Gender dimorphism of body mass perception and regulation in mice

The set-point theory of body mass regulation assumes that maintenance of an individual body mass is the result of a regulated mechanism(Cabanac, 2001). It is suggested that changes in body mass can be perceived by the body, which generates an endogenous signal, the ponderostat, which is correlated with the present body mass. Changes in the extent of the ponderostat signal following deviations from set-point body mass activate feedback control mechanisms,which in turn act on food intake and energy expenditure to regenerate the individual set-point mass (Cabanac et al.,1971; Bradley,1978; Keesey and Corbett,1984; Harris,1990; Cabanac,2001). Failure to identify the ponderostat(Macdonald, 2003; Berthoud, 2002), and the fact that changes in body mass are mainly reflected by changes in fat mass(Harris, 1990), led to the proposition that body mass regulation depends on a lipostatic feedback system,with leptin as a lipostatic signal(Friedman and Halaas, 1998). However, the widely accepted hypothesis of lipostatic regulation of body mass has been disputed (Speakman et al.,2002; Lacy et al., 2003). There is evidence that gravity and mass per se also influence body mass regulation. Under conditions of hypergravity induced by centrifugation, body mass of rats was decreased through loss in fat mass and energy expenditure was increased; energy intake initially decreased and returned to normal or increased levels in the long-term (Warren et al.,1997). Furthermore, it has recently been reported that intraperitoneal implantation of a metabolically inert weight into male deer mice Peromyscus maniculatus resulted in a loss of body mass(Adams et al., 2001).

Most mammals show gender dimorphisms in body mass and composition, and there is ample evidence that energy homeostasis and substrate utilization are regulated differently in males and females(Cortright and Koves, 2000). Very recently it was shown that male and female rats display differential sensitivity to central insulin and leptin levels with regard to food intake regulation (Clegg et al.,2003). However, even extensive discussions of body mass regulation do not address the possibility of gender differences in the basal regulatory mechanisms (Jequier and Tappy,1999).

We therefore tested the hypothesis that body mass per se is perceived and regulated, and that there might be gender differences in this regulation. In order to increase body mass artificially, we implanted inert weights corresponding to 10% of initial body mass into the abdominal cavity of mice. If mice are able to detect the increase in total body mass(M_TB) and there exists an individual set point of body mass we would expect them to show a compensatory decrease in biological body mass (M_BB, body mass exclusive of the weight load) to restore the set-point M_TB. To control for surgical stress and the volume effect of the implant, we included a sham operated group (SO)and a group with an implant of the same volume, but lower mass, corresponding to only 2–3% of body mass (LO), in addition to the 10% implant group(HI).

All experiments were performed according to the federal guidelines of Brandenburg Ministry of Agriculture and Environment. Male and female mice(FVB, in-house breed) were maintained under standard conditions at 25°C and a 12 h:12 h dark:light cycle. They had free access to water and standard rodent chow diet containing 19% protein, 4% fat and 50% carbohydrates(Altromin 1324 fortified, Lage, Germany). Mice were 7 months old at the time of weight load implantation. Body mass and energy intake were recorded regularly for at least 4 weeks prior to weight implantation. As basal values of physiological parameters the following measurements were used: energy intake, mean over 3 weeks; energy expenditure, mean over 2 weeks; body mass,mean over 1 week before weight implantation; body composition, immediately prior to implantation of the weight load. Because of technical limitations not all animals were subjected to the same measurements; animal numbers were 12–34 for each gender and parameter and exact numbers are reported in the figures and tables.

Animals were divided randomly into three treatment groups: a sham-operated group (SO), a group that received light weight loads (LO) and a group that received heavy weight loads (HI). LO and HI weight loads corresponded to 2–3% and 10% of initial body mass, respectively. Implants were composed of a rod-shaped core (1.4 cm length, 0.8 cm diameter) and a wax coating (Elvax Wax, Minimitter Co., Sunriver, OR, USA). The core of LO weight loads consisted of plastic tubing filled with cotton. HI weight load cores consisted of a metallic cylinder. For implantation of the weight loads, mice were anaesthetized with ketamin (1 μl g^–1; Ketamin Gräub,A. Albrecht, Aulendorf, Germany) and xylazinhydrochlorid (0.1 μl g^–1; Rompun: BayerVital, Leverkusen, Germany). Weight loads(see below) were disinfected and implanted into the abdominal cavity. The abdominal cavity was sutured using absorbable surgery thread (PGA Resorba,Resorba, Nürnberg, Germany), and skin was closed with clips (Becton Dickinson, Sparks, MD, USA) that were removed 1 week after the operation. SO animals were treated exactly like the implantation groups except that no weight was introduced. Weights were removed 14 weeks after implantation.

