Gestational Exposure to Perchlorate in the Rat: Thyroid Hormones in Fetal Thyroid Gland, Serum, and Brain

Abstract

Iodine is essential for production of thyroid hormones. Perchlorate is an environmental contaminant that interferes with iodine uptake into the thyroid gland to reduce thyroid hormone synthesis. As thyroid hormones are critical for brain development, exposure to perchlorate during pregnancy is of concern for the developing fetal brain. In this study we 1) define profiles of thyroid hormones in the maternal and fetal compartments of pregnant rats in response to inhibition of the sodium-iodine symporter (NIS) by perchlorate, 2) expand inquiry previously limited to serum to include fetal thyroid gland and brain. Perchlorate was added to the drinking water (0, 1, 30, 300, 1000ppm) of pregnant rat dams from gestational days (GD) 6–20. On GD20, blood, thyroid gland and brain were collected from the fetus and dam for thyroid hormone and molecular analyses. Thyroid gland and serum thyroid hormones were dose-dependently reduced, with steeper declines evident in fetus than the dam. The thyroid gland revealed perturbations of TH-action with greater sensitivity in fetus than dam. Thyroid hormones and thyroid hormone-responsive gene expression were reduced in the fetal cortex portending effects on brain development. These findings are the first quantitative assessments of perchlorate-induced deficits in the fetal thyroid gland and fetal brain. We provide a conceptual framework to develop a quantitative NIS AOP for serum thyroid hormone deficits and the potential to impact the fetal brain. Such a framework may also serve to facilitate the translation of in vitro bioactivity to the downstream in vivo consequences of NIS inhibition in the developing fetus.

Introduction

Epidemiological studies, animal models, and in vitro screening test systems have identified a large number of environmental contaminants with the potential to disrupt the thyroid system (Brucker-Davis 1998; Leemans et al. 2019; Noyes et al. 2019). These include chemical actions on thyroid hormone (TH) synthesis, metabolism, regulation and signaling. Thyroid hormones are required for normal somatic growth and organ system development and are especially critical for proper nervous system development (Bernal 2017; Williams 2008). As such, environmental chemicals that disrupt the thyroid system are of particular regulatory concern for potential negative effects on the developing brain (Gilbert et al. 2020).

The US Environmental Protection Agency (US EPA) is among a host of international regulatory bodies (e.g., OECD, EFSA, ECHA) that have adopted the Adverse Outcome Pathway (AOP) framework as a means to organize and integrate information from multiple data sources to inform scientifically sound regulatory practices (Ankley et al. 2010; Villeneuve et al. 2014). Anchored by a chemical’s interaction at a molecular target (molecular initiating event, MIE) on one end, and an adverse health outcome (AO) on the other, an AOP lays out the sequence of biologically plausible events (Key Events, KEs) and their interrelationships (KERs) that connect MIE to AO. A number of AOPs delineating thyroid system perturbations at several distinct target sites have been curated, each leading to species-specific developmental impairments – i.e., altered metamorphosis, swim bladder inflation, and brain function in frogs, fish, and mammals, respectively (Crofton et al., 2018; Haselman et al. 2020; Hassan et al. 2017; Nelson et al. 2016; Noyes et al. 2019; Rolaki et al. 2019). Serum THs, a common metric evaluated in clinical, epidemiological, and regulatory settings, occupy a pivotal node upon which many chemicals with distinct molecular targets (MIEs) may converge (Crofton 2018; Gilbert et al. 2020; Haselman et al. 2020; Hassan et al. 2019; Hassan et al. 2017; Noyes et al. 2019; O’Shaughnessy and Gilbert 2019; Rolaki et al. 2019).

One site of chemical interference with significant ramifications for thyroid hormone production is the sodium-iodide symporter (NIS). Iodine, an essential element for the synthesis of thyroid hormones and is actively transported into thyroid follicular cells via the NIS (Dohán et al. 2007; Ravera et al. 2017). A very potent NIS inhibitor is the environmental contaminant perchlorate (Wang et al. 2019; Wolff 1998). Perchlorate has been used commercially for many years as an oxidizer in solid rocket propellants, fireworks, and air bag inflation systems (National Research Council 2005; USEPA 2002). Being water soluble, the perchlorate anion (ClO_4^-) remains stable and persists in the environment. It is found in drinking water, food products, and infant formula (Abt et al. 2018) and been detected in human serum, placenta, amniotic fluid and breast milk (Blount and Valentin-Blasini 2006; Dasgupta et al. 2008; Kirk et al. 2005; Zhang et al. 2016). Predating the articulation of the AOP framework, mode-of-action (MOA) and biologically-based-dose-response (BBDR) models for perchlorate were published by US EPA in the context of regulatory action as early as 2002 (National Research Council 2005; USEPA 2002). The primary concern then, as now, is potential for NIS-mediated thyroid hormone suppression to affect the developing brain. A number of epidemiological studies have since reported associations between perchlorate exposure and serum thyroid hormones (Blount et al. 2006; Steinmaus et al. 2016; Taylor et al. 2014; Zhang et al. 2016), while others have further demonstrated associations between perchlorate exposure, adverse birth outcomes and child neurological development (Rubin et al. 2017; Taylor et al. 2014).

