Sensitivity of housekeeping genes in the suprachiasmatic nucleus of the mouse brain to diet and the daily light–dark cycle

Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

b r a i n r e s e a r c h 1 5 7 5 ( 2 0 1 4 ) 7 2 – 7 7

http://dx.doi.org/10.0006-8993/& 2014 El

Abbreviations: HKG

B-Act, beta actin; B

4A isoform 2; Gapdh

Ubc, ubiquitin C; YnCorresponding autE-mail address: j.1Present address

Research Report

Sensitivity of housekeeping genes in thesuprachiasmatic nucleus of the mouse brain to dietand the daily light–dark cycle

Jane K. Clealn, James N. Shepherd, Jasmine L. Shearer, Kimberley D. Bruce1,Felino R. Cagampang

Institute of Developmental Sciences, University of Southampton Faculty of Medicine, Southampton General Hospital(mailpoint 887), Southampton SO16 6YD, UK

a r t i c l e i n f o

Article history:

Accepted 21 May 2014

The endogenous timing system within the suprachiasmatic nuclei (SCN) of the hypotha-

lamus drives the cyclic expression of the clock molecules across the 24 h day–night cycle

Available online 2 June 2014

Keywords:

Housekeeping gene

Suprachiasmatic nucleus

High fat diet

Light–dark cycle

Mouse

1016/j.brainres.2014.05.03sevier B.V. All rights rese

, housekeeping genes;

2M, beta-2-microglobul

, glyceraldehyde 3-pho

whaz, tyrosine 3-monohor. Fax: þ44 23 8210 [email protected] (J.K.: Department of Metabol

a b s t r a c t

controlling downstream molecular pathways and physiological processes. The developing

fetal clock system is sensitive to the environment and physiology of the pregnant mother

and as such disruption of this system could lead to altered physiology in the offspring.

Characterizing the gene profiles of the endogenous molecular clock system by quantitative

reverse transcription polymerase chain reaction is dependent on normalization by

appropriate housekeeping genes (HKGs). However, many HKGs commonly used as internal

controls, although stably expressed under control conditions, can vary significantly in their

expression under certain experimental conditions. Here we analyzed the expression of 10

classic HKG across the 24 h light–dark cycle in the SCN of mouse offspring exposed to

normal chow or a high fat diet during early development and in postnatal life. We found

that the HKGs glyceraldehyde-3-phosphate dehydrogenase, beta actin and adenosine

triphosphate synthase subunit to be the most stably expressed genes in the SCN regardless

of diet or time within the 24 h light–dark cycle, and are therefore suitable to be used as

internal controls. However SCN samples collected during the light and dark periods did

show differences in expression and as such the timing of collection should be considered

when carrying out gene expression studies.

& 2014 Elsevier B.V. All rights reserved.

1rved.

SCN, suprachiasmatic nucleus; Atp5b, adenosine triphosphate synthase subunit;

in; Cyc1, cytochrome c-1; Can, Calnexin; Eif4a2, eukaryotic translation initiation factor

stphate dehydrogenase; Sdha, succinate dehydrogenase complex subunit A;

oxygenase/tryptophan 5-monooxygenase activation protein.Cleal).ism and Aging, The Scripps Research Institute, Jupiter, FL 33458, USA.

b r a i n r e s e a r c h 1 5 7 5 ( 2 0 1 4 ) 7 2 – 7 7 73

1. Introduction

Most mammals have an endogenous timing system in thesuprachiasmatic nuclei (SCN) of the hypothalamic region ofthe brain. It consists of an intracellular feedback loop coordi-nating the expression of molecular transcripts (termed as‘clock’ genes and clock-controlled genes) and their constitu-tive protein to oscillate across the 24 h day–night cycle (Buhrand Takahashi, 2013; Cagampang and Bruce, 2012). These24 h oscillations bring about rhythmic changes in down-stream molecular pathways and physiological processes, aswell as overt rhythmic changes in behavior such as the sleep–wake cycle, locomotor activity and feeding. Environmentalcues or Zeitgebers, such as the daily light–dark cycle, tem-perature and feeding cycles, are able to entrain this ‘clock’system to maintain its integrity and temporal coordination(Panda and Hogenesch, 2004; Roenneberg et al., 2007). Studiesare now emerging that similar clock systems are also befound in non-SCN neurons and in peripheral tissues, includ-ing the heart (Peirson et al., 2006; Young et al., 2001)and theliver (Davidson et al., 2004; Mohawk et al., 2012; Peirson et al.,2006). Nevertheless, signals from the ‘central’ clock in theSCN are required to maintain daily rhythms in these tissues.

