Photoperiodic effects on body mass, energy balance and hypothalamic gene expression in the bank vole

It is widely believed that body fatness and hence body mass are regulated by a neuro-hormonal system, which compares actual body fat with a target that is encoded in the brain (Schwartz et al.,2000; Mercer and Speakman,2001; Morgan and Mercer,2001). Differences between actual fatness and the target are then corrected through compensatory adjustments in both energy intake and expenditure (Kalra et al.,1999; Schwartz et al.,2000; Jeanrenaud and Rohner-Jeanrenaud, 2001; Mercer and Speakman, 2001):the so-called lipostatic regulation system (after Kennedy, 1953). Many components of this system have been characterised in small mammals, and their importance in the aetiology of human obesity was subsequently demonstrated(Montague et al., 1997). One component of the system, however, that has proved elusive is the nature of the encoded target fat mass to which compensatory adjustments are made. Understanding the nature of this target is important because genetic polymorphism within this target system, resulting in a failure to adequately regulate body fatness, is a potential contributory factor to human obesity(Barsh et al., 2000).

A model system that has proved useful in the characterisation of the differences between compensatory adjustments in body mass, which tend to keep body mass at a fixed target level, and anticipatory adjustments, which involve changes in the target itself (Mercer,1998; Morgan and Mercer,2001; Mercer and Speakman,2001), is the photoperiod-induced change in body mass of Djungarian and Siberian hamsters (Phodopus sungorus) (Mercer et al., 1995, 1997, 2000, 2001; Boss-Williams and Bartness,1996; Reddy et al.,1999; Adam et al.,2000). These animals demonstrate profound (30-35%) reductions in body mass (mostly fat) when exposed to short photoperiods. The generality of this model, however, is unknown, as is its applicability in other less profound photoperiod-induced body mass changes. In the current study, we sought to characterise the patterns of change in both energetic parameters and neuropeptide gene expression during photoperiod-induced changes in the body mass of the bank vole Clethrionomys glareolus, a small arvicoline rodent found throughout Europe.

Like many species of small mammal from the temperate zone(Iverson and Turner, 1974; Plunkett et al., 2000), bank voles reduce their body mass during winter(Klaus et al., 1988). This strategy is suggested to potentially offset some of the negative impacts of winter on energy balance, which include higher thermoregulation costs compounded with a reduction in food quantity and/or quality(Dark and Zucker, 1983). By reducing body mass, it is suggested that animals offset these increased energy demands by reductions in their resting metabolic rate, and the unit costs of thermoregulation and physical activity, thus reducing total daily energy requirements (Nagy and Negus,1993). After winter, body mass increases again in response to increased day length (Steinlechner et al.,1983). Two interesting questions arise concerning energetic adjustments during the seasonal cycle of body mass. First, what aspects of energy balance does the animal modify to effect the changes in body mass in response to photoperiod changes, and how are these signalled in the brain?Second, once the transitions have occurred, what is the magnitude of the effect of changes in mass on energy demands, what energetic adjustments does the animal perform to meet these demands and thus defend its body mass at the elevated level, and what changes in the brain encode the altered mass that is defended? This paper focuses predominantly on the latter of the two questions.

If, as is widely hypothesised, the decrease in mass under short photoperiods serves to save energy by reducing resting metabolism, then the increase in body mass under long photoperiods is likely to result in elevated resting metabolism. The extent of this increase will depend on the changes in mass of specific body components and their contribution to resting metabolic rate (RMR). Most studies suggest that adipose tissue plays a relatively small part in metabolic activity, when compared with organs such as the alimentary tract and liver (Field et al.,1939; Krebs, 1950; Schmidt-Nielsen, 1984; Selman et al., 2001; but see Geluso and Hayes, 1999; Speakman and Johnson, 2000). To accommodate the predicted rise in RMR, as voles move from low to high body mass, they may modify other components of expenditure, such as that devoted to physical activity (Stebbins,1984), to maintain a constant total energy expenditure. Alternatively, energy absorption may be increased, either through increased food intake (Dark and Zucker,1983) or by increasing absorption efficiency(Speakman and McQueenie, 1996; Nagy and Negus, 1993). Finally, animals may utilise a combination of these approaches.

