Actions of thyroid hormones and thyromimetics on the liver

Abstract

Thyroid hormones (triiodothyronine and thyroxine) are pivotal for metabolic balance in the liver and entire body. Dysregulation of the hypothalamus–pituitary–thyroid axis can contribute to hepatic metabolic disturbances, affecting lipid metabolism, glucose regulation and protein synthesis. In addition, reductions in circulating and intrahepatic thyroid hormone concentrations increase the risk of metabolic dysfunction-associated steatotic liver disease by inducing lipotoxicity, inflammation and fibrosis. Amelioration of hepatic metabolic disease by thyroid hormones in preclinical and clinical studies has spurred the development of thyromimetics that target THRB (the predominant thyroid hormone receptor isoform in the liver) and/or the liver itself to provide more selective activation of hepatic thyroid hormone-regulated metabolic pathways while reducing thyrotoxic side effects in tissues that predominantly express THRA such as the heart and bone. Resmetirom, a liver and THRB-selective thyromimetic, recently became the first FDA-approved drug for metabolic dysfunction-associated steatohepatitis (MASH). Thus, a better understanding of the metabolic actions of thyroid hormones and thyromimetics in the liver is timely and clinically relevant. Here, we describe the roles of thyroid hormones in normal liver function and pathogenesis of MASH, as well as some potential clinical issues that might arise when treating patients with MASH with thyroid hormone supplementation or thyromimetics.

Introduction

Thyroid hormones (triiodothyronine and thyroxine) are synthesized from iodide and tyrosine within the thyroid gland located in the mid-portion of the anterior neck. The production and release of thyroid hormones into the circulation are regulated by a complex feedback mechanism involving the hypothalamus, pituitary gland and thyroid gland (hypothalamus–pituitary–thyroid (HPT) axis)¹. This process initiates with the hypothalamus releasing thyrotropin-releasing hormone (TRH), which signals the pituitary gland to release the thyroid-stimulating hormone (TSH). Subsequently, TSH stimulates the thyroid gland to synthesize and release triiodothyronine and thyroxine into the circulation. The circulating levels of triiodothyronine and thyroxine exert negative feedback on the hypothalamus and pituitary gland, helping them maintain a delicate balance of circulating thyroid hormone levels. Triiodothyronine, the more potent form of thyroid hormone, binds with a higher affinity to thyroid hormone receptors (THRs) than thyroxine. On the other hand, thyroxine is the predominant type of thyroid hormone in the circulation and is mostly bound to proteins such as albumin, thyroxine-binding globulin and transthyretin^2,3. These proteins serve as buffers and reservoirs to maintain the serum concentration of free thyroxine (the circulating non-protein-bound form of thyroxine) within an optimal and narrow range.

Circulating free thyroxine and triiodothyronine are imported into hepatic cells through thyroid hormone transporters such as mono-carboxylate transporter 8 (MCT8) and organic anion transporting polypeptide 1B1 (OATP1B1)⁴. Once inside, intracellular thyroxine is converted to the more active triiodothyronine by deiodinase 1 (DIO1) and DIO2, whereas DIO3 converts thyroxine and triiodothyronine to reverse triiodothyronine and diiodothyronine, respectively. The expression and activity of these enzymes in the liver finely regulate the intracellular concentration of the potent ligand, triiodothyronine⁵. Additionally, DIO3-mediated synthesis of reverse triiodothyronine possibly reflects an adaptative response to prevent a hypermetabolic state due to triiodothyronine excess under a state of injury or starvation⁶.

Triiodothyronine can bind to two major nuclear THR isoforms, THRA and THRB, which are encoded on separate genes. THRs belong to the nuclear hormone receptor family and also bind to thyroid hormone response elements (TREs) located in the promoters of a target gene; therefore, they can be considered to be ligand-regulatable transcription factors. As such, in response to thyroid hormones, they regulate the transcription of hundreds of genes in the liver in positive and negative directions in almost equal numbers^7–10. During human fetal development, the thyroid gland remains non-functional until 10 weeks of gestation while the placenta concurrently has high DIO3 expression, so this early gestational period might represent a ‘silent’ period when epigenetic repression of thyroid hormone-responsive target genes by non-liganded THRs have a predominant role^11,12. In rats, THRA1 is the predominant hepatic THR expressed during fetal development, and THRB1 is expressed in the liver during the late fetal and neonatal periods¹³. An alternative splicing product of the Thra gene that does not bind thyroid hormones, Thra2, is abundant in the early fetal stages of rat liver development¹⁴. The Thrb gene expresses two isoforms (THRB1 and THRB2) due to alternative translation start sites. THRB1 and THRB2 share the same ligand binding domain and, hence, have identical binding affinity. THRB2 has restricted expression in the pituitary, hypothalamus, retina and cochlea, and influences their development in mice^14,15.

During the relative absence of thyroid hormones, such as during early gestation, THRs are thought to complex with corepressor proteins such as nuclear corepressor 1 (NCOR1) and silencing mediator for retinoid and THRs on TREs of positively-regulated target genes to maintain a state of transcriptional repression^16,17. This repression is facilitated by the recruitment of proteins such as histone deacetylase 3 that removes acetyl groups from acetylated histone proteins, generating a condensed chromatin structure that limits access of general transcription factors and RNA polymerase II (RNA Pol II) to the underlying DNA^18–20. In contrast, when triiodothyronine binds to THRs, conformational changes occur in THRs that lead to the release of corepressors and the binding of coactivators to THRs^21,22. Coactivators, such as the steroid receptor coactivators, can recruit proteins with histone acetyltransferase activity to form protein complexes that enable them to bridge with coactivator complexes on other promoters or with RNA Pol II complex located at the transcriptional start site^18–20. Histone acetyltransferases play a crucial part in the subsequent transcriptional activation process by adding acetyl groups to histone proteins. This acetylation causes a more relaxed chromatin structure, facilitating access of the general transcriptional machinery to the transcriptional start site to induce gene expression^23,24. Another coactivator, THR-associated protein (TRAP) 220 kDa component (also known as MED1), forms a TRAP complex that recruits proteins to help bridge it with the general transcription machinery²⁵. The formation of these coactivator complexes by the DNA-bound THRs facilitates RNA Pol II-mediated transcription and enables the integration of multiple signalling pathways within the cell. For example, TRAP 220 is a substrate for protein kinase A, protein kinase C and extracellular-signal-regulated kinase (ERK)²⁶. The details of transcriptional regulation by triiodothyronine are described in detail in other reviews^18–20. Besides the classic genomic action of triiodothyronine, several non-genomic actions of thyroxine and triiodothyronine have been described²⁷. Notably, thyroxine interacts with αvβ3 integrin at the cell surface to induce the serine–threonine kinase (MAPK–ERK) pathway, which further enhances nuclear THR-mediated transcriptional activity and induces several cytosolic signalling cascades in cell culture models^28,29.

Thyroid hormones exert substantial effects on virtually all tissues in the body, including the liver. The diverse hepatic actions of thyroid hormones have crucial roles in regulating metabolism and energy within the liver and the entire body^30,31. Consequently, thyroid hormones help maintain metabolic homoeostasis and facilitate adaptations to changes in the body’s energy requirements³¹. Thus, dysfunctions in the HPT axis contribute to metabolic disorders and liver-related diseases^30,31. Thyroid hormones have important roles in sexual development, puberty and fertility³². They directly affect the ovaries, testes and corpora cavernosa to regulate sexual development and metabolism in those tissues³². Additionally, thyroid hormones are known to increase the synthesis and secretion of sex hormone-binding globulin (SHBG)³³, which regulates the transport and biological availability of sex steroid hormones, primarily testosterone and oestradiol, which are responsible for exerting sex-specific metabolic effects in the body³⁴. Thus, the actions of thyroid hormones on the liver help maintain and coordinate sex hormone action and homeostasis throughout the body in tandem with its direct effects on the reproductive organs.

