Expression of the uncoupling protein 1 from the aP2 gene promoter stimulates mitochondrial biogenesis in unilocular adipocytes in vivo


J. Kopecký, Institute of Physiology, Academy of Sciences of the Czech Republic, 142 20 Prague, Czech Republic. Fax: + 420 2 475 2599, Tel.: + 420 2 475 2554,


Mitochondrial uncoupling protein 1 (UCP1) is a specific marker of multilocular brown adipocytes. Ectopic UCP1 in white fat of aP2-Ucp1 mice mitigates development of obesity by both, increasing energy expenditure and decreasing in situ lipogenesis. In order to further analyse consequences of respiratory uncoupling in white fat, the effects of the ectopic UCP1 on the morphology of adipocytes and biogenesis of mitochondria in these cells were studied. In subcutaneous white fat of both aP2-Ucp1 and young control (5-week-old) mice, numerous multilocular adipocytes were found, while they were absent in adult (7- to 9-month-old) animals. Only unilocular cells were present in epididymal fat of both genotypes. In both fat depots of aP2-Ucp1 mice, the levels of the UCP1 transcript and UCP1 antigen declined during ageing, and they were higher in subcutaneous than in epididymal fat. Under no circumstances could ectopic UCP1 induce the conversion of unilocular into multilocular adipocytes. Presence of ectopic UCP1 in unilocular adipocytes was associated with the elevation of the transcripts for UCP2 and for subunit IV of mitochondrial cytochrome oxidase (COX IV), and increased content of mitochondrial cytochromes. Electron microscopy indicated changes of mitochondrial morphology and increased mitochondrial content due to ectopic UCP1 in unilocular adipocytes. In 3T3-L1 adipocytes, 2,4-dinitrophenol increased the levels of the transcripts for both COX IV and for nuclear respiratory factor-1. Our results indicate that respiratory uncoupling in unilocular adipocytes of white fat is capable of both inducing mitochondrial biogenesis and reducing development of obesity.


adipocyte lipid binding protein

aP2-Ucp1 transgenic mouse

mouse with the expression of UCP1 from the fat-specific aP2 gene promoter


subunit IV of mitochondrial cytochrome c oxidase


mitochondrial uncoupling proteins


nuclear respiratory factor-1.

Increasing evidence suggests that respiratory uncoupling in white adipose tissue could prevent excessive accumulation of body fat. Part of the evidence comes from studies of mitochondrial uncoupling protein 1 (UCP1), an integral protein of the inner mitochondrial membrane and a well-established protonophore [1–3]. This protein is typically present only in brown fat [4–6] where it dissipates the energy of mitochondrial proton gradient and is essential for regulatory thermogenesis [1,7,8]. However, expression of UCP1 gene could be also induced in white fat depots of experimental animals by pharmacological compounds that reduce adiposity, e.g. β3-adrenoreceptor agonists [9–11], nicotine[12], or leptin [13]. Even in adult humans, relatively low levels of the UCP1 transcript could be detected in various fat depots. In abdominal fat, UCP1 mRNA levels are negatively correlated with obesity [14]. Accordingly, the expression of UCP1 gene from a highly fat-specific [15]aP2 gene promoter in transgenic aP2-Ucp1 mice [16] resulted in␣resistance against genetic [16] or dietary [17] obesity. The␣obesity resistance is induced by transgenic modification of white but not brown fat [3,8,18], and reflects reduction of all fat depots except for gonadal fat [8,16,18]. Ectopic UCP1 induces depression of mitochondrial membrane potential in adipocytes [19], increased energy dissipation [8,18] and depression of in situ lipogenesis [20]. The latter mechanism probably reflects insufficient supply of ATP by mitochondrial oxidative phosphorylation [20].

Besides UCP1, efficiency of oxidative phosphorylation in adipocytes may be also controlled by recently discovered UCP1 homologues, i.e. UCP2, UCP3, UCP5 [2,21–23], and even by an adenine nucleotide transporter [24,25]. All these proteins are probably present in mature brown adipocytes, while white adipocytes do not typically contain either UCP1 (see above), or UCP3 [2,26]. However, treatment with β3-adrenoreceptor agonists is capable of inducing not only UCP1 (see above) but also UCP3 [27] in white fat. In an obesity-prone strain of mice, UCP2 mRNA levels in white adipose tissue were lower than in mice resistant to diet-induced obesity [28,29] and a similar difference in UCP2 gene expression was observed in abdominal fat of normal and obese humans [30]. Moreover, a negative correlation between heat production in adipocytes and body fat has been found in humans [31].

