Metabolic consequences of timed feeding in mice

Physiology & Behavior 128 (2014) 188–201

Contents lists available at ScienceDirect

Physiology & Behavior journal homepage: www.elsevier.com/locate/phb

Metabolic consequences of timed feeding in mice Nurulaini Abu Shamsi, Mark David Salkeld, Leewen Rattanatray, Athena Voultsios, Tamara Jayne Varcoe, Michael James Boden, David John Kennaway ⁎ Robinson Institute, School of Paediatrics and Reproductive Health, University of Adelaide, Adelaide, 5005 South Australia, Australia

H I G H L I G H T S • • • • •

We studied the effects of feeding during the light or dark period in mice kept on long or short days. Body weight was unaffected by either long or short photoperiod or feeding time. Restricted feeding had the largest metabolic impact on mice exposed to long days versus short days. Glucose tolerance was impaired at the end of the light period in light fed compared to dark fed mice. Changes in liver gene rhythms did not correlate with changes in feeding time.

a r t i c l e

i n f o

Article history: Received 11 November 2013 Received in revised form 29 January 2014 Accepted 6 February 2014 Available online 15 February 2014 Keywords: Circadian Day length Restricted feeding Wheel running Glucose tolerance Clock genes

a b s t r a c t The time of day at which meals are consumed is known to impact on behaviour as well as physiological systems. In this study we investigated the behavioural and physiological effects of restricting access to food to the light or dark period in mice maintained on either long or short photoperiods. In both photoperiods, wheel running commenced upon the onset of darkness and was generally confined to the period of darkness. Provision of food during light provoked an anticipatory burst of activity several hours before feeding in both photoperiods. After 28 days on the feeding schedule, body weight was unaffected by either photoperiod or feeding time. Plasma insulin was increased and glucose and triglycerides tended to be lower in mice fed during the light period and sampled 2 h after lights off compared to the dark fed mice. Mice fed during the light while on long day length had improved glucose tolerance and whole body insulin tolerance when tested 2 h after lights on. This was not evident in mice kept on the short photoperiod. Because these observations were confounded by the time since their last meal, we undertook a study of glucose tolerance across 24 h in mice on the long photoperiod after a 2 hour food withdrawal. A clear rhythm of glucose tolerance was observed in mice fed during the light period with maximal glucose tolerance just prior to the expected presentation of food and minimal tolerance 2 h before lights off. By contrast, no rhythm in glucose tolerance was observed in the dark fed mice, but maximal glucose tolerance occurred 2 h before lights off. To investigate the evolution of the physiological adaptations, mice on this feeding/photoperiod regime were studied after 7 or 35 days. After 7 days the corticosterone rhythm was not different between light and dark fed mice, but by 35 days peak corticosterone secretion occurred a few hours before food presentation in both groups representing an 8 hour shift. The rhythm of expression of liver Bmal1 mRNA was similar in light and dark fed mice after 7 and 35 days on the schedule while the Per1, Per2, Nr1d1 and Dbp mRNA rhythms were delayed on average by 3.5 ± 1.1 h and 3.7 ± 0.9 h in light fed mice after 7 and 35 days respectively compared to dark fed mice. Rhythms of metabolically important genes were shifted in light fed mice compared to dark fed, by 5 h or became arrhythmic. This study shows that not only circadian rhythms facilitate metabolic control, but also different environmental events, including season and feeding opportunities, alter aspects of circadian and metabolic physiology. © 2014 Elsevier Inc. All rights reserved.

1. Introduction There is emerging evidence that the time of day at which food is consumed influences weight gain and metabolic function. In recent human weight loss studies, participants who voluntarily confined or were ⁎ Corresponding author.

http://dx.doi.org/10.1016/j.physbeh.2014.02.021 0031-9384/© 2014 Elsevier Inc. All rights reserved.

randomised to confine the majority of their energy intake to early in the day lost more weight [1,2] and had a greater improvement in insulin sensitivity, triglycerides and oral glucose tolerance [2] than those ingesting a similar diet later in the day. Similarly there is evidence to suggest that shiftworkers who have altered patterns of light exposure, sleep and meal times are at an increased risk of developing obesity and metabolic syndrome [3,4]. Furthermore, the duration of sleep alters

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

metabolism such that people who sleep less than 5 h or longer than 8 h a night are at increased risk of developing obesity and metabolic disorders [5]. The circadian timing system is implicated in the phenomena mentioned above. Rhythmicity within the suprachiasmatic nucleus of the hypothalamus is generated by a suite of genes, known as clock genes, which include Bmal1, Clock, Per1, Per2, Cry1, Cry2 and Nr1d1 (also called Rev erb a) [6]. Positive and negative feedback loops involving interactions of their protein products on their own gene promoters establish a near 24 hour cycle within the cells. Two of the proteins, BMAL1 and CLOCK also provide an output signal via interactions with promoters on other transcription factors and functional genes. The SCN rhythmicity is entrained to the environment by neural input from the retina and influences a wide range of physiological systems, utilising multisynaptic neural pathways (e.g., to the pineal gland, adrenal gland and liver) and hormonal routes (e.g., cerebrospinal fluid arginine vasopressin and prokineticin 2). It is well established however, that peripheral tissues like the liver and muscle also have the capacity to generate cellular rhythmicity via the same genes and that while generally the rhythms are entrained by the SCN, under certain circumstances they can operate in the absence of SCN cues. Cellular rhythmicity, while being driven by SCN cues and endogenous timing mechanisms, is also responsive to metabolic state. For example, the transcription factors PPARα [7,8] and PGC1α [9] bind to the promoters of core clock genes Bmal1 and Nr1d1, driving their transcription. At the protein level, the metabolic sensor, AMP activated protein kinase (AMPK) phosphorylates proteins involved in the negative arm of the cellular clock, targeting them for degradation and hence altering the phase of rhythmicity [10]. Other metabolically important protein kinases (GSK3β, MAPK) similarly target some core clock proteins for phosphorylation and subsequent degradation [11–15], whereas phosphorylation of PER2 and NR1D1 proteins by GSK3β is important for their stabilisation and nuclear translocation [16,17]. Alternatively, the nicotinamide adenine dinucleotide (NAD +) dependent histone deacetylase sirtuin 1 (SIRT1) which associates with and modulates the activity of CLOCK:BMAL1 driven gene expression [18] is also known to target PER2 protein for degradation [19]. These mechanisms, amongst others, intricately link nutrient state and circadian rhythms, thereby optimising the timing of metabolic processes. Disruption of rhythmicity through alterations to the various clock genes has a range of physiological consequences. Mice carrying a mutation in the core clock gene, Clock lack cellular rhythmicity in peripheral tissues such as the liver (but not central rhythmicity [20]) and have abnormal, strain dependent metabolic phenotypes, including obesity, hyperleptinemia, hyperlipidemia, hepatic steatosis, hyperglycemia, hypoinsulinemia [21], elevated levels of the adipokines, leptin and adiponectin [22,23], impaired glucose stimulated insulin secretion, impaired glucose tolerance and paradoxical improved insulin sensitivity [22,24]. Similarly, mice that lack central and peripheral rhythmicity (e.g., Bmal1 null mice) exhibit a more severe metabolic phenotype, in particular, altered fat deposition and adipokine secretion [25,26]. Laboratory rats and mice are nocturnally active and when food is available ad libitum they will consume 60–70% of their food during darkness when held on the traditional 12L:12D photoperiod [27,28]. When food is made available to rats and mice only during the resting light phase, the clock gene expression rhythm in the liver and other peripheral tissues is shifted by approximately 12 h within 5 to 7 days [29,30], but not the SCN [29] or its direct targets (e.g. pineal gland [31]). Manipulation of the time of feeding has been observed to alter body weight, metabolism and clock gene expression [32–34]. There have been no studies, however, on the influence of restricted feeding during long (summer) or short (winter) photoperiods on metabolic function of mice. In the current study we addressed several questions. (1) What is the impact of providing food access exclusively during the light or dark period on the body composition, plasma triglycerides, glucose and insulin