Disrupting the integrity and temporal coordination of theclock system, in the SCN and peripheral tissues, can leadto hormonal imbalances, sleep disorders, cardiometabolicdiseases, and susceptibility to cancer (Mahoney, 2010; Ranaand Mahmood, 2010; Rosenwasser, 2010; Takeda andMaemura, 2010). Malnutrition, even as early as the develop-mental period, can also have adverse effect on clock function.Studies have shown maternal dietary protein restrictionduring pregnancy adversely affected the quality of thesleep–wake cycle and locomotor activity rhythm in rat off-spring (Datta et al., 2000; Duran et al., 2005). Thus thedeveloping fetal clock system is sensitive to the environmentand physiology of the pregnant mother.

The gene and protein components of the endogenousmolecular clock, and their interaction have been well char-acterized (Buhr and Takahashi, 2013; Dibner et al., 2010;Rosbash, 2009). However, it remains to be elucidated howthe various environmental cues impact on clock function.One of the key techniques used to characterize the geneprofiles of the endogenous molecular clock system within theSCN is quantitative reverse transcription polymerase chainreaction (qRT-PCR). However, the accuracy of such data isdependent on normalization by appropriate internal controlreference genes, termed housekeeping genes (HKGs). Identi-fying stable control genes is important to ensure the validityof gene expression studies, as normalizing the gene ofinterest using the geometric mean of multiple internal con-trol genes increases the quality of the data (Vandesompeleet al., 2002). This controls for variables such as the amountor quality of starting material, enzymatic efficiencies, anddifferences between tissues or cells in overall transcriptionalactivity (Vandesompele et al., 2002).

We and others have previously shown that the expressionlevel of these endogenous HKGs varies according to tissuetype (Bruce et al., 2012; Hsiao et al., 2001; Sadek et al., 2012a,2012b) and developmental stage (Sellayah et al., 2008;

Warrington et al., 2000). They are constitutively expressedin the tissue and as they mediate basic cellular function werethought unlikely to vary due to the experimental conditionsbeing investigated. However, many HKGs commonly used asinternal controls, although stably expressed under physiolo-gical conditions, can vary significantly in their expressionunder certain experimental conditions (Radonic et al., 2004;Suzuki et al., 2000). For example in rat brain, the expression ofthe HKGs 18S rRNA and cyclophilin B (CypB) was altered inthe cortex and hippocampus following dietary restriction anddexamethasone treatment (Tanic et al., 2007). Other condi-tions such as hypoxia not only altered CypB but also thecommonly used HKG glyceraldehyde 3-phostphate dehydro-genase (Gapdh) and beta actin (B-Act) (Zhong and Simons,1999). It is therefore vital to establish the most stable HKGs inthe tissue of interest when selecting internal controls underthe experimental conditions being investigated.

In this study we investigated the expression of 10 classicHKG across the 24 h light–dark cycle in the SCN of mouseoffspring exposed to normal chow or high fat nutrition duringearly development and in postnatal life. The HKGs Gapdh,B-Act and adenosine triphosphate synthase subunit (Atp5b)were found to be the most stably expressed genes in the SCNregardless of diet or time within the 24 h light–dark cycle andare therefore suitable internal controls. However SCN sam-ples collected during the light and dark periods did showdifferences in expression and as such the timing of collectionshould be considered when carrying out qPCR studies.

2. Results

In SCN samples collected during the dark period, there was asignificant decrease in the expression of Atp5b (4.5-fold,p¼0.009), B-Act (3.6-fold, p¼0.02), Cyc1 (3.9-fold, p¼0.03),Eif4a2 (5.4-fold, p¼0.001), Ubc (3.4-fold, p¼0.01), Gapdh (4.7-fold, p¼0.04), and Ywhaz (2.8-fold, p¼0.04), and a trendtowards reduction in Sdha (2.8-fold, p¼0.08) and B2M(1.5-fold, p¼0.06) compared to samples collected during thelight period (Fig. 1). There was no significant effect of diet ongene expression (Fig. 1).

The 10 genes were ordered according to stability (Fig. 2)and the number of genes required for normalization deter-mined (Fig. 3). The pair-wise variation with the sequentialaddition of each reference gene indicated that two genes aresufficient as internal controls as the addition of a third genegave a V score below 0.15 indicating that the additional genehad no significant contribution to the normalization factor(Fig. 3). The housekeeping genes Gapdh, B-Act and Atp5bshowed the highest stability with all SCN samples analyzedtogether (Fig. 2) and when samples in the dark period wereanalyzed separately. For samples in the light period, Sdhareplaced Atp5b as the third most stable gene.

