Uncoupling Protein-3 Is a Molecular Determinant for the Regulation of Resting Metabolic Rate by Thyroid Hormone

Abstract

Thyroid hormones increase energy expenditure, partly by reducing metabolic efficiency. The control of specific genes at the transcriptional level is thought to be the major molecular mechanism. However, both the number and the identity of the thyroid hormone-controlled genes remain unknown, as do their relative contributions. Uncoupling protein-3, a recently identified member of the mitochondrial transporter superfamily and one that is predominantly expressed in skeletal muscle, has the potential to be a molecular determinant for thyroid thermogenesis. However, changes in mitochondrial proton conductance and resting metabolic rate after physiologically mediated changes in uncoupling protein-3 levels have not been described. Here, in a study on hypothyroid rats given a single injection of T₃, we describe a strict correlation in terms of time course between the induced increase in uncoupling protein-3 expression (at mRNA and protein levels) and decrease in mitochondrial respiratory efficiency, on the one hand, and the increase in resting metabolic rate, on the other. First, we describe our finding that uncoupling protein-3 is present and regulated by T₃ only in metabolically relevant tissues (such as skeletal muscle and heart). Second, we follow the time course (at 0, 6, 12, 24, 48, 65, 96, and 144 h) of both uncoupling protein-3 mRNA levels and mitochondrial uncoupling protein-3 density in gastrocnemius muscle and heart. In both tissues, the maximal (12-fold) increase in uncoupling protein-3 density was reached at 65 h. The resting metabolic rate[ lO₂(kg^0.75)⁻¹h⁻¹] showed the same time course, and at 65 h the increase vs. time zero was 45% (1.316 ± 0.026 vs. 0.940 ± 0.007; P < 0.001). At the same time point, gastrocnemius muscle mitochondria showed a significantly higher nonphosphorylating respiration rate (nanoatoms of oxygen per min/mg protein; increase vs. time zero, 40%; 118 ± 4 vs. 85 ± 9; P < 0.05), whereas the membrane potential decreased by 8% (168 ± 2 vs. 182 ± 4; P < 0.05). These data are diagnostic of mitochondrial uncoupling. The results reported here provide the first direct in vivo evidence that uncoupling protein-3 has the potential to act as a molecular determinant in the regulation of resting metabolic rate by T₃.

THYROID HORMONE (T₃) is a major regulator of energy expenditure in adult mammals (1, 2). Although the fact that thyroid hormones stimulate the metabolic rate and decrease metabolic efficiency has been known for many years, very little is known about the molecular mechanism via which these effects are elicited (3). It is presumed that T₃ regulates energy expenditure and efficiency by controlling, via its interaction with the different isoforms of nuclear receptors, the rate of transcription of genes encoding a subset of key proteins involved in energy metabolism within the cell (4). However, the identity and the relative contribution of these genes remain unknown. About 50 yr ago it was suggested that thyroid hormones increase the metabolic rate by uncoupling electron transport from ATP synthesis (5, 6). This hypothesis was subsequently discarded and was not thought to be physiologically relevant because 1) large concentrations of T₃ and T₄ were required; and 2) the effects seen with thyroid hormone in vitro were not observed in vivo. More recently, however, the uncoupling hypothesis has gained new support from the data of Rolfe and co-workers (7), who have shown that the leakage of protons back into mitochondria can account for a substantial portion of the energy requirement of the cell and from the discovery that uncoupling proteins (UCPs) are present not only in brown adipose tissue (BAT) but in almost all tissues (8–12).

