Optimization of an ex vivo model to test the ability of chemicals to disrupt thyroid hormone synthesis

TL;DR

Researchers developed an improved lab model of rat thyroid glands to study how chemicals can interfere with thyroid hormone production.

Contribution

An optimized ex vivo rat thyroid model was developed that maintains hormone production and responds to thyroid inhibitors.

Findings

The model produced thyroxine (T4) for up to 9 days while preserving tissue structure.

Methimazole significantly reduced T4 output and increased gene expression of key thyroid synthesis enzymes.

The model is sensitive to thyroid hormone synthesis inhibitors and can detect thyroid disruption.

Abstract

Thyroid hormone (TH) synthesis and secretion can be perturbed by endocrine disrupting chemicals (EDCs). Traditional 2D cell culture models are useful for studying specific effects on enzymes and transporters involved in hormone synthesis; however, they lack the structural organization required for hormone production. In this study, we refine the conditions for an ex vivo rat thyroid gland model that supports robust thyroxine (T4) production and to evaluate the inhibition of T4 output using a model inhibitor. Thyroid glands from postnatal day 6 were excised and cultured for 7 or 9 days. We assessed the influence of medium composition (RPMI or StemPro™-34), culture method (hanging drop or filter inserts), thyroid stimulating hormone (TSH) concentration and iodide supplementation by measuring T4 secretion into the culture medium, evaluating tissue morphology and quantifying expression of key genes in TH synthesis machinery. Cultivating on filter inserts in StemPro™-34 media with iodide addition and 5 mU/mL TSH was the superior condition for enhanced T4 production while preserving tissue morphology. By challenging the ex vivo model with the antithyroid drug methimazole (MMI) T4 release was markedly reduced after 9 days of culture, and gene expression levels of Tpo and Slc5a5(NIS) were significantly upregulated. Together, these results demonstrate that the optimized ex vivo rat model supports sustained T4 production and is sensitive to a TH synthesis inhibitor, making it a promising model for studying the thyroid gland and the effects of thyroid-disrupting chemicals.

•Optimized ex vivo rat thyroid gland model producing thyroxine (T4).•Tissue morphology and function retained over 9 days of culture.•Responds to stimulation with TSH and inhibition by methimazole.•Evaluation by T4 concentrations, histology and RT-qPCR.•Methimazole challenge confirmed model detects thyroid disruption.

Optimized ex vivo rat thyroid gland model producing thyroxine (T4).

Tissue morphology and function retained over 9 days of culture.

Responds to stimulation with TSH and inhibition by methimazole.

Evaluation by T4 concentrations, histology and RT-qPCR.

Methimazole challenge confirmed model detects thyroid disruption.

Figures

Tissue integrity after culturing in hanging drop or on filter inserts in RPMI or StemPro medium. H&E stained sections of postnatal day 6 rat thyroid glands cultured for 7 days. Samples represent two tested media (RPMI and Stempro™-34), using two culture methods: hanging drop and filter inserts. All samples were exposed to 1 mU/mL TSH. The quality of tissue sections varied within groups, and the most representative samples are shown. Overview images and a magnification insert are shown. Scale bar = 50 µm, applies to all images (n = 6).

Tissue integrity and T4 levels after stimulating with thyroid stimulating hormone (TSH) and supplementing with potassium iodide (KI). A) Histological sections of postnatal day 6 rat thyroid glands cultured for 9 days. Glands were exposed to either 1, 5 or 10 mU/mL TSH concentrations with or without 1 µM KI. Overview images and a magnification insert shown. Scale bar = 50 µm, applies to all images (n = 6) and the most representative samples are shown. The group exposed to 5 mU/mL TSH supplemented with KI is used in this figure and as the control group in Fig. 3. B) Total accumulated T4 concentr

Histological evaluation after exposure to the thyroperoxidase (TPO) inhibitor methimazole (MMI). Histological sections of postnatal day 6 rat thyroids cultured for 9 days. All samples were exposed to 5 mU/mL TSH and 1 µM KI from day 0; MMI exposure began on day 3. Overview images and a magnification insert shown. Scale bar = 50 µm, applies to all images (n = 6) and the most representative samples are shown. The control group in this figure is also used in [Fig. 2](#fig0010) A (5 mU/mL TSH supplemented with KI).

