 months
GD71	2.3	1	287	6.68	1.64	16.2	MMI for 5 months
GD80	0.01	1.2	226	5.25	3.1	12.2	MMI for 30 d

TSH normal range = 0.4–4.2 IU/liter; free T₄ normal range = 0.76–1.42 ng/dl; TPO-Ab normal value is less than 15 IU/ml; TSAb was measured using a biological assay (7). Cutoff of normal values was determined by the mean of at least five samples from normal subjects. The sd between these normal samples was determined, and this value was multiplied by two and added to the calculated mean. The cutoff value obtained in this way was arbitrarily considered equal to 1 arbitrary unit (AU). Normal range was therefore 0–1 AU. TR-Ab (RIA rTSHAb CT assay normal value <1 U/liter, values between 1.1–1.5 U/liter were considered uncertain or gray zone; Kronus Ltd., Boise, ID). Thyroid volume was calculated by using the volumetric ellipsoid method (height × width × depth × correction factor 0.524) for each lobe and then adding the obtained values. At diagnosis indicates untreated patients at first visit; remission of GD indicates patient untreated in the last 36 months after 18 months treatment with methimazole (MMI). NA, Not applicable; PTU, propylthiouracil; RAI, radioactive iodine; RX, treatment; TX, thyroidectomy.

Clinical tests

TSH was measured by an immunofluorometric assay (TSH ultra; normal range, 0.4–4.2 IU/liter; Perkin-Elmer Italia spa, Monza, Italy). Free T4 was measured by an immunofluorometric assay (normal range, 0.76–1.42 ng/dl; Perkin-Elmer Italia). Thyroid peroxidase antibody (TPO-Ab) was measured by TPO-Ab RIA (normal value, <15 IU/ml; Becton Dickinson Co., Franklin Lakes, NJ). TSAb was measured using a biological assay (7). Cutoff of normal values was determined by the mean of at least five samples from normal subjects. The sd between these normal samples was determined, and this value was multiplied by two and added to the calculated mean. The cutoff value obtained in this way was arbitrarily considered equal to 1 arbitrary unit.

Synthesis of 1

To a solution of 3-(chloromethyl)-4-methoxybenzaldehyde (300 mg, 1.625 mmol, 1.0 equivalent) and 2,6-dimethylphenol (218 mg, 1.787 mmol, 1.1 equivalent) in 10 ml acetonitrile was added potassium carbonate (1.1 g, 8.12 mmol, 5.0 equivalents). The mixture was heated to 150 C in a microwave for 30 min. Upon completion, the mixture was filtered and dried down to give 3-[(2,6-dimethylphenoxy)methyl]-4-methoxybenzaldehyde (400 mg, 91% yield) as a yellow solid. A portion of which (100 mg, 0.370 mmol, 1.0 equivalent) was taken up in ethanol (4 ml), and to it was added 2-amino-N-(pyridin-3-ylmethyl)benzamide (92 mg, 0.407 mmol, 1.1 equivalent) followed by ytterbium(III) trifluoromethanesulfonate (45.9 mg, 0.074 mmol, 0.2 equivalent). The mixture was heated to 80 C for 2 h. Upon completion, the mixture was dried down and chromatographed on silica gel with 0–30% EtOAc (ethyl acetate)/hexanes gradient elution to give the desired 2-{3-[(2,6-dimethylphenoxy)methyl]-4-methoxyphenyl}-3-(pyridin-3-ylmethyl)-2,3-dihydroquinazolin-4(1H)-one (110 mg, 62.0% yield) as a tan solid. ¹H nuclear magnetic resonance (400 MHz, DMSO-d₆) δ ppm 8.39–8.45 (m, 2 H), 7.60–7.67 (m, 2 H), 7.48 (d, J = 2.35 Hz, 1 H), 7.15–7.35 (m, 4 H), 6.94–7.02 (m, 3 H), 6.84–6.93 (m, 1 H), 6.60–6.69 (m, 2 H), 5.82 (d, J = 2.35 Hz, 1 H), 5.07 (d, J = 15.65 Hz, 1 H), 4.64 (d, J = 2.74 Hz, 2 H), 3.99 (d, J = 15.45 Hz, 1 H), 3.73 (s, 3 H), 2.12 (s, 6 H); liquid chromatography mass spectrometry: (electrospray +ve), m/z 480.2 (MH)⁺ (mass of molecular weight plus 1); HPLC: t_R = 5.05 min,UV₂₅₄ = 100%. High-resolution mass spectroscopy (electrospray ionization): m/z calculated for C₃₀H₃₀N₃O₃ [M+H]⁺ 480.2282, found 480.2291.