3. Discussion

Accurate qPCR data is dependent on normalization to coun-teract sample and experimental variation. The most commonapproach is to normalize mRNA data to the expression of an

0

200

400

600

800

1000

1200

Del

ta C

T

B-Act

Gapdh

Atp5b

0

200

400

600

800

1000

1200

Del

ta C

TSdha

B2M

Cyc1

0

200

400

600

800

1000

1200

Del

ta C

T

Canx

Ubc

Ywhaz

Eif4a2

LIGHT DARK LIGHT DARK

LIGHT DARK LIGHT DARK

LIGHT DARK LIGHT DARK

Fig. 1 – SCN delta Ct values for the housekeeping genes (a) Sdha, B2M, Cyc1, (b) B-Act, Gapdh, Atp5b and (c) Canx, Ubc, Ywhaz,Eif4a2. Dams and offspring were fed standard chow (C) or a high fat diet (HF). Samples were collected from the offspring at 6time points throughout the 24 h light-dark cycle.

b r a i n r e s e a r c h 1 5 7 5 ( 2 0 1 4 ) 7 2 – 7 774

endogenous control gene thought to be stably expressed.However numerous studies now show that such HKGs arenot stably expressed under all experimental conditions high-lighting the need to investigate the appropriate HKG for eachindividual study. We have demonstrated in this study thatthe expression of potential HKGs in the SCN does alterthroughout the 24 h light–dark cycle and may therefore beunder circadian regulation. Using the geNorm method we

have established that the most stable HKGs in the SCNregardless of dietary exposure or time within the 24 h light–dark cycle were Gapdh, B-Act and Atp5b. The geometric meanof these three HKGs could therefore be used for normal-ization of circadian related gene expression in the SCN.We did however observe that SCN samples collected duringthe light and dark periods show significant differences inexpression and as such the timing of collection should be

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Eif4a2 Ywhaz Ubc Canx B2M Cyc1 Sdha Atp5b B-Act Gapdh

Ave

rage

exp

ress

ion

stab

ility

(M)

Housekeeping gene

Fig. 2 – The most stably expressed housekeeping genes.A graph showing the average expression stability value (M)for each housekeeping gene ranked according to increasingstability with the most stable genes on the right(geNorm 3.4).

0

0.05

0.1

0.15

0.2

0.25

3rd 4th 5th 6th 7th 8th 9th 10th

Var

iatio

n (V

sco

re)

Additional control genes

Fig. 3 – The number of genes required for normalization.A graph showing the levels of variation in averagehousekeeping gene stability with the sequential addition ofeach housekeeping gene to the equation (for calculating thenormalization factor). Starting with the most stablyexpressed genes on the left with the inclusion of a 3rd, 4th,5th gene, etc. moving to the right (geNorm 3.4). The graphindicates that the two most stable genes create anormalization factor which is not significantly altered by theaddition of 3–8 more genes as they all have a V scorebelow 0.15.

b r a i n r e s e a r c h 1 5 7 5 ( 2 0 1 4 ) 7 2 – 7 7 75

considered when carrying out qPCR studies. Furthermore, itwould be interesting to examine in future studies whetherdifferences in expression during the light–dark cycle aremaintained or become stable in the SCN of animals kept inconstant light or constant darkness.

Gene expression studies are subject to inherent variationdue to factors such as the amount or quality of startingmaterial, enzymatic efficiencies, and differences betweentissues or cells in overall transcriptional activity. A commonapproach to control such variation is to normalize the geneexpression data using the geometric mean of an appropriatenumber of HKGs thus increasing the quality of the data(Vandesompele et al., 2002). In this study we establishedusing geNorm analysis that two HKGs could be used fornormalization of gene expression in the SCN. The consensus

of opinion is however the use of a third gene for increasedaccuracy of normalization (Meller et al., 2005).

When studying tissues or molecular components of thecircadian clock system it is critical that selected HKGs are notunder circadian control. However we did find that severalcommonly used HKGs have altered expression levels withregards to the time within the 24 h light–dark cycle the SCNsample was collected. We therefore recommend using thehousekeeping genes Gapdh, B-Act and Atp5b as internalcontrols. These appear to be the most stably expressed genesin the SCN regardless of diet or time within the 24 h light–dark cycle. It is recommended that all samples are collectedwithin either the light or dark period as there are cleardifferences in gene expression between these two lightingperiods. This could be due to the animal's sleep wake cycleand as such needs to be considered when designing a geneexpression study. There are other approaches to data normal-ization, and this includes searching for novel stable genesfrom a genome-wide background using microarray datacompiled from a set experimental condition (Hruz et al.,2011; McCall and Almudevar, 2012). Nevertheless, these novelcandidates will still need to be validated against commonlyused reference genes.