Inhibition of TH synthesis by the thyroperoxidase (TPO) inhibitor methimazole (MMI). A) T4 concentrations measured at day 5, 7 and 9. Graphs are based on means from 5 to 6 samples for all time points (mean ± SD, n = 5–6). All samples were exposed to 5 mU/mL TSH and 1 µM KI from day 0; MMI exposure began on day 3. Asterisk indicates a p-value ≤ 0.05. B) Gene expression of selected apoptosis-related genes (*Bax* and *Bcl-2*) and thyroid-specific genes (*Slc5a5*(NIS*)* and *Tpo*) (n = 5–6). All samples were exposed to 5 mU/mL TSH and 1 µM KI from day 0; MMI exposure began on day 3. Asterisk indic

Full text

Introduction

1

Thyroid hormone (TH) synthesis takes place in the thyroid follicles [1]. This process is regulated by the hypothalamic-pituitary-thyroid (HPT) axis, which maintains circulating levels of triiodothyronine (T3) and thyroxine (T4). Regulation is achieved through hormones such as thyroid stimulating hormone (TSH), which binds to and activates the TSH receptor (TSHR) on thyrocytes [2]. The subsequent synthesis of T3 and T4 involves multiple enzymes, transporters, and molecular components, many of which depend on the 3D architecture of the thyroid tissue to function correctly [3], [4]. A critical early step in TH synthesis is iodide uptake via the sodium-iodide symporter (NIS), making its expression essential for TH production [1]. Importantly, NIS expression depends on the 3D organization of thyrocytes, which establishes the electrochemical gradient necessary for NIS-mediated transport [5], [6]. Thyroperoxidase (TPO) is a critical enzyme in the TH synthesis pathway, as it both catalyzes the oxidation of iodide and its incorporation into tyrosyl residues of thyroglobulin, enabling the formation of T3 and T4 [1]. TPO is a common target for TH system disruptors (THSD), and due to its central role, inhibition of TPO can significantly impair hormone synthesis and lead to hypothyroidism [7].

Several in vitro models exist for assessing specific effects within disruption of TH synthesis, including assays that detect inhibition of TPO, disruption of TH transport, or interference with the TSHR [8], [9]. However, 2D cultures often lack the capacity for TH synthesis due to the absence of the complex tissue architecture and cellular interactions necessary for full thyroid function [3], [4]. Recent studies have shown promising results with models capable of producing TH in vitro, such as a 3D human thyroid microtissue model based on primary cells, a thyroid follicular cell model based on human-induced pluripotent stem cells (iPSCs), and a human thyroid organoid derived from embryonic stem cells [10], [11], [12]. While these models hold great potential, they rely on scarce material or are highly labor-intensive.

Ex vivo cultured thyroid glands have previously been explored as a TH synthesizing model using both animal and human tissue [13], [14], [15], [16], [17], [18], [19]. Amphibian thyroid glands respond to TSH and classical TH synthesis inhibitors with reduced T4 secretion [13]. Thyroid glands from mouse embryos differentiate similarly to in vivo glands and exhibit gene expression changes and decreased tissue T4 levels upon pharmacological manipulation [18]. In rats, fetal ex vivo cultures have shown increased T4 expression tissue by immunostaining and reduced T4 expression, when exposed to 6-propyl-2-thiouracil (PTU), a known TH synthesis inhibitor [16]. Furthermore, an adult rat ex vivo model has been shown to respond comparably to human thyroid slices in response to classical TH synthesis inhibitors PTU and methimazole (MMI), as evidenced by TPO inhibition and altered expression of genes related to the TH synthesis pathway [17]. However, among existing ex vivo models, only one, using tadpole thyroid glands, has been evaluated for its capacity to produce and secrete T4, a key functional readout [13], [14]. This represents a notable limitation, as measuring T4 output is essential for determining whether molecular and histological changes translate into effects on functional output and for enabling direct comparison with in vivo data.

To address the abovementioned shortcomings of models related to TH synthesis and secretion, we aimed to establish an ex vivo rat thyroid culture protocol capable of detecting THSD through measurement of T4 output in culture medium. One of the objectives was to design a model that can sustain TH synthesis function and enable chemical testing. To optimize the model, we evaluated various cultivation methods, media compositions, culture durations, TSH concentrations, and iodide supplementation to establish conditions suitable for detecting disruptions in TH synthesis.