Body mass and energy intake were measured 3 times per week throughout the experiment by weighing (BP 2100, Sartorius AG, Göttingen, Germany;detection limit 0.01 g). We define total body mass (M_TB)as the measured mass of the animals, i.e. including the implant, and biological body mass (M_BB) as the measured mass minus the implanted weight load.

Resting energy expenditure (E_RE) was defined as the mean of the 10 lowest daily values for energy expenditure(Klaus et al., 1998). Respiratory quotient (RQ; V̇_CO2/V̇_O2)and physical activity level (PAL; E_TE/E_RE) were calculated.

Body composition was measured at the time of surgery, prior to implantation, and when the weight load was removed after 14 weeks, using dual-energy X-ray absorptiometry (Lunar Piximus, Janesville, WI, USA) in anesthetized animals.

Most of the data are expressed relative to basal values measured 3 weeks(energy intake), 2 weeks (total energy expenditure, resting metabolic rate)and 1 week (body mass) prior to implantation, or at implantation of the weight load (fat mass, lean body mass), corresponding to 100%. Data are shown as means ± s.e.m. To achieve normal distribution, data for body mass values at 1 week after weight load and for energy intake were log-transformed. Baseline characteristics were compared using pooled or separate variances t-test for equality of means where applicable. Homogeneity of variances was tested by Levene's Test for equality of Variances.

Single measurement data were analyzed using analysis of variance (ANOVA),and Dunnett's-test was used as a post-hoc test for multiple comparisons against SO group. Development of body mass, energy intake, and energy expenditure were analyzed using repeated-measures ANOVA. Gender, mass implantation group and an interaction term between group and gender were included in the ANOVA models. To exclude any effects of post-surgical stress in these analyses, body mass and energy intake starting 2 weeks after surgery were included. Significant differences within the sexes were tested using one-way ANOVA including mass implantation group as independent factor. Significance was assumed at P<0.05. Analysis was performed using SPSS 8.0 (SPSS Inc., 1998, Chicago, IL, USA).

Phenotypic characterization of male and female mice before implantation revealed a number of gender differences(Table 1). Males showed higher levels of body mass, energy intake, absolute lean body mass, absolute fat mass, total energy expenditure and resting metabolic rate compared to females. RQ was similar in males and females but PAL was higher in females than in males. Initial body mass and gross energy intake were similar in all treatment groups of male and female mice (Table 2). In order to render the results more comparable between males and females, we subsequently normalized most of the data relative to basal measurements.

Fig. 1 shows development of M_TB, M_BB and energy intake throughout the experiment. Acutely after surgery, all groups showed a reduction in M_TB and M_BB(Fig. 1A,B). Body mass reduction after 1 week of weight load was more pronounced in HI mice than SO mice (M_TB, P<0.001; M_BB, P<0.001). M_TB of male HI mice was 99.6+1.1%,and was the same as M_TB of SO males (P=0.109),whereas M_TB of female HI mice was 102.7+0.8% of the initial value and significantly higher than in SO females(P<0.001). Male mice thus showed complete initial compensation of M_TB, in contrast to female mice, and this effect was still apparent 2 weeks after weight load (see Fig. 1A).

Further change in M_TB was influenced by gender(P<0.001) and by the mass of the implant (P=0.016). HI mice of both genders showed significantly increased M_TBthroughout the study period compared to SO mice(Fig. 1A).

Changes in M_BB were significantly affected by group(P=0.004) and sex (P<0.001). Up to week 4, M_BB was significantly reduced in HI males compared to SO,whereas from week 5 on, changes in M_BB were similar in all male mice. However, the M_BB of HI males was permanently slightly reduced, by about 3–5% compared to SO and LO mice. In females, M_BB was similar in all groups from week 2 on. Interestingly, from 8 weeks on, female LO mice showed the highest M_BB (Fig. 1B) resulting in M_TB levels comparable to those of HI females (Fig. 1A).