189

in mice kept on long or short photoperiods? (2) What is the effect of light and dark feeding on glucose, insulin, corticosterone and triglyceride rhythms in mice maintained on a long photoperiod for 7 or 35 days? (3) What is the effect of light and dark feeding on glucose tolerance across 24 h in mice maintained on a long photoperiod for 28 days? (4) What is the effect of light and dark feeding on liver gene rhythmicity in mice maintained on a long photoperiod? (5) If there are changes in glucose tolerance, what are the changes in the expression of an a priori selected set of key metabolic genes across 24 h? We predicted that since the difference in time of presentation of food under the 2 schedules was 8 h, physiological rhythms would also be altered by 8 h. 2. Material and methods 2.1. Animals Male C57BL/6 mice aged 4 weeks old were purchased from the Animal Resources Centre (Canning Vale, Western Australia) where they had been kept on a 12L:12D photoperiod and maintained on ad libitum standard chow and water. All the experimental protocols were approved by the University of Adelaide Animal Ethics Committee. 2.2. Experiment 1 To understand the impact of short and long photoperiods and restricted food availability on behaviour, two groups of 10 mice were entrained for four weeks to either 16 h of light and 8 h of darkness (16L:8D) or 8 h of light and 16 h of darkness (8L:16D) with food and water available ad libitum. Thereafter food was made available continuously (n = 2) or exclusively during the light (n = 4) or dark (n = 4) periods for each group. The time of day is designated as Zeitgeber time where ZT0 = time of lights off for both photoperiods. Food pellets were manually removed from or replaced into the cage-lid hoppers. Wheel running rhythmicity was monitored in individual mice housed in light-controlled chambers in cages equipped with 11.5 cm diameter running wheels and magnetic micro-switches. A data acquisition system (LabPro, Data Sciences, St. Paul, MN) was used to record the number of wheel rotations in 10 minute bins. The data was processed in Excel and visualised using the Actiview software package (MiniMitter, Bend, OR). 2.3. Experiment 2 Mice were group housed (n = 6 per cage) in light-controlled chambers and entrained to either 16L:8D or 8L:16D for 4 weeks with food and water available ad libitum. Thereafter at the transition to and from darkness, empty or filled lid food hoppers were exchanged to make food available exclusively during the light or dark period. The food intake was estimated by weighing the food left over in the hoppers or retrieved from the floor of the cages each week when the mice were moved to clean cages. After four weeks of timed feeding, intraperitoneal glucose tolerance tests (IPGTT) were performed 2 h after lights on following an overnight fast. Mice were injected with glucose (1 g/kg, Sigma, St. Louis, MO) and blood collected from the tail vein (5 μl) just prior to the injection and 15, 30, 60, 90 and 120 min post glucose injection. The blood glucose level was measured using a HemoCue Glucose 201 + Analyser (HemoCue, Angelholm, Sweden). Following the IPGTT the restricted feeding schedule continued. One week later, Intra Peritoneal Insulin Tolerance Tests (IPITT) were performed. Food was withdrawn from the cages at lights on for both photoperiods and 2 h later the mice were injected with insulin (0.75 m IU/kg, Actrapid) and blood collected just prior to the injection and 15, 30, 60, 90 and 120 min post insulin injection for glucose measurement. Following the IPITT the restricted feeding schedule continued.

190

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

Table 1 Primer sequences for the Real time RT-PCR analysis. Gene

Accession number Primers

β-Actin NM_007393 Bmal1

AB015203

Per2

NM_011066

Nr1d1

NM_145434

Clock

NM_007715

Per1

NM_011065

Dbp

NM_016974

Pck1

NM_011044

Pfkfb3

NM_133232

F R F R F R F R F R F R F R F R F R

CCACTGCCGCATCCTC AGAAGGAAGGCTGGAAAAGA GTCGAATGATTGCCGAGGAA GGGAGGCGTACTTGTGATGTTC AGGCACCTCCAACATGCAA GGATGCCCCGCTTCTAGAC TCCAGTACAAACGGTGTCTGAAA GCCAACGGAGAGACACTTCTTG ATCCCAGAGGGAGAACATTCAGA GGGTCTATTGTTCCTCGAAGCA TTTGGAGAGCTGCAACATTCC TGCTGACGACGGATCTTTCTTG AACTGAAGCCTCAACCAATCATG TGGCTGCTTCATTGTTCTTGTAC GTTCCCAGGGTGCATGAAAG AGGGCGAGTCTGTCAGTTCAA GCAAGAAGTTCGCCAATGC TCCGCGGTCTGAATGGTACT

Amplicon length 117 bp 101 bp 140 bp 101 bp 140 bp 142 bp 108 bp 107 bp 103 bp

mice on 8L:16D. We could not therefore justify ethically an extensive study of restricted feeding in mice on the 8L:16D photoperiod. Four week old mice were group housed and acclimatised to a 16L:8D photoperiod for 4 weeks in light-controlled chambers with ad libitum access to food and water. Thereafter food was made available throughout darkness or the light period for 1 week (Expt 3a) or 4 weeks (Expt 3b) as for Experiment 2. In Expt 3a, mice were euthanized at ZT14, ZT18, ZT22, ZT2, ZT6 and ZT10 (n = 5 at each time point) and the liver was collected and stored in RNAlater® (Ambion, Austin, TX) overnight at 4 °C and then kept at −20 °C for subsequent gene expression analyses. In Expt 3b, the mice and the remaining food were weighed weekly just before the food was made available. After the four weeks of timed food access, any food available in the cages was withdrawn 2 h before the mice were subjected to IPGTT at ZT14 (n = 8), ZT18 (n = 6), ZT22 (n = 5), ZT2 (n = 5), ZT6 (n = 5) and ZT10 (n = 5). After a further week of timed food access, the mice were euthanized in the fed state (i. e., food was available as previously programmed) at ZT14, ZT18, ZT22, ZT2, ZT6 and ZT10 and blood and liver tissue collected for analysis. 2.5. Hormone and metabolite assays

One week following the IPITT, the mice were euthanized by rapid decapitation 2 h after lights on (food was not withdrawn). Blood was collected into heparinised tubes, kept on ice briefly and centrifuged for 20 min at 4000 rpm at 4 °C and the plasma stored at −20 °C until assayed. The liver, epigonadal fat, retroperitoneal fat, adrenal gland, gastrocnemius muscle and pancreas were rapidly dissected and weighed.

Trunk blood plasma was assayed for insulin by radioimmunoassay (Millipore/Linco RIA kit) and glucose and triglycerides assayed using a Hitachi 912 automated sample system using kits from Roche Diagnostics, NSW, Australia. 2.6. RNA isolation and Real Time RT-PCR

2.4. Experiment 3 We observed improved glucose and insulin tolerance in mice fed during the light on 16L:8D but no difference between light and dark in

Liver RNA (50–90 mg) was extracted with TRI Reagent (Sigma) according to the manufacturer's instructions. The mouse liver was homogenised using the PowerLyzer24™ bench-top homogeniser

Table 2 Primer sequences for the GeXP analysis. Gene Insulin receptor

Insr

Protein kinase C, zeta

Prkcz

AMP-activated protein kinase (AMPK2)

Prkaa2

Rev erb alpha

Nr1d1

Insulin receptor substrate 1

Irs1

Glycogen synthase 2

Gys2

Glucose transporter type 2

Slc2a2

PPAR gamma, coactivator 1a

Ppargc1a

Peroxisome proliferator-activated receptor alpha

Ppara

6-Phosphofructo-2-kinase 3

Pfkfb3

Glycogen synthase kinase 3A

Gsk3a

Sirtuin 1

Sirt1

Glucokinase

Gck

Protein kinase AKT2

Akt2

Sterol regulatory element binding transcription factor 1

Srebf1

Phosphoenolpyruvate carboxykinase

Pck1

Beta actin

Actb

Cyclophilin

Pipb

Primer

Primer sequence

For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev For Rev

AGGTGACACTATAGAATATGTATGAAGGCAATGCCAAG GTACGACTCACTATAGGGATCGATCCGTTCTCGAAGACT AGGTGACACTATAGAATAAAGAAGGAGCTTGTCCACGA GTACGACTCACTATAGGGAGGAGTGTAAGCCAACCAGGA AGGTGACACTATAGAATATTGTCGATGAGGCTGTGAAG GTACGACTCACTATAGGGAATAAGCCACTGCAAGCTGGT AGGTGACACTATAGAATAAGTTCCCACAACAGCTGACA GTACGACTCACTATAGGGATTGTCATGGGCATAGGTGAA AGGTGACACTATAGAATAGGTAAGCTCTTGCCTTGCAC GTACGACTCACTATAGGGATAAGAGAGGACCGGCTTGTG AGGTGACACTATAGAATAATCATTCAGAGGAACCGCAC GTACGACTCACTATAGGGATGGGCTCTAGGTGGAATTTG AGGTGACACTATAGAATAAGGATCAAAGCAATGTTGGC GTACGACTCACTATAGGGACCAGTCCTGAAATTAGCCCA AGGTGACACTATAGAATAGTGGTGTAGCGACCAATCG GTACGACTCACTATAGGGAGCATCAAATGAGGGCAATC AGGTGACACTATAGAATATCTGGAAGCTTTGGTTTTGC GTACGACTCACTATAGGGATTCGACACTCGATGTTCAGG AGGTGACACTATAGAATATCCAGCAGAGGCAAGAAGTT GTACGACTCACTATAGGGATTCCTCACACACACCAGCAT AGGTGACACTATAGAATAATTTGCTTGTGGACCCTGAC GTACGACTCACTATAGGGATCAGCAAGTACACAGCCAGC AGGTGACACTATAGAATACTGTTGACCGATGGACTCCT GTACGACTCACTATAGGGAGCGTCATATCATCCAGCTCA AGGTGACACTATAGAATAAACCAACTTCAGGGTGATGC GTACGACTCACTATAGGGATGCCTTACAGGGAAGGAGAA AGGTGACACTATAGAATAAGGCCACGGTACTTCCTTCT GTACGACTCACTATAGGGACACTCTTCCCTCTCATCTGGA AGGTGACACTATAGAATACGGAGACAGGGAGTTCTCAG GTACGACTCACTATAGGGACTGGGGGATATGCTCTACCA AGGTGACACTATAGAATAATCATCTTTGGTGGCCGTAG GTACGACTCACTATAGGGACTTCCGGAACCAGTTGACAT AGGTGACACTATAGAATAAGCCATGTACGTAGCCATCC GTACGACTCACTATAGGGATTTGATGTCACGCACGATTT AGGTGACACTATAGAATATCCGTGGCCAACGATAAG GTACGACTCACTATAGGGATCCTCCTGTGCCATCTCC