Materials and methods

2

Test substances and medium composition

2.1

A 100 mM stock solution of methimazole (MMI) (Sigma) was prepared in dimethyl sulfoxide (DMSO) (Sigma). Bovine thyroid stimulating hormone (bTSH) (Creative Biomart) was prepared in a stock of 2 U/mL in MilliQ water aliquoted and placed in freezer until use. Stocks of bTSH were used within 6 months after preparation. Potassium iodide (KI) (Sigma) was prepared in MilliQ water at 1 mM. Solubility of all stocks was evaluated by visual inspection. Two different mediums were used: 1) StemPro™-34 medium with added StemPro™-34 nutritional supplement (Gibco), 50 U/mL penicillin, 50 μg/mL streptomycin (Gibco), 400 μg/mL geneticin (Fisher Scientific) and 2 mM L-glutamine (Gibco) inspired by [19] or 2) RPMI 1640 (ThermoFisher) supplemented with 50 U/mL penicillin, 50 μg/mL streptomycin, 10% charcoal stripped Fetal Bovine Serum (Gibco), 10 mM HEPES (Sigma), 1x non-essential amino acids (Gibco) inspired by [18].

Animals

2.2

Time-mated nulliparous 8- to 9-week-old Sprague Dawley rats (Crl: CD(SD) – bred by Charles River Europe, and distributed by SCANBUR (Karlslunde, Denmark) – arrived on gestation day (GD) 16–17 to the animal facilities at the Technical University of Denmark. Animals were housed individually under standard conditions in High-Temperature polysulfone (H-TempTM) cages (Tecniplast) with wood chip bedding, nesting material, and a wooden shelter (Tapvei). They were kept in a controlled environment with 12 h light/dark cycles (with one hour acclimatization period), 55 ± 10% humidity, 22 ± 5 ºC temperature, and with around 75 ventilation air changes per hour. Animals were fed a standard soy and alfalfa-free diet based on Altromin 1314 (Altromin GmbH). Tap water in Bisphenol A-free bottles (84-ACTBT0402PFS, 400 mL bottle with ring, Ultem, Scanbur) was provided ad libitum. The animal experiments were overseen by our in-house Animal Welfare committee, Animal Welfare Body at DTU, and were approved ethically by the Danish Animal Experiments Inspectorate (license number 2020–15–0201–00539). Dams and postnatal pups were decapitated under CO_2_/O_2_ anesthesia on postnatal day (PD) 6 (the morning following overnight birth was termed PD1). Thyroid glands were excised, the thyroid lobes dissected, lightly cleaned and collected in DMEM/F12 (ThermoFisher Scientific). Dissections were performed over a period of approximately 3 h.

Ex vivo culture

2.3

Two ex vivo experiments were conducted to improve and optimize the culture protocol. The first experiment evaluated culture methods, medium composition, initial TSH concentrations and MMI exposure. The second experiment evaluated TSH concentrations, iodide supplementation, and responses to MMI exposure. Each condition in both experiments included thyroid gland samples from six PD6 pups, comprising tissues from three females and three males.

In the first experiment, the two thyroid lobes from each animal were either cultured separately in a hanging drop culture or together on a Millicel® standing insert with a PTFE membrane (Sigma) placed in 1000 µL media in a 6-well plate for 7 days. The inserts allow for the tissue to be placed in the liquid-air interface with a thin film of liquid wrapping the tissue. The hanging drop cultures were performed as described previously [19] in 40 µL media per droplet. In the experiment, TSH concentrations of 0, 0.1, 1, and 10 mU/mL were tested. MMI exposure began on day 0 with concentrations of 10 or 100 µM. A control group was also included. All groups had vehicle concentration kept constant at 0.1% DMSO. Medium was changed every 48 h for both culture methods, and the samples were visually inspected to monitor tissue integrity. The medium from each well or hanging drop was collected and stored at −80°C for later hormone analysis. On day 7, one thyroid lobe from each animal was frozen at −80°C for later gene expression analysis and the other thyroid lobe from each animal was fixed for 2–3 h in 10% neutral buffered formalin. The fixed tissue was processed in Excelsior AS Tissue Processor (Thermo Scientific), embedded in paraffin, and stored until histological evaluation.

