The Year in Basic Thyroidology

Thyroid hormone signaling in human retinal organoids

In humans, two thyroid hormone receptor genes (THRA and THRB) encode three major thyroid hormone-binding thyroid hormone receptor (TR) isoforms (TRβ1, TRβ2, and TRα1). The major tissue-specific functions of TRβ are the regulation of the pituitary-thyroid axis, liver metabolism, and visual and auditory sensory activities. The major tissue-specific functions of TRα1 are the regulation of thermogenesis, heart activity, development of the intestine and bone, and neurogenesis and brain functions.

To understand the developmental regulation of cone photoreceptors in human retina, Eldred and colleagues generated human retinal organoids from stem cells (1). Three subtypes of human photoreceptor cones are defined by the visual pigments they express: blue-opsin (short wavelength, S), green-opsin (medium wavelength, M), and red-opsin (long wavelength, L). During the development of human fetal retina, S cones are generated first; L/M cones are specified later. The mechanism that controls the switch from S cones to L/M cones was unknown.

In their study, Eldred and colleagues studied the mechanism that controls the decision between the fate of S and L/M cones in human retinal organoids (1). They first validated that these human organoids have similar distributions, gene expression profiles, and morphologies of cone subtypes as in human retina. Moreover, the temporal switch of S cones to L/M cones in the human retinal organoid was similar to that of the human retina. In organoids in which the THRB function was completely ablated (i.e., both TRβ1 and TRβ2) by CRIPSR/Cas9 editing, only blue S cones were detected. After treatment of wild-type organoids with the thyroid hormone (T3), only green L/M clones were found. In contrast, treatment of THRB-function-deficient human organoids with T3 could not switch blue S cones to the green L/M cones. These results indicated that thyroid hormone signaling was necessary and sufficient for temporal switch between S and L/M fate specifications.

Next, they asked how T3 concentrations in the retina were regulated to control temporal switch between S and L/M fate specifications and found that the T3 concentrations were regulated by the temporal expression of DIO2 and DIO3 in the retina. In the early stage of human retina development, DIO3 was highly expressed to inactive thyroid hormones. The low thyroid hormone in the retina dictated the expression of S cones. In the later development, DIO2 was highly expressed to convert T4 to T3 to repress S clones and to induce L/M cones via TRβ. The importance of this work lies in its establishing stem cell-derived organoids as a model for determining the mechanism of human development with promising utility for therapy and vision repair.

Novel thyroid hormone actions on thermogenesis and heart regeneration

Hirose and colleagues found evidence for thyroid hormonal control of heart regenerative capacity during endothermy acquisition (2). The investigators asked why the adult mammalian heart loses regenerative capacity and discovered thyroid hormones as a surprising molecular link connecting the loss of mammalian cardiac regenerative potential to the acquisition of postnatal endothermy. They found that loss of heart regenerative capacity in adult mammals is triggered by increasing thyroid hormones: postnatal rise (>50-fold) of thyroid hormones coincided with cardiomyocyte cell-cycle exit and loss of regenerative capacity. However, in the adult heart with TRα1 deficiency, cardiomyocyte regenerative capacity was preserved. Further, high total serum T4 and high metabolism or body temperature was associated with low diploid cardiomyocytes and more loss in heart regenerative potential in the mammalian adult heart. In contrast, in the heart of neonatal mice and zebrafish, high content of diploid cardiomyocytes and low body temperature were associated with high regenerative potential. Experimental manipulation of serum thyroid hormone levels in mice and zebrafish led to an increase and a decrease of diploid cardiomyocyte content, respectively.

These findings were further supported by a cardiomyocyte ploidy analysis across 41 vertebrate species by this group (2). The investigators uncovered that certain monotreme, edentate, cetacean, and chiropteran species have unusually high percentages of diploid cardiomyocytes in the adult heart. Further, diploid cardiomyocyte abundance conforms to Kleiber's law—the 3/4-power law scaling of metabolism with bodyweight—and inversely correlates with standard metabolic rate, body temperature, and serum T4 level.

In view of others' findings that cardiomyocyte polyploidization is a major barrier for heart regeneration (3,4), these results suggest that the evolutionary increase of polyploid cardiomyocyte frequency and decline of cardiac regenerative potential may occur in parallel with the ectotherm-to-endotherm transition. As the major driver of energy metabolism and thermogenesis, thyroid hormones seem to be the key factors promoting both heart regenerative potential loss and the acquisition of endothermy.

