Reviewing Editor: Clara Alvarez

The Authors have nothing to disclose.

Correspondence to:
Dr. JC Galofré, Departamento de Endocrinología y Nutrición, Clínica Universidad de Navarra. 31080 Pamplona (Spain); Phone: +948-255400; E-mail: jcgalofre@unav.es

ABSTRACT
The association between thyroid hormones and energy expenditure (EE) is well established. Furthermore, an inverse relationship between obesity and EE is also well-known. Therefore, the enhancement of EE emerges as a plausible treatment for obesity. At present, no clinically approved drugs for this aim are available. However, the new thyromimetic compounds may fill in this gap. This review summarises the influence of thyroid hormones (TH) in obesity development, body composition changes and thermogenesis. The potential mechanisms involved in the variation of serum TH concentrations across the range of broad body mass index are briefly outlined. Finally, a synopsis of the thyroid mimetic compounds is presented.

1.0. Introduction
1.1. Obesity and thyroid hormones
Over the last few years an unprecedented increase in obesity prevalence has settled worldwide, especially in industrialised countries (1). The reasons for this pandemic are still a matter of debate, but they are certainly related to the profound changes associated with modern lifestyle that implies a deep transformation in energy balance. Albeit well known for decades that thyroid hormones (TH) play a key role in regulating energy homeostasis (2), the obesity pandemic has driven new interest in the relationship between TH and weight status (3). Weight loss is a typical sign of thyroid hyperfunction, whereas hypothyroidism is generally associated with weight excess. The relationship between weight and TH has been broadly spread by the media, and not always with precise information. Therefore, practitioners commonly deal with overweight patients who believe that small changes in thyroid function have significant impact on body composition. This type of patient usually blames his or her thyroid as the cause of obesity. And they might be right. If not fully right, at least partially right (4). In this scenario it is plausible to speculate that TH analogues may be used in the treatment of overweight and obese people in the future.

1.2. The normal thyroid function
There is a universal agreement that thyroid function is initially determined by serum thyrotropin (TSH) concentration. Unfortunately, the agreement is lacking as regards the definition of normal thyroid function as such (5-8). Wide TSH level variations are common when serum samples from different healthy subjects are compared, even if the analysis is performed within the same age range. For instance, there is compelling evidence that normal TSH levels increase with age, although a recent study found that TSH secretion is gender invariant and depends on age in women only (9). Therefore, a specific TSH normal range for different situations is needed, including ethnic origin, age sex, health status and, probably, body mass index (BMI) (10).

1.3. Thyroid function and body composition
More than a century has elapsed since the first clinical observation that hyperthyroid patients tend to lose weight and the reversibility of this tendency once treatment is established (11). However, unfortunately the explanation for this behaviour is still elusive in many aspects. Epidemiological data, albeit limited, generally show a higher prevalence of overt and subclinical hypothyroidism (~20%) in morbid obese individuals (12). Although the range of TH values may vary in different populations (regarding, for instance, dietary iodine intake or other factors), the usual finding is that TSH levels correlate with body weight (10,13,14). Many groups have reported that baseline serum TSH levels are usually in the upper limit (or slightly over it) of the normal range in euthyroid obese individuals (12,14-23). Additionally, in these subjects (even in euthyroid individuals), the increase in TSH concentrations is associated with elevated waist circumference and BMI (24,25). However, this is a non-consistent finding as several clinical observations show conflicting results (12,26-28). For instance, a small French study of 20 hypothyroid women and 17 controls found no significant differences in body composition, heart rate, energy metabolism, or muscular function between the treated and untreated groups. The authors concluded that the increase in circulating thyroxine (T4) does not appear to modify the body composition or muscular function of women (29). A similar conclusion is provided by a larger British study on a cohort of 401 euthyroid nonobese and obese subjects (27). Nevertheless, these are not unexpected findings because there are consistent observations showing that a reciprocal relation between weight and serum T4 levels is lacking. Furthermore, it has been found that those morbidly obese subjects with higher TSH concentrations exhibit higher levels of triiodothyronine (T3) (which is quite a constant observation) and, only in some studies, also high T4 (30). Additionally, in an Italian cohort of women an increase in free T3 and TSH levels was associated not only with BMI, but waist circumference and fat accumulation as well (31). This TSH profile was also observed in a sample of obese children (32). As expected, the decrease in TSH and free T3 associated with weight loss in obese women indicate a reduction in energy expenditure (EE) in response to caloric restriction (30). In line with these observations, another Italia group reported that weight loss after laparoscopic gastric banding induces a decrease in free T3, while free T4 increased and TSH remained steady despite the fact that all values were within the normal range both before and after surgery (33). The interpretation of the former results suggests that progressive central fat accumulation is associated with a parallel increase in free T3 levels (FIGURE 1), probably as an adaptive thermogenic phenomenon, and the regulation of TSH secretion by free TH is possibly impaired in obesity (31).