For the second experiment, the two lobes from each animal were cultured together on a Millicel® standing insert with a PTFE membrane placed in 1000 µL medium in a 6-well plate. In this experiment, TSH concentrations of 1, 5, and 10 mU/mL were tested with and without the addition of 1 µM KI from day 0. MMI exposure began on day 3 with concentrations of 10, 50, or 100 µM. A control group was also included, and the vehicle concentration was kept constant at 0.1% DMSO in all groups. The medium was changed every 48 h starting on day 3, and the samples were visually inspected to monitor tissue integrity. The medium was collected and stored at −80°C for later hormone analysis. On day 9, one thyroid lobe from each animal was frozen at −80°C for later gene expression analysis and the other was processed for histology, as described above.

T4 hormone analysis

2.4

T4 culture medium concentrations were measured by Thyroxine Competitive ELISA kit (Thermo Fisher Scientific) according to the manufacturer's protocol. The samples were diluted in assay buffer to allow for appropriate extrapolation through the T4 standard curve of the kit. All samples were analyzed in technical duplicates. ELISA plates were analyzed on Spark Multimode Microplate Reader (Tecan) with absorbance measured at 450 nm. Data were corrected for background absorbance measured at 590 nm in each well and culture medium samples that were included on all plates. Concentrations were extrapolated from a four-parameter standard curve for the T4 standard. If the covariance between technical duplicates exceeded 20%, the ELISA analysis for the individual sample was repeated.

Histopathology

2.5

Sections were cut at 4 µm thickness on a HM450 microtome. Two sections, representing two tissue depths per sample, were stained with Mayers hematoxylin and eosin (H&E) following standard protocols. The slides were scanned under a 40x objective in a Pannoramic Midi II digital slide scanner (3DHISTECH Ltd.). The resulting images were qualitatively evaluated by visual inspection for histological changes such as irregular follicles, signs of autolysis, epithelial cell height, decreased follicular lumen area, follicular epithelial cell hypertrophy, and hyperplasia.

RNA extraction, cDNA preparation, and RT-qPCR assay

2.6

Total RNA was extracted from ex vivo cultured thyroid glands. Glands were thawed on ice and homogenized in 200 μL Homogenization solution (Promega) with the addition of 20 µL 1-thioglycol (Promega) per 1 mL of homogenization solution, using a Tissuelyzer II (Qiagen, Hilden) at 30 oscillations per sec for 2 min. RNA was isolated with Maxwell® RSC simplyRNA Tissue Kits (Promega) using a Maxwell® RSC48 instrument (Promega, Denmark) following the manufacturer's protocol. RNA quantity and purity were determined with a NanodropOne spectrophotometer (Thermo Fischer Scientific). For purity an OD260/280 > 1.73, was accepted for further analysis.

The synthesis of cDNA and RT-qPCR protocols were as previously described [20], with some minor modifications. In short, cDNA was prepared from 300 ng RNA using the Omniscript RT kit (Qiagen) according to the manufacturer's protocol. cDNA was diluted (1:20) for all TaqMan assays, except for cDNA used for the TaqMan assay Slc5a5(NIS) (Rn00583900_m1), which was run separately with reference genes at a dilution of 1:6.

The gene-specific TaqMan Gene Expression Assays (Life Technologies) were used: Rps18 (Rn01428913_g1) and Sdha (Rn00590475_m1) as reference genes, Tpo (Rn00571159_m1) and Slc5a5(NIS) (Rn00583900_m1) due to their relation to TH synthesis and function, and Bcl-2 (Rn99999125_m1) and Bax (Rn01480161_g1) as apoptotic markers [21]. Samples were run in technical duplicates with a predefined acceptance criterion: the difference between duplicates should not exceed 0.5. The data were analyzed using the comparative Ct method [22] and normalized to the geometric mean of the two stably expressed reference genes, Rps18 and Sdha [23].

Data analysis

2.7

Data were assessed for normal distribution by applying the Shapiro-Wilks test. Variance homogeneity was confirmed by applying the Brown-Forsythe test. Data not normally distributed were log-transformed and reassessed. All data met the criteria and underwent analysis with one-way ANOVA followed by Dunnett's multiple comparison test. All statistical analyses were performed in GraphPad Prism 10 (GraphPad Software), and the significance level was set to ≤ 0.05 in all statistical tests.