Despite these epidemiological findings and the established role of thyroid hormones in brain development, rodent studies designed to interrogate a mechanistic link between NIS inhibition and TH-dependent neurotoxicity, have been largely inconclusive. Changes in locomotor activity were reported in offspring of rats exposed to perchlorate during pregnancy (York et al. 2004), whereas others failed to observe alterations in a suite of neurobehavioral tests (Gilbert and Sui 2008b; York et al. 2004; York et al. 2005b). Despite predominantly negative findings in neurobehavioral assessments, dose-dependent reductions in excitatory synaptic transmission in the hippocampus were observed in adult offspring, indicative of a permanent functional impairment in the brain (Gilbert and Sui 2008b). Although maternal serum thyroid hormones have routinely been significantly and dose-dependently reduced by perchlorate (Gilbert and Sui 2008a; York et al. 2004; York et al. 2003; York et al. 2005b), in our studies, serum thyroid hormone levels in the nursing neonate were only marginally suppressed (Gilbert and Sui 2008a). Fetal serum measures were not available, but the absence of hormone deficits in the early postnatal period suggests that prenatal thyroid hormone insufficiency may underlie the reported synaptic deficits. In studies where serum hormones have been reported in fetus/newborn rat, the assays often lacked sufficient sensitivity for accurate quantification of thyroid hormones in the sera at this age (York et al. 2004; York et al. 2003). In addition. these studies, including our own, failed to control iodine levels in chow and/or water - dietary iodine levels in commercial rat chow often far exceeding biological requirements (Fisher et al. 2013; Gilbert et al. 2013) and potentially masking detrimental effects of NIS inhibitors on hormone production (Clewell et al. 2003a; Clewell et al. 2003b; Merrill et al. 2005).

The present study was designed to refine the dose-response profiles of thyroid hormones in the maternal and fetal compartments and extend the range of observations beyond serum in response to gestational exposure to this classic NIS inhibitor. For the first time, in the same study, hormones are quantified in thyroid gland, serum, and brain founding the basis for development of a quantitative AOP for NIS inhibition and neurodevelopment.

Materials and Methods

Subjects

Pregnant Long–Evans (n=45) rats were obtained from Charles River (Raleigh, NC) on gestational (GD) 2 and housed individually in standard plastic hanging cages in an approved animal facility. All experiments were conducted with prior approval from the US EPA’s Institutional Animal Care and Usage Committee (IACUC) and were carried out in an Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) approved facility. Animal rooms were maintained on a 12:12 light:dark schedule and animals were permitted free access food. On arrival on GD2 pregnant dams were individually housed and placed on an iodine-controlled casein-based diet (D10001, Research Diets, Newark, NJ, 225 ng of iodine/gm of chow) and given free access to deionized water. This level of iodine is <25% of that found in typical rodent chows yet is more than sufficient for adequate hormone production in pregnant and non-pregnant rats (see (Fisher et al. 2013; Gilbert et al. 2013; Gilbert et al. 2011). On GD 6 animals were weighed and based on weight equally distributed to 1 of 5 dose groups. From GD6 to GD20 dams were administered 0, 1, 30, 300 or 1000 ppm ammonium perchlorate (ClO₄, CAS No. 77, purity 99.5%, Sigma, St. Louis, MO) in the drinking water. Doses were selected based on previous work from our laboratory (Gilbert and Sui 2008a) to produce a graded reduction in serum T4 in rat dams in the absence of any overt signs of systemic toxicity. Dam body weights were monitored throughout pregnancy.

Tissue Collection.

Dam blood was collected from tail vein on GD16 by briefly restraining animals in a Tail-Veiner restrainer (Braintree Scientific Inc, Braintree, MA) with their tails exposed and swabbed with a warm towel. A 21-gauge heparinized butterfly needle was inserted into the tail vein and 0.5–1mL of blood was collected into a serum microtainer tube. On GD20, dams were euthanized by decapitation without anesthesia and trunk blood collected. The uterine horn was removed and fetal trunk blood collected following decapitation and pooled from all progeny within a litter at that time. Samples were allowed to clot on ice, centrifuged and serum collected and stored at −80 C until analysis. All samples were collected between the hours of 8:00 am and 12:00 pm and each dose group consecutively sampled to control for diurnal fluctuations in thyroid hormones.

Thyroid glands were collected from dams and fetuses at the time of euthanasia on GD20 and flash frozen for gene expression and thyroid hormone analysis. Whole fetal brains were extracted from the skull, rinsed in cold phosphate buffered saline, blotted dry, and flash frozen in liquid nitrogen. Forebrains were collected from littermates for gene expression analysis, rinsed in cold phosphate buffered saline, and saved in RNAlater^™ (Invitrogen) according to the manufacturer’s recommendation. All samples subsequently stored at −80˚C until analysis.

Quantification Of Thyroid Hormones In The Serum, Thyroid Gland, Brain

Thyroid hormones were measured in serum, thyroid gland, and brain by Liquid Chromatography Mass Spectrometry (LCMS) using an AB Sciex (Framingham, MA) Exion AC UHPLC-Qtrap 6500+ Linear Ion Trap LC/MS/MS system as previously described (Hassan et al. 2017; Hornung et al. 2015; O’Shaughnessy et al. 2018a). Solvent-based calibration standards were used for quantitation over a range of 10 – 10,000 pg/mL, each curve used a minimum of five sequential points and correlation coefficients of the curves were ≥ 0.99. Two ion transitions were monitored for each target analyte, qualitatively identified based on retention time relative to the internal standard and calibration standard, and the ratio of the peak areas of the monitored ion transitions. The lower limit of quantitation (LLOQ) for each analyte was set to the concentration of the lowest calibration standard that gave an acceptable ion ratio, and acceptable recovery of ±30% of the spike amount. Each sample batch consisted of a method blank, a laboratory control sample (blank spike), and a continuing calibration verification sample prepared in solvent.

Serum TH. Triiodothyronine (T3) and thyroxine (T4) were measured in serum from dams on GD 16 and GD20, T4 was measured in serum from GD20 fetuses. Fetal (50μL) and dam serum (20μL) were aliquoted into Eppendorf tubes and spiked with 5μL of stable isotope internal standard containing ¹³C₆T3 and ¹³C₆T4 dissolved in methanol. Dam and fetal samples received 20μL or 50μL of the following, 1N HCL, 250μL H₂O, and 150μL of 50:50 H₂O: acetonitrile (v/v) with 0.1% formic acid. The sample were vortex mixed briefly and incubated at 37°C for 2 hours (h). Samples were cooled at room temperature and diluted with 0.1% aqueous formic acid to a total volume of 300μL dam and 750μL fetal. Samples were vortexed and put through solid phase extraction (SPE) using Evolute CX 1 mL SPE cartridges (Biotage, Charlotte, NC) with a vacuum manifold. Dam serum was processed through 10 mg SPE cartridge while 25 mg cartridge was used to process the fetal samples. Cartridges were first conditioned with methanol followed by 0.1% formic acid. Samples were then applied to the cartridge then washed with 0.1% formic acid followed by methanol. Thyroid hormones were eluted using 5% NH₄OH in to 15 ml polypropylene tubes and evaporated to dryness using nitrogen. The dried samples were reconstituted with 100μL of 25%:75% acetonitrile:water (v/v) with 0.1% formic acid and transferred into micro-sampling vails (Agilent Technologies, Santa Clara, CA). The LLOQ for serum T3 and T4 was 0.01 ng/ml.