Cell culture

The generation of a stable HEK-EM 293 cell line expressing TSHRs (HEKTSHR cells) was described previously (4). HEKTSHR cells were grown in DMEM supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 10 μg/ml streptomycin (Life Technologies Inc., Carlsbad, CA) at 37 C in a humidified 5% CO₂ incubator.

Thyroid tissue samples were collected from normal thyroid tissue from patients undergoing total thyroidectomy for thyroid cancer at the National Institutes of Health Clinical Center. Patients provided informed consent on an Institutional Review Board-approved protocol, and materials were received anonymously via approval of research activity through the Office of Human Subjects Research. Thyrocytes monodispersed with collagenase IV (Life Technologies) were plated in 10 ml DMEM with 10% FBS in 10-cm tissue culture dishes and incubated at 37 C in a humidified 5% CO₂ incubator, and after 24 h, the primary cultures of adherent thyroid cells were obtained as described previously (8). For determination of TPO mRNA expression, thyrocytes were seeded into 24-well plates at a density of 6 × 10⁴ cells per well. The cells were incubated in DMEM with 2% FBS for 48 h in the presence or absence of 1 and without or with the designated dilutions of GD sera.

Determination of cAMP production

HEKTSHR cells were seeded into 48-well plates at a density of 11 × 10⁴ cells per well in DMEM containing 10% FBS. Cells were cultured for 24 h before incubation for 30 min in Hanks' balanced salt solution (HBSS)/10 mm HEPES (pH 7.4). To determine basal cAMP production, cells were incubated for 1 h at 37 C in a humidified incubator in HBSS/HEPES containing 1 mm 3-isobutyl-1-methylxanthine (IBMX) (Sigma, St. Louis, MO) in the presence or absence of 1. To determine the effect of 1 on human TSH (hTSH) (Thermo Fisher Scientific, Fremont, CA) or GD sera activation of TSHRs, cells were incubated for 30 min with 1 in HBSS/HEPES without IBMX and subsequently for 45 min with hTSH or GD sera and 1 in HBSS/HEPES supplemented with 1 mm IBMX at 37 C. The levels of cAMP in cells incubated in HBSS/HEPES without IBMX were subtracted in all experiments. After aspiration of the incubation medium, cells were lysed using lysis buffer of the cAMP-Screen Direct System (Applied Biosystems, Carlsbad, CA) and total cAMP content determined as described by the manufacturer. The chemiluminescence signal was measured in a VICTOR3 V 1420 Multilabel Counter (PerkinElmer, Shelton, CT). Data analysis was performed with GraphPad Prism 4 for Windows.

Effect of 1 on [¹²⁵I]TSH binding

HEKTSHR cells were seeded into 24-well plates at a density of 2.2 × 10⁵ cells per well. Cell surface binding was measured after 18 h by incubation in 0.25 ml binding buffer (HBSS containing 2.5% milk powder and 0.2% BSA) containing 60,000 cpm bovine [¹²⁵I]TSH (Brahms Aktiengesellschaft, Hennigsdorf, Germany) without or with 30 μm 1 for 2 h at room temperature; nonspecific binding was measured in the presence of 1.8 μm unlabeled bovine TSH (Sigma, St. Louis, MO) (9). Cells were washed three times with 0.5 ml ice-cold HBSS and lysed with 0.5 ml 0.4 n NaOH, and the cell-associated radioactivity was counted in a 1470 Wallac Wizard γ-counter (PerkinElmer).

Effect of 1 on antibody binding

HEKTSHR cells were seeded into 12-well plates at a density of 4.4 × 10⁵ cells per well. After 24 h, cells were incubated with 30 μm 1 for 1 h in PBS and then harvested with 1 mm EDTA/1 mm EGTA in PBS, washed with PBS containing 0.1% BSA and 0.1% sodium azide and incubated in 1:200 dilutions of one of two antibodies that bind to different epitopes within the amino-terminal ectodomain of TSHR, 2C11 (epitope between amino acid residues 354 and 359) (Serotec, Raleigh, NC), or 4C1 (epitope between amino acid residues 381 and 384) (Santa Cruz Biotechnology, Santa Cruz, CA) in the presence of 30 μm 1 for 90 min. Cells were then washed and incubated with a 1:200 dilution of an Alexa-Fluor 633-conjugated F(ab)2 fragment of goat antimouse IgG (Molecular Probes, Carlsbad, CA) for 30 min. After a final wash, cells were fixed with 1% paraformaldehyde and analyzed by fluorescence-activated cell analyzer (FACSCalibur; BD Biosciences, San Jose, CA). Antibody binding was determined as the number of singlet cells and the level of fluorescence present in a gate that contained no cells in an analysis of cells incubated without primary antibody.

