The circadian pacemaker in the suprachiasmatic nuclei (SCN) controls endogenous near 24-h physiological and behavioral rhythms in metabolism, neuroendocrine function, and locomotor activity. Recently, we showed that vasoactive intestinal polypeptide (VIP) and its receptor, VPAC2 are critical to the intercellular communication between individual SCN neurons, and appropriate synchronization and phasing of these oscillatory cells. Mice defective in VIP signaling manifest grossly impaired circadian rhythms of SCN neuronal firing activity and are typically unable to maintain rhythmic wheel-running behavior in the absence of external time cues. Here we report that daily rhythms of metabolism and feeding behavior are also overtly altered in these animals. Under diurnal conditions (12:12-h light-dark; LD), metabolic and feeding rhythms are advanced in mice lacking either VIP or VPAC2 receptor expression, peaking in the late day, rather than early night, as observed in wild-type mice. When placed in constant light (LL), both VIP-deficient and VPAC2 receptor-knockout mice exhibit dampening of metabolic and feeding rhythms, which deteriorate after a few days. In addition, overall metabolic rate is greatly reduced in VPAC2-knockout mice, when compared with wild-type mice, regardless of lighting condition. The advancement of metabolic and feeding rhythms in these mice under LD suggests that these rhythms are less sensitive to masking by light. These results demonstrate that altering SCN function not only affects neuronal and wheel-running activity rhythms but also dramatically impairs temporal regulation of metabolism and feeding.
- vasoactive intestinal polypeptide
- VPAC2 receptor
behavioral and physiological cycles that approximate 24 h are maintained by endogenous circadian oscillators, which synchronize to environmental cues, such as light-dark transitions, food availability, and social interaction (8, 21, 31). In mammals, the most studied neural circadian pacemaker is localized to the suprachiasmatic nuclei of the hypothalamus (SCN). Destruction of the rodent SCN abolishes daily behavioral, metabolic, and neuroendocrine rhythms, and transplantation of fetal SCN tissue grafts to arrhythmic SCN-lesioned adult animals can rescue behavioral, but not hormonal, rhythms (26, 30). Many molecular components of the SCN pacemaker have been identified, and interactions among the core clock genes and their protein products have been extensively reviewed (12, 14, 23, 28).
Although many SCN neurons express core clock genes and can function as autonomous cellular oscillators, synaptic contact and intercellular communication are vital for the appropriate coordination of these neurons into a coherent central oscillator at the tissue level (37). A key signaling pathway in this process includes vasoactive intestinal polypeptide (VIP) and its receptor VPAC2, both of which are expressed by SCN neurons (2, 4, 25). Pharmacological studies suggest that VIP is also involved in synchronizing SCN circadian oscillations to light (24, 27). The overt impairments in circadian competence of transgenic mice with impaired VIP-VPAC2 signaling support the fundamental role for this neuropeptide in SCN timekeeping processes. Deletion of either the VPAC2 receptor (Vipr2−/−) or VIP [VIP/peptide histidine isoleucine (PHI)−/−] in mice leads to aberrant wheel-running rhythms and marked diminution in SCN molecular and cellular activity profiles (2–4, 6, 7, 10, 13, 18).
Recently, core clock genes have been implicated in the regulation of metabolism across a wide range of organisms (16, 35). For example, mice expressing a defective form of the Clock protein (Δ90 clock mice) exhibit disrupted wheel-running rhythms and reduced metabolic activity (29, 33). At present, it is unclear whether VIP-VPAC2 signaling pathways contribute to the temporal regulation of energy metabolism. Here, we examine the feeding behavior and metabolic phenotype of mice lacking either VIP or VPAC2 expression using indirect calorimetry, under both diurnal and circadian conditions. Importantly, both knockout genotypes show disrupted synchronization of feeding and metabolism with exogenous light-dark cycles, as well as greatly attenuated circadian rhythms in metabolism under constant light (LL) conditions.
MATERIALS AND METHODS
Vipr2−/− mice were obtained from Tony Harmar (University of Edinburgh), while VIP/PHI−/− mice were obtained from Chris Colwell and James Waschek (University of California at Los Angeles). The methods followed for generating these mice have been described previously (6, 10). Mice were introgressed onto a C57/BL6 wild-type (WT) background for seven generations and maintained as a colony at the University of Manchester. All studies detailed herein were licensed under the Animals Act of 1986 (Scientific Procedures) and received ethical approval from the University of Manchester animal welfare committee. Male mice 12–16 wk of age were used for all experiments and were housed at an ambient temperature of 20–22°C and maintained in a 12:12-h light-dark cycle (LD), unless stated otherwise.