Some aspects of the relationships between UCPs in white fat and adiposity remain to be clarified, namely the identification of the adipose cell type involved, and the underlying biochemical mechanisms. The first aspect relates to the occurrence of multilocular cells expressing UCP1 that are interspersed in white fat [9,10,32–38]. In large mammals, such as humans, typical brown fat depots do not exist in adults, however, some adipocytes equipped with UCP1 and containing many mitochondria probably remain present in white fat during adulthood [14,36–38]. However, developmental studies on these cells are scarce [37]. The induction of UCP1 in white fat by β3-adrenoreceptor agonists [9–11], or by cold exposure of animals [32,39–41], occurs in multilocular cells interspersed in white fat depots. Such cells may arise from transdifferentiation of unilocular white adipocytes, or reflect recruitment of brown fat precursor cells [9,10]. The possible role of UCP1 in conversion of unilocular into multilocular cells has not been studied.

Reduction of adiposity by respiratory uncoupling in adipocytes may be limited by mitochondrial oxidative capacity. Importantly, it has been shown in vitro[42] that the uncoupling, induced by ectopic UCP1 in HeLa cells, could induce mitochondrial biogenesis and upregulate its␣co-ordinating factor, the nuclear respiratory factor-1 (NRF-1). In animals treated with β3-adrenoreceptor agonists [43], the metabolic rate was relatively high and the treatment induced formation of mitochondria in the multilocular cells in white fat depots [10]. Also cold acclimatization induces mitochondrial biogenesis in brown fat, reflecting increased sympathetic stimulation of this tissue [32,40,41,44,45]. These data suggest that respiratory uncoupling in adipocytes is associated with mitochondrial biogenesis. However, possible existence of a causal link between these two processes requires further clarification.

The aim of this work was to characterize further the mechanism by which respiratory uncoupling in white fat reduces adiposity, namely with respect to morphology of adipocytes and mitochondrial biogenesis. It has been investigated whether ectopic UCP1 in white fat of aP2-Ucp1 mice can induce formation of multilocular cells depending on the age of the animals. The possibility that respiratory uncoupling may activate mitochondrial biogenesis has been also explored both in the transgenic mice and in 3T3-L1 adipocytes differentiated in cell culture.

Materials and methods

Animals and tissues

Control C57BL/6J male mice and their hemizygous aP2-Ucp1 transgenic littermates were identified by Southern blot analysis [16]. The mice were born and maintained at 20 °C with a 12-h light/dark cycle. After weaning at 4 weeks of age, mice were housed four or five per cage and had free access to a standard chow diet [17] and water. If not specified otherwise, animals were killed at 5 weeks (young mice) or at 7–9 months (adult mice) of age by cervical dislocation. Interscapular brown adipose tissue, subcutaneous dorsolumbar white fat [17], and epididymal fat were used for the experiments. Samples were stored at −70 °C for immunoblotting analysis, and in liquid nitrogen for isolation of total RNA.

Morphological studies

The animals were anaesthetized by intraperitoneal injection of thiopental (80 µL of 5% thiopental/animal) and whole animals were fixed by perfusion with paraformaldehyde (4% solution in 0.1 m phosphate buffer, pH 7.4) through the left ventricle (after the right atrium was opened). After perfusion, the tissues (see above) were dissected and fixed overnight by immersion in the same fixative for light microscopy and immunohistology, and in a mixture of 2% glutaraldehyde and 2% paraformaldehyde in 0.1 m phosphate buffer, pH 7.4, for 4 h, for ultrastructural study. Tissues for light microscopy and immunohistology were embedded in paraffin blocks. For ultrastructural studies small fragments were postfixed in 1% osmium tetroxide, dehydrated in ethanol, and embedded in an Epon/Araldite (Epon, Multilab Supplies, Fetcham, UK; Araldite, Fluka Chemie AG, Buchs, Switzerland) mixture. Semithin sections (2 µm) were stained with toluidine blue; thin sections were obtained with a Reichert Ultracut E (Reichert, Wien, Austria), stained with lead citrate, and examined in a transmission electron microscope, Philips CM10 (Eindhoven, the Netherlands). Immunohistological demonstration of UCP1 was carried out by the avidin–biotin peroxidase (ABC) method. De-waxed sections (3 µm) were processed through the following incubation steps: (a) 0.3% hydrogen peroxide in methanol for 30 min to block endogenous peroxidase; (b) 0.02 m glycine for 10 min; (c) normal rabbit serum 1 : 75 for 20 min to reduce nonspecific background staining; (d) polyclonal sheep antibodies against UCP1 isolated from rat brown adipose tissue, diluted 1 : 8000 in NaCl/Pi, overnight at 4 °C; (e) biotinylated rabbit anti-(sheep IgG) Ig 1 : 300 (secondary antibody) for 30 min (Vector Laboratories, Burlingame, CA); (f) ABC complex for 1 h (Vectastain ABC kit, Vector Laboratories); and (g)␣histochemical visualization of peroxidase using 3′,3′-diaminobenzidine hydrochloride chromogen (Sigma). The specificity of the method was tested by the omission of the primary antibody in the staining, and the use of preimmune serum instead of the first antiserum. Furthermore, tissues known to contain UCP2 and UCP3 (skeletal muscle, white adipose tissue, spleen, and kidney) but not UCP1 were tested. All tissues containing UCP2 and UCP3 showed negative results. The specificity of the anti-UCP1 Ig has been recently confirmed [23]. For immunohistochemical studies, three mice for each type of condition were used.