The gender differences are more obvious in Fig. 2, which shows the time needed to regain initial body mass. Females recovered initial body mass earlier than males (P=0.003). Male HI mice took longer to recover body mass compared to SO males (P=0.011). In females the weight of the implant had no effect on the time required for body mass recovery.

During the first week, energy intake was significantly decreased in both genders (Fig. 1C). There were significant effects of group (P=0.026) and sex (P<0.001)on reduction of energy intake after 1 week. Males decreased their food intake to lower levels than females. Energy intake of male mice after 1 week was 80.1±3.2% in SO, 61.5±1.7% in LO, and 54.5±3.1% in HI relative to basal values. Energy intake of female mice after 1 week was 77.8±2.8% in SO, 67.4±3.8% in LO, and 66.1±3.6% in HI relative to basal values. From the second week on, energy intake of males and females was similar among the implantation groups. Increased energy intake following weight load was more pronounced in females of all groups than in males (P=0.042), presumably because of slightly higher values in HI females. However, comparison of the different female groups did not reveal significant differences; P-values for a given time point were between 0.1 and 0.2 for comparison between HI and SO (Dunnett's-test).

When cumulative energy intake from weeks 1 to 14 was calculated, no significant effect of gender could be detected(Fig. 3). The mean cumulative energy intake of LO and HI males seemed to be lower compared to SO males but was only significant for LO males (P<0.05), probably due to lower variation of values in the LO group. Although HI females displayed higher cumulative energy intake compared to SO and LO females there are no significant differences between female groups.

Fig. 4 shows changes in body composition 14 weeks after weight load in relation to initial values. Increase in fat mass at 14 weeks after weight implantation was lower in males than in females (Fig. 4; P=0.046). SO and LO males increased their fat mass compared to initial fat mass whereas HI males maintained their initial fat mass, the differences reaching borderline significance (P=0.052). In contrast,females increased fat mass in all treatment groups and there were no differences between the groups (P=0.456). Lean body mass was not affected by gender or the implanted weight and did not change compared to initial values (Fig. 4B).

E_TE was mainly determined by E_RE(data not shown). E_RE was influenced by gender but not by the implanted weight load (Table 3). Absolute E_RE was lower in females than in males (P=0.006). Following weight load females increased E_RE to a higher extent than males relative to initial levels (P=0.001). Due to the high variability in E_RE, no significant effect of the implanted weight could be detected in male and female mice. Physical activity level (PAL) was significantly affected by sex (P=0.041) with higher values in females that were consistent with baseline measurements(Table 1), but it was not influenced by the implanted weight (Table 3). Respiratory quotient (RQ), i.e. substrate oxidation, was not affected by weight load or by gender (data not shown).

The present results show that male and female mice react differently to the augmentation of body weight by implantation of an artificial weight. In the medium and long term, HI males displayed slightly decreased M_BB compared to SO males; by contrast in females there was clearly no difference in M_BB between the treatment groups,except for a slightly increased M_BB in LO females in the second half of the study. Long-term changes in body mass could not be explained by measurable changes in energy intake and energy expenditure.

Both genders showed an acute reduction in food intake accompanied by a reduction in body mass, which was obviously due to post-surgical stress. Male mice carrying the heavy implant decreased M_BB during weeks 1 and 2 following weight load to an extent that M_TBcorresponded to basal values. This observation seems to comply with the set-point theory of body mass and a mass-dependent regulation of body mass. However, this effect was only transiently present as M_TBincreased thereafter and became significantly higher compared to SO males. In addition, the transiently observed recovery of the basal body mass was not present in HI females, which acutely lost less M_BB and regained it rapidly to levels similar to SO and LO females. HI males displayed a permanently lower M_BB following weight load compared to SO males, the difference accounting for about half of the implanted mass (5%). Although not significant, the permanence of the reduction seems to suggest a real effect. Therefore, in the long-term there might be only a partial compensation of the increased body mass and only in male and not female mice.

The partial compensation observed in males occurred mainly at the expense of fat mass; lean body mass was not changed. This seems reasonable considering that a certain muscle mass is necessary to move an acutely heavier body mass. In addition, since fat contains much more energy per mass unit compared to lean mass, fat loss due to negative energy balance would lead to a lower overall body mass reduction than a loss in lean mass.