WT, Vipr2−/−, and VIP/PHI−/−, mice (n = 7, 11, and 6 mice, respectively) were housed individually in calorimetric cages (Columbus Instruments, Columbus, OH). Mice were monitored for a minimum of 5 days in LD (intensity with lights on: ∼500 lux) followed by a minimum of 5 days in LL, during which time oxygen consumption (V̇o2), carbon dioxide production (V̇co2), and respiratory quotient (RQ) were measured every 10 min. Cages were not equipped with running wheels, and environmental enrichment was limited to bedding material. Standard rodent chow (3.7 kcal/g, 20% protein, 10% fat by energy; Beekay International, Hull, UK) and water were supplied ad libitum throughout these experiments. Measures of metabolic rate were carried out in four independent experiments. Illustrations of V̇o2 and V̇co2 traces in Fig. 1 depict group (genotype) averages taken from a representative experiment. The location and setup of our calorimetric cages did not allow for adequate monitoring of the health and well-being of the mice by facility staff under conditions of constant darkness; therefore, this free-running condition was not examined.
Activity at the food hopper was monitored within the calorimetric cages using a laser beam set across the opening of the hopper (n = 7 WT, 11 Vipr2−/−, and 6 VIP/PHI−/− mice). The number of beam breaks (minimum 1-s duration) was recorded for every 10-min time bin. Leading up to and during feeding studies, mice (n = 6 WT, 8 Vipr2−/−, and 7 VIP/PHI−/− mice) were housed individually, and food intake was monitored at four intervals in the diurnal cycle [zeitgeber time (ZT) 0, 6, 12, 18].
Analysis of metabolic rhythms.
Rhythms in metabolic activity were analyzed using curve-fitting software developed in-house. Recordings from individual animals during LD and LL portions of the experiment were analyzed separately. Initially, data were normalized such that they spanned a range of values between 100 and −100. The normalized data were then fit with the equation Y = A sin[B(x + C)] using the Newton-Raphson iterative method, where A equaled the amplitude of the rhythm, B equaled the period in radians per hour, and C determined the phase. Initial values of A, B, and C were estimated from the best-fitting curve of a series of >3,000 standard curves, which exhibited periodicities between 3 and 34 h and a range of different amplitudes and phases.
Data are presented as means ± SE, and generally, statistical significance was determined using one-way or two-way ANOVA with Bonferroni's post hoc test using the software suite Prism (GraphPad, San Diego, CA). Statistical differences in mean V̇o2 and V̇co2 were determined using general linear model analysis of covariance (ANCOVA) with mouse body weight as a covariant to metabolic gas exchange rate, using the SPSS software suite.
Reduced metabolic rate in mice lacking VIP or VPAC2 expression.
Metabolic rate was assessed in WT and transgenic mice by indirect calorimetry, with gas sampling every 10 min for a period of ∼10 days. When subjected to LD photoperiod, clear daily rhythms in oxygen consumption (V̇o2) and carbon dioxide production (V̇co2) were observed in WT mice, peaking in the first half of the night (Fig. 1). Rhythmic V̇o2 and V̇co2 cycles were also observed in VIP/PHI−/− and Vipr2−/− mice, although V̇o2 and V̇co2 levels were greatly reduced in both strains during the dark (active) phase of the cycle (V̇o2 WT: 3,110 ± 105 ml·kg−1·h−1,Vipr2−/−: 2,595 ± 85, P < 0.01, VIP/PHI−/− 2,789 ± 89; V̇co2 WT: 3,351 ± 137, Vipr2−/−: 2,660 ± 108, P < 0.05, VIP/PHI−/−: 2,808 ± 90, P < 0.05; Fig. 1, B and D). The statistical significance of this observed reduction was maintained when the V̇o2 readings for the VPAC2 knockout mice were averaged over the full 24-h period (V̇o2 WT: 2,884 ± 91,Vipr2−/−: 2,594 ± 109, P < 0.05). In contrast, VIP/PHI−/− mice were found to have a similar 24-h mean V̇o2 to those measured in WT mice (2,787 ± 111 ml·kg−1·h−1). Both strains of transgenic mice showed a reduction in RQ during ZT 12–24 and a slight increase during ZT 0–12, compared with WT mice.