Both hypothyroidism and hyperthyroidism have multiple effects on the transcription of genes involved in lipid, glucose and protein metabolism, as well as protein synthesis and detoxification in the liver³¹. When combined with the effects of thyroid hormones on other tissues, disturbances in these hepatic functions have profound, systematic effects on whole-body metabolism. For example, conditions such as obesity and dyslipidaemia frequently occur in hypothyroidism, whereas increased body temperature, decreased serum cholesterol and weight loss are commonly found in hyperthyroidism. Currently, strong epidemiological evidence supporting the association between thyroid hormone dysfunction and metabolic dysfunction-associated steatotic liver disease (MASLD) (formerly known as non-alcoholic fatty liver disease³⁵), that ranges from benign steatosis, to hepatic inflammation and fibrosis inflammation in metabolic dysfunction-associated steatohepatitis (MASH) (formerly known as non-alcoholic steatohepatitis³⁵), has generated increasing scientific and clinical interest in the metabolic actions of thyroid hormones in the liver. The next sections provide a detailed overview of the roles of thyroid hormones in major hepatic functions and pathogenesis of hypercholesterolaemia and MASLD, and the employment of thyroid hormones and thyromimetics to treat metabolic diseases.

Thyroid hormone actions in the liver

Thyroid hormone developmental actions on the liver

During mouse and human development, there is a transient surge in circulating thyroid hormone levels shortly after birth coupled with increased hepatic DIO2 activity to raise intrahepatic cellular triiodothyronine concentration and induce the expression of genes involved in lipid metabolism³⁶. However, circulating thyroid hormone concentrations and hepatic DIO2 gene expression decrease rapidly after birth, and the latter is not detectable in the adult liver. This parallel increase in circulating thyroxine level and intrahepatic DIO2 expression suggests that circulating thyroid hormone entering the liver has an important role in the regulation of the neonatal hepatic transcriptome. In this connection, the perinatal thyroxine surge transiently lowered histone 3 lysine 9 trimethylation (a post-translational modification) in mice and removed approximately 1,500 DNA methylation sites on the promoter regions of target genes, including peroxisome proliferator-activated receptor-γ (Pparg) and zinc-finger protein 125 (ref. 36).

The surge in circulating thyroxine at birth is associated with increased expression of hepatic genes involved in lipogenesis. Notably, loss of perinatal Dio2 expression in mice with hepatocyte-specific deletion of the Dio2 gene resulted in the repression of several thyroid hormone-regulated lipid metabolism genes in the fetal liver. Furthermore, this loss of perinatal DIO2 activity in hepatocytes persisted into adulthood and attenuated diet-induced and alcohol-induced hepatic steatosis in challenged mice^36,37 (Fig. 1). However, these findings stand in contrast to liver steatosis observed in the global Dio2-knockout mice³⁸, which probably have increased sympathetic activity and lipolysis compensating for the global inactivation of Dio2 and a decrease in Dio2-mediated thermogenesis in brown adipose tissue³⁸. These findings suggest that during evolution, the liver might have undergone epigenetic adaptations to cope with nutrient deprivation via triiodothyronine-mediated activation of pro-lipogenic genes such as Pparg during a narrow developmental window of perinatal Dio2 expression and activity³⁹. This transient expression of Dio2 helps fine-tune epigenetic regulation of lipid metabolism that persists even during adulthood in mice, and most likely occurs in humans³⁶. Although this adaptation is useful for storing energy and handling free fatty acid (FFA) flux in adulthood, it could have deleterious effects by causing hepatic steatosis and obesity under nutrient-rich conditions. Interestingly, this surge in DIO2 due to triiodothyronine also occurs in hepatic organoids generated from human-induced pluripotent stem cells and induces the expression of hepatocyte-specific transcription factors and cell markers such as hepatocyte nuclear factor 4α, albumin and apolipoprotein⁴⁰.

Fig. 1. A critical developmental window of Dio2 expression governs susceptibility to diet- and alcohol-induced hepatic steatosis in mice.

A perinatal surge in the expression of Dio2 at birth increases intrahepatic triiodothyronine (T₃) signalling, affecting the methylation status (differentially methylated region (DMR)) of several hepatic genes such as ATP-binding cassette subfamily G member 1 (Abcg1) and elongation of very long-chain fatty acids protein 7 (Elov7) involved in lipid and cholesterol metabolism secretion, which results in hepatic steatosis when an adult mouse is exposed to a high-fat diet or alcohol. By contrast, in mice, hepatocyte-specific deletion of Dio2 gene at embryonic stages results in hypermethylation of promoters of lipogenic genes, such as Abcg1 and Elov7, during development, which is sustained until adulthood, resulting in protection from diet or alcohol-induced hepatic steatosis. DIO2, deiodinase 2; TAG, triglyceride; WT, wild-type.

Thyroid hormone actions on adult hepatic metabolism

The liver is a major hub for lipid, carbohydrate and protein metabolism in the body. Thyroid hormones regulate both anabolic and catabolic metabolic pathways involved in lipid, carbohydrate and protein metabolism in the liver. The effects of thyroid hormone on hepatic metabolism are explained in the following sections.

Thyroid hormones and hepatic triglycerides

Triglycerides (TAGs) serve as the primary storage form for fatty acids in hepatocytes and play a crucial part in their transportation within the circulatory system⁴¹. In the fed or anabolic state, hepatic uptake of circulating FFAs from the diet, glycolysis and de novo synthesis of fatty acids induces TAG accumulation in the liver⁴¹. However, in the fasting state, hepatic TAGs undergo lipolysis to release FFAs that are oxidized within peroxisomes and mitochondria to generate ATP and ketones⁴¹. Additionally, intrahepatic TAGs can be exported out of the liver via TAG-rich VLDLs⁴¹. Although thyroid hormones influence hepatic lipid metabolism through both direct and central pathways^42–44, this Review focuses specifically on the direct effects of thyroid hormones on hepatic functions.

Thyroid hormones influence all aspects of hepatic lipid metabolism, including increases in expression of fatty acid translocase (also known as CD36) and fatty acid binding protein to promote the influx of plasma FFAs into the liver^42,45(Fig. 2). Thyroid hormones also stimulate hepatic de novo lipogenesis by increasing the transcription of lipogenic enzymes such as fatty acid synthase, acetyl-CoA carboxylase-α, malic enzyme and thyroid hormone responsive protein (THRSP, also known as SPOT14), which are among the primary THRB-regulated genes involved in this process^42,45(Fig. 2). Additionally, thyroid hormones indirectly enhance the expression of pro-lipogenic genes by stimulating other nuclear receptors such as sterol regulatory element binding protein 1, liver X receptors, carbohydrate-responsive element-binding protein and PPARG, all of which have important roles in regulating target genes involved in hepatic lipogenesis¹⁹. However, it is notable that although thyroid hormones stimulate the expression of de novo lipogenesis genes to increase fatty acid synthesis, they negatively regulate the expression of proteins that are required for the latter steps of TAG elongation and assembly, such as stearoyl-CoA desaturase and glycerol-3-phosphate acyltransferase⁴² (Fig. 2).

Fig. 2. Thyroid hormone regulation of hepatic triglycerides and cholesterol metabolism.