The alterations in energy balance and thus changes in body and fat mass in response to photoperiod are probably signalled predominantly in the hypothalamus, which has long been established as a site responsible for control of feeding behaviour, particularly the arcuate (ARC), ventromedial(VMN) and paraventricular (PVN) hypothalamic nuclei. Peptidergic neurons within these nuclei produce a number of neuropeptides involved in energy balance regulation (Schwartz et al.,2000), which can be divided into two categories. Anabolic neuropeptides, such as neuropeptide Y (NPY) and agouti-related protein (AGRP),increase energy intake and decrease energy expenditure(Stanley et al., 1986; Boss-Williams and Bartness,1996; Hagan et al.,2000), and as a result, body fat stores and thus body mass increase. Catabolic neuropeptides, such as melanocortins cleaved from the proopiomelanocortin (POMC) precursor molecule, as well as cocaine- and amphetamine-regulated transcript (CART) and corticotrophin-releasing factor(CRF), suppress food intake and increase energy expenditure(Bray et al., 1990; Kristensen et al., 1998; Tritos et al., 1998), which results in a decrease of amount of fat stored and body mass loss.

In the present study, we examined the effect of increased photoperiod on the energy balance and body composition of the bank vole. Simultaneously, we quantified hypothalamic mRNA levels for neuropeptides and melanocortin receptors known to be involved in short-term energy regulation, to determine their role in the maintenance of seasonally different levels of body mass.

Forty-two bank voles Clethrionomys glareolus Schreber 1780 were bred in the laboratory and maintained in a 12 h:12 h L:D photoperiod at a temperature of 20±2°C. Voles were weaned at 18-20 days of age and kept in single-sex groups of 4-6 animals. Rodent Maintenance Diet (CRM -Special Diets Services; Witham, UK) and water were provided ad libitum. At 7-10 months old, an age by which they had reached sexual maturity, the 20 male and 22 female voles were switched to a cold room at an ambient temperature of 6±2°C; the photoperiod remained 12 h:12 h L:D. After 10 days, voles were separated into individual cages (28 cm×11 cm×12 cm) and the photoperiod was switched to 8 h:16 h L:D. They were then matched for body mass and assigned to either the long-day (LD) or short-day group (SD). The LD group consisted of 11 males and 12 females(N=23) and the SD group 9 males and 10 females (M=19).

After 10 weeks of acclimation at 6°C in the short photoperiod, baseline measurements of all energetic parameters were obtained over the following 3 weeks. Body mass and food intake were measured daily for 7 days, during 4 days of which we obtained faecal samples for gross energy (GE) analysis. RMR and daily energy expenditure (DEE) were measured once for each animal. After baseline measurements were taken, the weight of each vole and of its food was recorded. The LD group was then placed in a room similar in all respects to the first except that the photoperiod was 16 h:8 h L:D (Day 0). Temperature remained at 6°C, as this is representative of natural conditions in Aberdeen at the same time of year. Over the next 12 weeks, all voles were subject to the following regime. Body mass and food intake were recorded every third day for the whole duration. Faeces were collected over a 4-day period in week 7, RMR was measured in week 8 and DEE in week 9, to allow any modifications due to the body mass increase to be determined.