Fig. 1. The reference range for serum concentration of thyroid hormones does not take into account special circumstances like weight, age, sex, iodine intake, etc. All these situations affect the reference range limits. There is a good amount of evidence showing that weight is directly correlated with serum Thyrotropin (TSH) and free triiodothyronine (FT3) levels, whereas free thyroxine (FT4) remains unchanged. The illustration summarises this idea. The means for normal serum FT4, FT3 and TSH levels are depicted by a blue, red and green line, respectively. The "normal" reference range remains useful for all three thyroid hormones in normal weight subjects (yellow shaded areas), while TSH and FT3 are usually below the "normal" reference range in underweight patients, as is the opposed to the obese. Thus a "new normal" reference range should be considered and adjusted for the subject's weight (red shaded area).

Some authors have studied the contribution of thyroid autoimmunity to the elevation in serum TSH levels of morbid obese subjects (34). They found that autoimmunity is not a major cause sustaining the high rate of subclinical hypothyroidism in these patients. Thus, in this population, the diagnosis of subclinical hypothyroidism, as assessed by an isolated high serum TSH level, remains questionable (3,34). Additional information has been offered by a study on patients with differentiated thyroid cancer. Resting EE (REE) and body composition were evaluated during the short-term of hypothyroidism previous to the whole body scan, and on TSH-suppressive LT4 treatment. The results were compared with healthy controls. REE was significantly lower, whereas the percentage of body fat was significantly higher in patients than in controls. It was concluded that LT4 treatment enhances REE, but the increase was not significantly different from controls. Short-term deprivation of TH has an impact on body composition and influences EE (35).

1.4. Thyroid function and energy consumption
EE is determined mainly by REE and physical activity (FIGURE 2).



Fig. 2. Body Mass Index (BMI) depends on the balance between Energy intake (ie Food intake) and Energy consumption. Energy consumption is determined mainly by Physical Activity (PA) and Resting Energy Expenditure (REE). REE depends on adaptative and obligatory thermogenesis, which is in part mediated by thyroid hormones. The evolution of BMI over the life span depends on the equilibrium of these factors.

REE depends on obligatory and adaptative thermogenesis. Controlling thermogenesis is one of the major tasks of T3 (36). TH are key regulators of metabolism, although it is uncertain which T3- responsive-energetic processes are most important for the determination of the basal metabolic rate (37). Around 30% of obligatory thermogenesis depends on TH, and this fraction is essential for the temperature homeostasis. This action (the rise in basal thermogenesis and therefore the obligatory thermogenesis) is driven through speeding up ATP turnover. T3 raises basal metabolic rate and promotes thermogenesis by inducing an increase in the mitochondrial respiratory chain activity. TH are also important for adaptative thermogenesis (38). Moreover, the most specific example of TH-dependent EE is not related to the basal metabolic rate, but rather to the adaptative thermogenesis (38). Adaptative thermogenesis is characterized by an uncoupling of oxidative phosphorylation in cold-exposed brown adipose tissue (BAT), which is dependent on locally generated TH (3). A decisive component in this process is the type 2 thyroxine 5'-deiodinase (D2), which converts T4 to the active metabolite T3. Particularly, D2 can increase local, intracellular T3 production from T4 without affecting serum T3 levels (39). It has traditionally been thought that the role of BAT in humans was limited. However, a number of recent studies have changed this view, opening new fascinating perspectives. By the use of PET scans, several authors have shown the presence of 18-FDG uptake in BAT areas as well as regions of hypermetabolism in relation with cold exposure (40-42).
TH induce changes in behaviour and physical activity. Overactive thyroid patients display more physical activity than normal subjects (43). This association is most likely related to the recognized relationship between TH action and body mass. In situations with a lack of physical activity there is an increase in serum reverse T3 (rT3) levels that denotes an increase in thyroxine 3 5-deiodinase enzyme (D3) activity, which converts T3 to the inactive metabolite rT3 (43). T3-driven mechanisms in tissues other than skeletal muscle may eventually prove to be important as well, such as the work performed by the heart, which may be responsible for up to 15% of the overall EE at rest (3,44).