These HKGs are involved in the basal cellular functionsand structures, metabolism and the cytoskeleton, and assuch should be constantly expressed in the cell regardlessof the time within the light or dark period. Atp5b encodes asubunit of the mitochondrial enzyme ATP synthase whichcatalyzes ATP synthesis during oxidative phosphorylation(Jabs et al., 1994). Gapdh is a key enzyme in glycolysis,breaking down glucose for energy and metabolism (Sirover,2005), while B-Act encodes a ubiquitous cytoskeletal proteinand is expressed in most cell types (Bassell et al., 1994). Gapdhand B-Act expressions in the hypothalamus and cortex of ratbrain have previously been shown to be stable followingtreatment such as dietary restriction and dexamethasone(Tanic et al., 2007). Interestingly B-Act but not Gapdh wasshown to be stable following a mismatch of pre and post-natal nutrition (Sellayah et al., 2008). In the human placenta,Gapdh and B-Act were found to be the commonly usedhousekeeping genes (Murthi et al., 2008). Atp5b has beenrecommended as a HKG as it is stably expressed inthe mammary gland across the estrous cycle phases(Hvid et al., 2011).

4. Conclusion

The present study assessed the suitability of several HKGs asinternal controls for qPCR experiments using SCN samplescollected at different time points across the 24 h light–darkcycle following pre and postnatal exposure to normal chow orhigh fat nutrition. We found that Gapdh, B-Act and Atp5b arethe most stable HKGs under these conditions and thereforesuitable for normalization of circadian related gene expres-sion in the SCN. Whether samples are collected within thelight or dark period should however be considered whencarrying out qPCR studies as HKGs will be influenced by thesleep–wake state of the individual.

b r a i n r e s e a r c h 1 5 7 5 ( 2 0 1 4 ) 7 2 – 7 776

5. Experimental procedures

5.1. Experimental animals and treatments

Animal procedures were carried out in accordance with theUK Animals (Scientific procedures) Act 1986. Female C57BL/6Jmice were maintained under a 12 h light-dark cycle (lights onat 07:00 h), and at a constant temperature of 2272 1C withfood and water available ad libitum. Dams were randomlyallocated to one of two diets: control (C; 7.4% kcal fat; SpecialDietary Services, Witham, Essex, UK), or a high fat diet(HF; 45% kcal fat; Special Dietary Services). Dams were fedthe designated diet 6 weeks pre-pregnancy, during pregnancyand lactation. Offspring were weaned at 3 weeks of age andwere randomly allocated to either the C or HF diet resulting intwo offspring groups; C/C and HF/HF. These offspring werefed the diets for the next 12 weeks. We have shown inprevious studies that this experimental mouse model willinduce a metabolic syndrome-like phenotype in the adultoffspring (Bruce et al., 2009; Elahi et al., 2009). This includesobesity, raised blood pressure, altered physical activity level,glucose metabolism and appetite, and substantial fat deposi-tion in the liver.

At 15 weeks of age, male offspring were killed by cervicaldislocation at 6 time points across the 24 h light–dark period,at 07:00 h (Zeitgeber time 0 or ZT0; dark to light transitionperiod), at 11.00 h (Zeitgeber time 4 or ZT4; 4 h into the lightperiod), at 15.00 h (Zeitgeber time 8 or ZT8; 8 h into the lightperiod), at 19.00 h (Zeitgeber time 12 or ZT12; light to darktransition period), 23.00 h (Zeitgeber time 16 or ZT16; 4 h intothe dark period), and at 03.00 h (ZT20; 8 h into the darkperiod). Brains were collected, snap frozen in liquid nitrogenand stored at �80 1C. The SCN region was micropunchedfrom coronal brain sections according to previous protocols(Chong et al., 1996) using a brain punch device (StoeltingEurope, IRL) and processed for RT-qPCR.