RMR was measured using an open circuit respirometry system (Servomex Ltd.,Crowborough, UK). Each respirometry measurement lasted for 2 h and food and water were not denied prior to measurements. Voles were placed in a sealed Perspex chamber within an incubator (INL-401N-010, Gallenkamp, Loughborough,UK) at a temperature of 25±0.5°C, which is within the thermoneutral zone of these animals (Peacock and Speakman, 2001). Air was dried using silica gel (BDH, UK) and pumped through the system at a rate of 600-800 ml min^-1 (Alexander Wright flowmeter DM3A, Zeal Ltd., London, UK). A 150 ml sample of dried excurrent air was subsequently passed though an oxygen analyser and the mean measurements recorded on a computer at 30 sintervals. Carbon dioxide was not absorbed prior to gas analysis to minimise error in the conversion of oxygen consumption to energy expenditure (Koteja,1996; Speakman,2000). RMR was calculated from the ten lowest consecutive readings, equivalent to 5 min within the respirometry chamber, using the appropriate equation from Speakman(2000), with values corrected for temperature and pressure.

DEE was measured using the doubly-labelled water (DLW) technique(Speakman, 1997). Individuals were weighed to ±0.01 g using a balance (Sartorius, Epsom, UK) and labelled with an intraperitoneal injection of approximately 0.2 g ofwater containing enriched deuterium (4.63 atom%) and oxygen-18 (9.44 atom%). The syringe was weighed before and after the injection (±0.0001 g; Ohaus Analytical Plus, Brooklyn, USA) to provide an accurate measurement of the amount of isotope injected. An initial 50-100 μl blood sample was collected by tail tipping 1 h after the injection, which was the time generally assumed to be required for the isotopes to reach equilibrium(Anbar and Lewitus, 1958; Nagy, 1983). Blood samples were immediately flame-sealed into 50 μl pipettes (Vitrex, Camlab Ltd.,Cambridge, UK) until analysis. A final blood sample was collected 24 h after the initial sample. Blood samples were vacuum-distilled into glass Pasteur pipettes (Volac, John Poulten Ltd., Barking, UK; Nagy, 1983) and the distillates used for mass spectrometric analysis of stable isotopes. Mass spectrometric analysis of deuterium enrichment was performed using H₂ gas. The H₂ was produced by reacting water, distilled from the blood, with LiAlH₄(Ward et al., 2000). For analysis of O¹⁸ enrichment in the blood samples, the water distilled from the blood was equilibrated with CO₂ gas using the small sample equilibration technique(Speakman et al.,1990). DEE was calculated using the single pool intercept method,equation 7.17 (Speakman,1997).

In week 12, 11 voles (five males and six females) from LD and ten voles(five males and five females) from SD were chosen at random. These animals were killed, in the middle of the light phase, by cervical dislocation and dissected into 20 components. Brains were removed, frozen on dry ice and stored at -80°C for further analysis. The contents of the stomach and intestines were removed, after which the length of the caecum, small and large intestines were measured with a ruler (±1 mm). All tissues were weighed(±0.0001 g; Ohaus Analytical Plus), dried in an oven at 60°C for 14 days, and then reweighed to determine the dry mass.

Levels of mRNA for neuropeptide and receptor systems were quantified by in situ hybridisation in coronal hypothalamic sections (Mercer et al., 1995, 1997). Antisense riboprobes complementary to rat NPY (Mercer et al.,1995), hamster POMC and AGRP(Mercer et al., 2000), rat CART (Adam et al., 2000), rat CRF (Mercer et al., 1995) and human MC3-R and MC4-R (Adam et al.,2000) were transcribed from cloned cDNA templates.