Revolutions/10 minutes

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

*

900 800 700 600 500 400 300 200 100 0

Baseline

Revolutions/10 minutes

6

900 800 700 600 500 400 300 200 100 0

Revolutions/10 minutes

900 800 700 600 500 400 300 200 100 0

Revolutions/10 minutes

Revolutions/10 minutes

10 12 14 16 18 20 22 Zeitgeber time

0

2

4

0

2

4

0

2

4

*

8

10 12 14 16 18 20 22 Zeitgeber time

42 days

6

Revolutions/10 minutes

8

7 Days

6

*

191

8

10 12 14 16 18 20 22 Zeitgeber time

900 800 700 600 500 400 300 200 100 0 14 16 18 20 22

0 2 4 6 Zeitgeber time

8

10 12

14 16 18 20 22

0 2 4 6 Zeitgeber time

8

10 12

14 16 18 20 22

0 2 4 6 Zeitgeber time

8

10 12

900 800 700 600 500 400 300 200 100 0

*

900 800 700 600 500 400 300 200 100 0

Fig. 1. Wheel running activity in mice kept on either 16L:8D or 8L:16D and fed exclusively during the light or dark period. Top panels: The left and right panels show actograms for wheel running activity of mice in 16L:8D, 14 days before and after (denoted by *) initiation of dark feeding (left panel) or light feeding (right panel). The data are the averages (n = 4) of the wheel revolutions completed in 10 minute intervals. The time on the x-axis is designated as Zeitgeber time where ZT0 = time of lights off. Each line represents 1 day of recording and the times of lights off are shown by the horizontal black bar at the top of the figures. The centre figures represent the average number of revolutions 1 day before (top), 7 days (middle) and 42 days (bottom) after initiation of the restricted feeding. The shaded areas show the data (revolutions/10 min) for the dark fed mice and the lines show data for the light fed mice. The horizontal arrows indicate the days of the experiment this data was extracted from. The scales are the same for all 3 graphs. Bottom panels: the same as the above data from mice in 8L:16D.

192

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

(MO BIO Laboratories Inc., Carlsbad, CA) at 6500 rpm (30 s × 2, 30 s). Residual contaminating DNA from all samples was removed using Ambion DNAfree™ kits (Applied Biosystems, Foster City, CA, USA). RNA (2 μg) was reverse-transcribed using Super-Script III (Invitrogen, USA) according to the manufacturer's instructions, with a total volume of 40 μl, which was made up to 100 μl following reverse-transcription. Amplification of cDNA was performed on a GeneAmp 7500 Sequence Detection System (Applied Biosystems, Foster City, CA, USA) in duplicate, using primers which have been previously validated and optimised in our laboratory (Table 1). The expression of each gene of interest within each sample was normalised against β-actin and was expressed relative to a calibrator sample with the use of the formula 2−(ΔΔCt) as previously described [35,36]. The term ΔΔCt is defined as: [CTgene of interest (unknown sample) − CTβactin (unknown sample)] − [CTgene of interest (calibrator sample) − CTβactin (calibrator sample)].

as above for conventional RT-PCR, potential contaminating DNA was removed from the RNA using an Ambion DNAfree kit according to the manufacturer's instructions. A GenomeLab™ GeXP Start Kit (Beckman Coulter Inc., USA) was used for the multiplex gene analysis. The RT reaction, PCR and GeXP analysis were carried out using the manufacturer's instructions, with the following modifications. The amount of internal control RNA (KANr) was decreased from 5.0 μl to 2.5 μl, the amount of DNA Size-standard-400 was decreased from 0.5 μl to 0.4 μl, the Sample Loading Solution was increased from 38.5 μl to 38.6 μl and the amount of PCR product from each sample loaded onto the GeXP analysis plate was increased from 1.0 μl to 1.5 μl. A validated multiplex gene set that included 16 genes of interest and 2 control genes was used (Table 2). Each RNA sample was run in duplicate from the PCR stage.

2.7. Multiplexed gene expression analysis

Body weight, organ weight, hormones, metabolites and gene expression data were analysed by univariate ANOVA (SPSS v20), using when appropriate photoperiod and time of food availability as the dependent variables followed by post hoc analysis using the Bonferroni correction for multiple comparisons. For the IPITT and IPGTT data, the areas

Livers from mice that had been fed exclusively during the light or dark period for 35 days (Expt 3b) were separately analysed using a novel technique called Gene Expression Profiling (GeXP) [37]. Following extraction

2.8. Statistical analysis

Body weight 30

Liver weight 1200

a

b

(mg/20g BW)

1000

(g)

20

10

800 600 400 200

0

Ad Lib Dark Fed Light Fed Ad Lib Dark Fed Light Fed 16L : 8D 16L : 8D 16L : 8D 8L : 16D 8L : 16D 8L : 16D

Pancreas weight

Epididymal / retroperitoneal fat pad weight

400

c (mg/20g BW)

(mg/20g BW)

150

0 Ad Lib Dark Fed Light Fed Ad Lib Dark Fed Light Fed 16L : 8D 16L : 8D 16L : 8D 8L : 16D 8L : 16D 8L : 16D

100

50

d

300 200 100 0

0 Ad Lib Dark Fed Light Fed Ad Lib Dark Fed Light Fed 16L : 8D 16L : 8D 16L : 8D 8L : 16D 8L : 16D 8L : 16D

e

Ad Lib Dark Fed Light Fed Ad Lib Dark Fed Light Fed 16L : 8D 16L : 8D 16L : 8D 8L : 16D 8L : 16D 8L : 16D

Adrenal weight

300

f

Gastrocnemius weight

(mg/20g BW)

(mg/20g BW)

6

4

2

200

100

0

0 Ad Lib Dark Fed Light Fed Ad Lib Dark Fed Light Fed 16L : 8D 16L : 8D 16L : 8D 8L : 16D 8L : 16D 8L : 16D

Ad Lib Dark Fed Light Fed Ad Lib Dark Fed Light Fed 16L : 8D 16L : 8D 16L : 8D 8L : 16D 8L : 16D 8L : 16D

Fig. 2. Body weight and body composition of mice kept on either 16L:8D or 8L:16D and fed ad libitum or exclusively during the light or dark period for 42 days. (a) Body weight, (b) relative liver weight, (c) relative pancreas weight, (d) relative combined epigonadal and retroperitoneal fat pad weight, (e) relative adrenal weight and (f) relative gastrocnemius weight. The data are the mean ± SEM (body weight in grammes and tissues in mg/20 g body weight) n = 6 mice per group. The horizontal bars indicate that the difference between the groups was significant at P b 0.05.

