| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
The Wenner-Gren Institute (V.G., A.M., B.C., J.N.), The Arrhenius Laboratories F3, Stockholm University, SE-106 91 Stockholm, Sweden; and Department of Cellular and Molecular Biology (H.G., B.V.), Karolinska Institute, SE-171 77 Stockholm, Sweden
Address all correspondence and requests for reprints to: Jan Nedergaard, The Wenner-Gren Institute, The Arrhenius Laboratories F3, Stockholm University, SE-106 91, Stockholm, Sweden. E-mail: jan{at}metabol.su.se.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1(-/-)ß(-/-) (TR-ablated mice)] had slightly decreased body temperature and much decreased basal metabolic rate, were still able to markedly increase metabolic rate in the cold, but were cold intolerant due to inadequate total heat production at low temperatures. A standard norepinephrine test showed that adrenergically induced thermogenesis could not be activated normally in the TR-ablated mice. This was not due to inadequate recruitment of brown adipose tissue, nor to the absence, decreased recruitment or dysfunction of the uncoupling protein-1. However, isolated brown fat cells were 10-fold desensitized, explaining the lack of response to standard adrenergic stimuli; cell culture experiments demonstrated that this desensitization was not an innate effect. Thus, the cold intolerance was probably not due to inadequate sympathetically induced nonshivering thermogenesis. Additionally, the results indicated that no metabolic effects of thyroid hormones could become manifest in the absence of nuclear TRs, that ligand-bound TRs were needed for euthermia and eumetabolism, but that TRs per se were not required for brown adipose tissue recruitment and uncoupling protein-1 gene expression. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Thyroid hormones are generally believed to exert their effects through nuclear hormone receptors, encoded by the thyroid hormone receptor (TR) and ß genes (6, 7, 8). The transcripts from the TRß gene give rise to several ligand-binding proteins (nuclear hormone receptors). Concerning the TR gene, the situation is more complex. Only one of the two major products of the TR gene, TR1, gives rise to a receptor; the other major product, TR2, does not bind ligand but retains DNA-binding properties; its function in vivo is unclear (9). In certain tissues, non-ligand-binding, non-DNA-binding proteins, TR
1 and TR2, may also be formed from the TR gene (10). Many of the earlier known effects of thyroid hormones have recently been ascribed to transcriptional regulation of specific target genes by the ligand-binding nuclear TRs (11, 12, 13, 14). However, mechanisms not involving transcriptional regulation have been suggested for some of the effects of thyroid hormones, especially in metabolism (e.g. see Refs. 5 and 15, 16, 17).
The availability of mice with different compound ablations of TRs (10, 18, 19) permits analysis of whether the nuclear TRs are involved in specific metabolic events and facilitates the identification of which receptors are involved. We have here used mice lacking all ligand-binding TRs [i.e. TR1(-/-)ß(-/-) mice, referred to below as TR-ablated mice] to delineate the significance of these receptors for metabolic control. Thus, the mice studied here express no T3- or T4-binding nuclear receptors, although the mRNAs for the TR2 and rev-erbA protein are still present at normal levels (20). The general effects of these TR ablations are relatively mild (11, 13, 21), especially as compared with the well-known grave effects of thyroid hormone deficiency during development (i.e. severe neurological, mental, and motor damage). The TR-ablated mice are approximately 25% smaller than wild-type and they exhibit disorders of e.g. the pituitary-thyroid axis, impaired bone maturation etc., but they are still able to grow to maturity, and the males are fertile (11, 13, 21). Metabolically, their blood glucose, free fatty acid, and ß-hydroxybutarate levels are normal (our unpublished observations).
In this study, we have used these TR-ablated mice to delineate to which extent and how the regulatory role of thyroid hormones on metabolism is mediated through the nuclear TRs. Theoretically, three outcomes should be distinguishable. Concerning thyroid hormone effects mediated by the nuclear TRs functioning as positive modulators, the TR-ablated mice should be expected to show traits of lack of thyroid responsiveness, i.e. their phenotype should be hypothyroid. Concerning thyroid hormone effects mediated by the unliganded thyroid receptors functioning as repressors, with thyroid hormones normally acting to alleviate this repression, a repression should not be evident in mice without these TRs (irrespective of whether thyroid hormone is present or not), and the mice should in this case be phenotypically euthyroid. Furthermore, the TR-ablated mice are hyperthyroidemic, with extremely high plasma levels of T3, T4, and TSH (19). Therefore, concerning possible thyroid hormone effects not mediated via the nuclear TRs, these mice should demonstrate hyperthyroid traits.
We conclude from this analysis that, concerning the metabolic parameters investigated, we find no evidence for non-nuclear-receptor-mediated effects of thyroid hormones. However, TR effects are manifest in distinct ways with respect to basal metabolism, to euthermia, and to nonshivering thermogenesis: concerning basal metabolism and body temperature set-point, the TRs function as positive modulators, whereas concerning recruitment of uncoupling protein (UCP) 1 and nonshivering thermogenesis, unliganded nuclear TRs function as repressors, an action that is normally reversed in the presence of thyroid hormones but which in the TR-ablated mice is inherently absent.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, heavy line). Similarly, mice with ablations of either the TR1 receptor, or of the TRß receptors, did not show cold sensitivity (not shown). However, in nearly all the mice lacking all hormone-binding nuclear TRs (the TR-ablated mice) (Fig. 1), the body temperature fell precipitously when the animals were exposed to 4 C. One animal of seven tested could defend its body temperature for 3 d, and only one animal could defend its body temperature for more than this time. [At normal ambient temperatures, the body temperature of the TR-ablated mice was 0.5–1 C lower than that of wild-type mice during all phases of the daily cycle (Fig. 1) (22).]
|
).
|
37 C) (filled arrow on x-axis in Fig. 2
).