Coronal hypothalamic sections (20 μm) of bank vole brains were cut on a cryostat and collected from the caudal extent of the ARC through to the rostral extent of the PVN onto two sets of eight slides, with six or seven sections mounted on each slide. The first set of slides spanned the region from approximately bregma -2.70 mm to bregma -1.46 mm according to the atlas of the mouse brain (Franklin and Paxinos,1997), and therefore contained the full extent of the ARC and the caudal part of the VMN. The second set of slides continued to bregma -0.58 mm,and thus contained the rostral part of the VMN, as well as the caudal and rostral extent of the PVN. Slides were fixed, acetylated (optional), and hybridised overnight at 58°C with ³⁵S-labelled riboprobes in hybridisation buffer (1-2×10⁷ c.p.m. ml^-1). For each riboprobe, the hybridisations were carried out simultaneously on 21 slides (one slide from each individual). The next day, slides were treated with ribonuclease A, desalted with a final high-stringency wash (30 min) in 0.1× SSC at 60°C, air-dried and exposed to autoradiography film(Kodak Biomax MR1, Amersham Pharmacia Biotech UK Ltd.) for 7-22 days,depending on the riboprobe used. Autoradiographic images were quantified using the Image-Pro Plus system (Media Cybernetics, Silver Spring, MD, USA), which determines the intensity and area of the hybridisation signal. Image analysis was performed on four or five sections from the ARC and the VMN, and two or three sections from the PVN. Data were analysed using a standard curve generated by simultaneous exposure of ¹⁴C autoradiographic microscales (Amersham Pharmacia Biotech, Amersham, UK), and the integrated intensity of the hybridisation signal was computed.

Data were analysed using Minitab version 11(Ryan et al., 1985). All data were tested for normality using the Anderson-Darling test and non-parametric tests were used for non-normal data. Organ mass data were analysed with general linear models (GLM), following Selman et al.(2001), and body mass was used as a covariate where appropriate. Repeated one-way analysis of variance(ANOVA) was used to analyse the effect of time on body mass and food intake,followed by least-squares regression (LSR) to determine daily variation or general trends. The effect of photoperiod and sex on hypothalamic neuropeptide and receptor gene expression was determined by two-way ANOVA. Tukey's test was employed post hoc when required. Finally, two-sample t-tests, paired t-tests or Mann-Whitney U tests were used where appropriate. All results had a significance level of P<0.05. We aimed for sample sizes of 12 animals in both male and female treatment and control groups; however, due to losses in the course of the acclimation period and the experiments we ended up with samples of 11 and 9 males and 12 and 10 females. Analyses were restricted to animals where we had complete data. We have previously determined the repeatability of RMR measurements using our respirometry system, and the day to day coefficient of variation on repeated measures is 7.7% (N=69 paired measurements: Król et al., 2003). With a sample size of 10 per group, and alpha set at 0.05, the power to detect a 10% difference in RMR between the control and treatment groups was 80%. For the repeated measurements comparing pre- and post-treatment samples the power for the same sample size was 95.4%. Using the estimated mean deviations in validation studies of the doubly-labelled water technique, we could detect differences of 13% between control and treatment groups with a power of 80% at alpha=0.05. Using day to day variation in food intake measures we estimated that we could detect differences of 5% in daily food intake at the same power and significance levels for N=9 per group. The corresponding discrimination abilities in the repeated-measures analyses were 10% for doubly-labelled water estimates and 3% for daily food intake measures.

There was no significant difference in any variable during the baseline period between voles in LD and SD groups (body mass: W=519, P=0.99; FI: W=535, P=0.49; DMD: W=80, P=0.47; RMR: T₄₁=0.26, P=0.80; DEE: T₂₅=1.22, P=0.24; AEAE: T₁₅=0.56, P=0.58).

There was a significant increase in body mass over time in both male and female bank voles switched to LD (F20,482=2.16, P=0.003; LSR F_1,24=105.7, P<0.001), which was not observed in the control SD groups (F20,434=1.23, P=0.23; Fig. 1). The change in body mass following the switch from SD to LD was divided into three distinct periods. An initial stable period preceded the increase-phase, which was then followed by the plateau-phase. These distinct phases were observed in both male and females but the durations and timings were slightly different between sexes. In males, the increase-phase began on day 3 and lasted for 21 days,during which time the voles gained 1.96 g (9.6% of the initial body mass). Female bank voles exhibited a similar mass increase of 1.95 g (11.1% of the initial body mass); however, the increase-phase was slightly shorter,beginning on day 9 and lasting only 15 days. In both sexes body mass reached a plateau on day 24 following the photoperiod manipulation. The rate of change in body mass during the increase-phase was significantly higher than the changes observed during the initial stable period and the plateau-phase for both sexes (male F_2,32=10.5, P<0.001; female F_2,35=34.3, P<0.001). The components of energy intake and expenditure for male and female voles in the treatment group,divided into the different phases of the response, are presented in Table 1.