TSH was analyzed by radioimmunoassay in the dam serum based on methods published in Louis et al. (2017) and Stoker et al. (2010). TSH was not determined in fetal serum due to lack of sufficient sample volume required for the assay. The limit of detection for TSH was 0.389 ng/ml, inter- and intra- assay variation <10%.

Thyroid Gland Hormones.

Thyroid gland hormone analysis was conducted using a modified method published in Hassan et al 2017. Thyroid glands were homogenized in Tris buffer using FastPrep-24 tissue and cell-lyser (MP Biomedicals, Solon, OH). Thyroid gland homogenates were thawed on ice and 50 μL (fetal) or 5μL (dam) of the homogenates were transferred into Eppendorf tubes containing 450 or 495 μL of protein digest buffer (PDB) made up of 100 mM Tris-Cl, pH 8.4, 2.5 mM 6-propylthiouracil (PTU), and 1% Triton X-100. Diluted homogenates, 300μL, were transferred to Eppendorf tubes and an equal volume of 20 mg/mL pronase (>3.5 units/mg, Sigma P8811) in PDB was added and the samples were vortexed and incubated at 37°C for enzyme digestion. After 2h and 24 h, 200 μL of the digested sample was removed, and to stop enzyme digestion, the samples were into another Eppendorf tube containing an equal volume of aqueous 4% formic acid and 20% acetonitrile. Prior to SPE, 5 μL of stable isotope internal standard solution (¹³C6-T2, ¹³C6-T3, ¹³C6-T4 at 100 ng/mL in methanol) (Isosciences, King of Prussia, PA) was spiked into the samples. The thyroid gland digests were transferred to 96-well 10 mg Evalute Express CX plate (Biotage, Charlotte,NC) and process through SPE. First the SPE columns were conditioned with 400 μL of methanol followed by 2% formic acid in H₂O (2×400 μL) and pressure was applied using a positive pressure manifold (PPM-96, Biotage, Charlotte, NC). The samples were then quantitatively transferred on to the preconditioned columns then low pressure was applied. The wells were washed with 400 μL of each of the following solutions, 50 mM ammonium acetate, 2% formic acid, and methanol, respectively. Samples were eluted three times with 200 μL of 5%NH₄OH in methanol into 96-well sample collection plate and evaporated to dryness with nitrogen. The dried samples were resuspended with 100μL of 5% acetonitrile: 95%water mixture with 0.1% formic acid. The LLOQ for monoiodtyrosine (MIT) and diiodotyrosine (DIT) was 0.10 ng/gm. The LLOQ for T4, T3 and reverse T3 (rT3) was 0.01 ng/gm.

Brain Thyroid Hormones.

Frozen whole brain samples were weighed and immediately transferred to a 15 ml centrifuge tube containing methanol 1 mM PTU and spiked with a mixed stable isotope solution containing ¹³C₆-T2, ¹³C₆-T3 and ¹³C₆-T4. THs were extracted from the brain using methanol, chloroform and an aqueous solution of 0.05% CaCl₂ followed by nitrogen evaporation. One ml of water was subsequently added to the tube and the total volume of the aqueous mixture was determined. Equal volumes of 80/20 (v/v) 4% formic acid (aq) and acetonitrile were added and the sample was vortexed then thyroid hormones were extracted using an Evolute 50 mg, 3 ml CX solid phase extraction (SPE) cartridge (Biotage, Charlotte, NC, USA) on a vacuum manifold. The SPE cartridge was conditioned with methanol followed by 2% formic acid. After sample application, the cartridge was washed with 50 mM ammonium acetate (pH 6), 2% formic acid and then methanol. Thyroid hormones were eluted using 5% NH₄OH in methanol, evaporated to dryness using nitrogen and the residue dissolved in 100 μl of 5% acetonitrile: 95% water with 0.1% formic acid for transfer to an amber liquid chromatography (LC) micro-vial. SPE reconstituted extracts were analyzed for thyroid hormone by stable isotope dilution LC/MS/MS as described above for serum. The LLOQ for both T4 and T3 in the brain tissue was 0.1 ng/g.

Gene Expression in Thyroid Gland and Fetal Brain by Quantitative Real Time-PCR.

Gene expression was assessed in the thyroid glands of dams and fetuses and in forebrain of fetus on GD20 according to standard procedures. Briefly, total RNA was extracted via TRIzol^® (Invitrogen) according to manufacturer’s protocol. RNA pellets were resuspended in nuclease-free H2O, and RNA concentrations measured on Nanodrop 1000 (Nanodrop Technologies). RNA samples were treated with DNase I (Promega, M6101) and quantified using the Ribogreen Quantitation Kit (ThermoFisher, R11490). DNase I treated RNA was reverse transcribed with the ABI cDNA Archive Kit (ThermoFisher, 4322171) and 25 ng equivalent cDNA was amplified in a 12 ml volume using ABI TaqMan Gene Expression Assays and ABI Universal Master Mix (ThermoFisher, 4304437). Amplification was performed on an ABI model 7900HT sequence detection system using standard Taqman cycling parameters. All samples were run in technical duplicate. Molecular probes for the thyroid gland were chosen for their role in hormone synthesis. Brain targets were based on those previously identified in fetal brain in response to hypothyroxinemia and hypothyroidism in propylthiouracil model (O’Shaughnessy et al. 2018b). All Taqman probes used for these analyses are summarized in Table 1. Beta-2-microglobulin (B2m) or β-actin was used as the endogenous control as neither showed a significant change between control and treated groups by single factor ANOVA. Data were analyzed by relative gene quantification using the 2_ddCt method (Livak and Schmittgen 2001).