Why is increased metabolism and thermogenesis positively selected for during mammalian evolution? Elevated metabolic rate and the evolution of endothermy likely offered early mammals survival advantages as they inhabited a temporally wide nocturnal niche and expanded into regions of colder climates previously unexplored by reptiles. The loss of proliferation-competent cells and the decline of regenerative capacity in the heart and other organs were seemingly not under selection pressure in animal evolution. Thus, these findings suggest that the loss of heart regenerative capacity in adult mammals is triggered by increasing thyroid hormones and may be a tradeoff for the acquisition of endothermy.

Thyroid hormone action on myelin repair

It is known that T3 action produces oligodendrocytes, the cells in the central nervous system (CNS) that make myelin. Currently, however, there is no approved myelin repair therapy for multiple sclerosis or other demyelinating diseases. Hartley and associates employed two enabling recent advances to address the question of whether thyroid hormone action in the CNS can have a pro-remyelination effect in a demyelinating in vivo model (5). The first advance is the thyromimetic prodrug Sob-AM2 that delivers substantially more of the thyromimetic sobetirome to the CNS from a systemic dose than an equivalent dose of the parent drug sobetirome. The second advance is a newly developed genetic mouse model of demyelination (iCKO-Myrf mice). This iCKO-Myrf mutant mouse features an inducible knockout of the myelin regulatory factor gene (Myrf), essential for mature oligodendrocyte survival. In this model, Myrf ablation can be induced by short-term treatment with tamoxifen, and Myrf knockout occurs only in mature oligodendrocytes, resulting in CNS-wide demyelination. Induced demyelination in iCKO-Myrf mice is accompanied by a robust and quantifiable motor disability that improves in response to remyelination.

Hartley et al. tested sobetirome and Sob-AM2 for therapeutic benefit in motor disability, and they found accelerated CNS remyelination via histology, magnetization transfer imaging, and electron microscopy to evaluate brain and spinal cord myelin status in iCKO-Myrf mice. They found that the CNS-penetrating thyromimetics provide significant improvement to motor disability and accelerate remyelination in both brain and spinal cord. Hartley et al. also showed that thyromimetic treatment generates a larger population of oligodendrocytes in demyelinated brain regions, thus providing support for the expected mechanism that thyromimetics promote remyelination by stimulated oligodendrogenesis. This study provides preclinical rationale and support for developing CNS-penetrating thyromimetics for the treatment of multiple sclerosis and other demyelinating disease.

Brain hypothyroidism and deiodinase

In the rat brain, most T3 is generated locally via deiodination of T4, a reaction catalyzed by the type-two deiodinase (D2) (6). The important role that D2 played in generating local T3 in the brain became evident in studies of mice with astrocyte-specific D2 inactivation (7). Despite having normal serum T3 levels, these animals exhibited anxiety-depression-like behavior and a hippocampal transcriptome typical of reduced thyroid hormone signaling (7). No D2 mutations have been reported in humans, but a single nucleotide polymorphism (SNP) from Thr92Ala-D2 is present in 12–36% of the population (8). This SNP does not seem to be associated with cognition or quality of life in euthyroid individuals (9). However, hypothyroid patients who are carriers of the Ala92-D2 SNP exhibited improved quality of life when T3 was added to T4 therapy (10). Even though some of these findings have not been reproduced in another study (9), at face value they suggest that Ala92-D2 carriers on T4 therapy may not produce sufficient amounts of T3 via D2.

Thus, to gain insight into these mechanisms, Jo and colleagues first studied cell lines stably expressing Thr92-D2 or Ala92-D2 tagged with fluorescence probes and found that Ala92-D2 causes endoplasmic reticulum (ER) stress and accumulates in the Golgi apparatus (11). In this setting, Ala92-D2 was found to produce ∼20% less T3 in the presence of different T4 concentrations. To further their understanding of Ala92-D2 actions, these investigators used CRISPR genome editing and created mouse strains carrying the Thr92-D2 or the Ala92-D2 gene. The Ala92-D2 mice were systemically euthyroid, growing and reproducing normally. However, they were sensitive to an anxiogenic environment and exhibited higher exploratory activity and more risk-assessment behavior. However, once settled in their home-cage, Ala92-D2 mice refrained from physical activity and exhibited sleepiness and impaired memory. Similarly, the brain of the Ala92-D2 mice exhibited generalized signs of ER stress and diminished thyroid hormone signaling in distinct areas, that is, the striatum, amygdala, and frontal cortex. Enhancing T3 signaling in the brain with the administration of T3 improved cognition. Further, restoring proteostasis with the chemical chaperone 4-phenyl butyric acid treatment eliminated the Ala92-D2 phenotype. In contrast, primary hypothyroidism intensified the Ala92-D2 phenotype, with only partial response to T4 therapy.