Dry masses of the dissected organs of male and female voles, culled after 12 weeks exposure to long-day photoperiod from both the treatment (LD) and control (SD) groups are presented in Table 2. The main effects of the photoperiod treatment were on the masses of the white adipose depots (subcutaneous, mesenteric and abdominal),which were all significantly larger in LD voles (N=11) compared to those in SD (N=10; Tables 2 and 3). The combined mass of white adipose depots was 0.89±0.13 g and 0.54±0.07 g in voles from LD and SD, respectively, an increase of 64.8%. Of the 15 lean tissues dissected,only the reproductive organs differed significantly in mass between voles from LD and SD groups. There was a large sex difference in the dry mass of the reproductive organs, with the organs of males being much larger than those of females (Tables 2 and 3). Sex differences were also apparent in the sizes of the components of the alimentary tract, liver, pelage and total dry mass of the empty carcass but photoperiod had no effect on any of these components (Tables 2and 3).

Across all groups the mean food intake was 6.0±0.15 g per day. There were no significant effects of sex (F_1,42=0.03, P=0.87), photoperiod (F_1,42=1.78, P=0.19) or time (F18,436=15.4, P<0.001, LSR F_1,18=0.88, P=0.36) on the level of food intake. To illustrate the constancy of food intake, the food intake of the treatment(LD) and control (SD) groups measured at 3 day intervals through the entire experiment and averaged across all individuals, are presented in Fig. 2. Total food intake over days -13 to 73, pooled across sexes, was 416.4±13.7 g in the LD group and 412.5±16.0 g in the SD group (T₄₁=0.19, P=0.85). There was also no difference in food intake between the three phases for male or female voles from the LD group(Table 1).

Dry food had an energy content of 17.6 kJ g^-1. DMD and AEAE were significantly higher in the LD group (81.2±0.80 and 83.8±0.79,respectively) compared to the SD group (78.6±0.57 and 81.2±0.69,respectively) following 7 weeks in LD (DMD: T₁₅=2.85, P=0.01; AEAE: T₁₅=2.57, P=0.02). Sexes were not analysed separately due to the small sample size. Measured AEAE was used to transpose food intake into assimilated energy (kJ). There was no significant effect of time (F18,436=15.4, P<0.001; LSR F_1,18=1.14, P=0.30), sex(F_1,42=2.11, P=0.15) or photoperiod(F_1,42=0.26, P=0.61) on the daily energy assimilation. Summing the total energy assimilation over days -13 to 73, voles in the LD group assimilated on average 5396±178 kJ compared with 5180±192 kJ by voles in the SD group. Interestingly, female voles in the LD group assimilated significantly more energy per day during the plateau-phase compared to the stable-phase(Table 1).

Across all groups RMR averaged 28.6 kJ day, approximately 40% of the energy assimilation rate. Despite the fact that the LD animals increased their body mass by approximately 10%, there were no significant effects of time, sex or photoperiod on RMR (Fig. 3A;time T=0.45, P=0.66; sex F_1,42=2.66, P=0.11; photoperiod F_1,42=2.88, P=0.10).