Morphometric evaluation of subcutaneous white fat of nine control and eight transgenic animals was performed both with light microscopy (semithin sections) and at the ultrastructural level. In case of light microscopy the surface area of about 130–170 cells for each animal was measured by an Image Analyzer KS100 IBAS Kontron (Karl Zeiss Jena, Germany), in order to calculate the diameter of the adipocytes. In the ultrastructural study four to six pictures for each animal (nine control and eight transgenic mice) were taken randomly at a final magnification of 11 300× by a CM10 PHILIPS EM (see above). The images were analysed by the IBAS morphometer in order to measure the lipid-free cytoplasmic surface area, the surface area of the mitochondria (µm2), mitochondrial density (i.e. number of mitochondria per 100 µm2 cytoplasmic area) and cristae density [i.e. total cristae length (pm) per mitochondrial surface area, per 100 µm2 cytoplasmic area].

Evaluation of UCP1 and cytochrome content, protein, and DNA concentration

Crude cell membranes (100 000 g) were prepared from tissue homogenates and used for quantification of the UCP1 antigen by immunoblotting using rabbit anti-(hamster UCP1) serum [46] and a standard consisting of mitochondria isolated from brown fat, as described previously [18]. As a second antibody, 125I-labelled donkey antibody against whole rabbit IgG (Amersham) was used, and radioactivity was evaluated using PhosphorImager SF (Molecular Dynamics). Protein concentration was measured using the bicinchoninic acid procedure [47] and BSA as standard. The membrane fraction was solubilized in the presence of 2% n-dodecyl β-d-maltoside (Sigma) and used for evaluation of mitochondrial cytochromes using a pseudo-dual-wavelength spectrophotometry [19]. Tissue DNA was estimated as described previously [8].

Isolation of adipocytes

Adipocytes were isolated from subcutaneous white fat of adult mice according to Rodbell [48]. Modified Krebs-Ringer bicarbonate (KRB) buffer was used, containing 118.5 mm NaCl, 4.8 mm KCl, 2.7 mm CaCl2, 1.2 mm KH2PO4, 1.1 mm MgSO4·7H2O, 25 mm NaHCO3, 5 mm glucose and 4% (w/v) bovine serum albumin (fraction V; BSA); pH 7.4. Adipose tissue (1–2 g) was collected from four␣mice, minced with scissors and digested in 5 mL KRB buffer containing 3 mg·mL−1 type II collagenase (C-6885, Sigma) while shaking at 37 °C for 90 min. The tissue was then filtered (250 µm) and floating adipocytes were washed three times in the KRB buffer in the absence of collagenase by centrifuging at 400 g for 1 min at 20 °C.

Differentiation of 3T3-L1 adipocytes

Cells of 3T3-L1 clonal line were differentiated in cell cultures as described previously [20]. When used for experiments (12–14 days after confluence), cultures contained 50–60% of differentiated adipocytes. Ten hours before use for RNA isolation (see below), a complete change of the medium was performed. 2,4-dinitrophenol (dissolved in 0.1% KOH) was added in some dishes at a 150-µM final concentration.