Because of their chromosomal location within a region of genetic linkage to obesity and diabetes as well as their pattern of expression, the recent identification of UCP2 and UCP3, both members of the mitochondrial transporter superfamily, has greatly stimulated research on the mechanism underlying energy expenditure and its control (9, 13, 14). Both proteins are homologous to the classic uncoupling protein, UCP1, which is located in BAT (8) and has been known and studied for over 20 yr. UCP2 is ubiquitously expressed (9, 10). UCP3, on the other hand, is expressed preferentially in skeletal muscle and BAT (11, 12). The biochemical activities and physiological role of the new UCPs are not well known. Because of their putative uncoupling properties, UCPs are good candidates for the role of molecular determinants in the control of energy metabolism by T₃. In fact, their homology to UCP1 and the demonstration that they lower mitochondrial membrane potential when transfected into yeast and mammalian cells are in favor of an uncoupling activity of these proteins (9–17). Recent studies show that mice overexpressing UCP3 are hyperphagic and lean with a decreased mitochondrial efficiency (18), whereas mice lacking UCP3 show a reduced mitochondrial proton conductance (19, 20). Some studies indicate that T₃ up-regulates UCP3 mRNA levels in skeletal muscle with an increased proton leak (15, 21, 22). UCP2 expression, on the other hand, is clearly up-regulated by T₃ in heart, but only weakly, if at all, in other tissues (15, 23). As skeletal muscle represents the majority of total metabolically active body mass and is endowed with significant mitochondrial capacity (24), the regulation of resting metabolic rate (RMR) by T₃ via UCP3 would be of great physiological relevance. However, direct evidence for a role for UCP3 in the regulation of resting energy metabolism that is exerted by thyroid hormone is lacking. In this study we sought to clarify this issue, first by studying the presence and T₃-mediated regulation of UCP3 mRNA and protein in tissues classically known to be either metabolically responsive (e.g. heart and muscle) or unresponsive (e.g. spleen) to T₃, and second by injecting a single dose of T₃ into hypothyroid rats, after which we measured the time course of changes in UCP3 expression (at mRNA and protein levels) and mitochondrial respiratory efficiency as well as in resting metabolic rate.

Materials and Methods

Materials

Thyroid hormone (T₃), 6-n-propyl-2-thiouracil (PTU), and iopanoic acid (IOP) were purchased from Sigma-Aldrich Corp. (St. Louis, MO).[α -³²P]Deoxy-ATP was purchased from Amersham Pharmacia Biotech (Milan, Italy), and a polyclonal antibody raised against the C-terminal region of the human UCP3 protein (AB3046) was purchased from Chemicon International (Temecula, CA).

Animals

Male Wistar rats (220–250 g) were kept, one per cage, in a temperature-controlled room at 28 C under a 12-h light, 12-h dark cycle. A commercial mash and water were available ad libitum. Hypothyroidism was induced by simultaneous injection of PTU and IOP as previously described (34) (P+I rats). Chronic hyperthyroidism was induced by giving seven daily ip injections of 15 μg T₃/100 g BW to hypothyroid rats; control rats (euthyroid and hypothyroid) received saline injections. The dose of T₃ and the treatment duration were chosen so as to provide us with hyperthyroid, but not thyreotoxic, animals (26, 27). At the end of the treatment (24 h after the last dose of T₃) rats were anesthetized by an ip injection of chloral hydrate (40 mg/100 g BW) and killed by decapitation. For time-course experiments, hypothyroid rats were injected with 25 μg T₃/100 g BW. This dose, given acutely, is the smallest single dose capable of inducing a significant change in RMR in rats (28). Animals were killed 6, 12, 24, 48, 65, 96, or 144 h after T₃ injection. Tissues were isolated and immediately either 1) processed for the preparation of the mitochondria or 2) frozen in liquid nitrogen and stored at −80 C for later processing. All experiments were performed in accordance with local and national guidelines regarding animal experiments.

Measurement of RMR

To determine the time course of oxygen consumption, sequential measurements were taken 1 h before and at various time intervals after T₃ injection. The RMR was measured using open circuit indirect calorimetry. For each measurement, one rat was placed in a respiration chamber (∼32 × 20 × 19 cm) with airflow measured using an O₂-ECO mass flow controller (Columbus Instruments International Corp., Columbus, OH). Measurements for RMR calculations were taken at 28 C between 1100–1600 h when the energy expenditure was at a low level with respect to any other period of the day. Details of this set-up and of the way of measuring have been published previously (25).