Results and discussion

3

Selecting the medium and culture method

3.1

The initial optimization of the model was guided by histological outcomes comparing RPMI and StemPro™-34 media in tissues cultured using either hanging drops or filter inserts. Substantial differences in tissue integrity and follicular morphology were observed between conditions (Fig. 1).Fig. 1. Tissue integrity after culturing in hanging drop or on filter inserts in RPMI or StemPro medium. H&E stained sections of postnatal day 6 rat thyroid glands cultured for 7 days. Samples represent two tested media (RPMI and Stempro™-34), using two culture methods: hanging drop and filter inserts. All samples were exposed to 1 mU/mL TSH. The quality of tissue sections varied within groups, and the most representative samples are shown. Overview images and a magnification insert are shown. Scale bar = 50 µm, applies to all images (n = 6).

Thyroid glands cultured using the hanging drop method exhibited poor tissue preservation, with widespread necrosis and an absence of distinguishable thyroid follicles. In contrast, samples cultured on filter inserts had improved structural integrity with clear follicular organization across both media types. However, RPMI-cultured samples displayed early signs of autolysis (cellular degeneration and single cell necrosis) in at least half of the samples assessed histologically. Notably, StemPro™-34 medium-cultured thyroids on filter inserts maintained the most regular follicular architecture, with well-defined colloid-filled lumens and the least severe changes of the follicular epithelium (vacuolation and/or hypertrophy observed in some of the samples) and minimal signs of autolysis, and these conditions were selected for further testing.

The poor structural preservation observed in the hanging drop differs from our previous report, where hanging drop cultures of ex vivo rat thyroids maintained normal histology [19]. However, it is important to note that the culture period in the previous study was only 2 days, compared to 7 days in this study. One plausible explanation for the poor tissue integrity observed here is the increased risk of dehydration and the accumulation of waste products within the small medium volume, especially with a longer culture period. Previous studies have successfully used 40–50 µL per drop without compromising tissue integrity [15], [19]. 48 h intervals for media change has been suggested for hanging drop cultures [15], more frequent changes, such as every 24 h, have previously been tested (unpublished data) but did not lead to improved tissue integrity. In contrast, cultures using filter inserts likely provided more efficient atmospheric gas exchange, preserving follicular architecture more effectively. The larger medium volume of 1 mL could ensure dilution of waste products. This observation aligns with reports from similar ex vivo systems, where filter inserts supported viable and functional tissue for extended periods [18], [24]. These findings suggest that filter inserts may represent a more robust option for long-term culture. Media composition also affected morphology. These results demonstrate that both culture method and medium composition contribute towards maintaining thyroid tissue ex vivo.

Selecting thyroid stimulating hormone concentrations for culture

3.2

With the medium and culture method established, the next parameter optimized was TSH concentration. The goal was to maintain tissue morphology comparable to that of normal thyroid tissue, while ensuring the model supported robust T4 production. This included achieving a measurable increase over basal hormone levels, enabling the reliable detection of inhibitory effects. TSH plays a central role in TH synthesis by binding to the TSH receptor (TSHR) and activating downstream pathways essential for iodide uptake, thyroglobulin iodination, and hormone release [1]. An initial range of 0, 0.1, 1 and 10 mU/mL TSH was tested (Figure S1), revealing a peak in T4 release on day 3 at concentrations of 1 and 10 mU/mL, followed by a gradual decrease in TH levels through to day 7. We hypothesized that this initial release of T4 was a release of pre-synthesized hormone stored in the follicular colloid; a phenomenon also reported in amphibian thyroid cultures [13]. The lowest TSH concentrations (0 and 0.1 mU/mL) did not significantly stimulate T4 production, leading to the selection of 1, 5, and 10 mU/mL for further analysis (Fig. 2). To improve the dynamic range of the test system, we increased the culture period to 9 days, thus diminishing the contribution from follicular TH release and focusing on de novo synthesis alone. To create optimal conditions, we added iodide to ascertain that iodide deficiency during prolonged culture would not affect total TH synthesis capacity.Fig. 2. Tissue integrity and T4 levels after stimulating with thyroid stimulating hormone (TSH) and supplementing with potassium iodide (KI). A**)** Histological sections of postnatal day 6 rat thyroid glands cultured for 9 days. Glands were exposed to either 1, 5 or 10 mU/mL TSH concentrations with or without 1 µM KI. Overview images and a magnification insert shown. Scale bar = 50 µm, applies to all images (n = 6) and the most representative samples are shown. The group exposed to 5 mU/mL TSH supplemented with KI is used in this figure and as the control group in Fig. 3. B**)** Total accumulated T4 concentrations measured at multiple time points over 9 days of culture. Glands were exposed to either 1, 5 or 10 mU/mL TSH concentrations, with or without 1 µM KI. The dotted line at 100 ng/mL is included for easier comparison. Graphs are based on means from 5 to 6 samples (each representing TH release from the two lobes form each animal) for all time points (mean ± SD, n = 5–6).

