The "TSH Receptor Glo Assay" - A High-Throughput Detection System for Thyroid Stimulation

Abstract

Background: To identify novel small molecules against the TSH receptor, we developed a sensitive transcription-based luciferase high-throughput screening (HTS) system named the TSHR-Glo Assay (TSHR-Glo).

Methods: This assay uses double-transfected Chinese hamster ovary cells stably expressing the human TSHR and a cAMP-response element (CRE) construct fused to an improved luciferase reporter gene.

Results: The assay was highly responsive toward TSH in a dose-dependent manner with a TSH sensitivity of 10(-10)M (10 ± 1.12 μU/ml) and thyroid-stimulating antibodies, a hallmark of Graves' disease, could also be detected. The assay was validated against the standard indicator of HTS performance - the Z-factor (Z') - producing a score of 0.895. Using the TSHR-Glo assay, we screened 48,224 compounds from a diverse chemical library in duplicate plates at a fixed dose of 17 μM. Twenty molecules with the greatest activity out of 62 molecules that were identified by this technique were subsequently screened against the parent luciferase stable cell line in order to eliminate false positive stimulators.

Conclusion: Using this approach, we were able to identify specific agonists against the TSH receptor leading to the characterization of several TSH agonist molecules. Hence, the TSHR-Glo assay was a one-step cell-based HTS assay, which was successful in the discovery of novel small molecular agonists and for the detection of stimulating antibodies to the TSH receptor.

Keywords: TSH receptor; high-throughput; screening; small molecule; thyroid-stimulating hormone.

Figure 1

TSH dose–response and HTS optimization. (A) The TSHR-Glo cells were tested against bovine TSH and a WHO international standard human TSH (81/565 obtained from NIBSC). Inset: shows the response curve using recombinant human TSH. (B) To optimize the use of the assay as an HTS method, we performed a similar TSH dose–response using different cell densities on a proxy-shallow 384-well plate and calculated the Z factor scores under each cell density. A Z factor score of 0.895 with 15,000 cells/well was adopted as the standard format of cell density for HTS screening [note: 1 ≥ z ≥ 0.5 is considered an excellent assay (23)]. Luciferase is measured on a luminescence based microplate reader after lysis of the cells with a one-step reagent and expressed as luciferase units (LU).

Figure 2

Effect of time and DMSO on TSHR-Glo assay. (A) Since this is a transcriptional based assay, we wanted to evaluate the effect of time. For this, the assay was performed with 10⁻⁹M (100 μU/ml) of TSH and time periods of 1, 2, 3, and 4 h without (0TSH) and with TSH treatment. Though a measurable difference in response to TSH was observed in 2 h, the response at 4 h was taken as the optimum time of incubation required for the screening because of the robust signal. (B) Since the library compounds were held in DMSO, it was imperative to test the effect of DMSO on the TSHR-Glo assay. The bars show the effect of increasing concentrations of DMSO on stimulation of the cells with TSH. We found that the bioassay could tolerate DMSO concentrations up to 1%. However, the concentration of DMSO in our actual screen was 0.1% DMSO.

Figure 3

Specificity of the TSHR-Glo assay. (A) The specificity of the assay was examined by dose–responses curves against TSH, hCG, and FSH hormones. As indicated, the assay showed no responses to hCG or FSH at the concentrations tested here but a dose-dependent increase in luciferase signal. (B,C) In order to test the heteroscedasticity of the assay, we measured response of increasing doses of TSH in an inter- and intra-assay format as indicated by graphs. Intra-assay exhibited very minimal variability compared to the inter assay though both had CVs <5%.

Figure 4

Response of TSHR-Glo against stimulating and blocking monoclonal antibody – further proof of the specificity of the assay was obtained by testing a stimulatory monoclonal antibody (MS-1) and blocking (non-stimulatory) monoclonal TSHR antibody (K1-70) in the assay. MS-1 stimulated luciferase activity in a dose-dependent manner, whereas K1-70 was unable to show any response in the assay even at the highest concentration.

Figure 5

Agonist activity measured in patient sera using TSHR-Glo. Graves’ disease patient and normal sera were diluted 1:5 in medium. Ten of the 12 GD (Graves’ disease) serum samples (denoted GD with dark gray bars) stimulated the production of cAMP in contrast to no stimulation observed with any of the normal (denoted NS with light gray bars) or rheumatoid arthritis (denoted RA with black bars) patient sera. Background is denoted by the two filled hatched bars, a human-stimulating monoclonal antibody (M22 shown as hatched bar), human-blocking monoclonal antibody (K1-70 shown as filled criss-cross bar) 1 μg/ml, and bovine TSH (gray bars) at increasing log concentrations.

Figure 6

Evaluation of HTS performance. (A) Z factor was calculated by taking the positive control and basal control responses of each duplicate plate and using the formula Z′ = 1 − (3 × SD of total signal + 3 × SD of basal signal)/(total signal − basal signal) as indicated by Zhang et al. (23). (B) Signal-to-background ratio (S/B): total signal was obtained from alternating wells of columns 1 and column 22 using 17 μM of positive control and background signal was were collected from the alternating wells of column 1 and column 22 treated with medium containing 0.1% DMSO. (C) Coefficient of variation (% CV), calculated as the SD from the wells in the basal medium divided by the wells of the positive control in all 137 plates.