Thyroid hormones regulate the expression of several lipid metabolic genes in the liver. Thyroid hormones increase intrahepatic fatty acid uptake from dietary sources and stimulate the expression of lipogenic genes in the liver. However, thyroid hormones increase autophagy-mediated triglyceride (TAG) hydrolysis, thereby decreasing the TAG content in hepatocytes and their secretion as VLDL into the blood. Furthermore, thyroid hormones increase the oxidation of fatty acids derived from autophagy and lipophagy to produce acetyl-CoA and ketones. Thyroid hormones also stimulate the expression of hepatic cholesterol biosynthesis genes. In addition, thyroid hormones concurrently reduce circulating cholesterol by suppressing the assembly and secretion of cholesterol esters, such as VLDL, increasing LDL and HDL cholesterol uptake and excreting them as bile via reverse cholesterol transport. Upward arrows denote thyroid hormone-induced proteins and pathways, and downward arrows denote thyroid hormone-suppressed proteins and pathways in the liver. FFA, free fatty acid; LDLR, LDL receptor; RCT, reverse cholesterol transport; SRB1, scavenger receptor class B member 1; TH, thyroid hormone.

In addition to their anabolic actions on lipid metabolism, thyroid hormones exert a more substantial effect on lipid catabolism. In this regard, thyroid hormones stimulate autophagy-mediated lipolysis of TAGs stored in fat droplets, known as ‘lipophagy’ in hepatocytes⁴⁶ (Fig. 2). This selective autophagy enables TAG stored within the hepatocytes to undergo lipolysis by lysosomal acid lipase and release FFAs into the cytosol⁴⁶. Notably, the initiation of lipophagy by thyroid hormones is intricately linked to mitochondrial fatty acid β-oxidation, as the loss of autophagy impairs ketone production by thyroid hormones⁴⁶. Furthermore, this process depends upon THRs and their interaction with NCOR1 (ref. 46). At the transcriptional level, thyroid hormones induce angiopoietin-like protein 8, which can target TAG stores and prime them for lipophagy⁴⁷. Thyroid hormones also induce mitochondrial fatty acid oxidation by directly stimulating the expression of key mitochondrial proteins such as carnitine O-palmitoyltransferase 1α, medium-chain specific acyl-CoA dehydrogenase, pyruvate dehydrogenase kinase isoform 4 (PDK4), and mitochondrial uncoupling protein 2 (refs. 42,45) (Fig. 2).

Thyroid hormones stimulate mitochondrial turnover to maintain efficient oxidative phosphorylation by increasing mitochondrial biogenesis and their removal by autophagy, referred to as ‘mitophagy’^48,49. Mitochondria biogenesis is partly controlled by thyroid hormone-mediated increases in the expression of proliferator-activated receptor-γ coactivator 1α (PGC1A), oestrogen-related receptor-α (ESRRA), peroxisome proliferator-activated receptor-α and mitochondrial transcription factor A^19,48,50,51. Additionally, thyroid hormones activate PGC1A and Forkhead box protein O1 (FOXO1) by stimulating NAD-dependent protein deacetylase sirtuin 1 (SIRT1) deacetylation activity to modify these transcription factors⁵². Furthermore, thyroid hormones increase mitochondrial fission and mitophagy, which are essential for removing damaged mitochondria from oxidative stress generated by the increased mitochondrial β-oxidation^53,54. These processes involve THRB-dependent increase in the levels of hepatic ESRRA, which, in turn, augment unc-51-like kinase 1 (ULK1) gene expression⁵¹ and AMP-activated protein kinase (AMPK) signalling to stimulate ULK1-mediated mitochondrial translocation and mitophagy⁵³. This thyroid hormone-induced phenomenon, termed ‘mito-hormesis’, ensures adequate pruning of damaged mitochondria and, together with thyroid hormone-mediated replenishment of new mitochondria, helps maintain mitochondrial turnover and quality control despite increased reactive oxygen species generation from elevated oxidative metabolism (Fig. 2).

Thyroid hormones and hepatic cholesterol metabolism

Thyroid hormones regulate serum cholesterol levels by their effects on the hepatic cholesterol biogenesis pathway, production of HDL, efflux of hepatic cholesterol in bile via the reverse cholesterol transport (RCT), and hepatic uptake of circulating cholesterol via LDL receptor (LDLR)^42,45,55 (Fig. 2). In the mevalonate pathway responsible for intrahepatic cholesterol synthesis, a key enzyme, 3-hydroxy-3-methylglutaryl-coenzyme A reductase, is transcriptionally stimulated by Thrb in rats^42,45. Thyroid hormones also indirectly regulate genes in the mevalonate pathway by inducing the expression of secondary transcription factors such as upstream stimulatory factor 2, sterol regulatory element-binding protein 2 (SREBP2)⁵⁶ and nuclear factor Y^42,45. Interestingly, hyperthyroidism reduces serum cholesterol levels in both rats and humans^57,58. This reduction in serum cholesterol by triiodothyronine seems to be multifactorial, and in mice and humans involves increased uptake of circulating LDL cholesterol (LDL-C) and HDL cholesterol due to augmented LDLR and scavenger receptor class B type 1 expression^55,59,60, decreased sterol O-acyltransferase 2 gene expression via miR181d⁶¹, reduced apolipoprotein B100 expression (required for re-esterification and packaging cholesterol into VLDL and LDL)^61–63, and inhibition of LDLR degradation due to the suppression of proprotein convertase subtilisin/kexin type 9 protein levels⁵⁷. Furthermore, thyroid hormones stimulate cholesterol excretion by RCT, which involves the induction of several hepatic genes such as cholesterol 7α-hydroxylase (CYP7A1)⁶⁴, ATP-binding cassette transporters subfamily G member 5 and member 8 (refs. 42,45) to increase biliary cholesterol excretion in mice and humans. Thus, a complex interplay exists among thyroid hormones, SREBP2, LDLR and CYP7A1, resulting in reduced serum cholesterol levels in patients with hyperthyroidism or euthyroidism who are treated with thyroid hormones or liver-selective thyromimetics.

Thyroid hormones and hepatic glucose metabolism

Thyroid hormones and THRB-selective ligands induce insulin resistance in mouse liver which can cause hyperglycaemia⁶⁵. Thyroid hormones increase hepatic glucose output by stimulating glycogenolysis and gluconeogenesis in rats and mice, and human hepatic cells^19,66–68. Thyroid hormones increase the expression of hepatic phosphorylase kinase, glycogen phosphorylase and lysosomal α-glucosidase and inhibit glycogen synthase genes to stimulate glycogenolysis⁶⁶ (Fig. 3). They also increase the transcription of gluconeogenic genes such as phosphoenolpyruvate carboxykinase⁶⁹, glucose-6-phosphatase⁷⁰, and PDK4 and SIRT1-mediated activation of FOXO1 to increase gluconeogenesis^71,72, and induce glucose transporter 2 expression to augment hepatic glucose output⁷³ (Fig. 3). Notably, the increased hepatic glucose output by thyroid hormones is counterbalanced by increased glucose uptake in the skeletal muscle to maintain serum glucose levels within the normal range. Although hyperthyroidism has been linked to glucose intolerance in humans, a population-based study in patients (n = 200) with euthyroidism showed that decreased levels of free thyroxine are also associated with elevated markers of insulin resistance (perhaps due to reduced glucose utilization by skeletal muscle), suggesting that optimal regulation of glucose homeostasis occurs within the normal range of thyroid function⁷⁴. Additionally, thyroid hormones inhibit intestinal farnesyl X receptor signalling to enhance glucagon-like peptide 1 secretion and improve glucose homeostasis in a hypothyroid mouse model and patients with hypothyroidism⁷⁵. Thus, the net effect of thyroid hormones on circulating glucose levels is governed by its opposing effects on the liver versus the gut and peripheral tissues.

Fig. 3. Thyroid hormone regulation of hepatic glucose metabolism.