Table 1.

Taqman Probes Used for qRT-PCR in Dam and Fetus Thyroid Glands and in Fetal Forebrain on GD20

Thyroid Gland Transcripts		Gene Name	ID
Iodine transporters	Slc5a5	Sodium-iodine symporter— NIS	Rn00583900_ml
	Slc26a4	Pendrin	Rn01469208_ml
Iodine recycling	Diol	Deiodinase-1	Rn00572183_ml
Thyroid hormone biosynthesis	Tpo	Thyroid peroxidase	Rn00571159_ml
	Tg	Thyroglobulin	Rn00667257_gl
Thyroid hormone regulation	TshR	TSH receptor	Rn00563612_ml
Thyroid gland morphogenesis	Nkx2.1	NK2 homeobox 1, transcription factor	Rn01512482_ml
	Pax8	Paired box 8, transcription factor	Rn00579743_ml
Fetal forebrain transcripts	Bdnft	Brain-derived neurotropic factor, total	Rn02531967_sl
	BmpT	Bone morphogenetic protein 7	Rn01528889_ml
	Camk4	Calcium-calmodulin kinase 4	Rn00664802_ml
	ThrB	Thyroid hormone receptor beta	Rn00562044_ml
Reference genes	βP2m	Beta 2-microglobulin	Rn00560865_ml
	β-actin	Beta actin	Rn00667869_ml

Statistical Analysis.

All data were expressed as mean ± standard error mean (SEM). Statistical significance of treatment effects on thyroid hormone concentrations were assessed with one-way analysis of variance (ANOVA) using SAS Version 9.2 (Cary, NC). When a significant treatment effect was detected (p<0.05) pairwise difference among groups were tested post hoc using Dunnett adjustment of multiple comparisons. For measures taken across time, repeated measure analyses were performed. Mean fold change in gene expression were set at 1.5-fold change and a 5% false discovery rate. To control for experiment-wise error, alpha was reduced to 0.020 for thyroid gland and 0.025 for brain by dividing 0.05 by the square root of number of targets examined.

Results.

Perchlorate exposure did not change dam body weight or litter size (Figure 1 A and B). Thyroid gland weights were increased in dams at the two highest doses (Figure 1C). Serum T4 was reduced in dams on GD16 at the two highest dose levels (Figure 2A), while T3 concentrations remained unchanged (Figure 2B). By GD20, decreases were evident in dam serum T4 and T3, and TSH was increased in the 300 and 1000ppm dose groups (Figure 2D–2F). Fetal serum concentrations of T4 were also reduced at these doses (Figure 2E).

Figure 1.

Mean (+/− SEM) dam body weight, litter size, thyroid gland weight (n=7–9/dose group). A) Dam body weight was not affected by perchlorate at any dose level. Results of repeated measures ANOVA revealed main effect of Gestational Day [F(5, 190)=15.2, p<0.0001] but no effect of Dose or Age X Dose interaction [p’s>0.72]. B) Litter size was not altered by perchlorate exposure [F(4,35)=1.65, p>0.18]. C) Dam thyroid gland weight was increased at the two highest dose levels [F(4,37)=18.43, p<0.0001]. * p<0.05 Dunnett’s mean contrast test following significant ANOVA.

Figure 2.

Serum thyroid hormones in dam and fetus. A and B) Serum T4 was reduced in dams on gestational day (GD) 16, n=5–6/dose group, [F(4,24)=32.0, p<0.0001] and GD20, n=7–9/dose group [F(4,38)=71.3, p<0.0001]. C and D) Serum T3 showed a trend to slight reductions on GD16 that failed to reach statistical significance [F(4,24)=2.62, p>0.062], with more robust reductions at the two highest dose groups evident on GD20 [F(4,23)=11.6, p<0.0001]. E) Similar to dam T4, GD20 fetal serum T4 was reduced, but to a greater degree than in dams, at the two highest dose groups n=7–9/dose group, [F(4,38)=241, p<0.0001]. F) Dam serum TSH on GD20 increased 5–10 fold above control levels at the two highest dose groups, n=7–9/dose group, [F(4,37)=50.1, p<0.0001].

Consistent with drops in serum T4, concentrations of thyroid hormone analytes were reduced in the thyroid gland of the dam and fetus on GD20. Dose-dependent reductions in MIT, DIT, T3 and T4 were seen at the 30, 300 and 1000ppm dose levels in the dam thyroid gland (Figure 3A). Analytes were also significantly reduced in the fetal thyroid gland at these same concentrations, with additional reductions evident at the lowest dose of 1ppm. These declines in substrates for thyroid hormone synthesis were most pronounced for DIT in the fetal thyroid gland (Figure 3C). Expressing these changes in hormone analytes as a percent of control clearly shows the larger impact on both magnitude of change and at lower doses of perchlorate on the fetal (Figure 3D) relative to the dam (Figure 3B) thyroid gland.

Figure 3.

Thyroid Gland Hormone Analytes. Mean (± SEM) thyroid hormones and precursors in dam and fetal thyroid glands. A,B) Thyroid hormone analytes were reduced in the thyroid gland of the GD20 dam at the 30ppm dose and above, expressed at ng/gm (A) or as percent of control (B). n=7–9/dose group [F(4,35), all p’s<0.0001]. C,D) Fetal thyroid glands had lower levels of all analytes relative to dams, and significant declines were evident at all doses of perchlorate tested, expressed as ng/gm (C) or as percent of control (D). n=5–9/dose group [F(4,36), all p’s<0.0001]. MIT monoidodotyroine, DIT diiodotyrosine, rT3 reverse T3.