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

3. Results 3.1. Experiment 1 For mice maintained in both 16L:8D and 8L:16D, wheel running activity was generally confined to the dark period and commenced almost immediately upon lights off (Fig. 1). In the 16L:8D photoperiod maximal running activity occurred early in the dark period and steadily declined throughout darkness. In the 8L:16D photoperiod, running activity was sustained for approximately 9 h, followed by sporadic running, until a short burst of running in the last 2 h before lights on. In the control mice fed ad libitum, this behaviour pattern was maintained throughout the 56 days of observation (data not shown). Restriction of food access to the dark period in 16L:8D did not change the pattern of wheel running dramatically although the amount of running had decreased by day 42. Mice provided with access to food only during the light period developed a clear biphasic pattern of running by day 42, with intense running activity for the first 4 h of darkness, a period of quiescence and then intense running commencing an hour before lights on. Feeding mice during darkness in 8L:16D resulted in consolidated running for the first 5 h of darkness, a long period of quiescence and then a burst of moderate intensity running starting approximately 2 h before lights on. A similar pattern of running activity was evident in the mice fed during the light phase, but the pre-dawn running intensity was very pronounced. Mice fed ad libitum while on the 16L:8D and 8L:16D photoperiods had patterns of wheel running similar to those fed during darkness (data not shown).

libitum fed and dark fed mice respectively (P b 0.05), while triglycerides were not different between the groups. 3.2.3. Glucose and insulin tolerance tests Mice maintained in 16L:8D and fed during the light period had a reduced (18.5%) total area under the blood glucose curve following glucose injection than that of dark fed mice (P b 0.05; Fig. 4). In mice maintained in 8L:16D there was no difference in glucose tolerance between groups. Mice maintained in 16L:8D and fed during the light period had a 30.5% reduction in the total area under the blood glucose curve following insulin injection than ad libitum fed mice (P b 0.05). In mice maintained in 8L:16D there was no difference in insulin tolerance between feeding groups. 3.3. Experiment 3 3.3.1. Glucose tolerance tests across 24 h Mice maintained in 16L:8D and fed during the light period for 28 days had the highest area under the blood glucose curve late in the

12

a

8 6 4 2 0 Ad Lib 16L : 8D

Dark Fed Light Fed 16L : 8D 16L : 8D

Ad Lib 8L : 16D

Dark Fed Light Fed 8L : 16D 8L : 16D

Insulin

3.2. Experiment 2 1.6

b

1.4

(ng/ml)

1.2 1.0 0.8 0.6 0.4 0.2 0.0 Ad Lib 16L : 8D

Dark Fed Light Fed 16L : 8D 16L : 8D

Ad Lib 8L : 16D

Dark Fed Light Fed 8L : 16D 8L : 16D

Triglyceride 2.0

c

1.5

(mM)

3.2.1. Food intake and body composition The amounts of food consumed across the experiment were comparable between photoperiods and between times of food presentation. For mice held in 16L:8D the estimated individual daily intake was 3.1, 3.1 and 2.9 g/day for mice fed during the light, dark or continuously respectively. For mice held in 8L:16D the estimated individual daily intake was 2.8, 2.9 and 2.9 g/day for mice fed during the light, dark or continuously respectively. Mice maintained on 16L:8D and fed during either the light or dark periods had a 17.2% and 21.0% reduction in liver weight compared to ad libitum fed mice (Fig. 2b, P b 0.05). The pancreas weight of mice fed during darkness was 28.8% heavier than that of ad libitum and light fed mice (Fig. 2c, P b 0.05). Body weight (Fig. 2a) and relative adrenal (Fig. 2e), combined epigonadal and retroperitoneal fat pad (Fig. 2d) and gastrocnemius weights (Fig. 2f) were unaffected. Mice maintained on 8L:16D and fed during the light period had a 21.1% and 26.5% decrease in relative fat pad weight (Fig. 2d) compared to ad libitum and dark fed mice respectively (Fig. 2, P b 0.05). Body weight and relative liver, pancreas, adrenal and gastrocnemius muscle weights were not affected.

Glucose

10

(mM)

under the curves (AUC) (from baseline for glucose and from zero for insulin) were analysed by univariate ANOVA. To determine whether the hormone, metabolite and IPGTT data and gene expression data were rhythmic (i.e., could be fitted to a sine curve) the data were analysed using CircWaveBatch; http://hutlab.nl/ [38].

193

1.0

0.5

3.2.2. Plasma metabolites and insulin Plasma glucose in mice maintained on 16L:8D was not different between the feeding groups (Fig. 3). Plasma insulin in the light fed mice was increased 3.2 fold and 3 fold over ad libitum fed and dark fed mice respectively (P b 0.05), while triglycerides were decreased by 55% compared to ad libitum fed mice (P b 0.05). Plasma glucose in mice maintained on 8L:16D and being fed during the light period tended to be lower (P = 0.058) than that in the ad libitum and dark fed mice. Plasma insulin was increased 3 fold and 1.3 fold over ad

0.0 Ad Lib 16L : 8D

Dark Fed Light Fed 16L : 8D 16L : 8D

Ad Lib 8L : 16D

Dark Fed Light Fed 8L : 16D 8L : 16D

Fig. 3. Plasma glucose, insulin and triglyceride levels of mice kept on either 16L:8D or 8L:16D and fed ad libitum or exclusively during the light or dark period for 42 days. (a) glucose (mM), (b) insulin (ng/ml) and (c) triglyceride (mM). The data are the mean ± SEM; n = 6 mice per group. The horizontal bars indicate that the difference between the groups was significant at P b 0.05.

194

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

16L : 8D

Glucose (mM)

20

8L : 16D 20

a

15

15

10

10

5

5

0

0 0 15 30

AUC (Arbitrary units)

2000

60

90

120

c

1500

1000

1000

500

500

0

Glucose (mM)

Dark fed

90

120

d

Light Fed

Ad Lib 15

e

10

10

5

5

0

Dark fed

Light Fed

f

0 0 15 30

1500

60

0 Ad Lib

AUC (Arbitrary units)

0 15 30 2000

1500

15

b

60

90

120

0 15 30 1500

g

1000

1000

500

500

60

90

120

h

0

0 Ad Lib

Dark fed

Light Fed

Ad Lib

Dark fed

Light Fed

Fig. 4. Glucose tolerance and insulin tolerance tests in mice kept on either 16L:8D or 8L:16D and fed ad libitum or exclusively during the light or dark period. (a, c) the time course of blood ), dark fed glucose (mM) following intra peritoneal administration of glucose (1 g/kg) and the area under the curve taken from zero in mice kept on 16L:8D for 35 days. Ad libitum fed ( ) and light fed ( ). The data are the mean ± SEM; n = 5 in each group. (b, d) the time course of blood glucose following intra peritoneal administration of glucose (1 g/kg) and the ( area under the curve taken from zero in mice kept on 16L:8D for 35 days. (e, g) the time course of blood glucose following intra peritoneal administration of insulin (0.75 m IU/kg) and the area under the curve taken from zero in mice kept on 16L:8D or (f, h) on 8L:16D for 42 days. The horizontal bars indicate that the difference between the groups was significant at P b 0.05.

light period, between ZT10 and ZT14, (Fig. 5). Mice fed during the dark period had the highest area under the blood glucose curve at ZT10, 6 h before the onset of darkness. While the data for the light fed mice could be fitted to a sine curve, the data for the dark fed mice did not change systematically. Interestingly, ANOVA (correcting for multiple comparisons) indicated that there was a significant difference between the 2 feeding groups at ZT14 (P = 0.001) with near maximal AUC for the light fed mice and lowest AUC for dark fed mice. The 2 hour fasting glucose levels varied rhythmically (P b 0.05) across the day such that highest blood glucose levels occurred at ZT14 and ZT11.5 in light- and dark-fed mice respectively and were significantly different between groups at ZT18 (P b 0.05). 3.3.2. Plasma glucose, insulin, triglycerides and corticosterone across 24 h Plasma glucose levels were rhythmic in mice fed during the light phase after 7 and 35 days, with peak levels occurring at ZT11.2 and

ZT11.6 respectively (Fig. 6). Interestingly overall glucose levels decreased across the experiment. A rhythmic change in plasma glucose levels was not evident in mice fed exclusively during darkness for 7 days, but after 35 days a clear rhythm was established with peak glucose levels occurring at ZT10.2 (Fig. 6). Plasma insulin was low at all times throughout the 24 h in light fed mice except at ZT2, soon after food was presented after 7 and 35 days of scheduled feeding. Plasma insulin secretion was rhythmic across 24 h in mice fed during darkness on both occasions, with peak levels occurring late in the dark period, at ZT20.5 and ZT21.1 after 7 and 35 days respectively. Mice fed during the light period for 7 days had a rhythm in plasma triglyceride levels with a peak at ZT6, 4 h later than the dark fed mice. After 35 days, however plasma triglyceride levels remained low and arrhythmic throughout the 24 h. In dark fed mice, plasma triglycerides were rhythmic after 7 days of scheduled feeding, with peak levels at ZT2. After 35 days of feeding levels were still rhythmic with the peak at a similar

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

20

a

average peak expression of these genes delayed by 3.5 ± 1.1 h and 3.3 ± 0.9 h in the light fed mice after 7 and 35 days respectively compared to dark fed mice. In addition, the amplitude of expression of these genes was damped in the light fed mice. Expression of Pck1 mRNA was rhythmic in light fed mice, with peak expression occurring at ZT16.5 and ZT15.2 after 7 and 35 days scheduled feeding (Fig. 8). Expression was also rhythmic in dark fed mice, with the peak occurring at ZT13.7 and ZT14 after 7 and 35 days. In both conditions the peaks occurred a few hours before the food was made available. For Pfkfb3 mRNA expression in the light fed mice, the highest expression occurred around the light to dark transition period, many hours before food presentation (Fig. 8), whereas for dark fed mice, the highest expression occurred just prior to feeding.