Preparation of mitochondria

Mitochondria from liver, heart, spleen, lung, and skeletal muscles (tibialis anterior and gastrocnemius) were isolated after homogenization in an isolation medium consisting of 220 mm mannitol, 70 mm sucrose, 20 mm Tris-HCl, 1 mm EDTA, 5 mm EGTA, and 5 mm MgCl₂, pH 7.4 (all from Sigma-Aldrich Corp.). After brief homogenization, samples were centrifuged at 700 × g, and supernatants were collected and transferred into new tubes with subsequent centrifugation at 10,000 × g. The final mitochondrial pellet was resuspended in a minimal volume of isolation medium and kept on ice. Mitochondria prepared for Western blot analysis were kept in the same medium supplemented with the following protease inhibitors: 1 mm benzamidine, 4 μg/ml aprotinin, 1 μg/ml pepstatin, 2 μg/ml leupeptin, 5 μg/ml betastatin, 50 μg/ml N-tosyl-l-phenylalanine-chloromethyl ketone, and 0.1 mm phenylmethylsulfonylfluoride (all from Sigma-Aldrich Corp.).

Northern blot analysis

Northern blot analysis was performed as described previously (21). To detect UCP3 mRNA, we used a 312-bp probe derived from rat UCP3 cDNA by PCR amplification using UCP3-specific oligonucleotide primers (see next section). A 764-bp β-actin PCR fragment generated using the oligonucleotide primers described in the next section, was used as the internal standard. The PCR fragments were labeled with [α-³²P]deoxy (d)-ATP by random priming. UCP3 mRNA levels were first determined and quantified in separate preparations from three rats; then the three samples were pooled. Data from the pooled samples are presented in the figures.

RT-PCR assays

One microgram of total RNA was reverse transcribed using 1 pmol oligo(deoxythymidine) primers (15 nucleotides; Sigma Genosys, Cambridge, UK), 2.0 U Superscript reverse transcriptase, 0.5 U ribonuclease inhibitor, and 1 mm dNTPs in reverse transcriptase buffer (all from HT Biotechnology Ltd., Cambridge, UK). The total volume was adjusted to 20 μl with distilled H₂O. The reaction was carried out for 1 h at 40 C. One quarter of the RT reaction mixture was used directly for the PCR reaction in a total volume of 20 μl, containing 0.2 U SuperTaq polymerase, 0.2 mm dNTPs, SuperTaq PCR buffer (all from HT Biotechnology Ltd.), and 300 nm of the relevant oligonucleotide primers (Sigma Genosys). These primers had the following sequences: UCP3 sense, 5′-ATGGATGCCTACAGAACCAT-3′; and UCP3 antisense, 5′-CTGGGCCACCATCCTCAGCA-3′ (cDNA nucleotide position, 545–856; GenBank accession no. U92069). They generated a fragment of 312 bp. As an internal control, the same cDNAs were amplified using β-actin oligonucleotide primers with the following sequences: β-actin sense, 5′-TTGTAACCAACTGGGACGATATGG-3′; and β-actin antisense, 5′-GATCTTGATCTTCATGGTGCTAGG-3′ (cDNA nucleotide position, 1552–2991; GenBank accession no. J00691), generating a fragment of 764 bp. Parallel amplifications (20, 25, and 30 cycles) of the same cDNA were used to determine the optimum number of cycles. After 30 cycles, a readily detectable signal within the linear range was observed. For the actual analysis, samples were heated for 5 min at 94 C, then 30 cycles were carried out, each consisting of 1 min at 94 C, 1.5 min at 50 C, and 1.5 min at 72 C. This was followed by a final 10-min extension at 72 C. One half (UCP3) or one quarter (β-actin) of the PCR reaction products were separated on a 2% agarose gel containing ethidium bromide, and the products were readily visualized.

Western immunoblot analysis

Analyses were performed using mitochondrial protein. Mitochondrial lysate was prepared by resuspending the mitochondria in SDS loading buffer, as described by Laemmli (29), followed by heating for 3 min at 95°C. Mitochondrial lysates containing 30μ g protein were loaded in each lane and were electrophoresed on a 13% SDS-PAGE gel. A polyclonal antibody against UCP3 (see Materials and Methods) and an antirabbit antibody were used as primary and secondary antibodies, respectively, in a chemiluminescence protein-detection method (NEN Life Science Products, Boston, MA). The protein concentration was determined by the method of Hartree (30). UCP3 protein levels were first determined and quantified in separate preparations from three rats, then the three samples were pooled. Data from the pooled samples are presented in the figures.