The reduced metabolic activity of VIP/PHI−/− and Vipr2−/− mice over ZT 12–24, typically the active phase of the day for mice, appeared to be due, in part, to an alteration in the timing of activity in these mice. To address this issue, V̇o2 and V̇co2 recordings for each animal were individually analyzed using curve-fitting software to generate a waveform representing the dominant rhythm in its daily metabolic gas exchange. Fig. 2 illustrates V̇o2 recordings overlaid with the resulting rhythm waveform from representative individuals of each genotype (genotype averages are illustrated in ⇓⇓Fig. 5). An advancement in the peaks of V̇o2 and V̇co2 are clearly visible in the VIP/PHI−/− and Vipr2−/− mice under LD. WT mice exhibited strong rhythms that peaked in the early night (V̇o2: ZT 15.7 ± 0.2 h; V̇co2: ZT 16.3 ± 0.3 h). In contrast, Vipr2−/− and VIP/PHI−/− mice exhibited peaks in V̇o2 and V̇co2 3–4 h earlier than those of WT individuals (V̇o2: ZT 12.5 ± 0.4 h and 11.9 ± 0.7 h, respectively; V̇co2: 12.7 ± 0.2 and 11.9 ± 0.7, respectively; all P < 0.01 vs. WT).
Disrupted feeding behavior in VIP-VPAC2 signaling-impaired mice.
A daily rhythm in feeding behavior was observed in WT animals, with both hopper visits (Fig. 3, A and B) and food intake (Fig. 3C) being maximal during the early night (ZT 12–18), similar to the pattern observed in metabolic rate. Unlike WT mice, which made over 70% of the visits to the food hopper at night (71.8 ± 2.6%), Vipr2−/− and VIP/PHI−/− mice exhibited a spread of food hopper activity throughout the day and night (Vipr2−/− 51.6 ± 1.6% at night, VIP/PHI−/−: 50.3 ± 3.1% at night), with the highest level of activity observed in the late day (ZT 6–12). Measures of food intake paralleled hopper visits, with both transgenic strains of mice showing the highest level of consumption between ZT 6–12 (Fig. 3C).
Although not statistically significant compared with WT mice, average daily food intake was reduced in Vipr2−/− mice (WT: 0.13 ± 0.01 g/g body wt, Vipr2−/− 0.10 ± 0.01 g/g body wt), yet it was not reduced in the VIP/PHI−/− mice (0.14 ± 0.004 g/g body wt) (Fig. 3D).
No significant differences in body weight were observed between the three mouse genotypes, although Vipr2−/− mice were somewhat larger than wild-type mice (WT: 28.2 g ± 0.8, n = 14; Vipr2−/−: 29.1 g ± 0.6, n = 22; VIP/PHI−/−: 27.1 g ± 0.4, n = 15), possibly reflecting the reduced metabolic rate in these mice. In contrast to the Vipr2−/− mice, VIP/PHI−/− mice exhibited a similar 24-h mean V̇o2 to WT mice.
Effect of constant light on metabolic rhythms.
Following a minimum of 5 days of monitoring metabolic gases under LD conditions, mice were switched to constant light (LL) for a further 5–6 days. Mean V̇o2 and hopper activity records collected from representative individuals across the period of LL are illustrated in Fig. 4. Robust circadian rhythms of V̇o2 were maintained in WT mice when switched to LL. V̇o2 rhythms in LL were also observed in mice lacking either VIP or VPAC2 receptor expression, although with much less consistency than WT mice. As with LD, patterns of hopper activity roughly corresponded with metabolic rhythms in all three genotypes under LL (Fig. 4).
Curve-fitting analysis of V̇o2 recordings collected in LD and LL were performed and averaged across genotypes (Fig. 5, A and B). Similar results were obtained from analysis of V̇co2 recordings (data not shown). Under LD, WT mice exhibited V̇o2 rhythms with a period of ∼24 h (Fig. 5C). In contrast, many Vipr2−/− and VIP/PHI−/− mice exhibited periods well below 24 h, although a statistically significant reduction in period was only detected in the Vipr2−/− mice (P < 0.05). This daily advancement of V̇o2 rhythm (for example, the Vipr2−/− mouse illustrated in Fig. 2A) suggests a disruption of the daily resetting of the endogenous clock in these mice. V̇o2 rhythms under LD were weaker in the Vipr2−/− and VIP/PHI−/− mice, as assessed by peak-to-trough amplitude, when compared with WT mice (Fig. 5D).
