INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THERMOGENIC ACTIVITY in brown adipose tissue (BAT) is the predominant source of heat for maintenance of body temperature (T_b) in small and newborn mammals in a cold environment and for arousal of hibernators from hypothermia (32). BAT thermogenesis is due to a unique mitochondrial protein, the uncoupling protein (UCP), a proton translocator that uncouples respiration from ATP synthesis (16). BAT further plays a role in total energy homeostasis and body weight regulation. When a highly palatable "cafeteria diet" is fed, BAT can be used to dissipate excess energy in rats and mice by the so-called diet-induced thermogenesis (25, 26). Denervation and excision of interscapular BAT were shown to result in an increase in total body fat (5). To investigate further the role of BAT in mammalian energy balance, Lowell and co-workers (19) created transgenic mice [UCP-diphtheria toxin A chain (DTA) mice] with largely abolished BAT function. This was achieved by using the promoter of the BAT-specific UCP to drive expression of a gene for DTA. UCP-DTA mice were found to have 70% decreased UCP-1 and BAT content, they were less cold resistant, and they developed obesity and insulin resistance (10, 19). The initial development of obesity in UCP-DTA mice was shown to be independent of hyperphagia, although hyperphagia developed later with increasing obesity (10, 19). A recent study by Melnyk et al. (20) showed that raising of UCP-DTA mice at 35°C prevented obesity. However, no data on energy expenditure are supplied. Although these studies have shown clearly the importance of brown fat in the regulation of energy homeostasis, the physiological basis for development of obesity in UCP-DTA mice is still unclear. Here we analyzed in detail the regulation of the energy budget in UCP-DTA mice by measuring food intake and assimilation; locomotor activity; metabolic rates, i.e., basal metabolic rate (BMR) and average daily metabolic rate (ADMR); and also T_b.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animals and animal handling. UCP-DTA transgenic mice and control litter mates (FVB) were kindly supplied by Dr. Andreas Hamann, Hamburg, Germany. Only adult, ~1 yr old, females were used in the experiments. Animals were housed in groups of two to four individuals at 24°C with a 12:12-h light-dark cycle and fed Altromin diet 1324 with 6.5% fat (Altromin, Lage, Germany). Food and water were available ad libitum.

Food consumption and assimilation. Over a 30-day period, food consumption was determined every 2 days and animals were weighed to the nearest 0.1 g. For determination of assimilation efficiency, animals were housed individually on a plastic grid that allowed collection of feces. Over a 3-day period, food consumption and feces production were recorded. Feces and food samples were dried to constant weight, and energy content was measured by bomb calorimetry. The percentage of assimilated energy was calculated by the difference in energy content of food consumed and feces produced. Energy loss through urine was assumed as 2% and subtracted (4).

Locomotor activity. Locomotor activity was registered in individually housed animals over a 2-wk period using passive infrared motion detectors mounted above the cage (SA209, Conrad Electronics) (27). Each movement of an animal was recorded as a 3-s impulse, which was considered as one event. The sum of events was recorded every 6 min.

Metabolic rate. Oxygen consumption of individual animals was measured in an open respirometric system with an air flow of 50 l/h and determination of oxygen content every 1 or 6 min using a paramagnetic oxygen analyzer (12). For determination of lower critical temperature, ambient temperature was increased stepwise from 20 to 37°C in 1-h intervals as indicated in Fig. 3. Oxygen consumption was determined every minute, and the five lowest values were averaged. Lower critical temperature was calculated as described previously (12). Lowest measured metabolic rates at thermoneutrality (30-32°C) were considered to represent BMR. ADMR was determined by recording oxygen consumption every 6 min over a 24-h period of animals subjected to the normal light-dark schedule and supplied with food ad libitum and apple as a water source.

Body temperature. Core T_b was measured as described previously using temperature-sensitive radio transmitters (Mini-Mitter) of ~1 g weight that were implanted into the visceral cavity of anesthetized animals (27). After a recovery period of 3-4 days, T_b was recorded every 6 min over a 1-wk period.