Measurement of membrane potential (Δψ) and respiration rate

Throughout we used freshly isolated mitochondria from gastrocnemius muscles obtained from rats treated as described above. The value of Δψ was determined from distribution of the lipophilic cation triphenylmethylphosphonium (Ph₃MeP⁺), which was measured using a Ph₃MeP⁺-sensitive electrode. A Ph₃MeP⁺-binding correction of 0.4 was applied, and Δψ was measured in the presence of nigericine so that the whole proton-motive force could be expressed asΔψ . Nonphosphorylating mitochondrial respiration was measured in the presence of oligomycin using a Clarke-type oxygen electrode as described previously (21). Δψ and respiration rates were measured under conditions in which a putative variation of adenine nucleotide translocase (ANT) densities and FFA levels was excluded by supplementing the incubation medium with 0.5 mm oleate and 15 μg/ml carboxyatractylate.

Statistical analysis

Data are expressed as the mean ± sd, and differences between means were assessed using paired t test.

Results

UCP3 mRNA and mitochondrial protein measurement in metabolically responsive or unresponsive tissues

As the first step in determining whether UCP3 might be a molecular determinant for the calorigenic effect of T₃, we searched for evidence of the presence of UCP3 mRNA and mitochondrial protein in tissues that are well known to be metabolically either responsive (skeletal muscle, heart, liver) or unresponsive (spleen, lung) to T₃. First, confirming data published previously, including our own (15, 21), Northern blot analysis (not shown) revealed that UCP3 mRNA was clearly detectable in skeletal muscle (gastrocnemius and tibialis anterior), but was undetectable in heart, liver, spleen, and lung. T₃ clearly up-regulated UCP3 mRNA levels by about 25-fold with respect to the hypothyroid state in skeletal muscles (both gastrocnemius and tibialis anterior). In heart, UCP3 mRNA as well as its up-regulation by T₃ of about 20-fold with respect to the hypothyroid state was detectable only when RT-PCR analysis was employed (not shown). At the protein level, UCP3 was abundantly expressed in mitochondria from skeletal muscles and, to a lesser extent, in mitochondria from heart (Fig. 1A), and it was clearly increased by T₃ in these tissues (∼10-fold in both gastrocnemius and tibialis anterior and ∼8-fold in heart, both with respect to the hypothyroid state; Fig. 1, A and B). In spleen and lung, on the other hand, UCP3 protein was barely present and was not regulated by T₃, whereas in liver it was not detectable (Fig. 1A).

Figure 1

UCP3 protein density in euthyroid (N), hypothyroid (P+I), and chronically hyperthyroid rats (P+I+T₃). A, Western immunoblot analysis of UCP3 protein levels in mitochondria from gastrocnemius and tibialis anterior muscle, heart, spleen, lung, and liver. Each lane contains 30 μg mitochondrial protein pooled from three rats. Each filter contains mitochondrial protein from euthyroid (N) gastrocnemius muscle as an expression control (indicated as C). B, Quantification of the data from three separate samples. Data are expressed as percentage of the UCP3 protein level in euthyroid (N) gastrocnemius muscle and are presented separately for each tissue in the three thyroid states indicated above the bars. Error bars represent sd.