Based on these findings, the authors concluded that disruption of cellular proteostasis and reduced Ala92-D2 activity may contribute to the failure of T4 therapy in carriers of Ala92-D2 SNP. Although presently there is no evidence that carriers of the Ala92-D2 polymorphism exhibit brain hypothyroidism, the study of the cerebral cortex of Ala92-D2 carriers revealed a transcriptional fingerprint, suggestive of neurodegenerative diseases, with activation of the ubiquitin pathway, mitochondrial dysfunction, inflammation, apoptosis, DNA repair, and growth factor signaling (12). Moreover, in African Americans, Ala92-D2 was associated with molecular markers known to underlie Alzheimer's disease, and with increased odds of developing Alzheimer's disease/dementia (13). The authors conclude that these findings need to be reproduced by other groups before Ala92-D2 SNP could be used to guide treatment of hypothyroid patients.

Defective organic anion transporting polypeptide 1C1 transporter in a patient with impaired brain functions

T4 transport across the blood–brain barrier is predominately mediated by monocarboxylate transport 8 (MCT8), while T4 uptake by astrocytes is facilitated by organic anion transporting polypeptide 1C1 (OATP1C1). MCT8 patients with Allan-Herndon-Dudley syndrome have been described; however, until now, patients with OATP1C1 defects have not been reported. While OATP1C1 knockout mice develop normally without any overt neurological abnormality, the knockout mice have nervous system-specific hypothyroidism.

Stromme and associates reported the first case—a 15.5-year-old girl who had a mutation in OATP1C1 that was associated with severe brain hypometabolism and juvenile neurodegeneration (14). The girl was developed normally in her first year of life, but she gradually exhibited dementia with spasticity and intolerance to cold. Brain imaging showed gray and white matter degeneration and severe glucose hypometabolism. Whole-exome sequencing of the 15.5-year-old girl identified a homozygous missense mutation in OATP1C1 in a highly conserved site with an aspartic acid 252 to asparagine (D252N) amino acid change. Functional characterization showed that T4 uptake capacity was reduced in the mutant as well as in the cell surface expression of mutated OATP1C1.

Based on these findings, the investigators treated the patient with Triac co-administered with a low dose of T4. The patient's condition improved after 6 weeks, with the patient resuming eye contact and becoming more alert, being able to swallow better, and having improvements in motor function and her cold intolerance. The identification of the first patient supported the link between defective OATP1C1 functions and neurodegeneration. Future studies in the identification of additional patients with OATP1C1 mutations will help advance an understanding of the mechanisms underlying the disease-related impaired T4 transport to the brain.

Thyrocyte death and autoimmunity

Autoimmune hyperthyroidism, commonly referred to as Graves' disease (GD), is caused by stimulating autoantibodies to the thyrotropin (TSH) receptor (TSHR). TSHR is a heterotrimeric guanine nucleotide-binding protein (G protein)-coupled receptor (GPCR). Previous studies showed that in contrast to stimulating TSHR autoantibodies, antibodies that recognize the linear epitopes in the cleavage region of the TSHR ectodomain (C-TSHR-Abs) induce apoptosis of thyrocytes.

Morshed and colleagues investigated how and why C-TSHR-Abs are capable of initiating a cascade of signaling events that lead to programmed thyrocyte cell death (15). Extensive analyses using live-cell imaging of intracellular vesical trafficking, endocytosis, proteomic arrays, and fluorescence quantification led to the findings that S-TSHR-mAb activates major G proteins and behaves similarly to TSH (15). However, C-TSHR-mAb failed to activate G proteins and did not undergo normal endosomal and lysosomal vesicular trafficking. Moreover, the accumulated C-TSHR-mAbs induced cellular stress, caused TSHR-biased danger signaling, generated ROS, and activated apoptosis, leading to exacerbation of thyroid autoimmunity. These findings suggest that G proteins in thyrocytes are required not only for GPCR signaling but also for vesicular trafficking and sorting, a process that is independent of cAMP, PKA, and CREB. Further, C-TSHR-Abs may orchestrate overt intrathyroidal, intraorbital, or epidermal inflammatory autoimmune reactions in GD.

Non-pumping pro-tumorigenic actions of the sodium iodide symporter

The sodium iodide symporter (NIS) is classically localized within the plasma membrane of thyrocytes and thyroid cancer cells, where it mediates iodide uptake and forms the basis of radioiodine therapy for thyroid disease (16 –18). Radioiodide uptake is generally lower in thyroid cancer cells than normal thyroid tissue (19). Intriguingly, intracellular non-membranous NIS expression has been reported to be elevated in thyroid and other nonthyroidal cancers (20). This observation led investigators in the Charis Eng laboratory to ask whether NIS can have a pump-independent function. Because thyroid cancer is one of the major components of Cowden's syndrome (21), a subset of which is caused by germline mutations in PTEN, the investigators studied the noncanonical tumorigenic role of NIS in thyroid cancer cells in relation to PTEN signaling.