Parameters used in DLW estimates of DEE are presented in Table 4. Across all eight groups [sex (2), photoperiod (2) and time (2)], the average DEE measured by DLW was 5.4% greater than the calculated energy assimilation rate. This discrepancy appeared to be greater in males than in females. Hence, the average DEE estimated by DLW across the four groups of males was 8.1% higher than the calculated energy assimilation, while in females it was only 2.7%greater. Despite the fact that the LD animals increased in body mass by approximately 10%, there were no significant effects of time or photoperiod on DEE (Fig. 3B; time T₂₆=0.73, P=0.47; photoperiod F_1,26=0.46, P=0.50). Female voles, however,exhibited a mean DEE 14% lower than male voles(F_1,26=5.66, P=0.03). The ratio of DEE to RMR varied between 2.5 and 3.0 in the groups during the baseline period(mean=2.79). During the final phase of the experiment, after the voles had increased in mass by 10%, the average was almost identical at 2.77 (range 2.5-2.9).

The distribution of NPY, AGRP, POMC, CART, CRF, MC3-R and MC4-R mRNAs in the bank vole hypothalamus (Fig. 4) was consistent with other rodent species. NPY, AGRP and POMC mRNAs were mainly expressed in the ARC. CART was widespread throughout hypothalamus and strongly expressed in the ARC. CRF and MC4-R gene expression were most dense in the PVN. The major sites of the gene expression for MC3-R were the ARC and the VMN.

Gene expressions of NPY, AGRP, CART, MC3-R in the VMN, POMC in the ARC and MC4-R were not affected by photoperiod or sex(Fig. 5). There was a significant effect of photoperiod (F_1,18=5.62, P=0.03) and sex (F_1,18=4.51, P=0.05) on CRF gene expression, with higher levels of expression in the PVN of voles exposed to LD, and mRNA levels higher in females than in males. Voles in the LD group also had upregulated levels of MC3-R mRNA in the ARC(F_1,18=4.66, P=0.05).

As anticipated from their seasonal responses in the wild(Klaus et al., 1988), both male and female bank voles increased their body mass following the switch from a short-day to a long-day photoperiod. This change involved significant increases in all white adipose tissue (WAT) depots, which is also consistent with previous studies of seasonal composition changes in small rodents(Dark and Zucker, 1986; Bartness et al., 1989). In males, the testes also increased in mass, indicative of a readiness for reproduction (Blank, 1992). In contrast to the changes in WAT there was no significant change in brown adipose tissue (BAT) mass. Similarly, McDevitt and Speakman(1994) found short-tailed field voles (Microtus agrestis) did not downregulate their BAT capacities when switched to a long photoperiod at constant low temperature. The difference in amplitude of body mass change between wild voles(Klaus et al., 1988) and our present study may be a consequence of a number of factors. In the present study only the photoperiod was altered, whereas in the wild, ambient temperature and changes in food quality and quantity may combine with changes in photoperiod to influence weight or reproductive cycles(Clarke, 1985). In addition, we followed the same individuals over time, but Klaus et al.(1988) trapped different samples of the population each month, and sex and age differences, which were not controlled for, may thus have contributed to the greater apparent effect.

Although the voles increased in body mass under long photoperiod, there was no effect of this change on RMR in either sex. This absence of an effect of body mass change is consistent with the suggested low metabolic intensity of adipose tissue (Field et al.,1939; Krebs, 1950; Schmidt-Nielsen, 1984; Selman et al., 2001) and the fact that WAT depots were the most significant contributors to the increases in overall body mass. We predicted a priori that RMR would increase because of the changed body mass, and that voles would need to respond to these changes by increasing their food intake or modulating other components of their energy budgets. The absence of a change in RMR meant that changes in food intake and in total daily energy expenditure were not required to sustain the increased body mass. It was not surprising, therefore, to find that total daily energy demands also did not differ between the SD and LD animals for both sexes and, although food was available ad libitum, voles did not increase their food intake concurrent with body mass increase. The overall impression then was that food intake, total DEE and RMR were almost completely unaffected by the photoperiod and the consequent body mass changes.