Triiodothyronine (T₃) increases hepatic glucose output by upregulation of gluconeogenesis and glycogenolysis processes. Triiodothyronine regulation of gluconeogenesis is mediated in several steps, which can involve direct regulation of both human and mouse gluconeogenic genes, such as Pepck and G6pc, by THRB and PGC1A transcriptional complexes. In addition, thyroid hormones were shown to activate SIRT1 deacetylase activity, which, by inhibiting AKT signalling and directly deacetylating Foxo1, increases FOXO1 nuclear localization and its binding to the promoter of gluconeogenic genes. Furthermore, thyroid hormones were shown to inhibit insulin signalling, thereby increasing gluconeogenesis. The effect of T₃ on glycogenolysis is via induction of phosphorylase kinase, glycogen phosphorylase and lysosomal α-glucosidase. At the same time, thyroid hormones suppress the expression of glycogen synthase, resulting in a net increase in glycogenolysis.

Thyroid hormones and hepatic protein metabolism

Thyroid hormones stimulate new protein synthesis in the rat liver^76–78. However, thyroid hormones also increase hepatic protein catabolism and uric acid synthesis in both humans and rats by inducing the gene expression of period circadian protein homologue 2 (Period 2) and enzymes involved in nucleotide metabolism^79,80.

THRs, deiodinases and thyroid hormone transporters in hepatic metabolism

There are five major cell types in the liver. These include hepatocytes (which are of parenchymal origin) and other cell types of non-parenchymal origin, such as hepatic stellate cells (HSCs), Kupffer cells (liver macrophages), liver sinusoidal endothelial cells (LSECs) and cholangiocytes. These cells cooperatively and coordinately form the hepatic sinusoid architecture and regulate its function. The predominant THR isoform expressed in both mouse and human hepatocytes is THRB. Mouse models expressing Thra and Thrb mutants show distinctive effects on lipid and glucose metabolism in hepatocytes and non-parenchymal liver cells. For example, a mouse model resembling the resistance to thyroid hormone-β (RTHβ) phenotype in humans, carrying a dominant negative mutation in Thrb in both alleles (Thrb^PV/PV), had increased hepatosteatosis due to decreased autophagy and increased expression of lipogenic genes in hepatocytes^46,81. Consistent with these results, patients with RTHβ due to THRB mutations also develop hepatic steatosis and insulin resistance^82,83. In contrast, mice carrying an analogous Thra1^PV/PV mutation and Thra-null mice both had decreased lipid accumulation in hepatocytes accompanied by increased hepatic insulin sensitivity^81,84. These findings suggest that there might be THR isoform-specific effects on lipid metabolism in hepatocytes.

HSCs undergo differentiation into fibroblast-like cells after liver injury. THRA is the major isoform expression in mouse HSCs and LSECs, and inhibits liver fibrogenesis⁸⁵. Intriguingly, THRA and THRB are expressed in Kupffer cells, and THRA exerts immune-modulatory plasticity in hepatic macrophages⁸⁶. Additionally, both THR isoforms are expressed in rat cholangiocytes, where they exert an antiproliferative effect on biliary growth⁸⁷. Thus, these findings suggest that there might be THR isoform-specific effects in different liver cell types, which would have important implications when designing THR isoform-selective agonist therapies for liver diseases. Currently, little is known about the activity of THRB isoform-selective thyromimetics in these other hepatic cell types, and whether they can decrease inflammation and fibrosis by acting on them. The diversity in THR isoform expression among different hepatic cell types suggests that thyroid hormones might potentially offer the most beneficial effects on MASH as they would target both THR isoforms, provided their side effects on susceptible extrahepatic tissues can be mitigated.

DIO enzymes, responsible for the local activation of thyroid hormones, are also differentially expressed in various hepatic cells⁸⁸. Adult hepatocytes typically express high Dio1 and low Dio3 gene expression levels in mice³⁶. Interestingly, Dio2 gene is transiently expressed in neonatal hepatocytes only at birth in hepatoblasts and gradually diminishes thereafter to reach undetectable levels in adult hepatocytes³⁶. However, during hepatosteatosis, mouse Dio1 expression and activity increase as a compensatory change to maintain intracellular triiodothyronine concentration⁸⁹. Concurrently, mouse Dio1 gene expression and intracellular triiodothyronine both decrease subsequently in MASH as the disease progresses, resulting in ‘intrahepatic hypothyroidism’⁸⁹. Unlike adult hepatocytes that primarily express Dio1 and Dio3, hepatic macrophages in mice express all three DIOs with increased Dio2 expression correlating with macrophage activation⁸⁶. HSCs express both Dio1 and Dio3, with increased expression of DIO3 associated with increased fibrogenesis in mice and humans⁹⁰.

MCT8 was the first thyroid hormone transporter identified in liver⁹¹. In patients with mutations in MCT8 and mouse models with Mct8 deficiency, the liver seems to be in a hyperthyroid state with several serum biomarkers of hepatic triiodothyronine action, such as SHBG, and ferritin levels being elevated⁹². This paradoxical finding might be due to increased compensation by other thyroid hormone transmembrane transporters in the liver, including L-type amino acid transporters, OATP1B1 and MCT10, and the sodium taurocholate cotransporting polypeptide, which is highly expressed in the liver⁴. Alternatively, it might be due to MCT8 acting as a thyroid hormone bidirectional transporter in hepatocytes⁹³, with its loss resulting in hepatocellular thyroid hormone accumulation and a local ‘hyperthyroid’ phenotype. The expression of specific thyroid hormone transporters can vary among the different hepatic cell types^94,95 and could be a potential target for modulating the metabolism and cellular function of specific types of hepatic cells. Notably, thyroid hormone transporters, including MCT8 and OATP1B1, are enriched in hepatocytes^95,96, whereas MCT10 is expressed predominantly in Kupffer cells⁸⁶.

Thyroid hormones and MASLD

Thyroid hormones have an essential role in the metabolic status of humans. Patients with overt hypothyroidism and subclinical hypothyroidism have an increased prevalence of obesity, hypertriglyceridaemia and hypercholesterolaemia¹⁹. In a study in individuals with obesity or overweight and primary thyroid dysfunction in Tehran, a higher prevalence of hypothyroidism was observed in individuals with obesity (11.6%; 4.0% overt and 7.6% subclinical) than in individuals with a normal BMI (8.2%; 1.1% overt and 7.1% subclinical)⁹⁷. Furthermore, hypothyroidism is associated with decreased insulin sensitivity in humans¹⁹. On the other hand, hyperthyroidism is associated with weight loss and reduced LDL-C and TAG levels in the serum together with decreased glucose tolerance in humans¹⁹.

MASLD is a liver condition associated with obesity, type 2 diabetes and dyslipidaemia⁹⁸. MASLD encompasses a spectrum of liver pathologies ranging from benign hepatosteatosis to MASH, characterized by hepatocyte apoptosis, inflammation and fibrosis, which, in turn, can lead to cirrhosis and hepatocellular carcinoma (HCC)⁹⁸. Both MASLD and MASH are frequently accompanied by extrahepatic manifestations such as atherosclerosis, heart failure with preserved ejection fraction, sarcopenia and chronic kidney disease⁹⁹. MASLD currently affects more than 30% of the adult population worldwide, and is an independent risk factor for type 2 diabetes, cardiovascular disease and chronic kidney disease⁹⁸.

Over the past 10 years, several large-scale epidemiological studies have demonstrated an increased prevalence of MASLD in both overt and subclinical hypothyroidism across different regions in the world, suggesting a marked association between MASLD and hypothyroidism^100–109. In a prospective case–control study, there was an increased risk for the development of ultrasound-diagnosed MASLD in individuals with hypothyroidism (n = 2,324), independent of other metabolic risk factors (adjusted HR 2.21, 95% CI 1.42–3.44), suggesting a causal relationship¹¹⁰. Furthermore, subclinical hypothyroidism (defined by TSH >4.5 mIU/l) and low–normal thyroid function (TSH levels 2.5–4.5 mIU/l) were identified as independent predictors of MASH (OR 1.61, 95% CI 1.04–2.50) and advanced fibrosis (OR 2.23, 95% CI 1.18–4.23) in humans (n = 525)¹⁰⁹. More importantly, subclinical hypothyroidism and low–normal thyroid function were associated with an increased risk of both all-cause and cardiovascular mortality in adults with MASLD (n = 10,144)¹¹¹. Collectively, these studies suggest that hypothyroidism and subclinical hypothyroidism not only increase the risk of developing MASH but also contribute to mortality from extrahepatic manifestations of MASH, particularly cardiovascular diseases.