RNA analysis

Total RNA was isolated from adipose tissue or adipocytes and analysed on Northern blots as described before [8]. Filters (GeneScreen™; NEN Life Science Products, Boston, MA) were subsequently hybridized with full-length cDNA probes for mouse UCP1 [8], UCP2 [8], human liver subunit IV of mitochondrial cytochrome oxidase (COX IV; ATCC, Rockville, MD), and aP2 [8]. Final hybridization with a ribosomal 18S RNA probe was used to correct for possible intersample variations within individual blots. Radioactivity was evaluated by PhosphorImager SF. Total RNA isolated from brown fat of cold acclimatized mice served as a␣standard. In the case of total RNA isolated from adipocytes, levels of the transcripts for COX IV and for NRF-1,␣respectively, were evaluated using real time quantitative RT-PCR [20], using the LightCycler (Roche Molecular Biochemicals, Mannheim, Germany) and LightCycler-RNA Amplification Kit SYBR Green I (Roche; cat. no. 2015137). Each PCR cycle consisted of 0 s at 94 °C, 8 s at 60 °C, and 20 s at 72 °C. Transcript levels were expressed relative to that of β-actin. Primers used for RT-PCR are specified in Table 1.

Table 1. Sequences of PCR primers.
GeneSense primer (5′−3′)Antisense primer (5′−3′)GenBank acc. no.
  • a

    The primers are specific for the isoform 1 of subunits IV of cytochrome c oxidase.


Statistical analysis

A two-way analysis of variance (anova) with post hoc multiple comparisons was used as described before [17]. Otherwise, statistical significance was evaluated using Student's t-test. The morphometric measurements were evaluated using the Kruskal–Wallis nonparametric test. All comparisons were judged to be significant at P < 0.05.


Fat-depot- and age-dependent differences
of adipocytes' morphology in white fat

Morphology of adipocytes (Fig. 1) and their UCP1 content (see below) were characterized in semithin sections of subcutaneous white fat and epididymal fat (not shown) of control and transgenic animals during ageing. In both fat depots of all the animal subgroups studied, unilocular adipocytes represented the most abundant cell type. Only in subcutaneous fat of young mice multilocular adipocytes were also detected, and these cells formed a substantial portion of mature adipocytes, with the ratio between multilocular and unilocular adipocytes of about 1 : 4 to 1 : 5 (Fig. 1). No multilocular cells were detected in either subcutaneous fat of adult mice (Fig. 1), or in epididymal fat of both age groups (not shown). Transgene had no effect on the ratio between multilocular and unilocular cells in subcutaneous fat of young animals, neither induced multilocular cells in white fat depots of adult mice [16]. The mean diameter of the unilocular cells present in subcutaneous fat of adult transgenic and control mice were 56 ± 4 µm and 63 ± 5 µm, respectively; the difference was not statistically significant.

Figure 1.

Immunohistology of subcutaneous white fat from young (A and B) and adult (C and D) control (A and C) and transgenic (B and D) mice. (A)␣The␣depot is composed of unilocular and multilocular adipocytes. Only multilocular cells are weakly stained for UCP1 (arrows). (B) The depot is composed of unilocular and multilocular adipocytes. Most of the multilocular and some of the unilocular cells (arrows) are intensely stained. (C)␣Only unilocular cells are present and they do not contain UCP1 antigen. (D) Only unilocular cells are present and most of them are intensely stained for UCP1; areas of cytoplasmic rim stained for UCP1 are thickened (arrows).

Age-related changes in the expression of UCP1 gene in white fat depots

The expression of UCP1 in subcutaneous and epididymal fat depots of control and transgenic mice of different ages was analysed by immunohistochemistry (Fig. 1) and by biochemical techniques, at both mRNA and protein level (Fig. 2). Immunohistochemistry revealed that multilocular cells found in subcutaneous fat of young mice of both genotypes contained UCP1. The intensity of immunohistochemical staining of brown fat cells was stronger in transgenic than in control mice, in agreement with expression of both UCP1 endogen and aP2-Ucp1 transgene in these cells␣[16]. In adult control mice, the unilocular cells in both subcutaneous (Fig. 1) and epididymal fat (not shown) lacked UCP1, while they were UCP1-positive in the transgenic mice. All unilocular adipocytes in transgenic mice contained UCP1. These findings thus confirmed our previous observations in aged transgenic animals [16]. The staining for UCP1 was always restricted to the cytoplasmic area in the vicinity of the plasma membrane, which was thicker in transgenic than in nontransgenic mice. Electron microscopy revealed that these thicker parts of the cytoplasm were rich in mitochondrial content (see below).

Figure 2.