Histological evaluation revealed clear TSH concentration-dependent changes in thyroid morphology. In the absence of iodide, increasing TSH levels led to pronounced activation of follicular epithelial cells and follicles, characterized by colloid depletion**,** irregular follicle shape, and increased epithelial cell height (columnar shape), indicative of heightened thyrocyte activity (Fig. 2 A). When 1 µM KI was added, the morphological changes were smaller, and epithelial cell height was less affected at higher TSH levels. T4 measurements (Fig. 2B) showed that 5 mU/mL TSH resulted in the highest T4 output, but that iodide surprisingly did not appear to affect T4 levels. Based on these findings, 5 mU/mL TSH with 1 µM KI was selected for subsequent experiments.

Histological evaluation after MMI exposure

3.3

With all culture parameters optimized, we decided to challenge the model with the antithyroid drug MMI because of its well-defined, single-target mechanism; inhibition of TPO activity [25].

Exposure to increasing concentrations of MMI after 9 days of tissue culture resulted in apparent histological alterations of the thyroid tissue (Fig. 3). Higher MMI concentrations induced irregular follicular architecture, colloid depletion, and increased follicular cell height as observed qualitatively by histological evaluation, similarly to the effects induced by 10 mU/mL TSH exposure (Fig. 3). Notably, these alterations align with histological findings reported in in vivo rat studies of PD16 pups exposed to MMI during development, including markedly irregular follicles, colloid depletion, reduced follicular lumen area, and follicular epithelial hypertrophy, supporting the physiological relevance of the observed responses [26]. Similarly, in an amphibian ex vivo model [13], histological comparisons between ex vivo and in vivo tissues revealed comparable responses to similar MMI concentrations, as both models exhibited follicular cell hypertrophy and hyperplasia. This further highlight the utility of ex vivo systems for modeling thyroid disruption.Fig. 3. Histological evaluation after exposure to the thyroperoxidase (TPO) inhibitor methimazole (MMI). Histological sections of postnatal day 6 rat thyroids cultured for 9 days. All samples were exposed to 5 mU/mL TSH and 1 µM KI from day 0; MMI exposure began on day 3. Overview images and a magnification insert shown. Scale bar = 50 µm, applies to all images (n = 6) and the most representative samples are shown. The control group in this figure is also used in Fig. 2 A (5 mU/mL TSH supplemented with KI).

Testing the model for thyroid hormone synthesis inhibition

3.4

In the first experiment, MMI exposure began on day 0, resulting in increased initial T4 release on day 2 compared to the TSH control (Figure S2), suggesting that TH synthesis inhibition may increase the follicular release of pre-synthesized THs. We previously observed the same phenomenon with PTU exposure in ex vivo cultured tissue (unpublished data). This prompted us to delay MMI exposure to day 3.

T4 measurements revealed a significant reduction in MMI-exposed samples compared to controls during days 5–7 and 7–9 (Fig. 4), confirming the inhibitory effect of MMI on TH synthesis. This effect was not observed on day 3–5, supporting the notion that follicular T4 release can mask the inhibitory effect on de novo synthesis. These two mechanisms, colloid TH release and de novo synthesis, should be carefully considered when using ex vivo culture models, as inhibitors of TH synthesis potentially could induce TH release in the early phases due to colloid release in response to inhibition, as was also observed in the amphibian model [13]. Exposure after the colloid release, however, may better detect the inhibitory effect on T4 synthesis. Therefore, longitudinal sampling appears essential to capture both the transient and sustained effects of THSDs accurately.Fig. 4. Inhibition of TH synthesis by the thyroperoxidase (TPO) inhibitor methimazole (MMI). A) T4 concentrations measured at day 5, 7 and 9. Graphs are based on means from 5 to 6 samples for all time points (mean ± SD, n = 5–6). All samples were exposed to 5 mU/mL TSH and 1 µM KI from day 0; MMI exposure began on day 3. Asterisk indicates a p-value ≤ 0.05. B) Gene expression of selected apoptosis-related genes (Bax and Bcl-2) and thyroid-specific genes (Slc5a5(NIS*)* and Tpo) (n = 5–6). All samples were exposed to 5 mU/mL TSH and 1 µM KI from day 0; MMI exposure began on day 3. Asterisk indicates a p-value ≤ 0.05.