Time course of changes in RMR, UCP3 mRNA, and protein after a single dose of T₃

To build on these results, we performed a second series of experiments in which, after a single injection of T₃ into hypothyroid rats, we followed the time course (at 0–144 h) of the changes induced in 1) RMR, 2) UCP3 mRNA and protein levels in gastrocnemius and heart, and 3) the coupling state of gastrocnemius mitochondria. UCP3 mRNA and protein levels as well as RMR were evaluated at the same time points after T₃ injection (viz. 6, 12, 24, 48, 65, 96, and 144 h) using the same animals together with parameters related to mitochondrial respiration. UCP3 mRNA levels were already elevated 6-fold at 6 h after the T₃ injection, with the maximal (∼25-fold) increase occurring at 24 h in both gastrocnemius and heart. At time points after 24 h, the amount of UCP3 mRNA in each tissue showed a decline toward its initial level, which was reached at 96 h (Fig. 2A). At the protein level, the increase in mitochondrial UCP3 content started between 12 and 24 h in both gastrocnemius and heart; the values reached a peak at 65 h (an increase of ∼12-fold) before starting to decline (Fig. 2B). What was very striking, and we believe of great importance, was that after T₃ injection the variations with time in UCP3 mitochondrial protein content (in both heart and gastrocnemius; see Fig. 3B) coincided closely with the induced changes in RMR[ lO₂(kg(0.75)⁻¹h⁻¹], the magnitude of which started to increase at 24 h, reached a peak at 65 h, then declined (Fig. 3C: at time zero the absolute value was 0.940 ± 0.007, and at 65 h the absolute value was 1.316 ± 0.026; P < 0.001).

Figure 2

Time course of the changes in UCP3 mRNA and protein levels after a single dose injection of T₃. A, Northern blot (gastrocnemius muscle) and RT-PCR (heart) analysis of UCP3 mRNA. β-Actin mRNA levels were measured as the internal standard. The measured time points (hours) are indicated above the lanes. For Northern blot analysis, each lane contains 20μ g total RNA pooled from three rats. For RT-PCR analysis, each lane contains PCR product derived from cDNA, for which 1 μg total RNA pooled from three rats was used. B, Western immunoblot analysis of UCP3 protein levels in mitochondria from gastrocnemius muscle, heart, and liver. Each filter contains mitochondrial protein from euthyroid (N) gastrocnemius muscle as an expression control (indicated as C). The measured time points (hours) are indicated above the lanes. Each lane contains 30 μg mitochondrial protein pooled from three rats. P+I, Hypothyroid.

Figure 3

Time course of the changes in UCP3 mRNA levels (A), UCP3 protein levels (B) in both gastrocnemius muscle (□) and heart (○), RMR (C; ▵), each after a single dose injection of T₃. Data in A and B are the means of three separate experiments; data in C are the means of four separate experiments; error bars represent sd. For gastrocnemius muscle only the error bars facing upward, and for heart only the error bars facing downward are depicted.

Assessment of mitochondrial energy coupling

We next set out to determine whether mitochondria that showed a higher UCP3 density might also possess a greater degree of uncoupling of respiration (a finding that would be consistent with the calorigenic effect of T₃ being exerted via UCP3). The experiments were conducted using mitochondria derived from the gastrocnemius muscle of the same animals in which changes in RMR and UCP3 expression were examined. We evaluated the mitochondrial coupling state by measuring nonphosphorylating respiration and Δψ in mitochondria obtained at 0, 65, and 144 h in four separate experiments. At 65 h, at which point we detected maximal UCP3 density, we observed that the nonphosphorylating respiration (nanoatoms of oxygen per min/mg protein) was increased significantly, by 40% (118 ± 4 vs. 85 ± 9 at time zero; P < 0.05), whereas Δψ (millivolts) was decreased by 8% (168 ± 2 vs. 182 ± 4 at time zero; P < 0.05). Mitochondria obtained at 144 h showed no significant differences in the respiratory parameters or inΔψ (both vs. time zero). An increase in the nonphosphorylating respiration rate accompanied by a decrease in Δψ are diagnostic of mitochondrial uncoupling. To assess whether total mitochondrial respiration was also in line with the changes observed in RMR, we measured (n = 4) state 3 and state 4 respiration on mitochondria at 0, 65, and 144 h. At 65 h, state 4 respiration was significantly increased by 67% (P < 0.05; 97.9 ± 9 vs. 58.7 ± 5.5 at time zero), then it declined, and at 144 h the value was not significantly different from that at time zero (70.0 ± 7.4 vs. 58.7 ± 5.5 at time zero). Nonsignificant variations were observed in state 3 respiration (264 ± 24 at time zero; 253 ± 25 at 65 h; 288 ± 27 at 144 h). The respiratory control ratio (RCR; state 3/state 4) values were 4.5 ± 0.4 at time zero, 2.7 ± 0.3 at 65 h, and 4.0 ± 0.3 at 144 h. A significant reduction of the RCR was observed at 65 h (P < 0.05).