Quantification of UCP1 expression in white adipose tissue depots during ageing. Analysis was performed in subcutaneous white fat (Sc-WF) and epididymal fat (Epid-WF) of control (open symbols) and transgenic (full symbols) mice of indicated ages (n = 6–8). Values are means ± SE. (Top) Results of Northern blot analysis of UCP1 transcript (1.4 kb). Analysis was not performed in epididymal fat of 3-week-old mice, due to the relatively low amount of the tissue (19 ± 6.0 and 19 ± 4.3 mg in control and transgenic mice, respectively), as compared with subcutaneous white fat (54 ± 6.5 and 49 ± 6.7 mg, respectively). In control mice, the UCP1 transcript could be detected only in subcutaneous white fat of 3-week- and 2-month-old mice (0.010 ± 0.005 and 0.020 ± 0.010 arbitrary units of UCP1 transcript, respectively). In 7-month-old transgenic mice, the values were 0.05 ± 0.01 and 0.023 ± 0.001 arbitrary units of UCP1 transcript in subcutaneous and epididymal white fat, respectively. Evaluation of the aP2 transcripts (0.65 kb) in adult control mice (not shown) indicated significantly higher levels in interscapular brown fat (0.78 ± 0.08 arbitrary units) than in white fat (0.19 ± 0.06, and 0.214 ± 0.02 arbitrary units, in subcutaneous and epididymal fat, respectively). (Bottom) Results of immunoblotting experiments with membrane fractions isolated from fat depots. In control mice, UCP1 could be detected in subcutaneous white fat of 1- and 2-month-old mice. The content of UCP1 in subcutaneous white fat and epididymal fat of 7-month-old transgenic mice was 1.31 ± 0.36 and 0.30 ± 0.1 µg UCP1 per mg membrane protein, respectively. All the differences between genotypes were significant.

Both Northern blot analysis and immunoblotting (Fig. 2) detected UCP1 expression in subcutaneous white fat of 3-week- to 2-month-old-control animals and in both fat depots of transgenic mice, regardless of age of the animal. The levels of UCP1 mRNA in subcutaneous fat of control mice were by one order of magnitude lower than in transgenic mice, while the corresponding difference in the specific content of UCP1 antigen (expressed relative to adipose tissue membrane protein) was only about twofold. In both fat depots of the transgenic mice, the levels of the UCP1 transcript and UCP1 antigen declined substantially during ageing (5- to 10-fold), and they were twofold to fourfold higher in the subcutaneous than in epididymal fat. In 3-week- to 2-month-old transgenic mice, levels of UCP1 transcript in subcutaneous white fat were approximately 30% of those in interscapular brown fat, while in the case of UCP1 antigen this value was about 10% (not shown). No UCP1 mRNA or antigen could be detected either in white fat depots of adult (4- to 7-month-old) control mice [16,18], or in epididymal fat of younger nontransgenic animals (Fig. 2).

The results document the absence of multilocular adipocytes in the epididymal fat in all the age groups studied, while in subcutaneous fat these multilocular cells completely disappear as the animals age. These results also indicated a higher content of transgenic UCP1 in unilocular adipocytes in subcutaneous than in epididymal fat and suggest that UCP1 is not capable of inducing conversion of a unilocular into a multilocular adipocyte.

UCP1-induced increase of mitochondrial biogenesis

Several independent approaches were used to investigate whether ectopic UCP1 could induce biogenesis of mitochondria in white fat. First, the transcript level of COX IV, a nuclear gene for one of the subunits of mitochondrial cytochrome c oxidase, was evaluated in total RNA isolated from subcutaneous and epididymal fat during ageing in mice (Fig. 3). Except for a decrease of COX IV mRNA in subcutaneous fat between the first and second month of age, the level of the transcript did not change significantly during ageing in either genotypes. However, as indicated by anova, there was a main effect of genotype in both depots, with transgenic animals showing higher levels of the transcripts. Within different ages and depots, most differences (over 1.5-fold; Fig. 3) were statistically significant. Interestingly, also the levels of the transcript for UCP2 were higher in transgenic than in control mice (Fig. 3). With both, COX IV and UCP2, the highest differences (up to threefold) were observed in epididymal fat. It is known that composition of subcutaneous white fat is quite heterogenous and mature adipocytes represent less than 50% of all cells contained in this fat depot [45]. Therefore, gene expression was also characterized in mature adipocyte fractions isolated from subcutaneous fat of adult mice. The upregulation of both COX IV(Table 2) and UCP2 (not shown) genes by UCP1 was confirmed. A possible effect [42] of the transgene on NRF-1 mRNA levels was also tested but no significant difference between the adipocytes isolated from control and transgenic mice could be observed (Table 2).