In the TR-ablated mice, the thermoregulatory behavior was altered in several respects (Fig. 2). At temperatures below 18 C, the metabolic rate was high in these mice, but it was significantly lower than in wild-type mice. As these animals at these low ambient temperatures were unable to defend their normal body temperature (Fig. 1), the lowered thermogenic capacity could not compensate the heat loss. The inability to generate a sufficient facultative thermogenesis to counteract heat loss must be due to a reduced capacity for shivering or nonshivering thermogenesis (see below).
Within the temperature range 18–25 C, there was a less marked difference between the TR-ablated and the wild-type mice. In this temperature range, the slope of the thermoregulatory line of the TR-ablated mice was identical with that in wild type, indicating that there was no alteration in the total insulation properties of the TR-ablated mice (a better insulation would have decreased the slope of the line). However, the thermoregulatory line was shifted slightly to the left, and at any given temperature within this range, the metabolism was slightly lower (<10%) than in wild type. Indeed, at an ambient temperature of 22 C, heart rate (a good indicator of metabolic rate) is 10–20% lower in TR-ablated mice than in wild-type mice (22). Because the line was left-shifted but had the same slope as in wild-type mice, the line extrapolated to a lower temperature (35 C) (open arrow on x-axis) than that of wild-type mice; thus, the defended body temperature of the TR-ablated mice was lower than that of wild-type mice (cf. also Table 1). This would explain the slight metabolic depression (including lowered heart rate) observed at ambient temperatures between 18 and 25 C.
|
Thus, within their thermoneutral zone (range of minimal metabolism), the TR-ablated mice were impressively hypometabolic, the metabolic rate being about 30% lower than that of the corresponding wild-type mice (Fig. 2). Because body temperature was only decreased by 1–2 C, a Q10-like effect cannot explain this large decrease in resting metabolic rate. Rather, the thermoregulatory metabolic curves indicate two parallel effects of TR ablation: a modest decrease in body temperature set point and a large decrease in basal metabolism.
Adrenergically Induced Nonshivering Thermogenesis in TR-Ablated Mice
It was evident from the experiments shown in Figs. 1 and 2 that the TR-ablated mice had a lower maximal facultative thermogenic capacity than the wild-type mice. Facultative thermogenesis consists of both shivering thermogenesis (from muscle) and nonshivering thermogenesis [from brown adipose tissue (BAT)]. One (or both) of these sources must therefore have been attenuated or eliminated in the TR-ablated mice. Based on expectations from earlier experiments on classical hypothyroid mice (24), a probable explanation would be a reduction in the capacity for nonshivering thermogenesis.
Wild-type mice housed under normal animal house conditions (18–22 C) readily survive an acute exposure to 4 C (Fig. 1). This ability is due to the fact that 18–22 C is so much below thermoneutral temperatures (cf. Fig. 2) that a modest but essential nonshivering thermogenic capacity will be recruited, i.e. the mice become partly acclimated to cold. This acquired capacity for nonshivering thermogenesis, together with shivering thermogenesis, allows for adequate heat production at 4 C (25). The cold intolerance of the TR-ablated mice could therefore be due to an inability of these mice to recruit nonshivering thermogenic capacity during acclimation to 18–22 C. To investigate the ability of the mice to recruit nonshivering thermogenesis as an effect of acclimation to cold, we therefore chose an acclimation paradigm consisting in a comparison between animals acclimated to 30 C (referred to in the following as thermoneutrality) and to 18 C [referred to as cold acclimation (Figs. 1 and 2)]. In wild-type mice, exposure to 18 C necessitated the maintenance of a metabolic rate that was double as high as the BMR (Fig. 2); in TR-ablated mice, a metabolic rate 3-fold higher than the BMR was required. Thus, acclimation to 18 C would be expected to markedly recruit nonshivering thermogenic capacity in both types of mice.
This acclimation paradigm did not affect the body weight of the animals (Table 1), nor was the difference in body temperature between the TR-ablated mice and wild-type mice—or the difference in basal metabolic rate [i.e. the rate measured at thermoneutrality (30 C)]—influenced by cold acclimation.
To quantify the capacity of the animals for nonshivering thermogenesis in BAT, the standard norepinephrine (NE) test was used (26). In wild-type mice, maintained and measured at thermoneutrality (30 C), the increase in oxygen consumption caused by injection of the standard dose of NE, 1 mg/kg body weight (27, 28), was small, but it was somewhat higher than the response to the injection of vehicle (saline) (Fig. 3A). Acclimation of such wild-type mice to 18 C caused a pronounced increase in the response to NE (Fig. 3B). Thus, as expected (29), a higher capacity for nonshivering thermogenesis had been recruited due to the acclimation to cold.
|
), the response to NE injection was not different from that caused by saline injection. Acclimation of the TR-ablated mice to 18 C did not cause any increase in response to NE; rather, the response was even lower than in the mice acclimated to 30 C (Fig. 3D) (and paradoxically even lower than that to saline injection). Thus, the increase in the magnitude of the response to a standard NE injection that occurs in wild-type mice after acclimation to cold was not found in the TR-ablated mice. Taken by itself, this would indicate that the TR-ablated mice had lost the ability to show cold-acclimation-recruited nonshivering thermogenesis (but see below). Nonshivering thermogenesis is generally considered to be ß3-adrenoceptor mediated. In accordance with this, the pattern of the response to the selective ß3-agonist CGP-12177 was similar to the response to NE (not shown). Thus, adrenergically induced nonshivering thermogenesis could not be activated in the TR-ablated mice under standard experimental conditions.