BAT biochemistry and carcass analysis. Animals were killed by cardiac puncture after anesthesia with CO₂. The interscapular fat pad was excised, and cytochrome c oxidase (COX) activity and UCP were determined as a measure of thermogenic capacity. Transgenic animals displayed large lipid accumulation in the interscapular fat pad, which made it almost impossible to discriminate between white and brown adipose tissue. We therefore prepared a homogeneous batch of the whole fat pad from transgenic and control animals by grinding the tissue to powder in liquid nitrogen. Aliquots of this powder were homogenized in a phosphate buffer. COX activity of the homogenates was determined polarographically as described (35). For quantitation of UCP content, Western blots of 10 µg total homogenate protein were performed as described (15) using an enhanced chemoluminescence Western blotting detection system (Amersham) according to the manufacturer's protocol. Signals were quantified by densitometric scanning of the films.

Body composition was determined of the same animals used for BAT analysis. The gastrointestinal tract was removed before carcass analysis. Carcass weight is therefore animal weight without the interscapular fat pad and gastrointestinal tract. Carcasses were dried to constant weight at 65°C and fat content determined by extraction of lipids with chloroform using a Soxhlet apparatus. Fat mass was calculated as the difference between dry weight before and after chloroform extraction. Lean body mass is carcass weight minus fat mass.

Statistical analysis. All results are represented as means ± SE. Statistical significance was assessed by unpaired Student's t-test when appropriate. When data showed no homogeneous variances, a parameter-free analysis (Mann-Whitney U test) was performed. Significant differences between control and transgenic animals were assumed at P < 0.05.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

BAT thermogenic capacity. In UCP-DTA transgenic mice, BAT is partially destroyed by directed expression of DTA in brown adipocytes only. The interscapular fat pad, which contains BAT, was almost three times heavier in transgenic animals compared with controls (Fig. 1A). This was mainly due to increased lipid deposition, because the protein content of this fat pad was not different between transgenics and controls (Fig. 1A). Thermogenic capacity of the tissue, however, was largely reduced in transgenic mice. The COX activity as an indicator of total respiratory capacity was only at 21% of control levels. UCP was even reduced to 15% of control values (Fig. 1B), corroborating previously published data (19).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Brown fat biochemistry of uncoupling protein-diphtheria toxin A chain (UCP-DTA) transgenic mice and control FVB mice. Total interscapular brown adipose tissue (BAT) depot was excised and homogenized for analysis. A: tissue weight and protein content. B: thermogenic activity assessed by measurement of cytochrome c oxidase (COX) activity and amount of UCP. Data are means ± SE of 7 control and 5 transgenic animals. au, Arbitrary units. * Significant difference (P < 0.05).

Body weight and composition, food assimilation, and locomotor activity. UCP-DTA mice were ~70% heavier than control animals, i.e., they showed a pronounced obesity as observed in earlier studies (10, 19). To a large extent this was due to an increased fat mass because carcass lipid content was 22.2% in transgenics compared with 14.5% in controls (Table 1). However, fat-free dry mass and corresponding water content were also increased, resulting in a significantly higher absolute lean body mass of UCP-DTA mice, even if relative lean body mass was decreased in the transgenics (Table 1). Food consumption on the other hand was not significantly different between transgenics and controls (Table 2). We determined the efficiency of food assimilation by measuring the difference of energy content in consumed food and produced feces. As shown in Table 2, this was also not different in transgenics and controls. This means that both groups assimilated approximately the same amount of energy per day. Total locomotor activity was not significantly different between UCP-DTA mice and controls (Table 2). As shown in Fig. 2, transgenic animals showed a tendency to a decreased nighttime activity. However, because of a rather high individual variability, there were no statistical differences.

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Locomotor activity of UCP-DTA transgenic mice and control FVB mice recorded by infrared detectors. Animals were housed at 24°C with a 12:12-h light-dark cycle. Activity was recorded over a period of 2 wk. Data are means ± SE of 12 control and 9 transgenic animals.

Metabolic rate. The thermoneutral zone of UCP-DTA mice had so far not been determined. We therefore measured metabolic rate, i.e., oxygen consumption at ambient temperatures between 20 and 37°C. As can be seen in Fig. 3, both control and UCP-DTA mice had lowest metabolic rates at 30-33°C and showed a pronounced increase in metabolic rate at temperatures below 29°C. Melnyk et al. (20) raised UCP-DTA mice at 35°C, which they claimed to be thermoneutral. In our hands, mice did not support temperatures over 33°C very well; they showed evident signs of heat stress. However, they might be able to adapt to 35°C when raised at this high temperature. At all measured temperatures, the weight-specific metabolic rates of transgenic animals were significantly lower than in control animals. From the slope of the increase in metabolic rate below 29°C and the values of lowest metabolic rate, we could calculate the lower critical temperature, which was 29.4 and 29.7°C in control and UCP-DTA mice, respectively, and there was no significant difference between both groups.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Resting oxygen consumption of UCP-DTA transgenic mice and control FVB mice at different ambient temperatures (Ta). LCT, lower critical temperature. Data are means ± SE of 11 control and 8 transgenic animals.