Quantitative RT-PCR

Total RNA was purified using RNeasy Micro kits (QIAGEN, Valencia, CA). cDNA was prepared using a High Capacity cDNA Archive Kit (Applied Biosystems). RT-PCR was performed in 25-μl reactions using cDNA prepared from 100 ng or less of total RNA and Universal PCR Master Mix (Applied Biosystems). mRNA levels were measured using primers and probes from Applied Biosystems. Quantitative RT-PCR results were normalized to GAPDH to correct for differences in RNA input as described (5).

Statistical analysis

Data are expressed as mean ± se. The data were analyzed by Student's t test or one-way ANOVA; P < 0.05 was considered significant.

Results

We synthesized and tested six analogs of NCGC00161856 (Supplemental Fig. 1, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org) and found that 1 was as potent and effective as NCGC00161856 as a TSHR antagonist using bovine TSH but was a better drug because of its increased solubility (Fig. 1). In this study, we characterize 1 as an antagonist of basal and hTSH stimulation of hTSHR in human cells. Figure 2A illustrates that in HEKTSHR cells, 1 inhibited basal cAMP production by TSHRs by 53% at 30 μm. Figure 2B shows that 1 inhibited cAMP production stimulated by an EC₅₀ concentration of hTSH by 53% at 30 μm. Figure 2C shows that the inhibition of cAMP production by 30 μm 1 was overcome at high doses of hTSH and was, therefore, competitive. Because 1 is a competitive antagonist, the extent of inhibition is dependent on the concentration of TSH. Competitive inhibition may have been caused by competition of hTSH and 1 for binding; however, 1 had no effect on [¹²⁵I]TSH binding to HEKTSHR cells (Fig. 3). The finding that 1 did not affect TSH binding was expected because we have previously shown that these small-molecule ligands do not affect TSH binding and appear to bind in the serpentine region of TSHR (4, 5, 8, 10, 11). Allosteric antagonists that inhibit signaling by G protein-coupled receptors are well known (12).

Fig. 1.

Synthetic scheme and structure of 1. See Subjects and Methods for details of the synthesis.

Fig. 2.

1 is an inverse agonist at TSHR and a competitive antagonist of TSH-stimulated signaling. A, HEKTSHR cells were exposed to the noted concentrations of 1 for 60 min in HBSS with 1 mm IBMX (Basal Signaling). Total cAMP levels were measured by ELISA. The data are from three experiments with duplicate samples and are presented as mean ± se. B, HEKTSHR cells were exposed to the indicated concentrations of 1 for 20 min in HBSS and then were incubated in HBSS with 1 mm IBMX, 1, and 1 mU/ml hTSH. After 40 min, the incubation was stopped, and total cAMP levels were measured by ELISA. The data are from two experiments with duplicate samples and are presented as mean ± se. C, HEKTSHR cells were incubated in HBSS (Control or) exposed to 30 μm 1 for 30 min in HBSS and then 1 mm IBMX, and various concentrations of hTSH were added. After an additional 45 min, the incubation was stopped, and total cAMP levels were measured by ELISA. The data are from two experiments with duplicate samples.

Fig. 3.

1 does not inhibit TSH or TSAb binding to TSHR. Specific binding of [¹²⁵I]TSH to HEKTSHR cells was performed in the absence or presence of 30 μm 1 as described in Subjects and Methods. The data are from two experiments with triplicate samples. There was no effect of 1 on TSH binding (P > 0.1). Binding of 2C11 or 4C1 antibodies to HEKTSHR cells was performed in the absence or presence of 30 μm 1 and analyzed by FACS as described in Subjects and Methods. The data are from two experiments with triplicate samples. There was no effect of 1 on antibody binding (P > 0.1).

Before testing whether 1 inhibited signaling by GD sera, we showed that 1 had no effect on the binding of two monoclonal anti-TSHR antibodies, 2C11 and 4C1 (Fig. 3). Although we could not test whether 1 had any effect on the binding of TSAbs in GD sera, we assume that because it did not inhibit binding of 2C11 or 4C1 that it does not inhibit TSAb binding also. We tested whether 1 would inhibit cAMP production in HEKTSHR cells stimulated by three GD sera that increased cAMP by more than 10-fold at 1:30 or 1:100 dilution (Fig. 4A), and 1 inhibited the increase in cAMP production stimulated by these three sera by 30–75%. In experiments in which 30 GD sera that increased cAMP production at a 1:30 dilution in HEKTSHR cells were tested, 1 inhibited cAMP production by 39 ± 2.6% (mean ± sem) in all sera tested (Fig. 4B). Thus, 1 is a general antagonist of TSHR activation of cAMP by TSAbs in GD patient sera in the model cell system.

Fig. 4.

