Optimization of an in vitro assay methodology for competitive binding of thyroidogenic xenobiotics with thyroxine on human transthyretin and albumin

Graphical abstract


Background
The structural resemblance of many anthropogenic chemicals to thyroid hormones (THs) warrants concern for the perturbation of TH-dependent processes in humans and other vertebrates [1]. Thyroxine (T4) and 3,5,3-triiodothyronine (T3) are THs involved in the regulation of many important physiological processes in the body including neurological and behavioural development, growth, metabolism, and respiration [2]. In mammals (including humans), circulating T4 concentrations are greater in plasma than for T3. T4 is the prohormone TH, and it is deiodinated via enzyme-controlled mechanisms to T3, which is the main ligand for the TH receptor leading to gene regulation [3]. The vast majority (>99%) of THs in plasma are delivered to target tissues by binding to TH transport proteins, which include albumin (ALB), transthyretin (TTR), and thyroxine-binding globulin (TBG). The TH binding affinities and dissociation constants of these three proteins vary, with ALB having the lowest relative affinity for T4 and the greatest concentration in plasma; and TBG having the highest affinity for T4, and the lowest concentration in plasma [4]. TTR is the only TH transport protein synthesized in the choroid plexus, and it thus plays a role in the movement of THs from the blood into the cerebrospinal fluid (CSF) which is particularly important in early development [3].
One potential mechanism of action of xenobiotic thyroidogenic substances is competition for TH binding sites on serum transport proteins, as this can interfere with delivery of THs to target tissues. A limited number of previously published studies report variations of the in vitro radiolabelled T4-TTR binding assay technique, with target chemicals including polychlorinated biphenyls, perfluorinated compounds, polybrominated diphenyl ethers, and some degradation products [1,[5][6][7][8][9][10]. However, the methods reported in these studies do not provide sufficient details to successfully replicate the procedure and we faced several challenges in attempting to do so. Therefore, the objectives of the present study were to refine, optimize, validate, and fully describe a competitive in vitro radio-ligandprotein binding assay that is used to investigate thyroidogenicity of chemicals via interaction between T4 and human TTR. Furthermore, we expanded this method to include interactions between T4 and human ALB which, to our knowledge, has not been adequately described elsewhere. The model xenobiotic chemical that we used as a competitor ligand is 4-hydroxy-2,2 0 ,4,5 0 -tetrabromodiphenyl ether (4-OH-BDE-49), which is often detected in human serum [11] and has previously been found to bind to human TTR isolated from human plasma with higher affinity than T4 in vitro [9].
A number of other methods have been developed for the purpose of investigating potential competition of xenobiotics with T4 for TTR, including the ANSA (8-anilino-1-naphthalenesulfonic acid ammonium)-TTR competitive fluorescence displacement assay [12], the TR-CALUX (thyroid hormone responsive chemically activated luciferase gene expression) assay [13], surface plasmon resonance-based biosensor assays [14], FITC-T4 (fluorescence probe fluorescein isothiocyanate associated to T4) assays [15,16], and HPLC (high performance liquid chromatography) assays [17]. Notably, Ouyang et al. [16] presents an enhanced FITC-T4 assay that has been miniaturized into a 96 well microplate to generate a high throughput method. While the miniaturized FITC-T4 method offers increased efficiency, the radiolabelled T4-protein binding assay described herein is more sensitive than the assay presented in Ouyang et al. [16], and by an order of magnitude for potent xenobiotic competitors (those with IC50s < 100 nM). This is particularly important for the detection of receptor binding events at low concentrations (e.g., those found in blood or other biological samples) and for reducing the likelihood of a type II error. Additionally, our method extends to include interactions with ALB, which increases the relevance of this assay as ALB is the most abundant protein in plasma and serum.

Sample preparation (Day 1)
The entire assay takes two days to complete for each batch of samples to be processed. The Day 1 procedure is described as follows. Prepare the solutions of the proteins, THs and competitive ligands to be used in the in vitro assay according to Table 1 and using materials specified in Table 2. First, thaw the working solutions of these components to room temperature if they have been placed in frozen storage. For the protein working solutions, prepare the day they are to be used based on the number of assay aliquots required (one vial is sufficient for approximately 45 assay samples). Take  The following steps describe the order of pipetting components for assay incubation mixtures.
1. Starting with the calibration curve samples (T4 as a competitor), pipette 5 mL of the lowest competitor concentration into all replicate assays. Continue for each increasing concentration of the ligand. Control samples receive 5 mL of DMSO in place of a competitor ligand. 2. For the first "replicate set" (e.g., 1.1-11.1 in Fig. 1), pipette 70 mL of the T4/ 125 I-T4 solution into each test tube, followed by 25 mL of the selected protein (e.g., TTR). This results in a total sample volume of 100 mL.
3. Centrifuge the samples at 300 Â g for 30 s to force the solution to the bottom of the tube. 4. Measure the initial radioactivity on the gamma counter. Cap the samples immediately, being careful not to disturb solutions, and place in the fridge at 4 C overnight [ 3 3 _ T D $ D I F F ] to reach protein-ligand binding equilibrium.
While the first replicate set is on the gamma counter, repeat Steps 2 to 4 for the subsequent replicate set(s) (e.g., 1.2 to 11.3). Follow the same preparation sequence for the xenobiotic competitor (s) being analyzed (e.g., 4-OH-BDE-49).