Glucose (mM)

15

10

5

0

AUC (Arbitrary units)

2000

195

b

*

*

1500

1000

500

0 14 18 22

2

6

10 14 18 22

2

6

10

Zeitgeber time Fig. 5. Glucose tolerance tests across 24 h in mice kept on 16L:8D and fed exclusively during the light or dark period. (a) Blood glucose levels in mice following intraperitoneal administration of glucose (1 g/kg) at ZT14, ZT18, ZT22, ZT2, ZT6 and ZT10. The data are the mean ± SEM; n = 5 for each time point, except at ZT10 (n = 6) and ZT14 (n = 8) at 35 days. Dark fed mice are shown as (●) and light fed mice as ( ). Note that the data has been double plotted (ZT14 through ZT10) to assist visualisation of the changes. The vertical grey bars indicate the period of darkness. (b) Area under the curves from zero for the glucose tolerance curves. * indicates the difference between the light fed and dark fed AUC at ZT14 was significant (P b 0.05).

time of day, but the amplitude was significantly reduced. The plasma corticosterone rhythm in the mice fed during the light period for 7 days was similar to that of the dark fed mice, with the peaks occurring at ZT14 and ZT14.6 respectively, just prior to the withdrawal of food. After 35 days on the schedule however the corticosterone rhythm of light fed mice had shifted such that the peak levels occurred at ZT20.7, just prior to the onset of feeding. Plasma corticosterone levels in dark fed mice were rhythmic after 35 days of feeding, with peak levels occurring at ZT14, just prior to the onset of darkness and feeding. 3.3.3. RT-PCR analysis of liver gene expression (time of day) The timing of the rhythm of expression of Bmal1 mRNA in the liver was not affected by 7 or 35 days of light or dark feeding, however after 35 days the amplitude of the expression rhythm was lower in light fed mice than dark fed mice at ZT22 and ZT6 (Fig. 7). The expression of Clock mRNA was rhythmic in mice fed during the light period, but unexpectedly not rhythmic in dark fed mice after 7 days (Fig. 7). As a consequence Clock mRNA levels in light fed mice were higher at ZT18 and lower at ZT22 and ZT2 than that in light fed mice. After 35 days of feeding Clock mRNA expression was rhythmic in both groups, with the peak expression occurring 5.3 h earlier in light fed than dark fed mice. The expression of Per1, Per2, (Fig. 7) Nr1d1 and Dbp mRNA (Fig. 8) was rhythmic in both light fed and dark fed mice, with the

3.3.4. Gene expression profile analysis of liver after 35 days of scheduled feeding Of the 16 genes analysed in the liver using this technique, for mice fed during the light period the mRNA expression pattern of 12 genes could be fitted to a sine curve; Nr1d1, Pck1, Pfkfb3, Prkaa2, Gck, Insr, Irs1, Ppara, Gsk3a, Sirt1, Gys2 and Pgc1a mRNA (Fig. 9). Note that for Gck and Sirt1, the signal peaks were low and the numbers of usable values were reduced, particularly at ZT22, ZT2 and ZT6 (see Supplementary Table 1.) Expression of Slc2a2, Pkcζ, Akt2 and Srebf1 mRNA was not rhythmic. The average time of peak expression for the 12 rhythmic genes was ZT21.3 ± 0.9, approximately 3 h before lights on and the feeding opportunity for that group. In mice fed only during darkness, the mRNA expression data for 12 genes could be fitted significantly to a sine curve; Nr1d1, Pck1, Pfkfb3, Slc2a2, Prkaa2, Gck, Ppara, Akt2, Pkcζ, Srebf1, Sirt1 and Gys2 (Fig. 9). The time of peak expression ranged from ZT10.7 for Nr1d1 to ZT22.9 for Pkcζ mRNA. Expression of Pgc1a, Insr, Irs1 and Gsk3a mRNA which were rhythmic in light fed mice were not rhythmic in this group. The average time of peak expression for the rhythmic genes was ZT16.5 ± 0.9, i.e. at the time of lights off and the onset of feeding. Two way ANOVA revealed that the overall expression of Sirt1 and Gys2 mRNA was lower in light fed than dark fed mice. There were significant time of day × feeding time interaction effects for Nr1d1, Pck1, Pfkfb3, Slc2a2, Gck, Pparα, Akt2, Srebf1, Sirt1 and Gys2 mRNA expression, further confirming the phase differences in the rhythmic expression of these genes between the feeding groups. Because there was a difference in glucose tolerance between the light fed and dark fed mice at ZT14, we analysed the gene expression data at this time by t-test after correcting for multiple comparisons (alpha = 0.008). Expression of Pck1, Pfkfb3, Akt2, Sirt1 and Gys2 mRNA was lower in the light fed mice compared to the dark fed mice, while Srebf1 mRNA expression was higher and there was a trend for expression of Ppara mRNA to be lower in light fed mice compared to dark fed mice (P = 0.014). 3.3.5. Liver gene expression (time from food presentation) To gain an understanding of the potential role of the initiation of feeding on the expression of the liver genes, the data from the light fed mice was re-analysed after shifting it forward by 8 h. Thus data collected at ZT18 from dark fed mice was paired with data collected at ZT2 from light fed mice and so on. Visual inspection of the graphical data (Supplementary Fig. S2) shows that the rhythms of Pck1, Gck, Pparα, Sirt1 and Gys2 mRNA expression are similar when plotted in relation to the time of food presentation. This was confirmed statistically by ANOVA which showed that the significant time of day × feeding time interactions originally observed for Pck1, Gck, Pparα, Sirt1 and Gys2 mRNA expression were lost. 4. Discussion In this study we addressed a series of questions about the physiological impact of providing access to food exclusively either during the dark

196

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

7 days

a

***

*** 14

12

12

10

10

8

8

6 3

6 3

Triglyceride (mM) Corticosterone (ng/ml)

35 days 16

14

Insulin (ng/ml)

Glucose (mM)

16

c

**

*** ***

**

*** ***

2

2

1

1

0 6

0 6

e

***

***

4

4

2

2

0 250

0 250

g

200

200

150

150

100

100

50

50

0

b

d

* **

f

h

* **

*

*

**

**

0 14 18 22 2

6 10 14 18 22 2

6 10

14 18 22 2

6 10 14 18 22 2

6 10

Zeitgeber time Fig. 6. Plasma glucose, insulin, triglyceride and corticosterone levels in mice kept on 16L:8D and fed exclusively during the light or dark period for 7 or 35 days. The data are the mean ± SEM; n = 5 for each time point, except at ZT10 (n = 6) and ZT14 (n = 8) at 35 days. Note that the data has been double plotted (ZT14 through ZT10) to assist visualisation of the changes. Plasma glucose (mM) from dark (●) and light fed mice ( ) after 7 days (a) and 35 days (b). (c, d) plasma insulin (ng/ml). (e, f) plasma triglycerides (mM). (g, h) plasma corticosterone (ng/ml). The vertical grey bars indicate the period of darkness. *P b 0.05. **P b 0.01. ***P b 0.001.

or light period in mice. These included the effect of restricting food access to mice held in both long (16L:8D) and short (8L:16D) photoperiods on wheel running behaviour, body weight and composition, glucose, insulin, triglycerides, corticosterone, glucose and insulin tolerance and liver gene expression. The main findings of the study are that restricted feeding had the largest impact on mice exposed to long days compared to short days. Restricted feeding during the light period in a long day photoperiod had no effect on food intake, body weight, adrenal, fat pad or muscle (gastrocnemius) weight, but decreased liver weight compared to ad libitum fed mice. Glucose tolerance (following a prolonged fast) and whole body insulin sensitivity tested 2 h after lights on were improved in mice fed during the light and when on the long photoperiod compared to ad libitum or dark fed mice. A high amplitude circadian rhythm in glucose tolerance (following a short fasting period) was evident in mice held on 16L:8D and were fed during the light but not in those fed during darkness. Glucose tolerance was significantly impaired in light fed mice at the end of the light period compared to that in dark fed mice. Analysis of liver clock gene expression (Bmal1, Clock, Per1, Per2, Nr1d1 and Dbp) after short (7 days) and long term (35 days) exposure to the altered feeding schedule resulted in an approximately 3.5 hour delay in the rhythms compared to dark fed

mice, while the rhythm of expression of a range of metabolic genes was delayed by approximately 5 h. Much of the interest in restricted feeding has centred upon the hypothesis that consuming meals at times which are out of synchrony with the circadian timing system may promote obesity. This in part stems from observations that the timing system is disrupted by a high fat diet [39] and from the study by Arble et al. [40] showing that mice given access to a high fat diet (60% kcal from fat) exclusively during the light period were heavier after 2 weeks than mice consuming the same diet during darkness. The effects of normal chow were not investigated. Bray et al. [28] conducted a restricted feeding study with a low fat diet (4.7% fat) for 9 days and reported that mice fed during the light phase gained more weight than those fed in the dark. Jang et al. [41] however, pair fed mice for 3 weeks with normal chow and a high fat diet and found no difference in weight gain between the restricted feeding groups on either diet. In the current study we observed no differences in weight gain between the restricted feeding groups in any of the trials. It is widely accepted that glucose tolerance varies across 24 h in humans [42] and rats [43], with the highest tolerance occurring soon after awakening, but to our knowledge there has only been one study