1 is an antagonist of GD sera stimulation of cAMP production. A, HEKTSHR cells were incubated without (control) or with 30 μm 1 for 20 min, and then 1 mm IBMX and 1:30 or 1:100 dilutions of each of three GD sera were added. After an additional 40 min, the incubation was stopped, and total cAMP levels were measured by ELISA. The data are from one experiment with triplicate samples and are representative of three experiments. The effects of 1 were significant by paired t test (P < 0.01). B, HEKTSHR cells were incubated without (control) or with 30 μm 1 for 20 min, and then 1 mm IBMX and 1:30 dilutions of 30 GD sera were added. After an additional 40 min, the incubation was stopped, and total cAMP levels were measured by ELISA. The data are presented as percent inhibition = 100 − (100 × samples exposed to 1/control). The data are representative of two experiments performed in duplicate.

To test the effect of 1 in a more physiologically relevant cell system in which an important thyroid-specific biological response can be measured, we next determined whether 1 inhibited TSHR-mediated up-regulation of TPO mRNA in human thyrocytes in primary culture, and 1 inhibited TSH-induced increases in TPO mRNA (data not shown). Figure 5 illustrates that for all 30 GD sera tested, 1 inhibited GD sera up-regulation of TPO mRNA by 65 ± 2.0% (mean ± sem). This effect of 1 was not because 1 is toxic to human thyrocytes because although 1 decreased the control (or basal) levels of TPO mRNA (Fig. 5) and the mRNAs of thyroglobulin and TSHR (Supplemental Fig. 2), it had no effect on the level of deiodinase type 2 mRNA (Supplemental Fig. 2). Thus, 1 is a nontoxic antagonist of TSHR activation by TSAbs in GD patient sera in human thyrocytes.

Fig. 5.

Inhibition of basal and TSAb-induced up-regulation of the expression of TPO mRNA by 1 in primary cultures of human thyrocytes. Thyrocytes were incubated in DMEM containing 2% FBS without (control) or with 30 μm 1 without (basal) or with 1:10 dilutions of GD sera at 37 C. After 48 h, the buffers were aspirated, the cells were lysed, and the levels of TPO mRNA were measured and normalized to GAPDH mRNA. The mRNA levels are presented as fold of basal levels (control) (A) and as percent inhibition = 100 − (100 × samples exposed to 1/control) (B). The data from two independent experiments with duplicate samples are shown.

Discussion

Graves' hyperthyroidism may be treated with antithyroid drugs, radioactive iodine, or surgery (13). Each of these therapies has been shown to be effective in many clinical trials and is well tolerated by patients, but there is a significant rate of relapse with each, especially antithyroid drugs and radioiodine. For example, in a randomized trial of 179 patients, these three options were equally effective as initial therapy, each was well accepted by over 90% of patients, but the relapse rate was approximately 40% for patients treated with antithyroid drugs, 21% for those treated with radioactive iodine, and 5% after thyroidectomy (14). None of these therapies is pathogenesis based; that is, none attempts to inhibit the production of TSAbs or TSAb activation of TSHR signaling. This type of therapy might be more easily titrated to achieve a euthyroid state and might be used to treat recurrences associated with antithyroid drugs and radioactive iodine. Inhibition of TSAb production would perhaps be the most desirable approach, but no benign treatment to accomplish this has been developed. Herein we report the further development of a TSHR antagonist that inhibits TSAb activation of TSHR. An orally effective TSHR antagonist, if shown to be without significant side effects, might be used as initial chronic therapy for GD or for recurrent disease, and its dosing could be varied to maintain a euthyroid state. To be used as therapy for GD, it would be important that an antagonist inhibit TSHR activation by the majority of TSAbs. It is noteworthy, therefore, that we found that 1 inhibited activation of TSHR by the sera from all 30 GD patients tested.

The observation that 1 is an allosteric antagonist at TSHR may explain its efficacy to antagonize stimulation by all GD sera tested. It has been shown that TSAbs, like TSH, bind primarily to the amino-terminal ectodomain of TSHR at multiple epitopes. It was unlikely that a small molecule like 1 would inhibit binding of TSAbs (or TSH) because it has been found that it is very difficult to inhibit protein-protein interactions with small molecules (15). Indeed, we showed that 1 does not inhibit TSAb (or TSH) binding to TSHR. Therefore, the mechanism of its inhibition most likely involves a restraint on the conformational changes needed for TSHR activation when 1 is bound at an allosteric binding site, and that may account for its ability to inhibit TSHR signaling by all GD sera.

In conclusion, 1 is an antagonist of activation of the TSHR by TSAbs and is a lead drug for the development of a therapeutic agent to treat the hyperthyroidism of GD.