Figure 3.

Quantification of mRNA for␣mitochondrial markers in white adipose tissue depots during ageing. Analysis of the transcripts for COX IV (0.9 kb), and UCP2 (1.7 kb) was performed using Northern blots in control and transgenic mice. For details and symbols, see Fig. 2. There was a main effect of genotype (anova) within each fat depot and type of transcript. Asterisks indicate significant differences between genotypes within the same age group.

Table 2. .  Quantification of gene expression in adipocytes. Levels of the transcripts were quantified by real time RT-PCR in adipocytes isolated from subcutaneous white fat of 7-month-old control (+/+) and transgenic (tg/+) mice and from 3T3-L1 adipocytes differentiated in cell cultures. 3T3-L1 adipocytes were incubated for 10 h in a cell culture dish with or without 150 µm 2,4-dinitrophenol before RNA isolation. Values are means ± SE (n = 6).
TranscriptmRNA level (arbitrary unit)
Isolated adipocytes3T3-L1 cells
  1. * P < 0.05.

COX IV0.66 ± 0.080.95 ± 0.15*0.30 ± 0.050.40 ± 0.05*
NRF10.016 ± 0.0050.010 ± 0.0043.8 × 10−3 ± 3.6 × 10−76.7 × 10−3 ± 8.7 × 10−7*

Further experiments were focused only on subcutaneous fat, as the size of this fat depot but not of the epididymal fat was reduced by the transgene in adult mice [16,17]. The content of mitochondrial cytochromes b, and a + a3, respectively, was established in subcutaneous white fat of young and adult mice (Fig. 4). A highly sensitive quantification of absolute amounts of the cytochromes was performed using a pseudo-dual-wavelength spectrophotometry [19]. While cytochrome b is contained in the bc1 complex, cytochromes a + a3 are integral parts of the cytochrome c oxidase in the inner mitochondrial membrane. When the content of the cytochromes was expressed relative to the mass of tissue, there was a main effect (anova) of age on cytochrome b content, and a main effect (anova) of the genotype; a higher content of cytochromes was present in young and/or transgenic mice. Within the same age, the only statistically significant difference was found with cytochrome b content in young mice (1.7-fold difference between genotypes; see Fig. 4). Similar results were obtained when the values were expressed relative to tissue DNA (not shown).

Figure 4.

Content of mitochondrial cytochromes in subcutaneous white fat during ageing. Cytochromes a+ a3, and cytochrome b, were quantified in control (open bars) and transgenic (solid bars) mice of indicated age group (n = 6–7). Values are means ± SE. In the adult animals, the content of cytochromes a + a3 could not be measured due to the limited sensitivity of the method [19]. See text for details.

Mitochondrial morphology was characterized by transmission electron microscopy in subcutaneous white fat of adult animals (Fig. 5), where only unilocular adipocytes were present in both genotypes (Fig. 1). In control mice (Fig. 5A–C), the peripheral rim of adipocytes was always thin with a few ‘white-type’ mitochondria. These mitochondria were elongated and their cristae were randomly oriented. The presence of ectopic UCP1 in transgenic mice (Figs 5D–F) was associated with increased size of mitochondria contained in a thick periplasmic rim of the adipocyte. Mitochondria were mostly oval or round, and the number of cristae per mitochondrion was relatively high. Some cristae were regularly oriented. Thus, most of the mitochondria in the transgenic mice showed an intermediate morphology between that found in white and brown adipocyte [45]. This suggests an activation of mitochondrial metabolism and induction of mitochondrial biogenesis in white fat of transgenic mice. Changes in the ultrastructural appearance were substantiated further by a morphometric analysis (Fig. 6). Mean surface area of mitochondria, mitochondrial density in lipid-free cytoplasmic area, and density of cristae in mitochondria were bigger in transgenic than in control mice. The differences were 1.48-, 1.53-, and 1.22-fold, respectively, and they were statistically significant (see legend to Fig. 6). Calculations based on the morphometric data indicated that 20.3% of the cytoplasmic area of unilocular white adipocytes in transgenic animals was occupied by mitochondria, as compared with only 9.6% in control animals.

Figure 5.

Transmission electron microscopy of subcutaneous white fat in adult mice. Parts of unilocular adipocytes containing cytoplasmic compartment with mitochondria are shown (bar corresponds to 1 µm). (A–C) Control mice; (D–F) transgenic mice.

Figure 6.