BAT Is Competently Recruited in TR-Ablated Mice
The thermogenic response to injected NE, as examined above, represents an acute activation of the thermogenic capacity already present in BAT. However, during the adaptive process of acclimation to cold, there is an increase in the total capacity of the tissue for performing thermogenesis, resulting in the increased response to NE seen in Fig. 3, B vs. A. This process is multifactorial and involves increased total cellularity of the tissue as well as advanced differentiation of the cells in the tissue; these slow, adaptive processes may collectively be referred to as BAT recruitment (30). In wild-type mice, acclimation to 18 C, as compared with a thermoneutral temperature (
30 C), leads to a marked recruitment of BAT (31). A reasonable explanation for the inability of the TR-ablated mice to demonstrate a cold-acclimation-recruited increase in the magnitude of the response to NE (Fig. 3, D vs. C) could be that this process of recruitment of BAT had become severely impaired. Therefore, the effect of acclimation to cold on recruitment-related parameters of BAT was quantified.
The interscapular brown fat pads of the TR-ablated mice were smaller than those of the wild-type mice, but the TR-ablated mice were also approximately 25% smaller than the wild-type animals (Table 1). To compensate for the smaller size of the animals, all parameters were therefore recalculated to U/kg0.75. Only these relative content values will be discussed in the following (although full documentation is found in Table 1). When this compensation for animal body weight was made, no significant difference was observed between the wet weight of interscapular BAT of wild-type and TR-ablated mice (Table 1).
As expected (32), massive lipid accumulation was observed in the BAT of wild-type animals kept at thermoneutrality (Fig. 4A), and cold acclimation caused a marked reduction in the lipid content (Fig. 4B). Notably, the lipid content of the BAT of the TR-ablated mice was lower than that of the wild-type mice at thermoneutrality (Fig. 4C) and became even lower as an effect of cold acclimation (Fig. 4D). Thus, despite the apparent inability of the tissue to produce heat when stimulated by NE, the visual appearance of the tissue had changed in the direction of a more recruited tissue, both at thermoneutrality and during cold acclimation. To verify this apparent recruitment, we also measured tissue cellularity and protein content.
|
). Surprisingly, at thermoneutrality, the DNA content in TR-ablated mice was higher than that in wild-type mice. Cold acclimation did not significantly alter DNA content in the TR-ablated mice (Table 1) and in the cold-acclimated wild-type and TR-ablated mice, the tissue cellularity was the same.
As expected, cold acclimation also caused a 2- to 3-fold increase in total protein content of BAT in the wild-type mice. In TR-ablated mice at thermoneutrality, the protein content of BAT was higher than in wild-type mice. In cold-acclimated, TR-ablated mice, the protein content was further increased, and, as with DNA, was then similar to that in the wild type (Table 1).
Thus, the same pattern was visible in the three parameters determined: at thermoneutrality, the BAT of the TR-ablated mice was more recruited than the tissue of the wild-type mice (less stored lipid, higher DNA content, higher protein content), and cold acclimation resulted in some further recruitment so that in the cold, there was no difference between levels in TR-ablated and wild-type mice.
Heat production takes place in the mitochondria. Mitochondriogenesis is an essential part of BAT recruitment (33) and is in general known to be dependent on thyroid status (34). Thus, if mitochondriogenesis was inadequate, despite increases in the other recruitment parameters, this could explain the lack of an augmented response to NE in the TR-ablated mice in the cold. We used COX IV (subunit IV of cytochrome c-oxidase) as a marker of mitochondrial density. In wild-type mice, the amount of COX IV protein was, as expected, markedly increased in the cold, becoming 4-fold higher than at thermoneutrality. In the TR-ablated mice at thermoneutrality, the COX IV content was about triple that in BAT of the corresponding wild-type mice, and tended to be further increased by acclimation to cold, to reach the same level as in the wild-type mice (Table 1). Thus, mitochondriogenesis did not depend on the presence of TRs, and a lack of mitochondria could not explain the lack of an augmented response to NE in the TR-ablated mice in the cold.
UCP1 Is Adequately Expressed in TR-Ablated Mice
Within the brown fat mitochondria, thermogenesis is fully dependent on the presence of UCP1 (35), and the amount of UCP1 is the limiting factor for cold-induced nonshivering thermogenesis (25). Thus, the lack of an augmented response to NE in the TR-ablated mice in the cold Fig. 3) may be due to a selective lack of UCP1, especially because the levels of UCP1 have been reported to be dependent on thyroid status (36, 37, 38, 39, 40, 41).
As expected, UCP1 mRNA levels in the BAT of wild-type mice were increased severalfold upon cold acclimation (Table 1). Notably, in the TR-ablated mice, the UCP1 gene was indeed expressed; in animals kept at 30 C, UCP1 mRNA levels tended to be higher than in the wild-type mice. Acclimation to cold resulted in a further induction of gene expression even in the TR-ablated mice (Table 1), and UCP1 mRNA levels were then similar to those in wild-type mice acclimated to the same temperature. Thus, despite the complete absence of functional TRs, the mice could express the UCP1 gene, and could modulate its expression in relation to acclimation temperature.
UCP1 protein levels (Table 1) were qualitatively parallel to UCP1 mRNA levels. The total amount of UCP1 in wild-type mice was low at thermoneutrality and was nearly 9-fold increased as an effect of cold acclimation. Again, even at thermoneutrality, the amount of UCP1 in the BAT of the TR-ablated mice was almost 3-fold that in the corresponding wild-type mice. Cold acclimation caused a further 3-fold increase in UCP1 amount to a relative level identical with that in wild-type mice (Table 1). Thus, the complete absence of TRs did not cause any obvious dysfunction in synthesis of UCP1 protein or in the final level of the protein after cold acclimation. An inability to recruit BAT and UCP1 itself was therefore not the reason for the lack of an augmented response to NE in the TR-ablated mice acclimated to cold.
UCP1 Is Fully Functional in the TR-Ablated Mice
Considering the absence of thermogenic response to NE, and the normal amount of UCP1 present, the question remained as to whether the mitochondria and UCP1 were functional (thermogenic) in the TR-ablated mice. To address this question, mitochondria were isolated from BAT of cold-acclimated wild-type and TR-ablated mice [at this temperature, recruitment is similar in the two groups (Table 1)]. In these isolated mitochondria, we examined membrane potential, oxygen consumption, and the coupling ability of GDP.