1.5. Thyroid hormones and age
Extreme longevity has been associated with an increase in serum TSH concentrations (45) with thyroid diseases being more prevalent in older people (46). However, it is not known whether the rise in serum TSH levels represents a physiological trend related to aging or the consequence of illness (47). Therefore, an elevation in serum TSH level may be the translation of two situations with opposite influence on longevity (48). A low metabolic rate (which is associated with high TSH level) is a longevity marker. Furthermore, caloric restriction slows down the aging process. On the other hand, illness is a known cause of elevation in TSH – there are a good number of pathological conditions associated with mild or subclinical hypothyroidism (43,49,50). Thus, mild hypothyroidism seems to be detrimental for young or middle aged subjects, whereas it may be harmless or perhaps beneficial for advanced aged individuals (48). Therefore, it looks as if some of the metabolic alterations related with obesity may regulate thyroid homeostasis; and, this situation seems to be mitigated with age (15).

2. Pathophysiology
Many teams of investigators have been speculating on possible mechanisms that could explain the relation between obesity and thyroid gland activity. However, as recently published, fat and energy economy in hypothyroidism and hyperthyroidism are not the mirror image of one another (11). The observed positive association between TSH and BMI could be due to alterations in TH activity or as a result of an alteration in the regulation of the hypothalamic-pituitary-thyroid (HPT) axis. The hypothesis that involves a direct effect of TSH is also plausible as the TSH receptor is expressed in adipose tissue (51). It has been published that circulating cytokines related with metaboli syndrome can suppress thyroid function either at hypothalamic or pituitary or thyroid levels (52). The contribution of autoimmunity or iodine deficiency to the rise in serum TSH levels in the obese population has been ruled out (53).
The more suitable contributing factor is the deregulation of the HPT axis in the obese population, since a direct relationship between TSH and BMI has been consistently observed (16). However, as aforementioned, there are conflicting data in the literature regarding the relationship between obesity and TH. Some studies, but not all, demonstrated low T3 and low T4 at higher body weight and BMI levels, whereas other authors found a direct relationship between free T3 and BMI. Therefore, there are a number of factors that contribute to free T3 levels in obese subjects. These factors could vary among different subjects with same BMI, like body composition, underlying thyroid diseases, iodine intake, etc.
In this scenario, it is then plausible that a neuroendocrine dysfunction resulting in an abnormal secretion rate of TSH could be the cause of elevated TSH concentrations in obese subjects. It has been observed that D2, which is the main pituitary deiodinase isoenzyme, and its activity, is the key point to release TSH under T3 control, but does not work appropriately in these individuals. This mechanism may be damaged according to the observation that pituitary D2 expression does not reach the normal range in obese subjects. In addition, an Ala92 D2 variant in humans induces obesity and an insulin resistant state in comparison with wild type 92Thr D2 (54). Consequently, a resetting of the HPT axis, and not merely insufficient TH levels, seems to be a key factor that shifts the EE equation in obese subjects. Fasting, for instance, induces profound changes in the HPT axis as well.
Studies in rodents have shown a dramatic down-regulation of TRH gene expression in the paraventricular nucleus (PVN) during fasting. Direct and indirect effects of decreased serum leptin, in addition to effects on increased local T3 concentrations in the hypothalamus during food deprivation, contribute to a decreased activity of TRH neurons in the PVN. Pituitary TSHβ mRNA expression also decreases during fasting, and this may be relatively independent of leptin and/or TRH, since leptin administration in this setting does not fully restore pituitary TSH expression, while it does restore TRH expression in the PVN. The observed decrease in serum TH concentrations is the result to some extent of a diminished thyroidal secretion of TH. The overall result of these complex HPT axis changes in various tissues during fasting is down-regulation of the HPT axis, which is assumed to represent an energy-saving mechanism, instrumental in times of food shortage (55).
Thus, it seems that the adipocyte-derived hormone, leptin, may be at the origin of this dysfunction (56) (see below), although other possibilities may also exist. Some investigators have suggested the existence of partially bioinactive TSH in obese subjects, although this hypothesis is very speculative (3). Other authors suggest that there may be certain TH resistance, as well as decreased T3 receptors in obese subjects (57).