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

2.0

2.0

1.5

1.5

1.0

1.0

0.5

0.5

0.0 2.0

Relative mRNA expression

7 days

a

c

0.0 2.0

** ** ***

** ** ***

1.5

1.5

1.0

1.0

0.5

0.5

0.0 2.0

0.0 2.0

e

*

**

*

** 1.5

1.0

1.0

0.5

0.5

0.0

0.0

1.5

g

1.5

***

***

***

35 days

b

*

d

*

***

1.0

1.0

0.5

0.5

0.0

*

**

*

**

f ***

1.5

197

h

***

***

***

***

***

0.0 14 18 22 2

6 10 14 18 22 2

6 10

14 18 22 2

6 10 14 18 22 2

6 10

Zeitgeber time Fig. 7. Liver clock gene expression in mice kept on 16L:8D and fed exclusively during the light or dark period for 7 or 35 days. The data are the mean ± SEM; n = 5 for each time point, except at ZT10 (n = 6) and ZT14 (n = 8) at 35 days. For each gene the highest expression for the dark fed mice is set at 1. Note that the data has been double plotted (ZT14 through ZT10) to assist visualisation of the changes. Expression of Bmal1 mRNA from dark (●) and light fed mice ( ) after 7 days (a) and 35 days (b). (c, d) Clock mRNA. (e, f) Per1 mRNA. (g, h) Per2 mRNA. The vertical grey bars indicate the period of darkness. *P b 0.05. **P b 0.01. ***P b 0.001.

addressing this phenomenon in mice. Nowell [44] conducted intra venous glucose tolerance tests under anaesthesia on mice kept on 12L:12D that had been fasted for 24 h or allowed access to food up to the time of the test. In the fasted state there were no differences in glucose tolerance (measured as the change in glucose 10 min after the glucose bolus) across the 4 time points studied, 0, 5, 12 and 17 h after lights on. When the tolerance tests were conducted in the fed state, glucose tolerance was highest 17 h after lights on, that is, during the active feeding period. Investigating circadian rhythms in glucose tolerance in animals provides a challenge, because the traditional procedure involves fasting the animals either overnight for 14–18 h or for 5–6 h [45] and so the length of time from the last meal to the time of the test may be a confounder. This was of particular importance in the first part of the current study in which we restricted feeding to either the light or dark period in 2 different photoperiods and conducted the tests at a set time of day (2 h after lights on). We found that mice fed during the light period had improved glucose tolerance compared to those fed during darkness. However, we were concerned that the food was

withdrawn 10 h before and 26 h before the tolerance tests for the light and dark fed mice respectively. While overnight fasting is commonly used (for example [24]), we are aware that it can produce outcomes in mice that are different from those observed following prolonged fasting in humans [45]. In an attempt to compensate for this issue for the study of the rhythm of glucose tolerance, we therefore modified the approach to withdraw food from the mice 2 h before each test. Mice fed during the light period of a 16L:8D photoperiod and tested in this manner showed a circadian rhythm in glucose tolerance with greatest glycemic control late in the dark period and highest glucose excursion late in the light period. Glycemic control is normally highest early in the activity/sleep period in humans and rats [42,43]. In our study we did note that wheel running activity in light fed 16L:8D animals was essentially absent for several hours, starting around 4 h after lights off and continued until the pre-dawn increase in activity, but whether the mice were sleeping is not known. Similar food restriction of rats in 12L:12D reversed the diurnal distribution of REM sleep [46], while 4 hour food access during the light period in mice also altered

198

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

1.5

7 days

a

***

***

1.0

1.0

0.5

0.5

0.0 1.5

0.0 1.5

c *** **

Relative mRNA expression

1.5

1.0

0.5

0.5

0.0 1.5

0.0 1.5

e

1.0

*** ***

***

*** ***

1.0

0.5

0.5

0.0

0.0

2.0

g

2.0

***

*** 1.5

1.0

1.0

0.5

0.5

0.0

***

***

f *** *** ***

*** *** ***

h **

1.5

***

d ***

*** **

1.0

***

35 days

b

**

0.0 14 18 22 2

6 10 14 18 22 2

6 10

14 18 22 2

6 10 14 18 22 2

6 10

Zeitgeber time Fig. 8. Liver gene expression in mice kept on 16L:8D and fed exclusively during the light or dark period for 7 (left panel) or 35 days (right panel). Data and animal numbers as in Fig. 7. (a, b) Nr1d1 mRNA. (c, d) Dbp mRNA. (e, f) Pck1 mRNA. (g, h) Pfkfb3 mRNA. **P b 0.01. ***P b 0.001.

the distribution of REM sleep [47]. This suggests that the relationship between sleep and maximal glucose utilisation may be preserved in light fed mice and shifted into the dark period in preparation for the meal at lights on. There has been considerable interest in the effects of restricted feeding on circadian rhythms on clock gene expression. In the case of mouse experiments, however, all studies have used a 12L:12D photoperiod and most have been of short duration. The apparent consensus of previous studies is that under these conditions scheduled feeding results in a 12 hour reversal of clock gene rhythms. For example, mice fed exclusively during the light period showed a complete (12 h) reversal of rhythmic liver clock gene expression [29,48], but no change in SCN rhythmicity [29]. Satoh et al. [49] fed mice (ICR strain) for 9 h during the light period (2 h after lights on until 1 h before lights off) while the “dark” fed mice were fed for 15 h, from 1 h before lights off until 2 h after lights on. After 7 days of this regimen there was a phase shift of 8–10 h in liver rhythmicity. Jang et al. [41] pair fed mice for 4 weeks during the light or dark phase to ensure that both groups had the same food intake and measured liver gene expression 2 h after lights on and off. Liver Bmal1, Per2 and Clock mRNA expression, but not

hypothalamic expression of these genes, was shifted significantly, but estimates of the degree of the phase shift were not possible due to only 2 time points being studied. Bray et al. [28] fed mice (FVB/N strain) during the light or dark for 9 days using an automated system and reported that food intake was highest within a few hours of food presentation in the light. In the current study we observed an increase in activity just prior to feeding as did Bray et al. [28] and although we did not have access to an automated food monitoring system, we noticed that the mice began eating immediately on presentation of the food at the beginning of either the light or dark periods. When liver clock gene expression (Bmal1, Per2, Cry2, Nr1d1 and Dbp) was analysed at 4 times across 24 h, there were phase delays of 8.3, 10.7, 9, 6.2, and 7.5 h respectively. In our study of restricted feeding of mice on a 16L:8D photoperiod, we predicted that feeding during the light period would shift liver clock gene rhythms by 8 h. However, even after 35 days of scheduled feeding, the rhythm of Bmal1 expression was clearly unchanged and maximal at the time of lights on. Similarly the rhythm of Per1, Nr1d1, and Dbp mRNA expression shifted by no more than 3.4 h with respect to the dark fed mice. Clock mRNA expression was unexpectedly arrhythmic in dark fed mice after 7 days but by 35 days a

2.5

1.5

N r1 d 1

**

2.0

1.5

P fk f b 3

**

G SK3a

S lc 2 a 2

2.0

1.5 1.0

1.0

1.5

1.0 1.0 0.5

0.5 0.5

0.5 0.0 2.0

0.0 3

P rk a a 2

** **

Ir s 1

In s r

S re b f1

1.5

** **

2.0

1.5

2 1.5 1.0

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

Relative mRNA expression

0.0 2.5

0.0 2.0

1.0 1.0 1

0.5

0.5

0.0

0

1.5

1.5

0.5 0.0

0.0

1.5

P gc1a

A k t2

**

**

**

P ck1

P kc3z

3

***

**

1.0

1.0

1.0

*

***

***

*

***

2 0.5

0.5

0.0

0.0

0

4

2.0

1.5

G ck

**

**

1.5

S ir t 1

*

1.5

2

1.0

1

0.5

0

0.0 6 10 14 18 22 2

0.0

P p a ra

3

14 18 22 2

0.5

1

6 10

G ys2

** *

*

1.0

1.0

0.5

0.5

0.0

0.0 14 18 22 2

6 10 14 18 22 2

6 10

** *

14 18 22 2

6 10 14 18 22 2

6 10

14 18 22 2

6 10 14 18 22 2

6 10

Zeitgeber time Fig. 9. Liver gene expression (GeXP) for 16 genes in mice kept on 16L:8D and fed exclusively during the light or dark period for 35 days. The data are the mean ± SEM for n = 4–8 for each time point, except Gck and Sirt1 mRNA (see Supplementary Table 1). For each gene the expression for the dark fed mice at ZT14 is set at 1. *P b 0.05. **P b 0.01. ***P b 0.001. 199