Discussion

The experiments presented here were designed to establish whether UCP3 might be considered a molecular determinant for the regulation of RMR by T₃. This would represent an important advance in our basic knowledge of the mechanism underlying the calorigenic effect of thyroid hormones. Indeed, although more than a century has passed since the first evidence was obtained of an increase in metabolic rate in subjects in whom thyroid extract had been given, the molecular basis of the variations in RMR consequent to T₃ treatment is still poorly understood. As mentioned in the introduction, UCP3, because of its putative uncoupling property, may represent a crucial determinant in explaining how thyroid hormones increase energy expenditure partly by reducing metabolic efficiency. However, despite a significant accumulation of data from various model systems, the uncoupling property of UCP3 is insufficiently proven (for review, see Ref. 31). UCP3 knockout and UCP3 transgenic studies could provide the best evidence that UCP3 may be involved in proton leak. Transgenic mice that overexpress human UCP3 in skeletal muscle (66-fold compared with wild type) are hyperphagic and lean, but, more importantly, mitochondrial proton conductance was increased 2- to 3-fold in these mice (18). These results would be consistent with an uncoupling activity of UCP3 in vivo and would explain why UCP3 transgenic mice remain lean despite the observed hyperphagia. However, as observed by Stuart et al. (31), it is important to consider that a 66-fold increase in UCP3 mRNA levels and the corresponding unknown increase in protein could lead to an alteration in the mitochondrial membrane integrity, which may account for the observed proton leak. This is in light of a report of artifactual uncoupling coincident with high levels of UCP1 expression in yeast mitochondria (32). Two recent studies (19, 20) reported that UCP3 knockout mice showed a reduced proton leak in isolated mitochondria from skeletal muscle. However, these animals are not obese, and in apparent contrast with the data on UCP3 overexpressing mice, these results indicate that the lack of UCP3 is not associated with obesity. The results of one of these reports (19), in which UCP3 knockout mice given a 4-d course of T₃ at 100 μg/100 g BW·d showed the same increase in RMR as wild-type control mice do not seem to support an important role for UCP3 in the T₃-induced increase in the metabolic rate. However, the contrast between these data and those reported by us in this paper may be apparent rather than real. Several possible reasons for the apparent discrepancy may be noted: 1) the dose used by the previous authors is very high, and nonspecific thermogenic mechanisms may have been activated; 2) other mechanisms, such as those involving ANT, may have been overstimulated at this dose; and 3) in UCP3-deficient mice, the deiodinase enzymes are fully active, and under such conditions, some of the injected T₃ will be converted into 3,5-diiodothyronine (T₂) (33). It should be considered, in fact, that not only T₃ is able to increase RMR, but also T₂ has the potential to both enhance RMR and stimulate mitochondrial respiration (1, 26, 33–36). Thus, in T₃-treated UCP3-deficient mice a putative effect of T₂ in enhancing RMR could not be excluded. The general conclusion may be that data concerning variations in UCP3 expression by genetic manipulation should be interpreted with caution. Perhaps the most significant omission in the data regarding UCP3 is the expression level of the protein itself in natural systems. In a recent report Jekabsons and co-workers (37) tested the hypothesis that UCP3 mRNA levels might show a positive correlation with RMR and proton leak in mice in various thyroid states (thyroidectomized, euthyroid, and after 6-d treatment with T₃). The researchers concluded that T₃ does not influence intrinsic mitochondrial properties and that variations in UCP3 mRNA levels may only partly explain the variations in RMR. However, by comparison with other data (mostly obtained in rats), some differences may be noted: 1) despite the fact that an association between RMR and the thyroid state of the animal is a universally recognized phenomenon, in their study RMR was not depressed in thyroidectomized mice (compared with sham-operated mice) even though it was enhanced by T₃; and 2) although numerous studies have demonstrated a clear-cut effect of T₃ on mitochondrial respiration (for reviews, see Refs. 1 and 2), Jekabsons and co-workers failed to show this in their animals. Some possible explanations for these discrepancies may be given. First, as discussed by Jekabsons et al. (37), the physiological response to altered thyroid states could differ between rats and mice. Second, in their study RMR was measured at 26–28 C, a temperature outside the thermoneutral zone for mice (30–32 C). In addition, the mice were housed for 5–8 wk at 22 C, a temperature that represents a cold stress (especially for thyroidectomized animals), and after this period the mice were cold-acclimated. This cold acclimation would have led to variations in mitochondrial activity that cannot be reversed within a few hours, and these variations might have minimized the variations induced by the T₃ treatment. Third, the researchers correlate UCP3 mRNA levels with RMR and proton leak, although it would be crucial to correlate the mitochondrial UCP3 protein density with RMR and proton leak. It remains essential therefore to establish a connection between a physiological condition and the amount of this protein. In this context, we correlate here the T₃-induced variation in both UCP3 mRNA and protein levels with that in mitochondrial respiration efficiency as well as with the change in the metabolic rate of the whole animal. It is evident from previous considerations and from the wealth of data in the literature that several factors have to be taken into account when studying the mechanisms underlying the calorigenic effect of T₃. Because of this, in the present study we established an animal model in which some “disturbing factors” are minimized. First, to exclude the occurrence of any changes in UCP3 expression secondary to variations in serum FFA levels (which are induced by T₃), the gastrocnemius and tibialis anterior were the skeletal muscles of choice, as it has been shown that a reduction in serum FFA levels results in a fall in the UCP3 mRNA level in the soleus, but not in the gastrocnemius or tibialis anterior (38). Second, we estimated the mitochondrial efficiency under conditions in which the influences of ANT and FFA, two factors known to be capable of uncoupling respiration (39, 40), were eliminated. Third, and most importantly, as some deiodinated products of T₃ may be effective in stimulating RMR (such as T₂), with a time course different from that elicited by T₃ (33), we chose to generate hypothyroid rats by simultaneous administration of PTU and IOP. This combined treatment produces hypothyroid animals and at the same time inhibits all three known types of deiiodinase enzymes, which should permit us to attribute the observed effects to the iodothyronines injected rather than to any of their deiodinated products (34).