Declines in serum hormone appear to follow deficits in component synthesis products in the thyroid gland. Expressing serum T4 concentrations as a function of thyroid gland T4 reveals that decreases >65% are required in T4 in the dam thyroid gland before any significant change is evident in the serum of the dam or fetus (Figure 4A). In contrast, fetal thyroid gland T4 reductions in excess of 90% are necessary before a decline in fetal serum T4 is realized (Figure 4B), possibly reflecting differences in secretion rates and homeostatic controls in maternal vs immature thyroid gland and the supplementation from the dam hormone supply (Fisher et al. 2014). However, with the greater degrees of thyroid gland hormone depletion achieved at the higher dose levels of perchlorate, fetal serum T4 is reduced more than dam serum T4 (Figure 4B). Although the relationship between maternal and fetal serum T4 appears linear, it requires extrapolation between 20 and 70% declines, indicating perhaps a less than optimal dose selection in our study (Figure 4C).

Figure 4.

Depiction of quantitative relationship of maternal thyroid gland T4 and serum T4 in the dam and fetus (A), fetal thyroid gland T4 and fetal serum T4 (B), and maternal serum T4 vs fetal serum T4 concentrations (C).

Gene expression within the thyroid gland for transcripts involved in iodine transport and thyroid hormone synthesis were evaluated, both qualitative and quantitative differences were apparent when comparing the dam and fetal thyroid glands. The relative expressions of Nis and TshR were altered in the thyroid glands of both dam and fetus. A more than 7-fold increase in relative Nis expression was evident in the dam at 30 and 300ppm with somewhat lower fold increase at the 1000ppm dose level (Figure 5A). A substantial increase in Nis expression was also induced in the fetal thyroid gland, but was of lower magnitude (4-fold) than the dam and limited to the two highest dose groups (Figure 5B). These differences may reflect lower levels of perchlorate exposure in the fetus or differential responses in serum TSH which we were not able to measure in the fetal serum. Interestingly, the iodine transporter pendrin (Scla26a) was also upregulated in the dam thyroid gland, whereas no significant increases in the expression of this gene transcript were evident in the fetal thyroid gland (Figure 5B). In contrast, downregulation of Tshr occurred at all doses in the fetal thyroid gland but was limited to the higher dose levels in the dam, while relative expression of Tg was reduced and Tpo increased in the fetal thyroid gland with no changes evident in these transcripts in the thyroid gland of the dam. Expression of the two differentiation factors in thyroid gland development, Pax6 or Nkx2.1 was not altered in the fetal thyroid gland suggesting that thyroid gland genesis was not perturbed at these doses of perchlorate. However, Pax8 was downregulated in the maternal thyroid gland in a non-monotonic fashion. Expression of Dio1, a deiodinase that serves to recycle iodine within the thyroid gland was not altered in the maternal or fetal thyroid glands. Fold-change expression from which these percent change from control values were derived is summarized in Supplementary Figure 1.

Figure 5.

Thyroid Gland Gene Expression. Mean (± SEM) percent change in expression of transcripts involved in thyroid hormone synthesis and regulation. A) In the dam thyroid gland, significant decreases in relative expression were observed for Pax8 [F(4,33)=4.61, p<0.046] and TshR [F(4,33)=5.75, p<0.0013], while upregulation was seen for iodine transporters Nis [F(4,35)=6.36, p<0.0007] and Pendrin/ Slc26a4 [F(4,31)=4.50, p<0.0056]. (B) Fetal thyroid glands exhibited significant upregulation in Nis [F(4,36)=13.34, p<0.0001] and Tpo [F(4,36)=6.20, p<0.0007], while reductions in relative expression were seen in Tg [F(3,36)=4.98, p<0.0027] and TshR [F(4,36)=7.81, p<0.0001]. Note difference in Y-axis between dam and fetus. Threshold fold cut-off of ±1.5 was used to control for false discovery rate. * p<0.05 Dunnett’s t-test following significant main effect of ANOVA at p<0.02 to control for the number of comparisons.

Perchlorate reduced thyroid hormones in the fetal brain at the two highest concentrations, with a greater degree of suppression in T4 than in T3 (Figure 6A, B). Interestingly, although not statistically significant, small increases in both brain T3 and T4 were evident at the two lower doses of perchlorate. Altered whole brain thyroid hormone concentrations were accompanied at the two highest concentrations by significant reductions in the relative expression in all four of the TH-responsive genes examined in the fetal forebrain. Relative expression of Bdnf_t, Camk4 and Thrb was reduced at all doses, and Bmp7 at the mid dose levels only (Figure 6C, Supplementary Figure 2).

Figure 6.

Fetal Brain Hormones and Gene Expression. Mean (± SEM) T4 (A) and T3 (B) in GD20 fetal whole brain, n=4–5/dose group. Significant reductions were evident at the two highest dose groups for both T4 [F(4,18)=19.9, p<0.0001] and T3 [F(4,18)=21.97, p<0.0001]. * p<0.05 Dunnett’s t-test. (C) Expression of thyroid-hormone responsive genes was downregulated by perchlorate in GD20 forebrain (n=7–9/dose group). Relative expression of Bdnf_tt [F(4,37)=6.68, p<0.0004], Camk4 [F(4,37)=9.91, p<0.0001] and TrhR [F(4,37)=12.0, p<0.0001] was reduced at all dose levels of perchlorate, while Bmp7 [F(4,37)=3.94, p<0.01] displayed a U-shaped dose response, decreases in expression evident at intermediate dose levels only. * p<0.05 following false discovery rate>1.5 fold change and a significant ANOVA with alpha set to <0.02 (see text).