Thyroid hormones and treatment of metabolic disease: pitfalls and opportunities

Treatment of overt hypothyroidism in patients (n = 52) with levothyroxine (thyroxine) replacement significantly (P < 0.05) improved lipid metabolism abnormalities in patients with hypothyroidism¹¹². Similarly, a study in patients with newly diagnosed hypothyroidism (n = 60) showed a decrease in serum total cholesterol and LDL-C levels after levothyroxine treatment¹¹³. In both rat and mouse models of MASLD and MASH, thyroid hormone treatment substantially decreased hepatic steatosis, increased mitochondrial fat oxidation and autophagy, and reduced inflammation and fibrosis^114,115. In a study in patients with type 2 diabetes with hepatosteatosis, low-dose levothyroxine therapy reduced hepatic fat content and increased insulin sensitivity⁴². In addition to the main thyroid hormones (triiodothyronine and thyroxine), a thyroid hormone metabolite, 3,5-diiodothyronine, also showed significant (P < 0.05) efficacy in improving hepatosteatosis¹¹⁶ and hepatic insulin resistance in mouse models of MASLD^117,118. However, major challenges facing thyroid hormone therapy for hepatic metabolic disease are the adverse off-target effects on tissues such as the heart, bone and skeletal muscle due to collateral activation of THRA, the predominant THR isoform in these tissues in humans¹¹⁹. Potential adverse effects include tachycardia, cardiac hypertrophy, reduced ejection fraction and increased cartilage and bone loss¹¹⁹. Additionally, suppression of the HPT axis is also a major concern⁷⁷.

To address these issues, other approaches to selectively target thyroid hormones to the liver have been developed. One strategy involves the conjugation of thyroid hormones to liver-targeting molecules. The first in this class is a glucagon-triiodothyronine conjugate that increases liver specificity and reduces extrahepatic adverse effects¹²⁰. Notably, glucagon–triiodothyronine conjugated therapy improves dyslipidaemia, atherosclerosis, body weight and hepatosteatosis in several dietary mouse models of obesity with no significant toxicities on the heart, bone and HPT axis¹²⁰. Similarly, liver-specific nanoparticle-mediated thyromimetic delivery is another upcoming strategy to provide selective beneficial effects on the liver while minimizing extrahepatic adverse effects¹²¹.

Thyromimetic actions in the liver

Thyromimetics for the treatment of hypercholesterolaemia and MASLD

To harness the beneficial effects of thyroid hormones on the liver (which are mediated mainly by THRB) while minimizing the adverse off-target effects from collateral activation of THRA in the heart, bone and skeletal muscle, several THRB-specific thyromimetics have been developed that specifically target THRB (Box 1). Two notable examples, GC-1 (sobetirome) and KB2115 (eprotirome), were designed to have affinity for THRB similar to that of triiodothyronine but decreased affinity for THRA. Both drugs were initially developed to treat hypercholesterolaemia and have undergone several clinical investigations. In a phase II study involving 184 patients with maximal statin therapy, treatment with eprotirome led to a mean LDL-C reduction of 32% at the highest dose (P < 0.001) and significantly reduced serum apolipoprotein B (P < 0.001), TAGs (P < 0.001) and lipoprotein (a) (P < 0.001)¹²². Importantly, few or no adverse effects were observed in tissues that predominantly expressed THRA (heart and bone), and there was no effect on serum TSH levels. The latter observation indicated that there were no adverse effects on the HPT axis, particularly as THRB is highly expressed in the pituitary and hypothalamus. Importantly, a phase III trial assessing the efficacy of eprotirome in patients with familial hypercholesterolaemia (236 patients randomized; 69 patients completed 6 weeks) was discontinued due to the development of bone cartilage damage in dogs¹²³. Furthermore, an analysis after 6 weeks of treatment showed that eprotirome significantly increased serum levels of aspartate aminotransferase (AST) (P < 0.0001), alanine transaminase (ALT) (P < 0.0001), conjugated bilirubin (P = 0.0006) and γ-glutamyl transferase (P < 0.0001), suggesting that there were adverse effects on the liver at the doses used in the study¹⁰⁹. Sobetirome also showed beneficial effects on LDL-C in a phase I trial; however, this drug was not pursued further for clinical development¹²⁴.

Box 1. On thyromimetics.

Thyromimetics are compounds which activate the nuclear thyroid hormone receptor. These thyromimetics have a chemical structure similar to that of the active hormone triiodothyronine with alterations specifically targeting the thyroid hormone receptor α or β. Some thyromimetics have also been modified to target the liver specifically. These alterations to the chemical structure of triiodothyronine lead to a reduced potency to activate the thyroid hormone receptor¹⁶⁵. This also explains why, for a similar chemical structure, the dose of thyromimetics is in milligrams, whereas thyroid hormone is supplemented in micrograms.

The current rapid increase in the prevalence of MASLD has renewed research interest in THRB-selective thyromimetics, especially given the previous lack of any FDA-approved treatment options. Preclinical studies have demonstrated that thyroid hormones stimulate lipophagy and β-oxidation of fatty acids, and reduce steatosis, inflammation and fibrosis in MASH^115,125. Preclinical studies have also shown the beneficial effects of THRB-specific thyromimetics, sobetirome, eprotirome, MB07811 (VK-2809), TG68 and HSK31679, on liver TAG levels^{65,114,126–129}. Building on the theoretical rationale for sobetirome and eprotirome, several new drugs have been developed for treating metabolic liver disease that target the THRB isoform selectively in the liver while minimizing off-target effects on THRA-expressing tissue (bone and heart) as well as extrahepatic tissues which express THRB (pancreas and pituitary)¹³⁰ (Table 1). Currently, VK-2809 has completed a 52-week phase IIb trial (NCT04173065), and resmetirom (MGL-3196) has completed a phase III trial in 966 patients¹³¹ and has been approved by the FDA for the treatment of MASH. Other thyromimetics (TERN-501, HSK31679 and ALG-055009) are also being investigated in clinical studies (Table 1).

Table 1. Completed and ongoing clinical trials of thyromimetics in MASLD and MASH.

Compound	Indication	Primary outcome	Timeline	Adverse effects
VK2809 (VOYAGE, phase IIb) NCT04173065 (ref. 160)	Biopsy-confirmed MASH (n = 229)	12-week reduction in liver fat content	Completed; 52-week biopsy data awaited	NA
Resmetirom (MGL3196) (MAESTRO-NAFLD-1 phase III)	MASLD (LSM >5.5 kPa and CAP >280 dB) (n = 972)	Week 52 TEAEs (not significant)	Published141; open-label extension (NAFLD-OLE)	Main TEAEs diarrhoea, nausea; TSH= and FT3=, FT4↓ (100 and 80 mg resmetirom per day)
Resmetirom (MGL3196) (MAESTRO-NASH phase III)	MASH (fibrosis stage F1B–F2–F3) (n = 966)	Week 52 MASH resolution (P < 0.001) fibrosis improvement (P < 0.001)	Published131; 54-month ongoing clinical outcomes study	No increased heart rate, no changes in BMD TSH= (100 mg resmetirom per day) ↓ (80 mg resmetirom per day) FT3=, FT4↓ (100 and 80 mg resmetirom per day)
Resmetirom (MGL3196) (MAESTRO-NASH OUTCOMES phase III) NCT05500222 (ref. 161)	MASH cirrhosis (well compensated) (n = 700)	Progression to decompensated cirrhosis	Ongoing	NA
TERN-501 (DUET phase IIa + FXR agonist) NCT05415722 (ref. 162)	Presumed MASH (biopsy and/or imaging) (n = 162)	12 weeks Reduction in liver fat content	Completed	NA
HSK31679 (phase II) NCT06168383 (ref. 163)	Biopsy-proven MASH (F2–F3) (n = 180)	52 weeks MASH resolution	Ongoing	NA
ALG-055009 (HERALD phase IIa) NCT06342947 (ref. 164)	MASH non-cirrhotic (n = 100)	12 weeks Reduction in liver fat content	Ongoing	NA

BMD, bone mineral density; CAP, controlled attenuation parameter (with FibroScan); FT3, free triiodothyronine; FT4, free thyroxine; FXR, farnesoid X receptor; LSM, liver stiffness measurement (with FibroScan); MASH, metabolic dysfunction-associated steatohepatitis; MASLD, metabolic dysfunction-associated liver disease; NA, not applicable; NAFLD, non-alcoholic steatohepatitis; TEAE, treatment-emergent adverse events; TSH, thyroid-stimulating hormone.