Mitochondrial morphometry in subcutaneous white fat of adult mice. Morphometric analysis of surface area of the mitochondria, mitochondrial density, and cristae density was performed in control (empty bars) and transgenic (solid bars) mice. Values are means ± SE. All the differences between genotypes were statistically significant (P = 0.023, P= 0.026, and P= 0.008 in the case of the mitochondrial area, mitochondrial density and cristae density, respectively).

Finally, in order to confirm that respiratory uncoupling in adipocytes may stimulate mitochondrial biogenesis, 3T3-L1 adipocytes differentiated in cell culture were used (Table 2). Some adipocytes were incubated with 2,4-dinitrophenol that was added to cell culture medium at a final 150 µm concentration. Previously, under similar conditions, a near maximal stimulation of fatty acid oxidation by 2,4-dinitrophenol was observed [20]. In the present experiments, 2,4-dinitrophenol induced a significant increase of the levels of transcripts for both COX IV and NRF-1.


It was found that ectopic expression of UCP1 in white fat depots of aP2-Ucp1 mice occurs in both forms of mature adipocytes, in multilocular and in unilocular cells. The multilocular adipocytes could be detected only in subcutaneous white fat of young but not adult mice, and they were absent from epididymal fat, regardless of either the age of the animals, or the genotype. Therefore, the results document further that the resistance against obesity brought by ectopic UCP1 in white fat of adult mice [16–18] reflects respiratory uncoupling [19] in unilocular white adipocytes [16]. A higher content of UCP1 in subcutaneous white fat compared with epididymal fat of the transgenic mice helps to explain the lack of the effect of the transgene on the size␣of␣the latter depot [16,17]; this is also associated with the differential effect of the transgene on in situ fatty acid synthesis in the two fat depots [20]. The results are in agreement with the hypothesis that induction of endogenous UCP1 acts locally, in concert with adrenergic stimulation [9], to reduce to a greater extent the adiposity of fat depots with high induction of UCP1 than in depots with low induction.

During mammalian ontogeny, recruitment of brown adipose tissue precedes the first appearance of white fat, and the timing of these events during perinatal development varies in different species [49]. Mice belong to a group of the altricial species, with the recruitment of brown fat during late period of the fetal development [46,49,50]. This study shows a dramatic decrease of the content of multilocular adipocytes expressing the UCP1 gene in subcutaneous white fat depot during ageing in mice. Also UCP1 expression in numerous fat depots of some other species (e.g. bovine [37] and human [51]) is restricted to early stages of development. Therefore, the disappearance of UCP1-producing cells from subcutaneous white fat of mice during ageing reflects a general trend for a localization of UCP1-based thermogenesis into a limited number of anatomical sites in adult animals.

It has been suggested that white adipocyte precursors might belong to brown fat lineage [9]. Inversely, most multilocular cells in white adipose tissue of rats treated with β3-adrenergic agonists originated from unilocular adipocytes and contained UCP3, while only a small fraction of novel multilocular adipocytes contained UCP1 [10]. As reported in this study, the expression of functional UCP1 in unilocular adipocytes of animals between 5 weeks and 9 months of age was not accompanied by the conversion of these cells into multilocular adipocytes. After prolonged (over 1 week) stimulation with β3-adrenergic agonists, the number of multilocular adipocytes containing UCP1 in rat white fat is still increasing, without further changes of the ratio between unilocular and multilocular cells (Zingaretti, M. C., Ceresj, E., Barbatelli, G. & Cinti, S., unpublished observation). All these experiments suggest that the expression of UCP1 (or UCP3) in unilocular adipocytes, in the absence of a contribution by other controlling factor(s), cannot convert unilocular into multilocular adipocytes. This is in agreement with the experiments on the emergence of brown adipocytes in white fat depots of mice, indicating involvement of at least four different genes [9].

In contrast with the inability of UCP1 to induce multilocular cells in white fat, the morphology of mitochondria and the mitochondrial content of the unilocular cells were affected by the transgene. The morphometric study of subcutaneous white fat of the adult transgenic animals demonstrated that the unilocular cells had a larger cytoplasmic area and contained more numerous and larger mitochondria with a relatively high cristae density, compared to control mice. Thus, the cytoplasmic area occupied by mitochondria was about twofold larger in the adipocytes of transgenic than control mice. The results of the morphometric analysis indicated induction of mitochondrial biogenesis by ectopic UCP1 in the unilocular adipocytes.