The Bax/Bcl-2 ratio, an indicator of apoptotic susceptibility [21], remained unchanged across MMI concentrations (Fig. 4). This suggests that MMI does not compromise tissue integrity, supporting the conclusion that the observed decrease in T4 is due to inhibition of hormone synthesis rather than decreased cell viability. This also suggests that the altered tissue morphology was a result of functional adaptation to TPO inhibition. Slc5a5(NIS) expression was significantly upregulated at 50 and 100 µM MMI, while Tpo gene expression increased only at 100 µM (Fig. 4). This could reflect a compensatory response to TPO inhibition and impaired TH synthesis, as the thyroid gland attempts to increase iodide uptake and hormone production through local feedback mechanisms.

Perspectives

3.5

In animal studies, exposure to MMI can induce a pronounced systemic hypothyroidism where low TH levels triggers an increased release of thyrotropin-releasing hormone (TRH) from the hypothalamus and pituitary release of TSH [27]. The elevated levels of TSH stimulates the thyroid gland via the TSHR to activate gene transcription, TH synthesis and secretion as well as thyrocyte proliferation and hypertrophy [27], [28], [29]. These changes reflect a systemic response mediated by endocrine feedback within the HPT-axis. In contrast, our ex vivo model uses isolated thyroid gland tissue, meaning that the observed effects arise from direct chemical interaction with the tissue, potentially activating paracrine or autocrine regulation of the gland. We found MMI-induced changes to both histology and gene expression, indicating compensational responses to TPO inhibition or direct effects of MMI. Similar paracrine regulation of TH synthesis has been reported in other ex vivo and in vitro systems [13], [30], together suggesting that perhaps MMI does not solely induce changes via HPT-axis activation, but also exerts direct effects on the thyroid gland. This distinction is important for our model because it suggests that the observed changes reflect intrinsic thyroid responses rather than systemic endocrine feedback, allowing for mechanistic studies of local processes and direct chemical effects within the gland.

Ex vivo thyroid models have proven valuable as an intermediate step between simplified in vitro assays and complex in vivo studies, enabling more reliable extrapolation of TH disruption. This has been demonstrated by comparing responses to TPO inhibition in in vitro and ex vivo models, using thyroid gland microsomes, and using this data to quantitatively predict reductions in circulating T4 levels in rats [31]. The study demonstrated a clear relationship between TPO inhibition and TH disruption using reference chemicals, such as PTU and MMI. In the future, our ex vivo model could similarly be integrated into quantitative extrapolation frameworks, supporting predictive toxicology.

Conclusion

4

This study demonstrates the feasibility of using an ex vivo rat thyroid model to detect TH disruption caused by TPO inhibition, by monitoring changes in T4 released into the culture medium. By integrating functional hormone output, histological integrity, and gene expression endpoints, the model provides a platform for mechanistic studies. Importantly, we have established parameters critical for maintaining healthy tissue over extended culture periods. The model responded predictably to TSH stimulation and MMI exposure, confirming its functional responsiveness.

CRediT authorship contribution statement

Eleni Barmpari: Writing – review & editing, Conceptualization. Louise Ramhøj: Writing – review & editing, Supervision, Formal analysis, Conceptualization. Mette Stub: Writing – review & editing, Visualization, Investigation, Formal analysis, Data curation, Conceptualization. Mikala Melchiors: Writing – review & editing, Writing – original draft, Visualization, Investigation, Formal analysis, Data curation, Conceptualization. Terje Svingen: Writing – review & editing, Supervision, Resources, Funding acquisition, Conceptualization. Anna Kjerstine Rosenmai: Writing – review & editing, Supervision, Formal analysis, Conceptualization. Anna Opstrup Bindel: Writing – review & editing, Investigation. Kieu-mi Tran: Writing – review & editing, Investigation.

Funding

The research was funded by the 10.13039/501100007036Danish Environmental Protection Agency and Centre on Endocrine Disrupters (CeHoS).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.