The selective THRB agonist, VK2809, previously known as MB07811, is a prodrug that undergoes conversion to VK2809A by cytochrome P3A4, an enzyme that is highly expressed in the liver¹²⁷. VK2809A, previously known as MB07344, has some similarities with sobetirome in its selective activation of the THRB isoform. In preclinical models of MASLD, VK2809 reduced steatosis by increasing β-oxidation of fatty acids, as evidenced by increased oxygen consumption rate and the presence of short and intermediate-length acyl-carnitine species¹²⁶. In a mouse model of glycogen storage disease type Ia (GSD Ia) in which glucose-6-phosphatase deletion increased glucose-6-phosphate levels to cause intrahepatic glycogen and TAG accumulation and severe steatohepatitis, VK2809 reduced intrahepatic lipid content and inflammation by stimulating autophagy, mitochondrial biogenesis and β-oxidation of fatty acids¹³². A phase IIb VOYAGE trial in 229 patients with biopsy-confirmed MASH, with the primary outcome of a reduction in intrahepatic lipid content after 12 weeks of VK2809 treatment, is completed (NCT04173065). However, the detailed findings of this study have not yet been published in a peer-reviewed journal, and the 52-week biopsy data are still pending (Table 1).

Resmetirom received FDA approval for the treatment of MASH after the completion of the phase III MAESTRO-NASH trial¹³¹. Resmetirom has selectivity for both the liver and THRB; however, the precise mechanism of its liver selectivity is not well understood. A study published in 2022 showed that the thyroid hormone transporter OATP1B1, which is highly expressed in the liver, enhances tissue-selective uptake of resmetirom⁹⁶. In cell culture studies, resmetirom also highly activated THRB (90% of maximal triiodothyronine potency) but only partially activated THRA (25% of maximal triiodothyronine potency)⁹⁶. Additionally, resmetirom induces the transcription of the major triiodothyronine-regulated gene, DIO1, and increases basal and maximal respiration in a similar manner to triiodothyronine⁹⁶. In the published phase III trial, 966 patients with fibrosis of stage F1B, F2 or F3 (as measured by FibroScan with controlled attenuation parameter >280 dB and liver stiffness measurement >8.5 kPa or biopsy-confirmed MASH) were treated with placebo or 80 mg or 100 mg resmetirom per day (1:1:1 ratio)¹³¹. The primary outcomes, MASH resolution with no worsening of fibrosis (placebo 9.7%, 80 mg resmetirom 25.9% and 100 mg resmetirom 29.9%) and fibrosis improvement (14.2%, 24.2% and 25.9%, respectively), were both achieved at 52 weeks (as shown in liver biopsy samples) in the study patients (P < 0.001) (Table 1). A decrease in intrahepatic lipid content, as found at 12 weeks in a phase II trial including 116 patients measured by MRI proton density fat fraction¹³³, was also substantial in the phase III trial after 52 weeks¹³¹. Additionally, in the phase II trial, resmetirom significantly reduced serum LDL-C (P < 0.001), lipoprotein (a), apolipoprotein B, apolipoprotein CIII and TAG levels¹³³. Notably, resmetirom caused similar decreases in LDL-C, lipoprotein (a), apolipoprotein B and TAG concentrations in 116 patients with familial hypercholesterolaemia in a 12-week phase II study¹³⁴.

In clinical studies, THRB-selective thyromimetics have shown effects on steatosis, inflammation and fibrosis. Thyroid hormones increase autophagy and fatty acid β-oxidation to reduce lipotoxicity¹¹⁵. These effects are mediated by THRB activation in the liver and presumably are employed by thyroid hormones and THRB-selective thyromimetics. However, the specific mechanism(s) underlying the reduction in inflammation and fibrosis in MASH by thyromimetics remains unclear. Similar to triiodothyronine, resmetirom stimulates oxygen consumption, suggesting that it reduces the intrahepatic levels of TAGs, fatty acids and lipotoxic metabolites generated from the latter to decrease cell injury and apoptosis⁹⁶. Thyroid hormones also increased AMPK signalling, stimulated autophagy and β-oxidation of fatty acids, and reduced inflammasome numbers in a dietary mouse model of MASH and resmetirom, which might have similar actions in the liver^20,31,88. These actions are mediated by THRβ as hepatic AMPK activation and autophagy are abrogated in mice lacking this THR isoform⁵¹.

Interestingly, resmetirom also reduced both MASH-induced phosphorylation of the signal transducer and activator of transcription 3 and expression of nuclear factor-κB in mice¹³⁵. However, it is not known whether resmetirom induces these effects solely via nuclear THRs or whether cytosolic THRB is involved. Notably, thyroid hormones also inhibit TGFB–SMAD-mediated signalling by decreasing direct THR interaction with SMAD on TGFB–SMAD target genes in mice¹³⁶. It has also been proposed that in mice, the induction of the expression of a regulator of G protein signalling 5 by resmetirom mediates its anti-MASH effects¹³⁵. In a mouse model of MASH, thyroid hormones reduced endoplasmic reticulum stress and hepatic inflammasome numbers by stimulating autophagy, which might be an additional mechanism¹¹⁵. Taken together, these findings point towards resmetirom affecting MASH by pathways related to lipid metabolism and inflammatory signalling. Thus, the issue of whether thyromimetics mediate their effect primarily through the reduction in hepatic lipid overload and associated lipotoxicity or have direct anti-inflammatory and anti-fibrotic effects in MASH remains unresolved. In support of a role for the latter, it is noteworthy that in mice, thyroid hormones and sobetirome can blunt bleomycin-induced lung fibrosis via intact PGC1A and PTEN-induced kinase 1 pathways¹³⁷. In a carbon tetrachloride mouse model of liver fibrosis, thyroid hormones also mitigated liver fibrosis by inhibiting pro-fibrotic SMAD signalling by TGFβ¹³⁶.

Given the multiple mechanisms for reducing inflammation and fibrosis by thyroid hormones, it is possible that each thyromimetic might have its own unique effect(s) on inflammation and fibrosis that is influenced by its selective uptake into the liver and its relative THRA or THRB specificity. The expression of THRA in macrophages and HSCs suggests that thyroid hormones have potential roles in modulating their plasticity during MASLD and MASH progression. In this connection, triiodothyronine induced anti-fibrotic action in mice by activating THRA-regulated pathways in HSCs, the major cell type involved in hepatic fibrosis⁸⁵. Interestingly, THRA-knockout mice exhibit more severe fibrosis than wild-type mice when fed a diet deficient in methionine and choline⁸⁵. As THRB-selective thyromimetics might exert weaker effects on THRA due to lower binding affinity for THRA, it remains uncertain whether they can induce anti-fibrotic action in HSCs during MASH. These observations highlight the complexity of the relationship between thyromimetics and the intricate pathways involved in inflammation and fibrosis, and emphasize the need for further research to elucidate the specific mechanisms and effects of different thyromimetics in the context of MASH.