As expected (35), the membrane potential of isolated brown fat mitochondria of the wild-type mice was very low (Fig. 5A), and this was associated with a high level of oxygen consumption (Fig. 5B), i.e. the mitochondria were, as expected, uncoupled. They could, as expected, be recoupled by addition of GDP (Fig. 5): membrane potential increased and oxygen consumption decreased—all of this demonstrating the presence and regulation of UCP1. However, the behavior of isolated brown fat mitochondria from the TR-ablated mice was indistinguishable from that of the mitochondria isolated from the wild-type animals: these mitochondria were also uncoupled when isolated, and could be equally well recoupled by addition of GDP (Fig. 5). Thus, UCP1 was both adequately expressed and fully functional in the mitochondria of the TR-ablated mice. As there was lack of an augmented response to NE in the TR-ablated mice in the cold (Fig. 3, C and D), the defect must be located at the level of the cells or the tissue.
|
Thyroid hormone-dependent regulation of adrenergic signaling has been observed in BAT. Alterations in the expression pattern of different types of adrenergic receptors and in the density of these receptors have been reported, as well as alterations in the coupling of the receptors to cAMP generation (24, 45, 46, 47, 48, 49). To clarify whether any of these reported thyroid hormone-dependent events were affected by TR ablation, the effect of different stimuli on oxygen consumption of isolated mature brown fat cells was analyzed.
The response of the isolated wild-type mouse brown fat cells to NE stimulation (Fig. 6A) was in good agreement with previous reports (43, 44). Brown fat cells could be prepared from the TR-ablated mice, and the maximal response to NE was similar to that of the cells from the wild-type animals. The EC50 of the response was, however, 10-fold shifted to the right, indicating a much decreased sensitivity to adrenergic stimulation in these cells (Fig. 6A).
|
), were desensitized similarly to the response to NE. The response to the cAMP analogs 8-bromo-cAMP (not shown) and adenosine-3',5'-cyclic monophosphorothioate, Sp-isomer (Sp-cAMPS) (Fig. 6C), the effect of which is independent of receptors, G proteins or adenylyl cyclase activation, was also shifted, although less markedly. The reduced sensitivity to ß3-adrenergic stimulation was thus pleiotropic and occurred both at the level of the ß3-adrenergic receptors and their interaction with the adenylyl cyclase, and in steps downstream from cAMP generation in the signal transduction cascade toward thermogenesis. There is, however, reason to think that the desensitization observed here is not a primary effect of TR ablation (see below).
Induction of Thermogenesis in TR-Ablated mice
If desensitization of the brown fat cells to adrenergic stimulation was also manifest in situ, this desensitization could be the reason for the inability of the TR-ablated mice to respond to NE injection with an increased oxygen consumption. In that case, it should be possible to induce a thermogenic response in intact animals by injection of a much higher dose of NE. To test this, a very high dose of NE (10 mg/kg) was injected into wild-type and TR-ablated mice, and the rate of oxygen consumption was followed (Fig. 7A). As can be seen in the compilation in Fig. 7B, in the wild-type mice, the response to this dose of NE was of similar magnitude to that to 1 mg/kg. Notably, in the TR-ablated mice, the dose of 10 mg/kg NE indeed induced an increase in oxygen consumption (Fig. 7). This response was of similar magnitude to (Fig. 7B) and lasted equally as long (Fig. 7A) as that in the wild-type mice. Thus, desensitization of the brown fat cells was indeed the reason for the inability of TR-ablated mice to respond adequately to exogenous adrenergic stimulation.
|
) and in situ (Fig. 7)], we used the PGC-1 (peroxisome proliferator activated receptor 
coactivator-1) gene as an in situ reporter gene. The expression of PGC-1 in BAT is induced via ß3-adrenoceptors (50). We found that a short-time cold exposure (3 h) increased PGC-1 levels from 1.0 (normalized) to 1.4 in wild-type animals. In the TR-ablated mice, the level was enhanced in the animals even before cold exposure (i.e. similarly to other recruitment characters) to 4.2, but cold exposure could further increase the level to 6.5, i.e. a similar 1.5-fold increase due to cold exposure as seen in the wild-type mice (means from three to four animals; similar data were obtained after 2 h cold exposure). Thus, the sympathetic nervous system was efficient in stimulating the tissue, despite the desensitization observed, and we can therefore assume that the nonshivering thermogenic capacity of the animals is being used when they are in the cold.
Adrenergic Desensitization in Brown Fat Cells of TR-Ablated Mice Is Not Innate
To clarify whether the reduced sensitivity to adrenergic stimulation in the freshly isolated brown fat cells from the TR-ablated mice was innate, or whether it was a consequence of the environment of the cells within the animal before isolation, brown preadipocytes from wild-type and TR-ablated mice were isolated and grown in culture under identical conditions. Although the cell culture system may not fully duplicate cell development in vivo, it makes it possibly to fully control external factors. Preadipocytes from wild-type animals proliferated (Day 3 in Fig. 8A) and then differentiated into mature adipocytes (51), as manifested by cell rounding and lipid accumulation in the cells (Day 6 in Fig. 8A). TR-ablated brown preadipocytes grew and accumulated lipids equally well, and the visual appearance of parallel cultures was practically identical (Fig. 8A). Thus, the lower in situ lipid content of brown fat cells in TR-ablated mice (Fig. 4) was not due to an innate defect in the ability of brown fat cells of the TR-ablated mice to accumulate lipids but was probably due to a different neuroendocrine environment of the cells in wild-type and TR-ablated mice.
|
TRs Are Not Essential for NE-Stimulated UCP1 Gene Expression
UCP1 gene expression is predominantly mediated adrenergically, via the ß3-receptor/cAMP cascade (51), similarly to thermogenesis. In the cultured brown adipocytes from both wild-type and TR-ablated mice, NE could induce UCP1 gene expression (Fig. 8B) and, in agreement with the lipolysis results, no difference in NE sensitivity could be observed (Fig. 8B). This confirms that adrenergic desensitization is not an innate characteristic of the brown fat cells lacking TRs, but is secondary to other alterations occurring within the intact organism.