Figure 7 depicts the quantitative relationships of concentrations of T4 in the thyroid gland, the serum, and the fetal brain. For clarity, variability estimates are omitted, but can be derived from previous figures. Declines in fetal thyroid gland thyroid hormones occurred at the lowest dose of perchlorate tested, slight increases in brain thyroid hormones at low doses that are not evident in serum T4, and precipitous drops in fetal serum and brain thyroid hormones at higher concentrations of perchlorate are summarized in Figure 7A. From Figure 7A, the sensitivity of fetal thyroid gland concentrations of DIT and T4 that occur at lower doses and precede subsequent declines in serum and brain T4 are clearly depicted. Transposition of these data to express fetal serum thyroid hormones and brain thyroid hormones as a function of fetal thyroid gland T4 is presented in Figure 7B, where initial nonsignificant increases, particularly of brain T3 appear at ~60–80% declines in glandular T4. Serum and brain thyroid hormones then decline dramatically as glandular supplies of T4 are reduced by >80%. Reductions in fetal brain thyroid hormones as a function of serum hormone declines reveal a similar pattern for fetal (Figure 7C) and dam (Figure 7D) serum T4.

Figure 7.

Depiction of quantitative relationships of thyroid hormones in serum, thyroid gland and brain. A) Mean percent of control change in thyroid gland synthesis analytes and thyroid hormones in fetal serum and brain as a function of maternal administered perchlorate dose. B) Fetal serum and brain thyroid hormones expressed as a percent of control change in T4 of the fetal thyroid gland. C and D) Fetal brain T4 and T3 expressed as a percent of control of Fetal (C) and Dam (D) serum T4. DIT diiodotyrosine.

Discussion

The dose response characteristics of maternal exposure to ammonium perchlorate on late term dam and fetal thyroid gland, serum and brain thyroid hormones and thyroid hormone-action are described in a rat model. At drinking water concentrations that were without overt toxicity on the health of the dam or her progeny, serum thyroid hormones were dose-dependently reduced, more severely so in the fetus than in the dam. Fetal thyroid gland thyroid hormone precursor analytes were also decreased at concentrations that were not affected in the dam and declines in thyroid gland DIT concentrations were the most sensitive indicator of thyroid disruption in both fetus and dam. Fetal brain thyroid hormones were also negatively affected at the two highest concentrations of perchlorate. Brain thyroid hormone decreases were accompanied by reductions in relative expression of thyroid hormone-responsive gene transcripts in the forebrain at all dose levels. The findings on serum hormone changes are in line with previous reports of perchlorate exposure in rat dams (Gilbert and Sui 2008b; York et al. 2004; York et al. 2005b). Greater relative reductions in the fetal serum were availed in the present study largely due to more sensitive methodology for hormone quantification, an extended dose range, and possibly a more restricted level of dietary iodine (Fisher et al. 2013; Gilbert et al. 2013). Reductions in thyroid hormone analytes reflect interference of iodine uptake through the NIS by perchlorate leading to reductions in thyroid hormone synthesis with consequent reductions in serum thyroid hormones, brain thyroid hormones, and thyroid hormone action in the brain of the developing fetus.

Thyroid Hormone Action in the Thyroid Gland.

NIS is the only protein that actively transports iodine into the thyroid gland. Perchlorate actively substitutes for iodine at the NIS, reducing iodine uptake while accumulating thyroidal perchlorate (Dohán et al. 2007; Ravera et al. 2017). In the thyroid gland of the perchlorate-treated rat dam increases in thyroid gland weight were accompanied by a dramatic upregulation of expression of Nis, reflecting a TSH-induced compensatory response to reductions in the availability of thyroidal iodine (Postiglione et al. 2002; Sun et al. 2021; Wolff 1998). Increases in the relative expression of Nis in the fetal thyroid gland were of lower magnitude and limited to the two highest dose levels. These findings could result from a more limited exposure to the fetus relative to the dam, but the greater relative decreases in fetal vs dam serum hormone concentrations would argue against such an interpretation. Rather, they suggest a more limited hormone reserve and adaptive capacity in the fetal thyroid gland when iodine transport to the thyroid gland is challenged by perchlorate. In contrast to Nis expression, a significant downregulation of Tshr expression in the fetal thyroid gland was observed at the lowest dose tested (1ppm) but was not evident until the 30ppm dose level in the dam. This gene encodes the TSH receptor whose activation is typically seen as a signal to increase synthesis and release of thyroid hormones from the thyroid gland (Postiglione et al. 2002; Wolff 1998). A downregulation in expression of Tshr might be anticipated if TSH levels surge as a consequence of declining serum T4. Tshr upregulation directly activates transcripts critical for hormone synthesis machinery (Tpo and Nis), but only indirectly modulates Tg expression (De Felice et al. 2004; Postiglione et al. 2002). In the fetal thyroid gland, downregulation of Tshr at all dose levels seems paradoxical in the face of increased expression of both Tpo and Nis, although consistent with the reductions in Tg expression. This pattern may represent differences in the temporal dynamics of expression patterns of these transcripts in the developing fetal thyroid gland under challenge by perchlorate, of which we have interrogated only a single timepoint. The dam thyroid gland also exhibited downregulation of Tshr expression, but these reductions were restricted to the higher doses and were not accompanied by any change in Tpo or Tg. Although the significance of a downregulation of thyroidal Tshr may be unclear, these data nonetheless reveal a greater sensitivity in the fetal relative to the maternal thyroid gland to Tshr, Tpo and Tg expression, but a less robust compensatory increase in fetal thyroid gland Nis expression in response to perchlorate exposure.