3.0. Associations related to thyroid hormones and obesity
3.1. Insulin resistance
The association between thyroid disease and glucose metabolism is well documented (58). Insulin resistance with hyperinsulinemia are important features of the metabolic syndrome and generally accompanies obesity (59). Insulin resistance has been related with both ends of the thyroid dysfunction spectrum, although the poorly understood mechanisms might be different in each case (60). Hyperthyroidism is a well-known cause of hyperglycaemia. The explanation for this relationship could be an unopposed activation of gluconeogenesis (61). Moreover, hyperthyroidism, even in subclinical forms, is associated with a reduced insulin half-life, most likely because of accelerated insulin degradation (58,60,62). On the other hand, hypothyroidism-associated insulin resistance could be the result of diminished tissue sensitivity to insulin. Accordingly, glucose disposal is then reduced in this situation (58). In hypothyroidism, insulin resistance is counterbalanced by a parallel reduction in gluconeogenesis, making it habitually irrelevant without clinical consequences (58).

3.2. Thyroid and adipokines
Several groups have studied the relationship between leptin and the pituitary-thyroid axis (15,22,62-67). Interestingly, TSH and leptin display similar serum concentration profiles in their circadian rhythm, which may indicate a leptin regulatory effect on TSH secretion. Considering that leptin is an indicator of fat mass, the observed association between serum leptin levels and thyroid function is interesting. Unfortunately, once more the results of these studies fall under discrepancy (56,63,65,68,69), making difficult the elucidation of the reasons for this relationship. Different investigators have found almost all possible combinations between thyroid function and leptin. Some of them have associated hypothyroidism with serum leptin levels below (69-71), above (72) or in the normal (56,73) range. Something similar has been shown in hyperthyroid subjects where high serum TH concentrations have been linked with low (71), high (70,74) or normal (69) serum leptin levels. As aforementioned, some authors have hypothesised on the role of leptin in the modulation of the pituitary-thyroid axis. This has been demonstrated in premenopausal obese women (67), despite conflicting findings. A recent study was also completed in a sample of premenopausal women with hyperthyroidism or hypothyroidism. Treatment of thyroid dysfunction – not associated with changes in BMI of % of body fat – did not influence serum leptin but did affect serum ghrelin. The authors concluded that thyroid status itself, in the absence of alterations in BMI and % body fat, exerts an important influence on circulating ghrelin but not leptin (75). In our experience (15), in agreement with former studies (76,77), there is a significant positive correlation between circulating leptin and TSH levels in a sample of obese men. On the other hand, we also observed that the correlation between leptin and age was negative (15).
Nowadays, there is no clear explanation to disclose the reasons for these disagreements between different studies. We can speculate that the catabolic condition of thyrotoxicosis is very similar to the fasting state, and both situations seem to lead to a reduction in both serum leptin and TSH secretion. On the other hand leptin itself directly stimulates TRH (38) secretion, and subsequently TSH and TH. A rise in serum TSH levels is usually interpreted as a hypothyroid status, but may also be the result of an effort to stimulate the thyroid and, therefore, the induction of gland overactivity. In addition, leptin has been shown to have a direct inhibitory effect on several components involved in TH production from thyrocytes (62); and, leptin may directly affect the sensitivity of the thyrotroph or the thyrocyte (4).
Data regarding the relationship between ghrelin levels and thyroid function exist, but are scarce. Serum ghrelin levels are increased by 32% in the hypothyroid state and became normalized after L-thyroxine replacement. Therefore, serum ghrelin levels are reversibly increased in hypothyroid patients (78). In agreement with this information, it has been found that hyperthyroid subjects present low serum ghrelin levels (79). In addition, circulating ghrelin has been significantly correlated with age, fasting, glucose and TSH, but not with BMI (79). In any case, it seems that there are complex mechanisms involved in these observations that are worth of clarification.

4.0 Thyroid hormones and mimetic compounds in the treatment of obesity
Obesity is defined as an excessive accumulation of body fat. Therefore, the ideal treatment for patients with obesity aims to achieve a negative caloric balance, not only to reduce weight, but also to improve body composition by decreasing as selectively as possible the percentage of body fat. However, caloric deprivation usually also leads to reduction of both fat tissue and fat-free mass. According to this, a desirable therapeutic response consists in not allowing a fat-free mass wasting superior to 25% of total body weight. Energetic balance, macronutrient proportion and physical activity are significant factors involved in the control of body composition in obesity treated subjects. Loss of fat-free mass due to a reduction in muscle tissue is at least in part, responsible for a fall in resting EE, which contributes to the frequent phenomenon of weight regain. Intense caloric deprivation is associated with a decrease in plasma leptin, T3 and sometimesT4 concentrations and a rise in rT3 levels in what has been considered as an adaptation process to reduce metabolic needs (80). This T3 decrease has also been involved in the weight loss-associated REE fall (81). In an attempt to maintain or promote further weight loss and avoid weight regain, different trials with TH supplementation have been carried out (82). The purpose of this treatment would be to increase fat loss through enhancing oxygen consumption and fatty acid oxidation without having either TSH suppression or side effects on muscle, central nervous system, bone or cardiac function.