When transferred to LL conditions, the period of daily V̇o2 rhythms lengthened in WT mice (LD: 24.2 ± 0.1 h, LL: 25.4 ± 0.3 h, P < 0.05), whereas period decreased in both transgenic strains and were significantly shorter than those observed in WT mice (Vipr2−/− LL: 23.0 ± 0.3, VIP/PHI−/− LL: 23.3 ± 0.3 h, both P < 0.01 vs. WT). Daily metabolic rhythms of WT and Vipr2−/− mice, but not VIP/PHI−/− mice, were also reduced in amplitude under LL (Fig. 5D).
Here, we provide the first examination of the loss of VIP or the VPAC2 receptor on metabolic rhythms in mice. This signaling pathway plays a key role in the regulation of circadian rhythms by the SCN pacemaker. Specifically, VIP-VPAC2 signaling is crucial for appropriately timed cellular rhythms in SCN neurons (2–4, 18), and loss of either the receptor or its peptide ligand in mice leads to a profound disruption of circadian rhythms in wheel-running behavior when housed in constant darkness (6, 10). The current findings add to our understanding of circadian rhythmicity in these mice and demonstrate that they also exhibit altered rhythms in metabolism and feeding behavior that are apparent even under diurnal conditions, with daily peaks of metabolism that are considerably phase-advanced relative to WT mice. This is also the first study to examine the effects of constant light on circadian rhythms in mice with disrupted VIP-VPAC2 signaling and demonstrates that such animals continue to express low-amplitude circadian metabolic rhythms in LL, albeit with a much shorter period than WT mice.
Altered clock function is often “masked” under LD due to the ability of light to suppress activity in nocturnal species such as mice. For example, wheel-running activity in mice with deficient VIP-VPAC2 signaling is largely restricted to the dark phase of the day (6, 10, 13, 32), indicating that light is a strong repressor of this behavior and that such masking effects of light are maintained in animals lacking VIP-VPAC2 signaling. In the present study, metabolic and feeding rhythms were advanced 3–4 h in both transgenic strains under LD, suggesting that light does not overtly mask these behaviors. These findings are consistent with the abnormal phasing of wheel-running rhythms that have been observed in Vipr2−/− and VIP/PHI−/− mice upon release from LD to constant darkness (6, 13). Why light should strongly repress wheel running but not feeding and metabolic activity is unclear. Nonetheless, it raises some questions about how representative wheel-running activity may be of other behaviors, especially in animals with altered circadian phenotypes.
Increasing evidence suggests that VIP and VPAC2 are important in the responsiveness of the SCN to light. Exogenous VIP application phase-shifts wheel-running rhythms in rodents (24) and SCN neuronal discharge rhythms in vitro (27) in a photic-like manner. Further, in contrast to WT mice, light pulses given during the subjective night do not induce expression of the clock genes per1 and per2 in the SCN of mice lacking the VPAC2 receptor (10). However, components of the light-response cascade upstream of these clock genes are robustly activated in the SCN of these animals throughout the circadian cycle (13). Our observation of advanced feeding and metabolic rhythms in VIP/PHI−/− and Vipr2−/− mice held under LD are consistent with an alteration in the response of the SCN to light in these mutant animals. Our current findings and previous work on wheel-running behavior indicate that VIP-deficient and VPAC2 knockout mice remain responsive to environmental lighting cues and can entrain their physiological rhythms to LD cycles, albeit with an abnormal phase angle. However, the timing of metabolic rhythms in Vipr2−/− and VIP/PHI−/− animals housed in LD was considerably more variable than in WT animals, with some individuals exhibiting periods substantially shorter than 24 h. This implies that such individuals free run though the LD cycle (for example, the Vipr2−/− mouse shown in Fig. 2). This possibility has never been investigated in detail and would require longer-term monitoring of metabolic rhythms under LD.
Our observation that VIP/PHI−/− and Vipr2−/− mice retain metabolic rhythms when housed in LL, without the extensive lengthening of circadian period seen in WT mice, suggests that the SCN circadian clock in these mice may be less responsive to photic conditions. The presence of VIP and VPAC2 receptor expression in the retina is consistent with such an interpretation (11, 17). However, as mentioned above, photic stimuli robustly activate c-Fos and other components of the light-responsive intracellular cascade in the SCN of Vipr2−/− animals at the same intensities that these stimuli elicit such changes in the WT SCN (13). It is also notable that under DD conditions, rhythmic Vipr2−/− and VIP/PHI−/− mice wheel-run with periods of ∼22–22.5 h. Under LL conditions, we observe periods of ∼23.2 h in the metabolic rhythms, suggesting that the endogenous period of these animals is lengthened in LL but not to the same magnitude as seen in WT mice. This raises the possibility of a role for VIP and the VPAC2 receptor in the parametric effects of light on the SCN clock.