Some insight into this pattern of thyroidal gene expession observed with perchlorate exposure may come from a NIS knockout mouse model where passive diffusion of iodine occurs through a different iodine transporter, pendrin (Ferrandino et al. 2017). Under high dietary iodine conditions in a mouse lacking NIS, expression of genes encoding Tpo and pendrin (Slc26a4) were upregulated, with a corollary downregulation of Tg expression. The authors postulated that in the absence of NIS, compensatory increases in pendrin provide a means for iodine to enter the follicle, prompting increases in TPO to promote an immediate organification of iodine onto thyroglobulin, while downregulation of Tg expression favors formation of T4 over other tyrosine residues (Ferrandino et al. 2017; Vassart et al. 1985). Paralleling these observations in the NIS KO mouse, perchlorate induced an upregulation of pendrin in the dam, but this was not seen in the fetal thyroid gland. However, as described above, the fetal but not the dam thyroid gland displayed increases in Tpo and reductions in Tg expression, similar to that reported in NIS KO mouse. It is possible that an upregulation of pendrin expression we observed may have permitted some passive diffusion of iodine into the maternal thyroid gland under high dose perchlorate exposure, but this avenue was not available to the fetus. The lack of this passive route of entry of iodine to the fetal thyroid gland coupled with decreased iodine uptake into fetal circulation from perchlorate blockade of placental NIS (see below) may contribute to the greater deficits in fetal thyroid gland hormone synthesis despite the compensatory changes in Tpo and Tg expression that were realized.

Overall, findings in fetal thyroid gland gene expression are consistent with reduction of thyroid hormone analytes observed at lower doses in the fetal relative to the dam thyroid gland. These findings also support previous reports of a greater vulnerability of the fetus over the dam to iodine deficiency (Gilbert et al. 2013; Schröder-van der Elst et al. 2001; Sun et al. 2021) and predicted by physiological based pharmacokinetic (PBPK) models of perchlorate in the rat fetus (Clewell et al. 2003b). They stand in contrast to conclusions of a National Academy of Science report on the Health Effects of Perchlorate (National Research Council 2005) where thyroid status of the fetus was only marginally affected and to a lesser extent than the dam (page 126). Their evaluation was based largely on serum hormone data collected in the Argus contract laboratory and subsequently reported by York and colleagues (York et al. 2005a; York et al. 2004; York et al. 2003; York et al. 2005b; Yu et al. 2002). Although information on assay performance is limited in these reports, the commercial RIA utilized as stated ‘according to manufacturer’s instructions’ would not be able to detect T4 levels below 5.0 ng/ml, the lowest calibrator evaluated. The seminal work from Morreale de Escobar and colleagues over 30 years ago using custom-designed highly sensitive RIAs and the LCMS methods afforded in this and other reports from our laboratory, consistently find serum T4 in euthyroid fetus/newborn pups of ~5 ng/ml (Calvo et al. 1992; Calvo et al. 1990; Gilbert et al. 2013; Gilbert et al. 2021; Hassan et al. 2017). As such, the modest declines in serum T4 reported in the fetus relative to the dam in the Argus laboratory studies may be an artifact of a limited dynamic range afforded for fetal serum by the RIA test methodology employed.

Thyroid Hormone Action in the Placenta.

The placenta also expresses NIS, but to a much lower extent than the thyroid gland (Roti et al. 1983; Sun et al. 2021). Although not addressed in the present study, perchlorate action at the placental NIS may reduce maternal iodine supplies to the fetus while at the same time increasing perchlorate exposure to the developing fetal thyroid gland (Clewell et al. 2003b; Sun et al. 2021; Zhang et al. 2016). Upregulation of Nis and pendrin gene and protein expression are observed in placenta under conditions of iodine deficiency, although pendrin may have a more predominant role in placental iodine transfer to fetus than does Nis (Sun et al. 2021). Future work examining perchlorate concentrations in fetal and maternal tissues including the placenta will elucidate the relative contribution of placental vs glandular sites of action for perchlorate on fetal serum thyroid hormone production.

Thyroid Hormone Levels and Thyroid Hormone Action in the Brain.

Maternal exposure to perchlorate reduced fetal whole brain thyroid hormones and suppressed expression of thyroid hormone-responsive genes in the forebrain. The data expressing the relationships between fetal thyroid gland and fetal serum T4 on fetal brain hormones are instructive. It appears that more 80–85% reduction in fetal thyroid gland or serum T4 is necessary before declines in whole brain measures of T3 and T4 become evident for this diet of iodine and strain of rat. This however is incongruent with the measures of thyroid hormone action where reductions in relative gene expression in fetal forebrain were observed at all dose levels. Furthermore, the effects on gene expression evident at the lowest doses were present when brain concentrations of T3 and T4, paradoxically, appeared to be marginally increased. Although not statistically significant, it is possible that these increases in brain hormone may derive from alterations in brain thyroid hormone transporter expression or deiodinase activity, both of which exhibit differential ontogenies in different brain regions and whose dynamics can be directly regulated by thyroid hormones themselves (Groeneweg et al. 2020; Stepien and Huttner 2019). Whole brain thyroid hormone measures also lack the precision of regional or cellular levels that drive TH-dependent gene expression in the fetal cortex (Morreale de Escobar et al. 2004). Alternatively, changes in gene expression patterns observed here in GD20 fetal forebrain may stem from an earlier more compromised brain thyroid hormone status, prior to the onset of fetal thyroid gland function, when maternal hormone supply is the sole source available to the fetus. Although requiring replication, the current data suggest that measures of TH-action in the form of gene expression in the fetal forebrain appear better able to detect potential thyroid-dependent neurological consequences of perchlorate exposure than either thyroid hormone concentrations in the serum or the whole brain. Nonetheless, that thyroid hormones were altered in the fetal serum and brain, albeit at higher dose levels than readouts of thyroid hormone action, increases confidence in the interpretation of transcriptional changes observed despite the very small number of genes examined in the present study.

A Construct for Building a Quantitative AOP of NIS Inhibition and Fetal Brain Development.