200

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201

rhythm was apparent that was 5 h later than that of light fed mice. A possible explanation of this smaller than expected shift could be the gorging feeding pattern reported in light fed mice (and inferred here from plasma insulin concentrations) combined with the short 8 hour dark photoperiod used as opposed to the 12L:12D photoperiod used in other studies. Together this would move the timing of food consumption in the 2 groups closer together, and hence reducing the shift in gene expression observed. The only clock gene studied that did shift its rhythm by approximately 8 h was Per2, which was highest 2 h after feeding onset in both groups. It has been established previously that temperature can alter the expression of Per2 [50]. Although we did not study body temperature in the current study, increased activity provoked by the presentation of food at the commencement of the light period in mice [47] and rats [46] coincides with a large increase in core body temperature. In our study there was a consistent burst of wheel running activity just prior to food presentation in the light fed mice and this may have facilitated the induction of Per2 mRNA transcription. It is interesting that the core clock genes described here and metabolic genes (discussed below) are perturbed by differing degrees by the photoperiod/feeding protocol, being out of synchrony with each other by several hours. This may be as much due to the increased complexity of the longer day length as the ongoing conflict between the Zeitgebers of food access and light exposure. A deeper understanding of the longer term interacting effects, as opposed to synchronised peripheral/feeding rhythms in antiphase with photoperiod/central rhythms, as observed in 12L:12D protocols, could offer insights to minimise health risks to shiftworkers, who likewise experience unbalanced light:dark photoperiods. In an attempt to understand the basis for the differences in glucose tolerance we observed between light fed and dark fed mice, we analysed the expression across 24 h of an a priori selected panel of liver genes that are involved in various metabolic functions. Of the 16 genes that were assayed by GeXP and reported here, all except Sirt1, Akt2 and Prkcζ mRNA have been shown previously to be rhythmically expressed in the livers of mice in 12L:12D (see [9,51]). In the current study, mice kept on a 16L:8D photoperiod and fed only during darkness showed rhythmic expression of all the genes except for Insr, Irs1, Gsk3a and Pgc1a mRNA (the expression of these genes did vary significantly across 24 h but did not fit a sine curve). The peak expression of the genes generally occurred late in the light phase or during early darkness, consistent with the role of many of them in producing proteins/ enzymes involved in glucose and lipid metabolism. To our knowledge this is the first report of Sirt1 mRNA rhythmicity in the liver; previous studies conducted in 12L:12D have indicated that only SIRT1 protein has a circadian rhythm [19], presumably due to post-transcriptional processing. Similarly this is the first report of rhythmic expression of Akt2 mRNA in the liver. In the mice maintained on a 16L:8D and fed only during the light period, the time of ZT14 marked the time of greatest glucose intolerance, particularly as compared to their dark fed counterparts, who in this period evidenced strong metabolic control. In the light fed animals at this time, plasma insulin levels were trending higher and cortisol lower compared to dark fed animals, which may be playing a role in the observed reduced expression of Sirt1 [52], Pck1, [53], Pgc1a [54] and the increased expression Srebp1 [55]. Similarly, the trend for lower expression of Ppara/Pgc1a gene expression (perhaps due in part to low corticosterone [56]) may explain the lower expression of Gsys2 [57] and Akt. While it could be hypothesised that these changes may play a role in reducing insulin sensitivity and slowed glucose clearance observed at this time, these changes would need to be investigated at the protein level, due to the importance of post translational modification in the signalling and activity of these proteins. In the mice fed during the light phase, the genes that were expressed rhythmically in dark fed mice and retained rhythmicity in light fed animals had their peak expression 5 h later. Interestingly the robust rhythmicity of Slc2a2 and Akt2 mRNA expression was lost in the light fed

mice. If this was reflected in the pattern of protein accumulation, this would suggest that there may have been a loss of rhythmic hepatic glucose transport and downstream insulin signal transduction. Future studies could address this possibility. Similarly the significantly lower expression of Sirt1 mRNA prior to darkness at 0700 h in the light fed mice may also help to explain the poor glucose tolerance at this time. 5. Conclusion Restricting access to a normal chow diet to the light period does not alter food intake, body weight or increase adiposity in mice maintained on either long or short day lengths. Feeding during the light period in a long photoperiod does, however, result in large and systematic changes in glucose tolerance such that glycemic control is poorest at the end of the light/rest period in contrast to the strong plasma glucose control at this time in mice fed during darkness. Contrary to previous results of experiments conducted in 12L:12D, feeding during the light period in 16L:8D does not phase shift liver clock gene expression by the predicted 8 h. Indeed even after 35 days of scheduled feeding, the time of peak expression was shifted on average only 3 h, or not at all. Rhythms in expression of liver metabolic genes were also only shifted on average by 5 h or became arrhythmic when mice were fed during the light period. Of the genes studied only the rhythms of Gck, Pck1, Ppara, Sirt1 and Gys2 mRNA shifted by approximately 8 h, potentially reflecting their roles in the altered timing of maximal glucose tolerance. Together this study presents further evidence that not only circadian rhythms facilitate metabolic control, but also different environmental events, including season and feeding opportunities, alter aspects of circadian and metabolic physiology. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.physbeh.2014.02.021. Conflict of interest statement No conflict of interest was declared. Acknowledgements This research was supported in part by a grant from the National Health and Medical Research Council (NHMRC) of Australia (Grant # 399131 and 1029869). DJK is a Senior Research Fellow of the NHMRC. References [1] Garaulet M, Gomez-Abellan P, Alburquerque-Bejar JJ, Lee YC, Ordovas JM, Scheer FAJL. Timing of food intake predicts weight loss effectiveness. Int J Obes 2013;37:604–11. [2] Jakubowicz D, Barnea M, Wainstein J, Froy O. High caloric intake at breakfast vs. dinner differentially influences weight loss of overweight and obese women. Obesity 2013;21:2504–12. [3] Ekmekcioglu C, Touitou Y. Chronobiological aspects of food intake and metabolism and their relevance on energy balance and weight regulation. Obes Rev 2011;12:14–25. [4] Wang X-S, Armstrong MEG, Cairns BJ, Key TJ, Travis RC. Shift work and chronic disease: the epidemiological evidence. Occup Med 2011;61:78–89. [5] Killick R, Banks S, Liu PY. Implications of sleep restriction and recovery on metabolic outcomes. J Clin Endocrinol Metab 2012;97:3876–90. [6] Boden MJ, Varcoe TJ, Kennaway DJ. Circadian regulation of reproduction: from gamete to offspring. Prog Biophys Mol Biol 2013;113:387–97. [7] Canaple L, Rambaud J, Dkhissi-Benyahya O, Rayet B, Tan NS, Michalik L, et al. Reciprocal regulation of brain and muscle Arnt-like protein 1 and peroxisome proliferator-activated receptor alpha defines a novel positive feedback loop in the rodent liver circadian clock. Mol Endocrinol 2006;20:1715–27. [8] Gervois P, Chopin-Delannoy S, Fadel A, Dubois G, Kosykh V, Fruchart JC, et al. Fibrates increase human REV-ERBalpha expression in liver via a novel peroxisome proliferator-activated receptor response element. Mol Endocrinol 1999;13:400–9. [9] Liu C, Li S, Liu T, Borjigin J, Lin JD. Transcriptional coactivator PGC-1alpha integrates the mammalian clock and energy metabolism. Nature 2007;447:477–81. [10] Jordan SD, Lamia KA. AMPK at the crossroads of circadian clocks and metabolism. Mol Cell Endocrinol 2013;366:163–9. [11] Sahar S, Zocchi L, Kinoshita C, Borrelli E, Sassone-Corsi P. Regulation of BMAL1 protein stability and circadian function by GSK3beta-mediated phosphorylation. PLoS One 2010;5:e8561.