4.1. Thyroid hormones
Recently, a systematic review has been published on the results obtained with TH treatment in obese patients submitted to caloric deprivation (82). The paper includes 14 studies designed to test the efficacy of T3 administration in obese treated patients. However, the heterogeneity in the quality and design of these trials prevented drawing any firm conclusions. T3 doses ranged from 18 to 117 mcg/70 kg of body weight. Significant weight loss was only achieved in five out of 26 comparisons and TSH and T4 values, when assessed, were systematically reduced. No firm data on protein breakdown due to catabolic effects of TH were obtained. No studies performed body composition evaluations to investigate if there were changes in the fat or fat-free mass compartments. Only few studies measured REE, nitrogen balance and 3-methylhistidine urinary excretion, showing no consistent results. No clear variations in heart rate were seen. Therefore, those studies did not demonstrate any sustained benefit on weight loss, whereas TSH suppression and a decrease in T4 values are compatible with inhibition of the HPT axis due to T3-induced subclinical hyperthyroidism. T3 administration increases leptin gene expression following caloric restriction in obese rats. Nevertheless, the contribution of this mechanism to weight loss or maintenance is unlikely since no increase in circulating leptin levels or significant weight reduction were observed in that experimental model (83).

4.2. Dextrothyroxine and TRIAC
Use of TH analogues have been tried in the past with the aim of taking advantage of beneficial actions such as fat mass reduction or control of hyperlipidemia while avoiding side effects on bone, brain and heart. Clinical experiences using dextrothyroxine for hyperlipidemia therapy have been unsuccessful (84). However, a natural TH metabolite, triiodo-thyroacetic acid, has shown thermogenic capacity in brown adipocytes in culture (85,86), though no clinical studies have confirmed its efficacy in the treatment of obesity.