Figure 4 shows the BMR in comparison to ADMR at 24°C and at thermoneutrality (30°C). Total oxygen consumption per animal was not different between the two groups (Fig. 4A). However, transgenic animals were much heavier than controls, and UCP-DTA mice showed a reduction in BMR and ADMR at thermoneutrality of almost 50%, when expressed on a weight-specific basis (Fig. 4B). It can be argued that UCP-DTA mice have an increased body fat content, i.e., an increase in metabolically less active tissue. Therefore, we also calculated metabolic rate per gram of lean body mass (Fig. 5). Although controls have a higher percentage of lean body mass, absolute lean body mass is still higher in transgenics (Table 1), resulting in reduced metabolic rates per gram of lean body mass.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Metabolic rates of UCP-DTA transgenic mice and control FVB mice. A: oxygen consumption per animal. B: oxygen consumption per gram body weight. Data are means ± SE of 11 control and 8 transgenic animals. BMR, basal metabolic rate (resting oxygen consumption rates at thermoneutrality, i.e., 30°C); ADMR, average daily metabolic rate (determined by 24-h measurements of oxygen consumption at 24 and 30°C, respectively).

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Metabolic rates of UCP-DTA transgenic mice and control FVB mice expressed per gram lean body mass. A lean body mass of 68.0% and 63.4% of body weight was assumed for control mice and transgenics, respectively (see Table 1 for details). Further explanations in legend of Fig. 4.

Body temperature. Because the lowered ADMR could not be attributed to a decreased locomotor activity of UCP-DTA mice, we measured morning rectal temperature over a period of 2 wk. As can be seen in Table 2, rectal temperature was significantly reduced by 0.9°C in the transgenic animals. To obtain a more continuous information on core T_b, we implanted temperature-sensitive radio transmitters into the visceral cavity of control and transgenic animals, allowing continuous records of T_b without disturbance of the animals. Figure 6 depicts a double daily plot of T_b rhythms of a control and a transgenic mouse throughout 7 days. A pronounced daily rhythm with higher temperatures at night (activity phase) could clearly be seen in control animals. T_b of UCP-DTA mice was regulated at a significantly lower level. Figure 7 shows the T_b range of control and transgenic animals. It is evident that, although the overall range of T_b was similar, UCP-DTA mice showed a pronounced downshift of T_b range of ~0.9°C, corroborating the results from rectal temperature measurement. Mean T_b and daily minima and maxima of T_b were significantly lower in UCP-DTA mice.

View larger version (33K): [in this window] [in a new window]   Fig. 6.   Exemplary recording (double daily plot) of core body temperature (Tb) of 1 control and 1 UCP-DTA mouse over a consecutive 7-day period. Core Tb was measured telemetrically with implanted radio transmitters; animals were kept at 24°C. Black bars correspond to night (light off) period.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Core Tb ranges of control and UCP-DTA mice measured telemetrically with implanted radio transmitters over a 7-day period. Shown are mean Tb ± SE and the range of Tb with daily minima and maxima ± SE of 6 control and 4 UCP-DTA mice kept at 24°C. All mean values are significantly different between transgenics and control mice (P < 0.01).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Many states of altered energy balance, such as obesity, are associated with changes in BAT thermogenesis (13). Already in 1979 it was proposed by Rothwell and Stock (25, 26) that increased BAT thermogenic activity represents a prevention of obesity when a highly palatable high-calorie diet is fed. This hypothesis of the so-called diet-induced thermogenesis was based mainly on circumstantial evidence until Lowell and co-workers (19) created the UCP-DTA transgenic mice, which indeed became obese because of largely ablated BAT function. Here we tried to analyze which compartment of energy metabolism was directly affected by the BAT malfunction.

UCP-DTA mice had the same net energy intake as controls; neither food intake nor the assimilation capacity was increased. The fact that they did not show hyperphagia is somewhat contrary to previously published data (10, 19, 20). This might be due to experimental conditions. Also it should be pointed out that we used adult, already obese animals that had a stable body weight over the investigation period and thus a balanced energy budget. Total locomotor activity was also not significantly changed in transgenics. Interestingly, they seemed to display an attenuated circadian activity pattern; however, because of large individual differences this was not statistically significant (Fig. 3).