Two potential thyroid hormone-responsive elements (TREs) have been identified by footprint analysis in the promoter region of the rat ucp1 gene (52, 53). Analysis of mutations in these sites has demonstrated that both of these TREs contribute to thyroid hormone responsiveness (53). We therefore examined whether regulation of expression of the ucp1 gene by T3 was affected by the TR ablation.
To examine the direct effect of T3 on UCP1 gene expression, we cultured cells without serum—but in the presence or absence of added T3—during the last 2 d before the experiment. In cells of wild-type mice, there was only a barely detectable effect of the presence of T3 (Fig. 8C) on UCP1 gene expression. NE could induce UCP1 gene expression even in cells cultured without T3. However, there was a dramatic synergistic effect of NE and T3 together (Fig. 8C). This was in agreement with the data of Bianco et al. (40).
In the cultured TR-ablated cells, T3 was fully without ability to induce UCP1 gene expression. In contrast, NE by itself could induce a very appreciable increase in UCP1 mRNA levels (Fig. 8C) and very clearly, the presence of T3 was here of no influence on NE-induced ucp1 expression (Fig. 8C): there was no synergistic effect. Thus, in the absence of TRs, thyroid hormone no longer functioned as a regulator of ucp1 expression. This pattern of regulation supports the tenet that unliganded receptors function in vivo as repressors of UCP1 expression (54). Furthermore, it could be concluded that the absence of repression by unliganded TRs is not in itself sufficient to induce UCP1 gene expression, and the effect of NE does not in itself require TRs; rather, the absence of repression allows the NE stimulation to become manifest.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1(-/-)ß(-/-), here referred to as TR-ablated mice) had a slightly decreased body temperature and a much decreased basal metabolic rate. Ablation of all hormone-binding TRs was both necessary and sufficient to prevent normal cold tolerance. The reason for the cold sensitivity of the TR-ablated mice was an inadequate heat production at low temperatures. A standard NE test showed that adrenergically induced thermogenesis could not be activated normally in the TR-ablated mice. This lack of response was not due to inadequate recruitment of BAT (the site of adrenergic thermogenesis), nor to the absence, decreased recruitment or dysfunction of the UCP1 itself. However, brown fat cells both in situ, and when freshly isolated, were 10-fold desensitized, explaining the apparent lack of response to standard exogenous adrenergic stimuli; cell culture experiments demonstrated that this desensitization was not an innate effect of TR ablation (but probably due to differences in the in situ environment between brown fat cells in wild-type and TR-ablated animals). Thus, the cold intolerance of the TR-ablated mice was probably not due to a lack of capacity for sympathetically induced nonshivering thermogenesis but rather (see below) to a decreased capacity for shivering. Although some of these effects may be indirect (e.g. secondary to lack of TR during development), we think that they can in general be interpreted as being direct consequences of the absence of TRs.
TRs and Body Temperature
From the present and previous publications (10, 18, 20, 22), it is clear that the different TRs have distinctive roles in the central regulation of body temperature. The absence of TRß has in itself no effect. The absence of TR1 leads to a slightly (
1 C) lowered body temperature observable only during the resting phase. The ablation of all mRNA isoforms derived from the two TR genes [TR(o/o)ß(-/-)], i.e. including the non-ligand-binding, non-DNA-binding TR1 and TR2, leads to a decrease in body temperature of 4 C (10). In contrast, ablation of all full-length mRNAs [TR(-/-)ß(-/-)] (18) or only hormone-binding TRs [TR1(-/-)ß(-/-)] (22) (and Figs. 1 and 2) leads only to a slight decrease in body temperature ( 1 C), but this decrease is not dependent upon time of day (Fig. 1). Thus, the non-ligand-binding TR2 gene product does not seem to influence body temperature control.
We demonstrate here that the modestly lowered body temperature in TR-ablated mice is regulated and defended. There was a clear daily rhythm in the body temperature of the TR-ablated mice (Fig. 1), and the body temperature was regulated independently of environmental temperature between 18 and 30 C (Table 1). At 30 C, energy requirements for maintaining constant body temperature are at a minimum, and at this ambient temperature a higher body temperature could easily have been metabolically achieved and defended. The lower body temperature in the TR-ablated mice is thus not a hypothermia. Rather, because the lowered body temperature of TR-ablated mice is defended, it represents a state of anapyrexia (reversed fever) (55): it is the body temperature set-point that is decreased (cf. Ref. 56).
A Significant Fraction of Basal Metabolic Rate Requires the Presence of TRs
When metabolism was analyzed in the thermoneutral zone, a very marked dependence on TRs for control of basal metabolic rate was evident: their presence increased metabolic rate by almost 50% (Fig. 2). Despite the severely depressed basal metabolism, the TR-ablated mice did not become obese, even when kept at thermoneutrality (Table 1). This suggests that the mechanisms controlling the balance between energy expenditure and food consumption are not affected by lack of TRs and demonstrates that low energy expenditure per se does not automatically lead to obesity.
The present investigation indicates that TRs positively mediate the effects of thyroid hormone on basal metabolic rate, presumably through altered expression of specific enzymes. The target for thyroid hormones must in this respect be peripheral energy-dissipating processes, such as stimulation of futile cycles of some kind, e.g. an increased Na+ permeability of the plasma membrane, an increased Ca2+ permeability or perhaps an increased proton permeability over the mitochondrial membrane (5). An investigation of the proton permeability properties of mitochondria from liver and skeletal muscle of the TR-ablated mice may be helpful in advancing our understanding in this respect.