In view of the above, we believe that the present data permit us to suggest a role for UCP3 as a regulator of RMR in vivo, and that they point to the occurrence of the following steps. 1) Thyroid hormone (T₃) regulates both UCP3 mRNA levels and mitochondrial UCP3 protein density only in metabolically active T₃-responsive tissues (with the exception of the liver, where UCP3 is neither expressed nor induced by T₃. If this organ could play a role in thermogenesis, other mechanisms should be operative). 2) T₃ enhances the expression of UCP3 in mitochondria from skeletal muscle as well as from heart. 3) The resulting higher density of UCP3 elicits a mitochondrial uncoupling (under conditions that exclude the involvement of the other putative uncoupling pathways) with a subsequent increase in state 4 mitochondrial respiration, but with no change in state 3 and a decrease in Δψ, which is in agreement with the data reported previously (18, 22). 4) In terms of time course, the changes we observed in mitochondrial UCP3 protein content and mitochondrial respiratory parameters (state 4, RCR, and Δψ) after T₃ injection coincided very well and could explain the time course of the induced changes in RMR. Hence, on the basis of the available evidence we conclude that UCP3 has the potential to act as a molecular determinant for regulation of the RMR by T₃.

Acknowledgements

This work was supported by the Ministero dell’ Università e delle Ricerche Scientifiche e Tecnologiche (MURST-COFIN 2000 Protocol MM05C48114).

Abbreviations

• ANT

  Adenine nucleotide translocase

• BAT

  brown adipose tissue

• PTU

  6-n-propyl-2-thiouracil

• IOP

  iopanoic acid

• RCR

  respiratory control ratio

• RMR

  resting metabolic rate

• T₂

  3,5-diiodothyronine

• UCP

  uncoupling protein

• Δψ

  membrane potential.