Although the changes in body mass observed in our voles in response to photoperiod mimic those observed in the wild, in terms of direction, the negligible energetic consequences of these changes call into question the common interpretation that reduced body mass in the wild is an adaptation to reduce energy demands during winter when food is in short supply. Our data suggest that voles could sustain significantly greater masses (+10%) with no energetic penalty. An alternative explanation is that the reduction in body mass during winter may reduce the risk of predation when cover due to vegetation is relatively sparse. Lighter individuals may be more adept at escaping predators because they are not slowed by carrying large fat stores(McNamara and Houston, 1990; Witter and Cuthill, 1993; Gosler et al., 1995).

Male voles in LD had a 14% higher daily energy expenditure than females in LD, possibly reflecting higher activity levels in males(Norrdahl and Korpimäki,1998; Marczinski et al.,1998).

In bank voles, photoperiod had no effect on gene expression of the anabolic neuropeptides NPY and AGRP, or the catabolic neuropeptides CART and POMC. Gene expression of melanocortin receptors MC3-R in the VMN and MC4-R in the PVN were also not altered by photoperiod. The only changes related to LD exposure were elevated CRF in the PVN, and the melanocortin receptor MC3-R in the ARC.

The relationship between photoperiod and gene expression of hypothalamic neuropeptides and receptors has been extensively studied in male Siberian hamsters (Reddy et al., 1999;Mercer et al., 2000, 2001), a species which,similar to bank voles, exhibits its maximum body mass during long-days in summer and reduces body mass during short-days in winter. All the experiments conducted on the hamsters to date have involved a transfer of animals from LD to SD. In contrast, the bank voles in the current study were transferred from SD (8 h:16 h L:D) to LD (16 h:8 h L:D). Despite this difference in experimental protocol, comparison between bank voles and hamsters is appropriate since the changes in body mass induced by photoperiod are reversible and closely mimic the physiological changes normally expressed during the life cycle of these species; although hamsters also exhibit an increase in fat free mass as well as fat mass. To allow ready comparison of the effects of photoperiod upon hypothalamic gene expression between studies and species, we have compiled these data in Table 5, where the gene expression of LD animals is expressed relative to SDs for both bank voles and Siberian hamsters.

The sensitivity of hypothalamic gene expression to photoperiod in bank voles had similarities to the changes reported for Siberian hamsters. However,three of the mRNA species studied (CART and POMC in the ARC and MC3-R in the VMN) were regulated differently in the two species. There is strong evidence that both POMC and CART are powerful physiological anorexic signals(Kristensen et al., 1998; Lambert et al., 1998). Differential regulation of these neuropeptides in bank voles and hamsters may be related to differing effects of LD on food intake. Bank voles showed no changes in food intake in response to photoperiod(Fig. 2), whereas hamsters exposed to SD had significantly lower food intake than those in LD(Knopper and Boily, 2000; Mercer et al., 2001).

Both bank voles and hamsters exposed to LD had an elevated gene expression of melanocortin receptor MC3-R in the ARC. The role of this receptor in the regulation of energy balance has been demonstrated by the genetic knockout of MC3-R (Chen et al., 2000). Mice lacking MC3-R are characterised by reduced food intake, normal body mass and elevated adiposity relative to wild-type littermates. Bank voles exposed to LD had higher levels of MC3-R mRNA than those in SD, but they had unaltered food intake. This suggests the changes in MC3-R in the ARC may not be related to short-term energy regulation, which is consistent with recent data on hypothalamic gene expression in LD and SD food-restricted hamsters suggesting that MC3-R expressed in the ARC is not involved in short-term energy homeostasis (Mercer et al.,2001). MC3-R levels in the ARC therefore remain potentially part of a putative coding mechanism for the long-term 'appropriate' body mass. The elevated levels of CRF in the bank voles transferred to LD were probably unrelated to energy regulation but more closely connected to the function of CRF as a regulator of the hypothalamic-pituitary-adrenal axis(Bray et al., 1990). Gender-specific patterns in the neural mechanisms controlling hypothalamic-pituitary-adrenal axis activity, as indicated by sex-differences in the expression of genes encoding CRF in the hypothalamus of rats, have been reported (Patchev et al.,1995; Patchev and Almeida,1996). In accordance, we found that female voles had higher levels of CRF gene expression in the PVN compared to males. This difference may be due to the role of CRF in the regulation of reproduction(Rivest and Rivier, 1995; Lin et al., 2001), as also indicated by the direct effect of estradiol on CRF gene expression(Vamvakopoulos and Chrousos, 1993, 1994).