Monitoring of thyromimetic efficacy by SHBG with thyromimetics

SHBG is specifically produced in the liver and released into the circulation, where it binds to sex hormones. As thyroid hormones increase the production of SHBG¹³⁸, serum SHBG levels have been suggested to serve as a biomarker of thyroid hormone action in the liver. SHBG levels have previously been used clinically to distinguish between thyrotropinoma (pituitary tumour that secretes TSH) and RTHβ as the levels rise in the former and are unchanged or decreased in the latter. In a phase II trial, which included 116 patients (55% women), resmetirom induced a twofold increase in serum SHBG levels in both women and men¹³³. Interestingly, the primary outcomes of reduction in intrahepatic lipid content and improvement in liver enzymes and histology were most pronounced in the patients with the highest serum SHBG levels. In a phase III trial of resmetirom in patients with MASH and fibrosis, patients with SHBG levels of >120% compared with baseline had a significantly (P < 0.001) greater likelihood of MASH resolution¹³¹. Therefore, serum SHBG levels might serve as markers to guide dosing, measure compliance, and predict outcomes in patients treated with thyromimetics. However, it is currently not known whether the induction of higher levels of serum SHBG affects the measurement of circulating oestrogen and testosterone levels and whether this has clinical implications. Additionally, serum cholesterol and ferritin levels have been proposed as markers of thyromimetic activity in the liver, but they might be affected by multiple systemic factors and lack the specificity of SHBG¹³⁹.

Effects of thyromimetics on the HPT axis and endogenous thyroid hormone production

As THRB is highly expressed in the hypothalamus and pituitary, it is possible that central activation of THRB by THRB-selective thyromimetics suppresses the HPT axis and inhibits TSH synthesis and release into the circulation to potentially cause central hypothyroidism, as shown for sobetirome in preclinical models¹⁴⁰. So far, eprotirome and resmetirom are the only two thyromimetics whose effects on the HPT axis have been described in peer-reviewed publications as treated patients reportedly did not exhibit any significant changes in serum TSH and triiodothyronine levels^123,133. Notably, in a phase III trial of resmetirom, a significant reduction in TSH (P < 0.001) (least squares mean difference −0.18 mU/l) was observed in patients treated with 80 mg resmetirom per day, but no significant reduction was seen in those treated with 100 mg resmetirom¹³¹ (Table 1). These findings most likely indicate that these drugs do not cause major suppression of the HPT axis at the therapeutic doses used in the studies. However, caution is warranted since eprotirome is the first drug to show significant dose-dependent decreases in free thyroxine concentrations (P < 0.0001)¹²³. In patients treated with 50 μg eprotirome per day (n = 24), there was a decrease of 2.8 pmol/l (0.2 ng/dl) in free thyroxine concentration from baseline, and in patients treated with 100 μg eprotirome (n = 22), there was an even greater decrease in free thyroxine concentration of 4.1 pmol/l (0.3 ng/dl) from baseline. Notably, with resmetirom treatment, there was an 18% decrease in free thyroxine levels from baseline in patients treated with 100 mg resmetirom per day in the phase III study¹³¹ (Table 1). In a phase III safety study of resmetirom¹⁴¹, the dose was reduced by 20 mg whenever there was a decrease in free thyroxine of >30% to <9 pmol/l (0.7 ng/dl). This decrease in free thyroxine levels occurred in 2.4% of patients treated with 100 mg resmetirom per day (n = 323) and in 0.6% of those treated with 80 mg resmetirom per day (n = 322).

The cause(s) of this decrease in free thyroxine levels by thyromimetics is not known. It is possible that they have an effect on the HPT axis that is not reflected by the serum TSH levels. However, TSH is generally considered a more sensitive marker of HPT axis status than free thyroxine¹⁴². Another possibility is that thyromimetics, similar to triiodothyronine, might induce hepatic DIO1 gene expression and activity to increase the conversion of circulating free thyroxine to free triiodothyronine, resulting in decreased serum levels of free thyroxine. This mechanism might explain the observed decrease in serum reverse triiodothyronine, the preferred substrate for DIO1, that was observed in patients treated with resmetirom¹³¹. Given that thyromimetics might potentially have mild adverse effects on the HPT axis, it would be prudent to use them judiciously with careful monitoring of serum free thyroxine, free triiodothyronine and TSH levels, and perhaps even reverse serum triiodothyronine levels, in patients with actual or potential limited thyroid reserve, such as those with subclinical hypothyroidism, inadequately treated hypothyroidism or a history of autoimmune thyroiditis. Such monitoring is essential, as serum concentrations of thyromimetics cannot be measured in clinical laboratories, so their circulating concentrations and potential effects on other tissues cannot be adequately assessed, and the thyroid status of the treated patients can only be inferred from the serum parameters of the HPT axis.

Extrahepatic effects of thyromimetics on heart and bone

In a 52-week phase III trial, patients treated with resmetirom showed adverse effects of mild to moderate diarrhoea (in 23.5% and 31.2% of those receiving 80 mg and 100 mg resmetirom per day, respectively) and nausea (in 11.9% and 18.2%, respectively), compared with diarrhoea in 13.8% and nausea in 7.9% of patients receiving placebo¹⁴¹. Diarrhoea and nausea occurred more frequently during the first 12 weeks of treatment, and their incidence did not increase any further after 12 weeks¹⁴¹. There were no increases in tachycardia or arrhythmias, suggesting THRA activation in the heart, found after resmetirom treatment in either the phase II or the phase III trials^133,141. Resmetirom treatment also slightly decreased blood pressure by 2–3 mmHg in all treatment groups compared with baseline in the phase III trial¹⁴¹. In the phase III trial, no increases in fractures or substantial changes in bone mineral density T-scores were seen after 52 weeks. Although there would be expected to be little or no THRA activation in bone by a liver-selective THRB agonist, the follow-up period of these studies might not have been sufficient to determine whether long-term treatment decreases bone mineral density and/or increases fracture risk. The bone cartilage damage observed in dogs following treatment with eprotirome¹²³ indicates that the effects of resmetirom on bone need to be monitored regularly and chronically in patients taking this drug until this issue is entirely resolved. Thus, longer follow-up studies are needed to determine whether resmetirom has any deleterious long-term effects on bone. Similarly, no significant changes in heart rate, dysrhythmias and blood pressure were seen following resmetirom treatment in the study¹⁴¹; however, further studies are needed to assess safety fully, particularly in patients who might have pre-existing cardiovascular disease¹⁴¹.

Thyroid hormones and thyromimetic effects on HCC

HCC is complex and multifaceted. Although there might be an increased risk of breast, lung and prostate cancer associated with hyperthyroidism^143,144, there is no convincing causal link between elevated serum thyroid hormone levels and HCC in humans. A study showed that patients with HCC and cirrhosis might have higher serum free-triiodothyronine levels than patients with cirrhosis alone; however, free thyroxine and TSH were similar in both groups¹⁴⁵. Additionally, in patients with HCC, elevated serum free-thyroxine levels were associated with larger tumour size and poorer prognosis than normal free-thyroxine levels¹⁴⁶. An early study suggested that hypothyroidism might inhibit hepatoma growth in rats, suggesting thyroid hormone-dependent growth in some types of hepatomas¹⁴⁷. The actions of thyroid hormones on lipogenesis, cell proliferation and the upregulation of genes associated with HCC aggressiveness, such as those encoding furin and lipocalin 2, suggest that thyroid hormones might have pro-oncogenic effects¹⁴⁸. Thyroid hormones binding to the plasma membrane integrin β3 may also promote liver cancer progression^28,29.