Unperturbed Recruitment of BAT in Mice Lacking TRs
The results presented here concerning BAT recruitment and function were rather unexpected, in that they seemed at first sight to indicate that BAT is not a positive target organ for thyroid hormones, despite numerous papers concluding the contrary (for review see Ref. 4). BAT in the TR-ablated mice even showed certain secondary signs of a higher degree of recruitment than in wild-type animals, especially at thermoneutrality (Fig. 4; Table 1). This is principally similar to what has been observed in hypothyroid animals (57, 58) and may be due to an increased sympathetic activity, similar to that observed as an effect of classical hypothyroidism (59, 60). The final level of recruitment obtained after cold acclimation was not different between wild-type and TR-ablated mice. This is physiologically unexpected because a higher demand for thermogenesis would be expected at all ambient temperatures below the lower critical temperature (Fig. 2).
To reconcile the competent recruitment and functionality of BAT in TR-ablated mice with the known positive interaction between thyroid hormones and adrenergic hormones in this tissue, transcriptional repression by unliganded TRs may be invoked, as discussed below.
Adrenergic Desensitization in Brown Fat Cells from TR-Ablated Animals Is Not Innate
Mature brown fat cells isolated from TR-ablated mice were thermogenically competent but were 10-fold desensitized to NE (Fig. 6A). This could at first sight be interpreted as a bona fide consequence of the absence of TRs because signaling through the cAMP pathway is reduced in tissues obtained from hypothyroid rats (24, 41, 45, 46, 47, 48, 49). However, the sensitivity to NE in TR-ablated preadipocytes differentiated in culture was not decreased (Fig. 8B). Thus, the desensitization seems rather to be secondary to alterations in the in situ environment of the cells before isolation. A plausible explanation may therefore be that the cells from the TR-ablated mice have been under enhanced sympathetic stimulation, as compared with the situation in wild-type animals, and have therefore become desensitized. Because sympathetic activity is increased in BAT in classical hypothyroidism (59, 60), it appears probable that the desensitization is physiologically induced and is due to a higher chronic sympathetic tone in TR-ablated than in wild-type animals, even at thermoneutrality. A very similar desensitization phenomenon is observed in brown fat cells from cold-acclimated animals (61, 62), a condition where the brown fat cells are (also) chronically sympathetically stimulated. It is counterintuitive that the brown adipocytes should become desensitized when their thermogenic capacity is as most needed (i.e. in the cold), but empirically, desensitization is often the way cells respond to prolonged receptor stimulation. Apparently, in the cold, the animals are competent to further increase their sympathetic stimulation of the tissue to overcome the cellular desensitization. Similarly, according to the PGC-1 data presented here, the tissue can be sufficiently stimulated sympathetically in the TR-ablated mice to overcome the desensitization.
TRs and UCP1 Recruitment
The promoter of the UCP1 gene contains cAMP-responsive elements (63), and the gene is thought to be predominantly under the control of adrenergic receptors. However, in addition to the cAMP-responsive elements, two TREs have been identified in the UCP1 gene, both located within a critical upstream enhancer element at -2.3 kb. Transactivation experiments have shown that both TREs contribute to T3 responsiveness, and that one of them, dnTRE (-2348 to -2334), is necessary for potentiation of the cAMP effect by T3 (52, 53, 54, 64). Our results on UCP1 expression in situ, in the BAT of the TR-ablated mice, indicated that TRs were not essential for either the basal (at thermoneutrality), or for acclimation-induced UCP1 expression (Table 1) in vivo. Similarly, in cultured cells from TR-ablated mice (Fig. 8), NE could efficiently induce UCP1 expression, demonstrating that adrenergic stimulation was competent even in the absence of nuclear TRs.
Notably, in cells lacking nuclear TRs, an equally high NE responsiveness was manifest whether or not T3 was present. In contrast, in wild-type cells grown in the absence of T3-containing serum, T3 addition was a prerequisite for maximal NE-stimulated UCP1 gene expression (Fig. 8). Taken together, this demonstrates that, in wild-type cells, unliganded receptors repress both basal and NE-induced UCP1 gene expression and that thyroid hormones are necessary for the release of the repressor effect, thus expanding the conclusions from studies on regulation of UCP1 expression in transfected HIB-1B cells (54).
No Observable Nuclear TR-Independent Effects of Thyroid Hormones
The TR-ablated mice examined here are hyperthyroidemic, i.e. they have highly elevated blood levels of T4 and T3 (19) [and probably diiodothyronines (T2)]. The data presented here can therefore be used to analyze whether effects of hyperthyroidemia can be mediated in a alternative way, not via nuclear TRs. The question is thus to which extent the known effects of hyperthyroidemia in wild-type rodents (i.e. classical hyperthyroidism) can become manifest in animals lacking nuclear TRs.
Concerning body temperature, classical hyperthyroidism leads to hyperpyrexia, i.e. a defended increase in body temperature of 1–2 C (65, 66). Evidently, in the absence of TR, hyperthyroidemia does not have this effect: the body temperature is decreased (Figs. 1 and 2 and Table 1).
Classical hyperthyroidism also leads to a marked increase in basal metabolic rate in mice (66). However, despite the high hyperthyroidemia in the TR-ablated mice, we observed a very marked metabolic depression in these mice (Fig. 2).