4.3. Selective thyroid hormone receptor activation
The progressive knowledge in the mechanisms of TH action at the cellular level has opened new possibilities on the therapeutic application of selective TH receptor (THR) activation. THR encoding genes are differentially expressed in various tissues. Different studies carried out in mice with inactivation of different THR isoforms (87,88) and data from patients with resistance to TH (89) have demonstrated that different THR forms are responsible for tissue-specific responses to TH. THR alpha (THRα) is mainly present in the brain and heart where it regulates cardiac function, whereas THRβ is preferentially detected in liver, where it is responsible for the effects of TH on lipid metabolism (90). The THRβ 1 is the predominant systemic form, while the pituitary form, which controls TSH secretion, is THRβ 2. These investigations have opened new perspectives to look for selective stimulation of particular TH actions for treatment of some diseases such as dyslipidemia and obesity. A rise in REE as well as a reduction in fat mass and an improvement in insulin sensitivity and lipid profile represent interesting objectives to achieve in obesity treatment following selective thyroid receptor activation. At the same time, classical side effects due to TH overexposure such as muscle wasting, bone loss, nervousness, hypertension and cardiac dysfunction (arrhythmias, heart failure) should be avoided.
Changes in the molecular structure leading to selective receptor binding with a particular THR isoform as well as the capacity of different compounds to be taken up by specific tissues explain the specificity of actions (91). Special efforts have been made to develop THRβ selective modulators due to their preferential effect on liver metabolism.
A selective thyromimetic compound, GC-1 (3,5-dimethyl-4-4-hydroxy-3-isopropylbenzyl phenoxy acetic acid -sobetirome-) with 10-fold preferential action on THRβ 1 over THRα 1, was able to induce an increase of EE of 5-10% while provoking only mild tachycardia in mice (92). The drug also caused a 4% body weight reduction in cynomologous monkeys treated for 7 days with no evidence of muscle wasting (93). Affinity of GC-1 for THRα1 is 10 times lower than T3 (94), accounting for the attenuation in heart rate stimulation. Accordingly, the increase in heart rate related to the rise in EE was less than that seen following treatment with T3. In addition, GC-1 administration to primates has been followed by an increase in oxygen consumption and body weight reduction (93). More recently these results have been confirmed in female rats treated for 6 weeks with either T3 or CC-1. The animals showed a rise in oxygen consumption similar to that found after T3 treatment, whereas GC-1 induced a 20% reduction in fat mass without increasing food intake. In contrast with T3 administration, no changes were seen in heart mass; and skeletal muscle mass was minimally affected by GC-1. These results suggest that GC-1 may have a promising role as an antiobesity agent, since it reduces fat mass without increasing food intake and facilitates the control of dyslipidemia without having deleterious effects on heart or bone mass (95).
Another THRβ agonist, KB-141, is 10-fold more selective for stimulating metabolic rate and 30-fold more selective for cholesterol lowering than for positive chronotropic effects. However, accumulation in the liver is much less than that seen with CC-1. To maintain or broaden this therapeutic window, reassuring the lack of cardiac side effects is essential before considering its use in humans. KB-141 has shown to induce weight loss, cholesterol and Lp(a) reduction. However, this compound decreases TSH values leading to secondary hypothyroidism in territories not accessible to selective THRβ agonists such as the brain (96).
GC-24 is another THRβ agonist that has demonstrated ability to reduce body fat accumulation and to prevent liver steatosis in rats with diet-induced obesity and increasing EE (97). These effects were seen without any change in food intake or cardiac weight, suggesting that this compound does not act significantly on myocardium. GC-24 also reduced the glucose response to a glucose load, improved insulin sensitivity and normalized the previous hypertriglyceridemia. However, GC-24 reduced only marginally total cholesterol levels and did not have any effect on free fatty acids or IL-6 levels. GC-24 increased the expression of Cpt 1, Sd and Acc 1 in BAT, suggesting that the drug has significant thermogenic effects as well as effects on EE. GC-24 displays a 40-100 fold preference for THRβ over THRα (98).
KB 2115 is the only compound that has been administered to human subjects. The drug was effective in achieving a 40% reduction in total and LDL-cholesterol plasma concentrations after 14 days of treatment (TABLE 1). Mechanisms seem related to an increase in bile acid synthesis.
Interestingly, KB 2115 therapy was associated with a dose-dependent reduction in total and free T4 levels without inducing any variations in TSH concentrations (99). No changes in metabolic rate or body weight were observed.

Table 1. Effects and characteristics of different selective thyroid hormone receptor agonists.



EE: Energy expenditure. BAT: Brown adipose tissue. THR: Thyroid Hormone Receptor. Asterisks indicate the drugs that have been tried in human subjects.

Recently, a less selective thyromimetic, 3,5-diiodothyropropionic acid (DITPA), has been assessed in patients with stable congestive heart failure. DITPA was able to reduce body weight, total and LDL-cholesterol values and triglycerides to a lower extent. However, the drug induced suppression of the hypothalamic-pituitary thyroid axis and increased bone turnover, whereas it was ineffective in improving cardiac symptoms (100,101).
Experimental studies suggest that several obesity complications such as liver steatosis and type 2 diabetes can be improved by TH mimetics (96). Whether this is due to a direct effect or to an indirect action derived from fat mass loss remains unclear.

4.4. Bile acids
Dietary supplementation with bile acids can regulate EE and TH activation via changes in D2 expression, an enzyme involved in BAT thermogenic pathways that contribute to EE (102,103). Kaempferol, a polyphenolic molecule of dietary sources has been shown to increase skeletal myocyte oxygen consumption by rising cAMP generation and inducing protein kinase A activation. Also, this compound is able to influence the expression of genes involved in thermogenesis, such as UCP-3, and to up-regulate D2 gene expression, prolonging its half-life (104). These pathways may also represent a target with clinical application in the treatment of obesity and other metabolic disorders. Selective manipulation of TH actions represents a promising therapeutic tool for the treatment of obesity and some of its complications. Future investigations to design new compounds with selective effects on different metabolic pathways may lead to the generation of valuable drugs to be used as monotherapy or in combination with already established modes of therapy for obesity, fatty liver, type 2 diabetes or dyslipidemia. Nevertheless, more studies are needed to gain more knowledge about adequate dosage, time and mode of administration, and especially on safety matters regarding integrity of other organs sensible to TH administration, such as the heart, bone, HPT axis and the central nervous system.

Acknowledgements
We thank David Holzweiss for his critical reading of the manuscript.