Energy intake was not affected by the additional weight in the long-term. After the initial reduction, male mice of all groups returned to their initial levels of food intake by the second week after implantation. One could suggest that the lack of an increase in energy intake in HI males that compensated for the acute weight loss led to the slightly decreased M_BBcompared to SO males. However, HI females also showed normal energy intake but displayed comparable M_BB to SO and LO females. This suggests that males are either more sensitive in their perception of an increased body mass or that they showed a higher stress response to the additional weight compared to females. There are some studies to support the first suggestion. Firstly, our observations in male HI mice confirm the results of Adams et al. (2001),who described a compensatory decrease in M_BB following intraperitoneal implantation of 1–3 g weights into male deer mice Peromyscus maniculatus. This reduction was paralleled by a reduction in food intake, which was significant only for the heaviest implant group (3 g). The authors suggested the existence of a `mechanical set point', referring to the loading of the musculoskeletal system that leads via unknown mechanisms to perception of the animal's body mass. However, they did not show a time course of body mass change or energy intake and investigated male mice only for a period of 5 weeks. The overall reduction in food intake they described could thus have resulted from the initial, stress induced reduction after surgery rather than being the result of a regulated process to approach the set-point mass. In addition, the authors described that the reduction of M_BB in the 3 g implanted group corresponded to about 1.5 g. This represents a M_BB reduction of half the implanted mass, similar to our study. Therefore, the conclusions drawn from our study could also be applied to the study by Adams et al.(2001).

Secondly, several studies have examined the effect of hypergravity induced by centrifugation, thereby multiplying the mass load on the body. Overall they show that hypergravity leads to an acute decrease in body mass in rodents,associated with acutely decreased energy intake and increased maintenance feed requirements (Wade et al., 1997, 2002; Warren et al., 1997). The decrease in body mass resembles that of male HI mice in our study. In fact,most of these studies were done in male rodents only. One study that included male and female rats showed that female rats raised in hypergravity (2.5 g) showed a much less pronounced body mass reduction than male rats (Wubbels and de Jong,2000), suggesting a stricter maintenance of M_BB in females similar to our results.

Thirdly, after i.p. implantation of a telemeter with a weight corresponding to about 12% of initial body mass into male Swiss Webster mice, M_BB was acutely reduced by the mass of the implant but was increasing above basal levels by the end of a 14 day period(Perry et al., 2000). However,a control group was not included in this study.

It seems that at least in male rats and mice an acutely increased body mass is partially compensated by a reduction in M_BB, although to a lower extent than would be expected from the mass of the implant if the maintenance of a mass-specific set-point is assumed. This suggests that a mass-specific set-point regulation of body mass is not accurate or might be impaired by still unknown competing mechanisms. The fact that males in general needed longer to recover their initial M_BB than females could point to a higher stress susceptibility of male mice.

We expected an increase in energy expenditure in HI mice, considering the increased energy demand needed for carrying the additional weight. However, E_RE was only slightly and not significantly increased in HI animals 2 weeks after weight implantation. The failure to detect significant group differences in energy expenditure is probably due to the small sample size, but also to the high intra-individual variability of energy expenditure. It should be emphasized that only very small effects on energy expenditure could be expected, since the maximum weight load was only 10% of body mass, i.e. around 2–3 g. This is not much considering that mice normally display daily mass fluctuations in the range of 1–2 g according to our own measurements (not shown). Interestingly, physical activity was apparently also not affected by implantation, as evident from the PAL values. However, in mice implanted with a telemetry transmitter, their willingness to practice in running wheels decreased after i.p. implantation(Perry et al., 2000),suggesting a decrease in physical activity. In contrast, subcutaneous implantation in rats of telemetry transmitters corresponding to 15% of the animals' initial mass did not impair activity levels during a 5 h measurement period (Moran et al., 1998)supporting the observation that energy expenditure was not influenced by the additional weight. These different observations suggest that the site of implantation (intraperitoneal versus subcutaneous) might have an impact on physical activity and overall energy expenditure. Intraperitoneal implantation seems rather more likely to impair physical performance and energy expenditure than subcutaneous implantation. In addition, studies in humans carrying additional weights suggested that an increase in energy expenditure occurs only if a certain mass threshold is obtained, usually higher than 10% of initial body mass(Maloiy et al., 1986; Jones et al., 1987). Possibly,artificial augmentation of body mass by 10% in mice in the present study lies below this threshold and thus fails to increase energy expenditure due to increased physical work required to move the body.