Concerning BAT recruitment, the literature is not fully concordant but many observations indicate that BAT is atrophied (less active) in classical hyperthyroid wild-type animals (as compared with euthyroid animals) (67, 68), probably due to indirect effects of enhanced thyroid thermogenesis (less extra heat from BAT is then needed) (69). This atrophy is not observed in the TR-ablated mice and the hyperthyroidemia thus does not induced a hyperthyroid phenotype in this respect in TR-ablated mice. Instead of an atrophy, rather a tendency to a recruitment of BAT is observed in the present situation of hyperthyroidemia in the absence of TRs. This recruitment is probably a secondary effect, as discussed above, and due to enhanced sympathetic activity in the animals. Indeed, the cell culture experiments (Fig. 8) indicated that no positive effect of added thyroid hormone on UCP1 gene expression could be observed in brown adipocytes that lack TRs.
Thus, with respect to the parameters investigated here, we find no evidence that any effects of hyperthyroidemia can be mediated via other pathways than the established nuclear TRs. Although the present system does not allow us to directly investigate whether the same conclusion can be drawn concerning the hypo-to-euthyroidemic transition, we think this would be a reasonable extension of the data presented.
Unliganded Nuclear TRs: Repressors or Positive Modulators
Based on the data presented here, it is also possible to differentiate between two modes of action of the nuclear TRs. If the unliganded receptors function as repressors, the characteristics observed in the TR-ablated mice would be similar to those observed in euthyroid animals (as the repression would be eliminated in the TR-ablated mice); if the receptors function as positive modulators, the parameters would be similar to those observed in classical hypothyroid animals.
Classical hypothyroidism is associated with a lowered body temperature (anapyrexia) (70), i.e. the same effect as TR ablation (Figs. 1 and 2 and Table 1). Hypothyroidism is also associated with a decreased basal metabolic rate (e.g. Ref. 71), also as seen in the TR-ablated mice (Fig. 2 and Table 1). However, whereas BAT recruitment, including UCP1 content, is similar in TR-ablated and wild-type mice (Table 1), UCP1 content is reduced in classical hypothyroidism and cannot be recruited (72). Thus, for body temperature control and control of basal metabolic rate, the nuclear TRs function as positive modulators, yielding characteristics similar to classical hypothyroidism in the TR-ablated mice, whereas regarding UCP1 gene expression, they function as repressors, yielding euthyroid characteristics in the TR-ablated mice.
Why Are TR-Ablated Mice Cold Sensitive?
In view of the above results, an absence of nonshivering thermogenesis is not the most likely explanation for the dramatic acute cold sensitivity of the TR-ablated mice exposed to 4 C (Fig. 1). However, even in wild-type animals, during acute exposure to severe cold (before acclimation has taken place), the heat-producing capacity of BAT present in animals adapted to normal room temperature is in itself insufficient to compensate for heat loss at 4 C, and the deficit is initially produced through shivering thermogenesis (25). In the TR-ablated mice, nonshivering thermogenesis would, therefore, also have to be supplemented by shivering thermogenesis during acute exposure to cold. However, shivering demands high muscular performance. Notably, in the TR-ablated mice, the soleus muscle is markedly slower than in controls (73) [similarly, hypothyroidism is associated with decreased muscular performance (74, 75, 76, 77, 78)]. Thus, the cold intolerance of TR-ablated mice may rather be caused by insufficient muscular shivering capacity than by inadequate development of nonshivering thermogenesis.
The Possible Significance of the Non-Hormone-Binding TR Gene Products
The present study has examined the metabolic significance of the hormone-binding TRs (1, ß1, -2, -3). However, there could also be metabolic effects of the non-hormone-binding gene products that can be generated from the -gene by alternative promoter usage, i.e. TR2 and TR1 and TR2 [of the latter two, only TR2 has been confirmed as an endogenously occurring protein (13)]. In the TR-ablated mice used here, TR2 and TR2 mRNAs will still be generated. The function of TR2, the mRNA of which is high in BAT (79, 80), is unknown in this context, but as it inhibits thyroid hormone action (81), it may have important regulatory roles, both in BAT specifically and in thermoregulation and metabolism in general. A delineation of the function of the 2-protein (and the gene products) may evolve by contrasting results obtained through further study of the mutant mice lacking the 2-products and the 2-containing mutant mice examined here.
Conclusions: Modes of Action of Thyroid Hormones in Metabolism
From the perspective of the present study, it is clear that distinct molecular modes of action of thyroid hormones can be discussed.
First, thyroid hormones may be discussed to act through pathways alternative to those mediated by the nuclear TRs studied here. However, there is no evidence in our experiments that, in the absence of nuclear TRs, the hyperthyroidemia can induce the metabolic characteristics of hyperthyroidism.
When acting through the nuclear receptors, thyroid hormones may act as derepressors. In systems under this type of control, unliganded TRs act as transcriptional repressors recruiting corepressors with histone deacetylase activity, and the effect of thyroid hormone is to relieve this inhibition and allow recruitment of coactivators with histone acetyl transferase activity (12, 21, 82, 83). The absence of TRs thus results in the absence of repression. This mode of TR action explains why effects of TR ablation are generally relatively mild, as compared with total thyroid hormone deficiency (11, 21). In the present investigation, the repressor action is evident for e.g. UCP1 gene expression and possibly also for BAT recruitment in general.