The stimulatory effect of UCP1 on mitochondrial content and biogenesis was also supported by differences in the level of the transcripts for COX IV, in both whole adipose tissue and isolated adipocytes, as well as by differences in the content of mitochondrial cytochromes between two genotypes. That UCP2 was upregulated in aP2-Ucp1 mice was somehow surprising and suggested that UCP1 and UCP2 function differently in adipocytes. This supports the idea that both UCP2 and UCP3 are linked to fatty acid oxidation [53] that is elevated by respiratory uncoupling in adipocytes [19]. It is not clear why the COX IV and UCP2 transcript levels in both white fat depots of transgenic mice change very little with age whereas the UCP1 antigen content strongly decreases during the same time. Nevertheless, all the approaches indicated a moderate induction of mitochondrial biogenesis by ectopic UCP1 in unilocular adipocytes. The resulting increase of mitochondrial content was evidently smaller than that induced in multilocular adipocytes by β3-adrenoreceptor agonists [10,44], or due to adrenergically mediated stimulation of mitochondrial biogenesis that occur in cold acclimatized animals [32,39–41]. The relatively high potency of the adrenergic stimulation could be explained by the complex effect on gene expression in adipocytes. It may be also speculated that the effect of adrenergic system on mitochondrial biogenesis represents a compensation for decreased efficiency of energy conversion in adipocytes with upregulated UCP1 gene expression.

It has been found by Zhou et al. [13] that adenovirus-mediated hyperleptinemia in rats depletes adipocyte fat while upregulating UCP1, UCP2, and genes for enzymes of fatty acid oxidation. On the other hand, genes for lipogenic enzymes, aP2, and the transcription factor PPARγ were downregulated. To achieve such a transformation of adipocytes may be useful for treatment of obesity [13]. Results of our present and the previous [20] study on white fat of adult mice suggest that UCP1 alone could initiate the ‘transdifferentiation’ program, including an increased expression of the genes controlling oxidative capacity (COX), as well as that of UCP2, and depression of genes engaged in fatty acid synthesis.

The molecular mechanism for the induction in mitochondrial biogenesis by ectopic UCP1 in HeLa cells was shown to involve up-regulation of NRF-1 [42]. In our experiments, an increase of NRF-1 mRNA level was detected in 3T3-L1 adipocytes incubated with 2,4-dinitrophenol but not in adipocytes isolated from white fat of transgenic compared to control mice. Therefore, NRF-1 may function as a critical␣component of the energy-sensing mechanism that co-ordinates expression of mitochondrial genes in adipocytes. However, stimulation of NRF-1 expression in mice may be only transient and can already have taken place before the experiments are carried out.

The levels of UCP1 transcript in white fat depots of adult transgenic mice were expected to reflect the activity of aP2 gene promoter that is contained in the aP2-Ucp1 transgene. However, in both subcutaneous and epididymal white fat of control adult mice, the aP2 gene transcript levels were quite similar, and they were about fourfold lower than in their interscapular brown fat (see legend to Fig. 2). This suggests a differential postranscriptional control of the transgene expression in various white fat depots, resulting in higher UCP1 content in subcutaneous than in epididymal fat. Differential post-transcriptional control of the endogenous UCP1 gene and the transgene, respectively, may also explain why the difference in UCP1 mRNA levels between transgenic and control mice is much higher than that in UCP1 antigen levels (see Fig. 2). Our results showed the profound fat-depot- and age-dependent differences in transgene expression that may be relevant for other studies, where the aP2 promoter is used to direct the expression of various genes into adipose tissue in mice (see also patent no. US5476926).

In conclusion, our results indicate that respiratory uncoupling per se is capable of inducing mitochondrial biogenesis in vivo. They also support the hypothesis that respiratory uncoupling in unilocular adipocytes of white fat depots may reduce adiposity and prevent the development of obesity.


This research was supported by the Grant Agencies of the Czech Rep. (311/99/0196) and the Acad. Sci. of the Czech Rep. (A5011710), COST-918 (to J. K.) and by grants from the University of Ancona, Italy (Cofin 1998 to S. C., and Contributo Ricerca Scientifica Finanziata dalla Università anno 2000 to S. C. and G. B.). We thank Dr B. B. Lowell (Harvard Medical School, Boston, MA) for the mouse UCP2 cDNA, and Dr D. Ricquier (CNRS/CEREMOD, Meudon, France) for polyclonal sheep antibodies against UCP1 isolated from rat brown adipose tissue, and Dr A. Kotyk (Institute Physiol., Acad. Sci. of the Czech Rep.) for critical reading of the manuscript.