Feng and colleagues reported that cytoplasmic NIS protein interacts with LARG (leukemia-associated RhoA guanine exchange factor) to activate RhoA, thereby enhancing cell migration and invasion in thyroid cancer cells (22). These data supported their working hypothesis that NIS has a pro-tumorigenic function besides its classical iodide pump function in thyroid cancer cells. Importantly, the team found that PTEN siRNA-mediated knockdown results in increased NIS protein levels, especially cytoplasmic NIS, in several thyroid cell lines with exogenous NIS expression, independent of its mRNA expression. Further, they elucidated that PTEN loss activates AKT-mTOR signaling to inhibit NIS glycosylation, subsequently impairing NIS membrane localization. These signaling perturbations resulted in mis-localization of NIS to the cytoplasm, endowing it with the ability to interact with the cytoplasmic protein LARG, which, in turn, increases intracellular NIS stability through these interactions. In addition, their data also indicate that PTEN signaling impacts the glycosylating enzyme DPAGT1 to then regulate NIS glycosylation. Importantly, they found that PI3K/AKT/mTOR inhibitors not only increase plasma membrane NIS through enhancing NIS glycosylation but also decrease intracellular NIS, indicating these pharmacologic agents' therapeutic role in thyroid cancer, from the standpoint of clearance of the pro-tumorigenic non-pump-function of NIS.

The uncovering of a non-pump function of NIS has added a venue for further exploring the biology of NIS and its actions in thyroid carcinogenesis (23), providing one possible mechanism to explain the more aggressive behavior in cancers with lower radioiodide uptake. Some studies have investigated the potential application of NIS gene therapy in the treatment of iodide refractory thyroid cancers and some nonthyroidal cancers (24). As such, therapeutic strategies should go beyond just increasing NIS protein levels: Effective therapies should promote NIS localization back to the plasma membrane.

Clone originality of anaplastic thyroid cancer

Anaplastic (undifferentiated) thyroid cancer (ATC) is commonly found to co-exist with differentiated thyroid cancers (DTC) such as papillary and follicular thyroid cancer, suggesting that ATC could develop from DTC. However, whether ATC evolves from DTC has been a long-standing issue of debate. Dong and colleagues performed whole-exome sequencing of five ATC tumors with co-existing DTC and matched normal tissue to determine whether there is a shared common ancestor (25). Phylogenic analysis revealed that in every ATC-DTC pair, a significant number of somatic mutations and somatic copy-number variations were shared, including multiple known cancer drivers. Clonal analysis further demonstrated that each ATC-DTC pair shared a common ancestor, with some pairs diverging early in their evolution and others in which the ATC seems to arise directly from a sub-clone of DTC. Although the precise lineal relationship requires further clarification, based on the genetic relationship, this study suggests a shared origin of ATC and DTC. These findings also raised the consideration as to whether a more aggressive treatment of DTC would be justified to prevent DTC from transitioning into deadly ATC.

Targeting c-MYC in ATC

c-MYC is a critical oncogene, frequently over-expressed in human ATC. c-MYC is one of the most sought-after drug targets for cancer therapy. However, targeting c-MYC has presented a huge challenge as c-MYC lacks binding pockets on its surface owing to its disordered protein structure. A strategy that has been used to target c-MYC is the use of small-molecule inhibitors (e.g., JQ1) to disrupt the interaction of the bromodomain and extraterminal domain (BET) family of proteins with acetylated histones on the chromatin (26).

Zhu and colleagues investigated the effect of targeting c-MYC transcription program via suppression of its super-enhancers on chromatin by using an inhibitor of the activity of BET proteins, for example, BRD4, PLX51107 (PLX). They further tested the efficacy of combined treatment on human ATC cell-derived xenograft tumors with an MEK inhibitor, PD0325901 (PD) (27). They found that the BET inhibitor PLX or the MEK inhibitor PD as a single treatment effectively suppressed the growth of xenograft tumors. However, the combination treatment was far more effective. Mechanistically, PLX and PD converge to potently inhibit the binding of BRD4 to the c-MYC promoter to suppress c-MYC transcription, and PD further suppresses MEK signaling to inhibit tumor growth. Recent studies have uncovered mutations of genes involved in chromatin modifiers and transcription regulators in ATC (28). Thus, targeting transcription programs that are critical for cancer progression as shown in the present studies is an effective strategy for novel treatment approaches for ATC.