To assess energy expenditure, we measured BMR, i.e., lowest oxygen consumption at thermoneutrality, and ADMR at thermoneutrality (30°-32°C) as well as normal housing temperature (24°C). None of the parameters were different between transgenics and control when expressed per animal. However, the transgenics were much larger animals with higher lean body mass as well as body fat. Weight-specific metabolic rates were thus consistently reduced in UCP-DTA mice by up to 50%. In a recent discussion the problem of normalizing metabolic data for animals of different body mass was addressed (14). It has been shown for a long time that the BMR of mammals can be normalized using body weight raised to the power 0.75 (17). However, this is applicable only on an interspecies basis for animals of roughly similar body composition. The problem with an obesity model is the increased body fat, which represents metabolically inert tissue. On the other hand, the excess fat mass must be kept at normothermia and carried around by the animal, resulting in higher energetic costs. This is important taking into account that UCP-DTA did not show a reduced locomotor activity. It is well known from human studies that overweight and obese people have an increased energy expenditure, because energy expenditure is closely related to lean body weight (22). If we take lean body mass as the basis for normalizing metabolic rate, it is clear that UCP-DTA mice displayed a reduced energy expenditure compared with controls because their lean body mass was also increased.

The reduced metabolic rate was clearly not due to a decreased locomotor activity in transgenic mice but rather to lowered T_b and thus lower energy expenditure by thermogenesis for maintenance of endothermia. The energy savings in reducing T_b by only 0.9°C could be ~20% if mice are living at their respective lower critical temperature (29.4 and 29.7°C). This would reduce the gradient between T_b and ambient temperature from 6.9 to 5.7°C (i.e., by nearly 20%) and, accordingly, basal energy requirements. Indeed, the reduction of metabolic rate was most evident at thermoneutrality. Interestingly, Melnyk and co-workers (20) could abolish the development of obesity in UCP-DTA mice by raising them at 35°C. At this ambient temperature, UCP-DTA mice must have a higher T_b than the 35.5°C measured in our study and thus a higher metabolic rate. Because 35°C is above the thermoneutral zone we measured, they might also be subjected to heat stress, which could prevent development of obesity.

Perspectives

The possible significance of shifts in T_b regulation for the development of obesity has so far not been thoroughly investigated. However, it was shown that ob/ob mice which have a defective leptin gene and develop massive obesity displayed a significant hypothermia (21). Treatment of these animals with recombinant leptin normalized T_b and resulted in loss of body weight (21). It has also been reported that leptin treatment of suckling rats abolished states of hypothermia, resulting in a reduction of fat stores (30). Leptin was shown to enhance sympathetic outflow to BAT (2, 11) and increase UCP mRNA levels in BAT (29, 34). These findings provide a link between leptin action on energy metabolism and BAT thermogenesis. Recently, we also showed that leptin gene expression in fat of Siberian hamsters is low in short photoperiod, which is a prerequisite for hamsters to show daily torpor, i.e., prolonged states of hypothermia (18).

It seems that the important role of brown fat in energy metabolism is not due alone to its UCP function. Transgenic mice lacking UCP were found to be more cold sensitive but did not develop obesity (6). However, these mice showed an increased, compensatory expression of UCP-2, an uncoupling protein expressed in many tissues with high homologies to the brown fat UCP (8). Another transgenic mice model lacking norepinephrine and epinephrine was reported to have a hypoactive BAT thermogenesis. These mice were hyperphagic but did not become obese because of an elevated BMR (31). It could be speculated that brown fat has an effect on overall metabolic rate independent of the sympathetic nervous system-UCP axis, maybe due to unknown secreted factors acting on other tissues.

Under our experimental conditions, adult UCP-DTA mice represent a model of maintained obesity solely due to a reduction of obligatory heat production, i.e., basal energy expenditure. BAT probably does not play a role in the development of human obesity. Reduced energy expenditure, however, has been shown to be a risk factor for weight gain in humans (23). Moreover, also in humans concomitant interindividual variations in T_b and metabolic rate have been reported (24). The recent finding that uncoupling proteins similar to UCP are expressed in tissues other than brown fat (1, 8, 9, 33) opens new research perspectives on the role of T_b regulation and energy expenditure in the development and subsistence of human obesity.