Separation procedure (Day 2)
Label a series of clean test tubes corresponding to the samples prepared on Day 1 and repeat these labels on Bio-Spin 1 columns. Work on one replicate set at a time, leaving the remaining columns and incubation mixtures in the fridge until ready for processing. Prepare spin columns for use according to Bio-Spin 1 column instructions (invert column to re-suspend gel, remove cap and tip, centrifuge in test tube at 1000 Â g for 2 min, discard drained packing buffer). Process the incubation mixtures as follows: 1. Extract the first sample of the replicate set from the incubation tube using a 100 mL pipette and load the sample to the head of the corresponding column, being careful not to disturb the gel bed. Release the emptied pipette tip into the incubation tube. 2. Repeat for each sample in the replicate set, and centrifuge samples for 4 min at 1000 Â g. 3. Measure the radioactivity of the eluate fraction, as well as the residual radioactivity in the incubation tubes containing the discarded pipette tips.
Repeat Steps 1 to 3 for each of the remaining replicate sets. The separation procedure involves size exclusion chromatography, whereby the protein-bound ligand is eluted from the Bio-Spin 1 columns and unbound ligand remains in the columns. The size exclusion limit of the chromatography columns is 6000 g/mol, and thus any compound smaller than this size will travel through the polyacrylamide beads (a longer route through the column) while larger molecules will move around the beads (a shorter route through the column). Measuring the radioactivity of the emptied incubation tubes and pipette tips accounts for nonspecific binding and any potential transfer loss.
Note: Previous versions of this method [1,[8][9][10] suggest an additional step involving rinsing of the columns with Tris buffer, followed by a second round of centrifugation to elute all protein-bound ligand. This step was found to be unnecessary as the present method is optimized to the sample volumes and Bio-Spin 1 columns used.

Precautions and tips
TTR is a labile substance and is susceptible to denaturation. It is thus important to handle proteins with care when preparing solutions and processing samples to ensure they remain intact. Do not vortex solutions containing TTR. Prepare working solutions of TTR or ALB only on the day of use to ensure that the protein structural integrity is maintained. Additionally, keep solutions containing proteins on ice during sample preparation and processing also to prevent denaturing. Incubation mixtures should be left in the fridge until ready for processing, and analyzed quickly so as to reduce time on the column. Use a temperature controlled centrifuge set to 4 C to maximize protein integrity.
[ ( F i g . _ 1 ) T D $ F I G ] Furthermore, it is recommended that Eppendorf LoBind 1 Polypropylene test tubes be used to prepare incubation samples, in order to minimize potential loss of proteins and ligands from adsorption to the vessel.
Care must also be taken when handling radiolabelled [ 3 4 _ T D $ D I F F ] T4. All components of the assay involving 125 I-T4 must be completed in a radioactivity-licensed laboratory by trained personnel. Wear the necessary personal protective equipment, and work under a fume hood for all steps involving pipette work. The proportion of 125 I-T4 added to samples can be varied depending on signal strength, aiming for >10,000 cpm per sample for initial readings. Adjust the amount of Tris-EDTA added to each sample accordingly.

Data analysis
Calculate the percent of protein binding for each sample by dividing the radioactivity of eluate on Day 2 by the initial radioactivity measured on Day 1, minus any transfer loss (i.e., % binding = eluate cpm/[Day 1 cpm À cpm of emptied incubation tube and pipette tip]). Express results with the logarithmic competitor concentration on the X-axis, and mean AE standard deviation of percent binding compared to controls on the Y-axis (Fig. 2). Generate IC50 estimates in GraphPad Prism 7.02 (GraphPad Software, Inc., La Jolla, CA), using the [inhibitor] vs. response À variable slope equation (Table 3), defined as: The inhibition constant (K i ) for each competitor can be determined based on Cheng and Prusoff (1973) and following GraphPad guidance [18,19]. In a homologous assay (i.e., with T4 as the competitor) the cold ligand and radioligand can be assumed to have the same binding affinities. As such, the K i and the equilibrium dissociation constant (K d ) for T4 are determined using the IC50 of T4 and the concentration of radioligand in samples:  The relative potency for a competitor is defined as (IC50) T4 /(IC50) competitor , and thus the above calculation can be re-arranged as: (K i ) competitor = (K i ) T4 /(relative potency) competitor Note that at low concentrations of radioligand, the K i for a competitor approximates the corresponding IC50, which is the case herein.

Method validation
A method calibration curve should be generated in duplicate each time an assay batch is performed, using the series of prepared concentrations of the natural ligand as the competitor (e.g., 4 to 2048 nM T4 for TTR, plus negative controls). To ensure instrument performance, the gamma counter should be calibrated prior to each use, and instrument blanks (i.e., empty test tubes) can be included at various points during measurement of sample radioactivity. Serving as a positive control, each assay should include triplicate samples with the IC50 concentration of a known potent competitor. For example, the IC50 of 4-OH-BDE-49 for TTR is [ 3 5 _ T D $ D I F F ] 12.2 nM (Table 3); other potential competitor ligands could include 4-OH-CB-187 (CAS 158076-68-7) [9] and tetrabromobisphenol A   [1,8]. Negative controls should also be included for each assay batch, where the competitive ligand is absent and replaced with only DMSO. An additional method validation test that can be included in the assay procedure is a test of column performance using a sample prepared as a negative control but containing Tris-EDTA buffer in place of TTR. With this control, the ligands are in an unbound state and therefore should remain in the column after centrifugation, resulting in background radioactivity in the eluate. All assays should include triplicate samples, and should be conducted twice on separate days to include inter-and intra-day replicates (n = 6 total). Interlaboratory reproducibility of this in vitro assay was confirmed by comparing T4-TTR calibration data produced in our National Wildlife Research Centre (Ottawa, Canada) laboratory to those produced in the laboratory of Dr. Timo Hamers (Department of Environment & Health, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands) (Fig. 3).