When occupied with thyroid hormones, TRs can act positively through positive TREs, and the thyroid hormone/receptor complex thus has an essential role in certain functions. In this type of regulation through TRs, a repression/impairment is seen in the absence of either thyroid hormone or TRs. The regulation of basal metabolism and, apparently independently, of body temperature set point, are clearly such functions, in which the thyroid hormone/TR complex is necessary to obtain eumetabolism and euthermia.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1(-/-)ß(-/-) mice (referred to as TR-ablated mice in the text), TR1(-/-) (20) and TRß(-/-) (84) strains were crossed (for details, see Ref. 19). Briefly, the wild-type animals were derived from TR1(+/-)ß(+/-) crosses and were kept as a substrain. Because deficiency for all TRs is incompatible with fertility (19), TR1(-/-)TRß(+/-) females were crossed with male TR1(-/-)TRß(-/-) mice to yield TR-ablated mice. To minimize genetic drift between the substrains, the breeding was done with mice that originated from two to three generations after the TR1(+/-)ß(+/-) intercrosses, and were used until their fertility declined. To avoid further genetic drift, the two substrains were intercrossed after about 2 yr and new substrains derived. The TR1(-/-) and TRß(-/-) mice have distinct genetic backgrounds; the former represents a cross between 129/Ola and BALB/c, the latter 129/sv and C57/BL-6. Although both these strains have been successfully backcrossed for nine to 11 generations to C57/BL-6J, intercrossing of these homogeneous strains to yield TR-ablated mice resulted in very low progeny numbers (see Ref. 85 for details). This necessitated the use of corresponding strains of TR1(-/-)ß(-/-) and wild-type that all have the mixed genetic background of the 129/Ola, 129/Sv, BALB/c, and C57/BL-6 strains. The genotype of all TR1(-/-)ß(-/-) animals was confirmed before the experiments. All experiments were approved by the North Stockholm Animal Ethics Committee. Age-matched animals were kept in climate-controlled rooms on a 12-h light, 12-h dark cycle, were fed standard diet (BeeKay Feeds; B&K Universal, Stockholm, Sweden) ad libitum and had free access to water. All studies on animals acclimated to different temperatures were performed after at least 1 month of acclimation.
Whole body oxygen consumption measurements were performed in an open circuit system, principally as in Dicker et al. (28), with a flow rate of 0.4 liters/min. Oxygen consumption and ambient temperature data were collected every 6 sec via MacLab/2e (AD Instruments Pty. Ltd., Castle Hill, Australia).
BAT Dissection and Analysis
After the indicated acclimation periods, the mice were killed by CO2 anesthesia followed by cervical dislocation. The interscapular brown fat depots were quantitatively dissected out, frozen in liquid nitrogen and kept at -70 C until analysis. Both pads were weighed. The left pad was used for RNA isolation and analysis, the right for DNA and protein analysis.
Total RNA was isolated from the left pad using Ultraspec (Biotecx, Houston, TX), according to the manufacturer’s instructions, and Northern blotting was performed as previously described (86). Probes for UCP1 (87) and COX IV (88) were those earlier described. The clones for UCP2 and UCP3 cDNA were obtained from Genome Systems Inc. (St. Louis, MO) as expressed sequence tag clones nos. 1040737 and 482847, respectively, and the identity of these clones was confirmed by sequencing.
The right pad of the interscapular BAT depot was homogenized in 0.5 ml saline with protease inhibitors [100 µg/ml Pefabloc (Roche Molecular Biochemicals, Indianapolis, IN); 10 µg/ml aprotinin (Sigma, St. Louis, MO) and 10 µg/ml pepstatin (Sigma)], sonicated for 15 sec at 40 W and centrifuged for 5 min at 350 x g. This partially fat-depleted material was collected and used for further analysis. DNA was measured fluorometrically with H33258 (Roche), in principle as described (89). Protein concentration was determined by the method of Bradford (90) with BSA as standard. Total DNA and protein contents of the tissue were recalculated from the total weight of the pads. Western blotting was performed principally according to (91) but with horseradish peroxidase-conjugated secondary antibodies (Sigma) and the ECL detection system (Amersham Biosciences, Buckinghamshire, UK). The primary antibodies used were rabbit-antirat UCP1 antibodies (92) and mouse-antirat COX IV antibodies (Molecular Probes, Eugene, OR).
BAT mitochondria were isolated, and mitochondrial respiration and membrane potential were measured principally as in (35). Mature brown fat cells were isolated by collagenase digestion and oxygen consumption was measured principally as in Refs. 42 and 43 .
Cell Culture
The cell isolation and culture procedures were performed principally as described previously (89). The isolation procedure included a centrifugation step that differentiates between lipid-loaded cells (i.e. mature brown adipocytes that float) and lipid-depleted cells (i.e. brown preadipocytes that sediments); during early cell culture, no lipid droplets accumulate (cf. Fig. 8, top) and no mature brown adipocytes are therefore present. The culture medium contained 10% newborn calf serum; this serum contained 3 nM total T3 (free 11 pM) and 100 nM total T4 (free 22 pM). For analysis of lipolytic capacity, culture medium was collected and used for the analysis of glycerol content with the Glycerol Analysis kit (Roche).
|
ACKNOWLEDGMENTS |
|---|
ngberg for technical assistance. |
FOOTNOTES |
|---|
Present address for V.G.: Pharmacological Research 3, Novo Nordisk, DK-2769 Måløv, Denmark.
Present address for H.G.: Biovitrum AB, SE-112 76 Stockholm, Sweden.
Present address for A.M.: MediHerb Research Laboratories, Chemistry Department, University of Queensland, Brisbane, Queensland 4072, Australia.
Abbreviations: BAT, Brown adipose tissue; BMR, basal metabolic rate; COX IV, subunit IV of cytochrome c-oxidase; NE, norepinephrine; PGC1, peroxisome proliferator activated receptor coactivator-1; TR, thyroid hormone receptor; TRE, thyroid hormone-responsive element; UCP, uncoupling protein.
Received for publication July 8, 2003. Accepted for publication November 10, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
) and in TR-ablated (
) mice that had been living at 20–22 C. At each ambient temperature, measurements were performed for 2.5 h (at 34 and 36 C, the period in the chamber was decreased to 1.5 h, to reduce the heat stress). Metabolic rates at the different ambient temperatures were defined as the lowest, stable metabolic rate observed for at least 4 min. Determinations in wild-type and TR-ablated mice were performed in alternating order. For each animal, at least 3 d was allowed between measurements at different temperatures. To compensate for the differences in the body weight of wild-type and TR-ablated mice (Table 1
, A) and oxygen consumption (B) of mitochondria isolated from brown fat of wild-type (