Maintenance of photoperiod-induced body mass changes in bank voles were not associated with major changes in activity of hypothalamic neuropeptide and receptor systems involved in short-term energy homeostasis. This suggests that both LD and SD bank voles perceived themselves to be in energy homeostasis,despite their differences in body mass. The energy balance measurements we have made in both states confirm that this was indeed the case. The implication is that seasonal cycles of body mass in bank voles (and Siberian hamsters) reflect changes in the 'appropriate body mass level' around which homeostasis occurs, rather than changes in activity of anabolic and catabolic neuropeptides directly involved in maintaining this homeostasis. The genes involved in coding the appropriate body mass remain obscure, apart from the possible involvement of MC3-R in the ARC.

The main focus of the present paper was the energetic consequences of differences in body mass that were stimulated by a change in photoperiod and how these might relate to differences in hypothalamic gene expression for transcripts known to be linked to short-term energy balance. Our data show that these consequences are undetectable at the resolution of our existing methods for measuring energy balance, and that despite their elevated body mass and body fatness, the voles perceive themselves after a period at LD to be in energy homeostasis. Between the periods of homeostasis, prior to the change in photoperiod and after stability in body mass has been regained,there is a period of dynamic change in body mass(Fig. 1). Our data show that during this phase food intake levels remain unchanged, strongly implicating changes in expenditure (either RMR or physical activity) as the mechanism used to increase body mass. Recent studies of collared lemmings (Dicrostonyx groenlandicus) suggest that reductions in RMR facilitate mass gains after photoperiod changes (Powell et al.,2002), consistent with the absence of differences in food intake during this phase in our study. Our future work will focus on this dynamic phase to elucidate in more detail the whole-body energy balance and central signalling mechanisms that underpin the body mass increase.

 
• AEAE

  apparent energy assimilation efficiency

 
• AGRP

  agouti-related protein

 
• ANOVA

  analysis of variance

 
• ARC

  arcuate nucleus

 
• BAT

  brown adipose tissue

 
• BM

  body mass

 
• CART

  cocaine- and amphetamine-regulated transcript

 
• CRF

  corticotrophin-releasing factor

 
• DEE

  daily energy expenditure

 
• DLW

  doubly-labelled water

 
• DMD

  dry mass digestibility

 
• FI

  food intake

 
• FM

  faeces mass

 
• GE

  gross energy

 
• GLM

  general linear model

 
• L:D

  light:dark

 
• LD

  long day

 
• LSR

  least-squares regression

 
• NPY

  neuropeptide Y

 
• POMC

  pro-opiomelanocortin

 
• PVN

  paraventricular nucleus

 
• RMR

  resting metabolic rate

 
• SD

  short day

 
• VMN

  ventromedial nucleus

 
• WAT

  white adipose tissue

W.P. was supported by a BBSRC studentship, and E.K. by a BBSRC grant 1/S12830 to J.R.S. and J.G.M. We thank the animal house staff in the Zoology Department of Aberdeen University for their help in the care taking of the animals; Peter Thomson and Nicola Goodship for technical assistance with the mass spectrometry analysis; and David Brown and Christine Horrocks at the Rowett Research Institute for their help with GE measurements. Finally, we would like to thank Sally Ward and David Tidhar for their helpful comments on earlier versions of the manuscript.