On the other hand, there seems to be an association between hypothyroidism and HCC¹⁴⁹, which might, in part, be due to the association between hypothyroidism and MASLD and the increased risk of progression of MASLD to HCC⁷⁷. Thyroid hormones also cause regression of preneoplastic tumours and HCC in rats¹⁵⁰. Additionally, thyroid hormones downregulate the expression of oncogenes such as cyclin-dependent kinase 2, cyclin E and phosphoretinoblastoma protein, induce metabolic reprogramming¹⁵¹ and upregulate tumour suppressor genes in cultured HCC cells¹⁴⁸. Decreased expression of THRA, THR and thyroid hormone-regulated target genes in HCC suggests that THRs might have tumour-suppressive effects. Furthermore, the high frequency of somatic point mutations of THRA and THRB found in HCCs^152,153 suggests that deficient or dominant negative THR activity might be associated with hepatic oncogenesis. In this connection, transgenic overexpression of the dominant negative oncogene, v-ErbA, which blocks normal thyroid hormone-mediated transcription, caused HCC in mice¹⁵⁴. Thyroid hormone might also induce metabolic reprogramming in HCC by changing the Warburg effect of fuel utilization in tumours from glycolysis to fatty acid oxidation¹⁵⁵. Additionally, the beneficial effects of thyroid hormones on autophagy, lipid metabolism, lipotoxicity, inflammation and fibrosis in MASH might also help prevent or delay the onset of HCC. Indeed, the association between the progression of MASH and the risk of HCC has been well-established, although it can occur even without cirrhosis⁷⁷. Thus, reversal of MASH progression by thyroid hormones or thyromimetics would presumably reduce the risk of HCC. Notably, the progression of HCC was inhibited by triiodothyronine and the thyromimetic TG68 in rat models of HCC, suggesting that thyroid hormone and/or thyromimetics could be useful clinically for the treatment of HCC^151,156.

Thus far, there have not been any careful preclinical or clinical studies of thyromimetics on HCC or other cancers. Although hyperthyroidism might increase the risk of some cancers other than HCC^143,144, there has not been convincing clinical data for HCC. There is some evidence suggesting that hypothyroidism might contribute to HCC¹⁴⁹; however, more confirmatory studies are needed. Besides circulating thyroid hormone levels, local intrahepatic thyroid hormone concentration could also have a role. Indeed, similar to MASLD⁴², DIO1 activity is decreased in HCC compared with normal liver tissue from patients¹⁵⁷. Although there might be dysregulation in other cancers, increased expression of DIO3 has not been reported in HCC despite being reported in hepatic haemangiomas¹⁵⁸. DIO2 is poorly expressed in adult liver and so far has not been reported to undergo changes in expression in HCC. Thyroid hormones and thyromimetics can be effective for treating MASLD, which frequently precedes HCC; therefore, they might help prevent or reduce progression to HCC. However, further studies are needed to determine whether thyroid hormones and/or thyromimetics have direct beneficial effects in HCC. It would also be important to weigh the risks versus benefits of thyromimetic therapy for MASH in patients with a high risk of cancer as well as pre-existing or active cancer, given the association of hyperthyroidism with certain types of cancers.

Conclusion

Thyroid hormones are essential in maintaining glucose, lipid and protein homeostasis in the body. Disturbances in circulating and/or intra-cellular thyroid hormone levels can disrupt these metabolic pathways in the liver and other tissues. The increasing prevalence of obesity and type 2 diabetes has contributed to an epidemic of MASLD, with more than 30% of adults affected worldwide. In preclinical models, thyroid hormones decrease steatosis, inflammation and fibrosis by induction of autophagy and β-oxidation of fatty acids in the liver. In a clinical study, low-dose levothyroxine also decreased hepatosteatosis in patients with type 2 diabetes⁴². THRB-selective and liver-selective thyromimetics also show promise by decreasing liver fat content, serum ALT and AST levels, and fibrosis without adverse effects on bone or heart tissue^131,133.

Notably, no serious effects on serum TSH and free triiodothyronine Published online: xx xx xxxx levels have been reported for thyromimetics that are currently being clinically evaluated for the treatment of MASH, although slight suppression of free thyroxine levels was observed following treatment with both eprotirome and resmetirom^134,141. The mechanism for this decrease in serum levels of free thyroxine is unclear. Nevertheless, these findings raise the possibility that the use of thyromimetics could be problematic in patients with low thyroid reserve, such as those with subclinical hypothyroidism.

Serum SHBG seems to be a reliable marker of thyroid hormone activity in the liver¹⁵⁹. It holds potential prognostic value for predicting response to thyromimetic therapy in patients with MASH¹³¹. Although other markers of thyroid hormone action in the liver, such as serum cholesterol and ferritin levels, might offer less precision, they can still confirm thyromimetic effectiveness in the liver. Both hypothyroid and hyperthyroid responses to thyromimetics can potentially occur. Depending upon their ability to cross the blood–brain barrier, they might cause TRH and/or TSH suppression, leading to secondary or tertiary hypothyroidism, as the hypothalamus and pituitary primarily express the THRB2 isoform. Although THRB-selective thyromimetics do not seem to have much activity in tissues that predominantly express THRA, such as the heart and bone, they still could have thyromimetic activity in other tissues besides the liver that express high levels of THRB, such as the brain, kidney, lung, gut, retina and cochlea, especially if they accumulate within these tissues. Thus, it is possible that excessive dosing of thyromimetics could create local ‘hyperthyroid-like’ states in those tissues. Although no significant adverse effects involving these tissues have been reported for resmetirom, careful studies need to be performed to be ensure that there are no subtle adverse effects that could be problematic in patients with pre-existing conditions involving these tissues.

The effects in the tissues mentioned above highlight the need for the development of tissue-specific serum markers to assess the level of thyroid hormone activity in vulnerable tissues, considering that most tissues have a mixture of THRA and THRB. Such markers would also be useful to prevent thyroid hormone overactivity in tissues expressing both THR isoforms, such as the lung and gut. In this connection, a small percentage of patients reported nausea and diarrhoea as adverse effects of resmetirom treatment (nausea in 11.9% and 18.2% of patients receiving for 80 mg and 100 mg resmetirom per day versus 7.9% of those receiving placebo, and diarrhoea in 23.5% and 31.2% versus 13.8%, respectively)¹⁴¹; however, it is not known whether inappropriate thyromimetic activity caused these symptoms. Moreover, the lack of commonly available tests for measuring circulating thyromimetic concentrations hinders the determination of these concentrations alongside routine thyroid function tests. Developing such tests to measure serum levels of tissue-specific markers and thyromimetics would be necessary for a more comprehensive understanding of the effects of thyromimetics on thyroid function.

Although concerns persist about deleterious extrahepatic effects of THRB-selective and liver-selective thyromimetics, exploring the potential benefits of thyroid hormones and thyromimetics in conditions such as type 2 diabetes, obesity, heart failure with preserved ejection fraction, sarcopenia and chronic kidney disease associated with MASLD are also important. Further research, alongside reliable serum markers of thyroid hormone action in specific tissues, will play an important part in determining the broader therapeutic potential of thyroid hormones and thyromimetics on these complications, which substantially contribute to the morbidity and mortality of patients with MASH.

Key points.

• Hypothyroidism is associated with metabolic dysfunction-associated steatohepatitis (MASH).

• Deiodinase 1 mRNA and protein expression and activity are downregulated as MASH progresses to cause ‘intrahepatic’ hypothyroidism.

• Increased lipogenesis and decreased fatty acid β-oxidation cause hepatosteatosis and lipotoxicity that lead to inflammation and fibrosis in MASH.

• Thyroid hormones increase autophagy of lipids (lipophagy), β-oxidation of fatty acids and mitochondrial turnover to reverse inflammation and fibrosis.

• Thyroid hormones or thyromimetics are effective therapeutic agents for MASH in mouse and human studies.