Data on changes of energy expenditure and physical activity following weight implantation remain contradictory, depending on the species, implant mass, and site of implantation, and need further investigation. In addition,direct measurements, e.g. by use of cages equipped with infrared beams, could be more useful for evaluation of physical activity than a calculation of PAL.

The gender-specific differences in body mass point to different strategies in males and females to cope with a situation affecting energy demands and body mass. There are a several rodent and human studies to support this suggestion. In addition to gender-specific responses in energy balance under hypergravity conditions as mentioned above, different catecholamine responses to space flight, i.e. in microgravity, have been reported in male and female astronauts (Stein and Wade,2001), pointing to gender-specific hormonal effects on body mass regulation. Gender specific responses were also reported in rats subjected to a change in energy expenditure by forced and voluntary exercise. Male rats decreased body mass under the influence of forced exercise whereas female rats increased energy intake and thereby maintained their body mass(Nance et al., 1977; Cortright et al., 1997). Interestingly, similar effects were observed in humans subjected to different levels of weekly fitness training. Despite an increased daily energy expenditure, men did not increase energy intake to compensate for the energy loss (Stubbs et al., 2002a),while women at least partially increased their energy intake(Stubbs et al., 2002b). In the present study, HI females showed a more efficient gain in body mass (both M_TB and M_BB) compared to HI males,although energy intake was not significantly increased in HI females compared to SO and LO females. This implies a higher efficiency of energy utilization and conservation of females compared to males. This seems reasonable considering the main evolutionary responsibility of females for reproduction(Cortright and Koves, 2000; Hoyenga and Hoyenga, 1982) and the necessity to maintain adequate energy stores throughout gestation and suckling period. It was suggested that this higher efficiency is a result of higher selection pressures on females during evolutional development(Cortright and Koves,2000).

Another interesting, gender-specific phenomenon is our observation that LO females showed increased M_BB in the second half of the study period compared to SO and HI females and also displayed the highest mean fat increase after 14 weeks. Obviously, in females but not in males the implant volume has an impact on energy balance. It is conceivable that the implant leads to an abdomen distension similar to that experienced during gestation, causing metabolic responses to maintain and increase energy stores rather than to maintain a certain body mass set-point, which in our model would mean to decrease energy stores. During gestation this regulation would ensure an adequate energy supply for the offspring. It is therefore possible that there are two different systems competing with each other: mass-specific set-point regulation of body mass overlapping with a volume-specific response that increased energy resources supposedly by mimicking gestation. The latter should be female-specific and hence would not appear in males.

We found gender-specific responses following an acute, artificial body mass increase. Perception of body mass per se and a compensatory decrease in M_BB following artificial body mass augmentation appeared only partially and in males rather than in females. This compensation happened mainly at the expense of fat mass; lean body mass was maintained at basal levels in both genders. Long-term energy intake was not affected by the weight implantation. Also measurements of energy expenditure revealed no significant influence on body mass, possibly because the sensitivity of the measurement was insufficient to detect the range of actual changes. In females the volume of the implanted weight seemed to have a more important impact on body mass development than the actual weight. It therefore seems that body mass per se might not be a major player in the set-point regulation of body mass. In females especially, the distension of the peritoneum induced by the volume of the weight could have caused the observed gender-specific body mass development. The real influence of this suggestion could be further investigated by implanting weights into other sites of the body different from the peritoneum. Overall, our results do not support the set-point theory of body mass control with weight per se being regulated. It seems that different compartments of body mass are perceived and regulated differently. Notably, the strict conservation of lean body mass compared to the evident changes in body fat mass suggests that the ponderostat is rather linked to lean body mass, thus arguing against a predominant lipostatic regulation of body mass.

We thank Carola Plaue for excellent technical assistance and Dr Cornelia Koebnick for statistical advise. We are grateful to Dr Sylvia Ortmann and Andreas Subke for the design and manufacture of the metal implant cores,respectively.