Intriguingly, WT and Vipr2−/− mice exhibit a dampening of metabolic rhythms when housed in LL, whereas VIP/PHI−/− mice do not. Since Vipr2−/− mice still produce VIP, it is possible that VIP signaling through a non-VPAC2 receptor-dependent pathway may contribute to the effects of constant light on the amplitude of metabolic rhythms. An additional possibility is that these findings result also from the absence of PHI signaling in VIP/PHI−/− mice (for a review, see Ref. 34), although there are presently no known specific PHI receptors in the mammalian CNS.
The basis of the altered metabolism and feeding behavior in mice lacking VIP or VPAC2 is unclear. Neuronal output from the SCN of these mice is blunted (3, 4), and this likely has consequences for temporal regulation in SCN-recipient brain areas such as the paraventricular hypothalamus or dorsomedial hypothalamus, which are implicated in ingestive behavior and metabolism. Since VIP and VPAC2 receptors are also expressed in peripheral tissues, it is also possible that transgenic impairment in this pathway alters peripheral regulators of metabolism. The combination of alterations in central and peripheral mechanisms most likely accounts for the metabolic disturbances seen in this study. Daily food intake and metabolic rate were significantly reduced in mice lacking the VPAC2 receptor, when compared with WT mice. Interestingly, this pronounced reduction was not observed in the VIP/PHI−/− mice. Although two groups have previously reported that administration of VIP can reduce feeding in rats (15, 36), a direct role for VIP and VPAC2 in the regulation of feeding behavior remains speculative. Nevertheless, it is possible that the differences we observe in food intake between the two strains of transgenic mice may involve feeding-related action of VPAC2 mediated through pituitary adenylyl cyclase-activating peptide (PACAP) signaling. The VPAC2 receptor binds PACAP with as equal affinity as VIP (9). PACAP knockout mice are reported to show reduced carbohydrate feeding and reduced expression of NPY mRNA in the arcuate nucleus (22), and central administration of PACAP decreases fast-induced feeding in rats and mice (5, 19, 20). Further, a recent study indicates that VPAC2 receptors still become activated in the VIP/PHI−/− mice, possibly due to endogenous PACAP signaling (3).
Daily metabolic rhythms are certain to reflect feeding-related arousal and activity, although feeding appears to be less responsive to light in the knockout mice. Sheward et al. (32) show that when Vipr2−/− mice are held on a restricted feeding schedule, rhythmic expression of clock genes in the liver corresponds with the time of feeding. Administration of a VIP antagonist can block food-induced increases in circulating ACTH and corticosterone in fasted rats (1). VPAC2 knockout mice lack a robust diurnal rhythm of corticosterone, although production of the steroid remains responsive to restricted feeding in these mice (32). This is in line with our own observations that WT and Vipr2−/− mice exhibit similar metabolic and feeding responses to a 48-h fast.
In conclusion, we demonstrate altered metabolic and feeding rhythms in mice with deficient VIP-VPAC2 signaling under both diurnal and constant lighting conditions, consistent with the proposed roles of this pathway in regulating the SCN clock. Additionally, we highlight differences in the circadian and metabolic profiles of VIP/PHI−/− and Vipr2−/− mice and provide evidence that the VPAC2 receptor may contribute to the regulation of feeding and metabolism independently from its role in the circadian clock.
Perspectives and Significance
The work presented here further links endogenous time-keeping pathways with key physiological outputs, namely, that of feeding behavior and metabolism. Examination of metabolic and feeding parameters has also provided a unique insight into the phenotype of VIP/PHI−/− and Vipr2−/− mice under diurnal conditions that is not apparent from monitoring their locomotor activity. These data suggest that the reinforcing properties of wheel running on behavioral rhythms can alter underlying circadian phenotypes; a finding that merits attention in future circadian studies.
Chris Colwell and Jim Waschek kindly supplied breeding stocks of vasoactive intestinal polypeptide-deficient mice, while Tony Harmar donated breeding stock of Vipr2−/− mice. We thank Rayna Samuels for her expert technical assistance and the Biotechnology and Biological Sciences Research Council for financial support.
↵* D. A. Bechtold and T. M. Brown contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society