The description of complex biological processes as simplified linear chains of biological events can be helpful in organizing data from multiple information streams, formulating weight of evidence tenets for risk-based decisions, identifying testable hypotheses, and revealing significant gaps in our knowledge. The construction of AOPs is challenging when attempting to detail the dynamic interactions of the developing endocrine superimposed on the complexities of the developing nervous system. Our data on thyroid gland hormone analytes interface with existing perchlorate PBPK models providing a direct quantitative link between radioactive iodine uptake inhibition to altered synthesis in the dam and fetal thyroid gland as described by Clewell et al. (2003b). A schematic summarizing exposure and dosimetry properties of perchlorate in the dam and the fetus interfacing with an AOP for NIS inhibition in the dam and fetus to affect TH-mediated effects in fetal brain is presented in Figure 8. Maternal exposure to perchlorate (Dosimetry Factor, DF 1–3) inhibits NIS in the dam thyroid gland. The first Key Event in an AOP is the Molecular Initiating Event (MIE), in this case a reduction in iodine uptake (MIE_thyroid) which compromises the synthesis and availability of release of thyroid hormones to the serum (KE 2_dam). The relationship between reductions in thyroid hormone synthesis and subsequent declines in serum thyroid hormones can be expressed quantitatively (KER1). Perchlorate-induced declines dam serum T4 (KE 2_dam) limit the availability of transfer of T4 from the maternal serum to the fetus (KER2).

Figure 8.

Conceptual Model of Dosimetry and Adverse Outcome Pathway for NIS inhibition in the Dam and Fetus. The left-hand side of the schematic depicts exposure and dosimetry factors (DF), the right-hand side depicts the AOP, top for dam, bottom for fetus. Maternal exposure to perchlorate inhibits NIS in the dam thyroid gland reducing iodine uptake (MIE_thyroid=KE 1_dam) and consequent synthesis/release of T4 to the serum (KE 2_dam). The relationship between reductions in synthesis reflected in reductions in serum hormone is KER1. With less T4 present in the serum of the dam, less is available for transplacental transfer to the fetus, depicted as KER2. NIS transporters in the placenta (MIE_placenta) also permit the transfer of perchlorate and limit the amount of iodine transfer from dam to fetus (DF4). NIS inhibition at the fetal thyroid gland (MIE_thyroid=KE 1_fetus) results in a similar sequence of events as in the dam to reduce hormone synthesis in the gland for release to the fetal serum (KE 2_fetus and KER3). The convergence of KER2 and KER3 dictate thyroid hormones found in fetal circulation and these relationships are altered by temporal dynamics of hormone regulation over the course of pregnancy and fetal thyroid gland development. Reduced fetal brain hormone concentrations (KE 3_fetus) alter thyroid hormone-dependent signaling in the fetal brain (KE 4_fetus). Direct KERs can be described from serum to brain hormone (KER4) and from brain hormone to hormone action (KER5). In the absence of brain hormone measures, an indirect relationship from gland or serum hormone analytes to hormone-mediated signaling in brain (KER6) can be described. ClO₄ perchlorate; DF Dosimetry Factor; MIE Molecular Initiating Event; KE Key Event; KER Key Event Relationshi;p; NIS Sodium Iodide Symporter; TH thyroid hormones; MIT monodidotyrosine; DIT diiodotyrosine.

The relationships between perchlorate exposure, serum perchlorate, thyroid gland perchlorate concentration, and iodine uptake reflect the PBPK components of the model. Perchlorate in the serum of the dam is also passed to the fetal serum via the placenta (DF4), and from the fetal serum to the fetal thyroid gland (DF6).

NIS transporters in the placenta (MIE_placenta) not only permit the transfer of perchlorate to the fetal serum and thyroid gland, but also limit the amount of iodine available for transport from the dam to the fetal serum for subsequent uptake by NIS in the fetal thyroid gland. NIS inhibition at the fetal thyroid gland (MIE_thyroid) results in a similar sequence of KEs as in the dam to reduce serum hormones in the fetus (KE2_fetus and KER3). The convergence of KER2 and KER3 dictate thyroid hormones found in fetal circulation, and these relationships are altered by the temporal dynamics of thyroid hormone regulation over the course of pregnancy and fetal thyroid gland development. Here we provide data for a single timepoint in late pregnancy when the fetal thyroid gland is fully functional. Fetal brain concentrations of thyroid hormones (KE3_fetus) are reduced when serum concentrations decline with negative impact on TH-mediated gene transcription in the fetal brain (KE 4_fetus). Direct KERs can be described from serum to brain thyroid hormones (KER4) and from brain thyroid hormones to thyroid hormone action (KER5). In the absence of brain thyroid hormone measures, an indirect relationship from serum hormone concentrations to thyroid hormone action in the brain (KER6) can also be described, but it comes with greater uncertainty. It appears from the present findings that fetal thyroid analytes (T4 and DIT) may serve as more sensitive predictors of potential adversity (i.e., altered TH-action in brain) than either brain or serum T4 (indirect KER6).

In conclusion, these data provide the first quantitative information for the action of the NIS inhibitor perchlorate spanning maternal exposure to fetal serum, thyroid gland, brain hormones and thyroid hormone action in the fetal brain of the rat. They complement previous perchlorate PBPK models based on radioactive iodine uptake (RAIU) inhibition, a measure reflecting the ability of the thyroid gland to sequester iodide. RAIU inhibition is used as a surrogate for the NIS-mediated iodide transport required for thyroid hormone synthesis (Clewell et al. 2003a; Clewell et al. 2003b; Merrill et al. 2004). Our findings expand these models by directly measuring thyroid gland hormone synthesis products and extend observations beyond pharmacokinetics to downstream responses in hormone concentrations in thyroid gland, serum, and fetal brain. We present a conceptual model encompassing perchlorate dosimetry in the placenta, thyroid gland and serum coupled to a NIS-mediated AOP with outcomes in the fetal brain. Studies extending exposures into the postnatal period are underway to evaluate hormone changes and neurodevelopmental outcomes in exposed offspring. Integrating toxicokinetic information with an AOP for NIS inhibition will aid the alignment of exposure and AOP-based effect models, facilitate the extrapolation of in vivo rodent data to humans, and permit a more accurate translation of in vitro derived estimates of NIS inhibition to biological action in vivo (Hallinger et al. 2017; Handa et al. 2021; Hassan et al. 2020; Hassan et al. 2017; Wang et al. 2019).