N.A. Shamsi et al. / Physiology & Behavior 128 (2014) 188–201 [12] Harada Y, Sakai M, Kurabayashi N, Hirota T, Fukada Y. Ser-557-phosphorylated mCRY2 is degraded upon synergistic phosphorylation by glycogen synthase kinase-3 beta. J Biol Chem 2005;280:31714–21. [13] Spengler ML, Kuropatwinski KK, Schumer M, Antoch M. A serine cluster mediates BMAL1-dependent CLOCK phosphorylation and degradation. Cell Cycle 2009;8:4138–46. [14] Sanada K, Okano T, Fukada Y. Mitogen-activated protein kinase phosphorylates and negatively regulates basic helix-loop-helix-PAS transcription factor BMAL1. J Biol Chem 2002;277:267–71. [15] Sanada K, Harada Y, Sakai M, Todo T, Fukada Y. Serine phosphorylation of mCRY1 and mCRY2 by mitogen-activated protein kinase. Genes Cells 2004;9:697–708. [16] Iitaka C, Miyazaki K, Akaike T, Ishida N. A role for glycogen synthase kinase-3beta in the mammalian circadian clock. J Biol Chem 2005;280:29397–402. [17] Yin L, Wang J, Klein PS, Lazar MA. Nuclear receptor Rev-erbalpha is a critical lithiumsensitive component of the circadian clock. Science 2006;311:1002–5. [18] Nakahata Y, Kaluzova M, Grimaldi B, Sahar S, Hirayama J, Chen D, et al. The NAD +− dependent deacetylase SIRT1 modulates CLOCK-mediated chromatin remodeling and circadian control. Cell 2008;134:329–40. [19] Asher G, Gatfield D, Stratmann M, Reinke H, Dibner C, Kreppel F, et al. SIRT1 regulates circadian clock gene expression through PER2 deacetylation. Cell 2008;134:317–28. [20] Kennaway DJ, Owens JA, Voultsios A, Varcoe TJ. Functional central rhythmicity and light entrainment, but not liver and muscle rhythmicity, are Clock independent. Am J Physiol 2006;291:R1172–80. [21] Turek FW, Joshu C, Kohsaka A, Lin E, Ivanova G, McDearmon E, et al. Obesity and metabolic syndrome in circadian Clock mutant mice. Science 2005;308:1043–5. [22] Kennaway DJ, Owens JA, Voultsios A, Boden MJ, Varcoe TJ. Metabolic homeostasis in mice with disrupted Clock gene expression in peripheral tissues. Am J Physiol 2007;293:R1528–37. [23] Kennaway DJ, Owens JA, Voultsios A, Wight N. Adipokines and adipocyte function in Clock mutant mice that retain melatonin rhythmicity. Obesity 2012;20:295–305. [24] Marcheva B, Ramsey KM, Buhr ED, Kobayashi Y, Su H, Ko CH, et al. Disruption of the clock components CLOCK and BMAL1 leads to hypoinsulinaemia and diabetes. Nature 2010;466:627–31. [25] Lamia KA, Storch KF, Weitz CJ. Physiological significance of a peripheral tissue circadian clock. Proc Natl Acad Sci U S A 2008;105:15172–7. [26] Kennaway DJ, Varcoe TJ, Voultsios A, Boden MJ. Global loss of Bmal1 expression alters adipose tissue hormones, gene expression and glucose metabolism. PLoS One 2013;8:e65255. [27] Paschos GK, Ibrahim S, Song W-L, Kunieda T, Grant G, Reyes TM, et al. Obesity in mice with adipocyte-specific deletion of clock component Arntl. Nat Med 2012;18:1768–77. [28] Bray MS, Ratcliffe WF, Grenett MH, Brewer RA, Gamble KL, Young ME. Quantitative analysis of light-phase restricted feeding reveals metabolic dyssynchrony in mice. Int J Obes (Lond) 2013;37:843–52. [29] Damiola F, Le Minh N, Preitner N, Kornmann B, Fleury-Olela F, Schibler U. Restricted feeding uncouples circadian oscillators in peripheral tissues from the central pacemaker in the suprachiasmatic nucleus. Genes Dev 2000;14:2950–61. [30] Wu T, Jin Y, Ni Y, Zhang D, Kato H, Fu Z. Effects of light cues on re-entrainment of the food-dominated peripheral clocks in mammals. Gene 2008;419:27–34. [31] Wu T, Jin Y, Kato H, Fu Z. Light and food signals cooperate to entrain the rat pineal circadian system. J Neurosci Res 2008;86:3246–55. [32] Glad CA, Kitchen EE, Russ GC, Harris SM, Davies JS, Gevers EF, et al. Reverse feeding suppresses the activity of the GH axis in rats and induces a preobesogenic state. Endocrinology 2011;152:869–82. [33] Salgado-Delgado R, Angeles-Castellanos M, Saderi N, Buijs RM, Escobar C. Food intake during the normal activity phase prevents obesity and circadian desynchrony in a rat model of night work. Endocrinology 2010;151:1019–29. [34] Vollmers C, Gill S, DiTacchio L, Pulivarthy SR, Le HD, Panda S. Time of feeding and the intrinsic circadian clock drive rhythms in hepatic gene expression. Proc Natl Acad Sci U S A 2009;106:21453–8.

201

[35] Kennaway DJ, Varcoe TJ, Mau VJ. Rhythmic expression of clock and clock-controlled genes in the rat oviduct. Mol Hum Reprod 2003;9:503–7. [36] Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) Method. Methods 2001;25:402–8. [37] Rai A, Kamath R, Gerald W, Fleisher M. Analytical validation of the GeXP analyzer and design of a workflow for cancer-biomarker discovery using multiplexed geneexpression profiling. Anal Bioanal Chem 2009;393:1505–11. [38] Oster H, Damerow S, Hut RA, Eichele G. Transcriptional profiling in the adrenal gland reveals circadian regulation of hormone biosynthesis genes and nucleosome assembly genes. J Biol Rhythms 2006;21:350–61. [39] Kohsaka A, Laposky AD, Ramsey KM, Estrada C, Joshu C, Kobayashi Y, et al. High-fat diet disrupts behavioral and molecular circadian rhythms in mice. Cell Metab 2007;6:414–21. [40] Arble DM, Bass J, Laposky AD, Vitaterna MH, Turek FW. Circadian timing of food intake contributes to weight gain. Obesity 2009;17:2100–2. [41] Jang H, Lee G, Kong J, Choi G, Park YJ, Kim JB. Feeding period restriction alters the expression of peripheral circadian rhythm genes without changing body weight in mice. PLoS One 2012;7:e49993. [42] Van Cauter E, Polonsky KS, Scheen AJ. Roles of circadian rhythmicity and sleep in human glucose regulation. Endocr Rev 1997;18:716–38. [43] la Fleur SE, Kalsbeek A, Wortel J, Fekkes ML, Buijs RM. A daily rhythm in glucose tolerance: a role for the suprachiasmatic nucleus. Diabetes 2001;50:1237–43. [44] Nowell N. Circadian rhythm of glucose tolerance in laboratory mice. Diabetologia 1970;6:488–92. [45] Ayala JE, Samuel VT, Morton GJ, Obici S, Croniger CM, Shulman GI, et al. Standard operating procedures for describing and performing metabolic tests of glucose homeostasis in mice. Dis Model Mech 2010;3:525–34. [46] Roky R, Kapas L, Taishi TP, Fang J, Krueger JM. Food restriction alters the diurnal distribution of sleep in rats. Physiol Behav 1999;67:697–703. [47] Szentirmai E, Kapas L, Sun YX, Smith RG, Krueger JM. Restricted feeding-induced sleep, activity, and body temperature changes in normal and preproghrelindeficient mice. Am J Physiol Regul Integr Comp Physiol 2010;298:R467–77. [48] Bur IM, Zouaoui S, Fontanaud P, Coutry N, Molino F, Martin AO, et al. The comparison between circadian oscillators in mouse liver and pituitary gland reveals different integration of feeding and light schedules. PLoS One 2010;5:e15316. [49] Satoh Y, Kawai H, Kudo N, Kawashima Y, Mitsumoto A. Time-restricted feeding entrains daily rhythms of energy metabolism in mice. Am J Physiol Regul Integr Comp Physiol 2006;290:R1276–83. [50] Kornmann B, Schaad O, Bujard H, Takahashi JS, Schibler U. System-driven and oscillator-dependent circadian transcription in mice with a conditionally active liver clock. PLoS Biol 2007;5:e34. [51] Pizarro A, Hayer K, Lahens NF, Hogenesch JB. CircaDB: a database of mammalian circadian gene expression profiles. Nucleic Acids Res 2013;41:D1009–13. [52] Xiong S, Salazar G, Patrushev N, Alexander RW. FoxO1 mediates an autofeedback loop regulating SIRT1 expression. J Biol Chem 2011;286:5289–99. [53] Granner DK, Andreone TL. Insulin modulation of gene expression. Diabetes Metab Rev 1985;1:139–70. [54] Daitoku H, Yamagata K, Matsuzaki H, Hatta M, Fukamizu A. Regulation of PGC-1 promoter activity by protein kinase B and the forkhead transcription factor FKHR. Diabetes 2003;52:642–9. [55] Fleischmann M, Iynedjian PB. Regulation of sterol regulatory-element binding protein 1 gene expression in liver: role of insulin and protein kinase B/cAkt. Biochem J 2000;349:13–7. [56] Lemberger T, Saladin R, Vazquez M, Assimacopoulos F, Staels B, Desvergne B, et al. Expression of the peroxisome proliferator-activated receptor alpha gene is stimulated by stress and follows a diurnal rhythm. J Biol Chem 1996;271:1764–9. [57] Mandard S, Stienstra R, Escher P, Tan NS, Kim I, Gonzalez FJ, et al. Glycogen synthase 2 is a novel target gene of peroxisome proliferator-activated receptors. Cell Mol Life Sci 2007;64:1145–57.