5.2. RNA extraction and cDNA synthesis

Total RNA was isolated from SCN samples (n¼4�6 per timepoint) using Trifast reagent (Peqlab, Erlangen, Germany)according to manufacturer's instructions. RNA quantification,quality and integrity were determined via (260/230 and 260/280ratios and concentrations) spectrophotometer and agarose gelelectrophoresis. Total RNA (1 mg) was reverse transcribed intocDNA using reverse M-MLV transcriptase (Promega, South-ampton, UK). Samples were incubated at 37 1C for 1 h followedby an enzyme deactivation step at 75 1C for 10min. The cDNAsynthesized was then diluted to a concentration of 5 ng/mlbefore amplification.

5.3. Quantitative reverse transcription polymerase chainreaction (qRT-PCR)

qRT-PCR was performed using primers and Perfect Probe(mouse geNorm kit, PrimerDesign, UK) to measure the levelof expression of 10 HKGs: Atp5b, B-Act, Gapdh, as well as beta-2-microglobulin (B2M), cytochrome c-1 (Cyc1), calnexin (Canx),eukaryotic translation initiation factor 4A isoform 2 (Eif4a2),

succinate dehydrogenase complex, subunit A (Sdha), ubiqui-tin C (Ubc), and tyrosine 3-monooxygenase/tryptophan5-monooxygenase activation protein (Ywhaz). All sampleswere measured in triplicate using a Roche Light-Cycler 480real-time PCR system. The cycle parameters were 95 1C for10 min, followed by 50 cycles of 95 1C for 15 s, 50 1C for 30 sand 72 1C for 15 s. The cycle number at which the increase influorescence and thus PCR product crossed a pre-definedthreshold was calculated by the second derivative methodand recorded as a Cp value for each sample.

5.4. Statistical analysis

Cp values were transformed into relative quantification datausing the deltaCT method. Data were compared by two-wayANOVA with factors, diet and day/night time using IBM SPSSStatistics 20 (SPSS Inc., Chicago, IL, USA). A significant differ-ence was accepted with po0.05 (trend po0.1). To determinethe stability of the HKGs, a computer algorithm called geNormis used (Vandesompele et al., 2002). In the present study, weused qbasePLUS software version 3.4 (Biogazelle BE, Belgium)to process the transformed data. From this, a stability measure(M) was generated by geometric averaging of multiple targetgenes and mean pair-wise variation of a gene from all othertarget genes in a given sample. It is based on the principle thatthe expression ratio between two ideal control genes can beobserved in all samples independent of tissue type or experi-mental conditions. Control genes with the lowest M values aredeemed the most stable. In order to identify the most stableand therefore most suitable HKG across different tissue type,the geNorm program introduces a pairwise variation (V) whichdetermines the number of HKG required for accurate normal-ization per experiment. Pairwise variation, V(n/nþ1) deter-mines the benefit gained from additional HK genes. A Vscore of 0.15 or below indicates that the additional gene hasno significant contribution to the newly calculated normal-ization factor and is therefore not needed. In most cases,geNorm will recommend the use of three reference genes as avalid method of an accurate normalization strategy, comparedwith a single non-validated reference gene.

Conflict of interest

The authors declare no conflict of interest. All the authorshave approved the final article.

Acknowledgments

This work was supported by the Biotechnology and BiologicalSciences Research Council (BB/G01812X/1) and Diabetes UK(BDA 11/0004377).

r e f e r e n c e s

Bassell, G.J., Taneja, K.L., Kislauskis, E.H., Sundell, C.L., Powers, C.M.,Ross, A., Singer, R.H., 1994. Actin filaments and the spatialpositioning of mRNAS. Adv. Exp. Med. Biol. 358, 183–189.

b r a i n r e s e a r c h 1 5 7 5 ( 2 0 1 4 ) 7 2 – 7 7 77

Bruce, K.D., et al., 2009. Maternal high-fat feeding primessteatohepatitis in adult mice offspring, involvingmitochondrial dysfunction and altered lipogenesis geneexpression. Hepatology 50, 1696–1698.

Bruce, K.D., Sihota, K.K., Byrne, C.D., Cagampang, F.R., 2012.The housekeeping gene YWHAZ remains stable in a model ofdevelopmentally primed non-alcoholic fatty liver disease.Liver Int. 32, 1315–1321.

Buhr, E.D., Takahashi, J.S., 2013. Molecular components of theMammalian circadian clock, Handbook of ExperimentalPharmacology3–27.

Cagampang, F.R., Bruce, K.D., 2012. The role of the circadian clocksystem in nutrition and metabolism. Br. J. Nutr. 108, 381–392.

Chong, N.W., Cagampang, F.R., Coen, C.W., Campbell, I.C., Powell,J.F., 1996. Rapid identification of novel genes expressed in acircadian manner in rat suprachiasmatic nuclei. Neuroreport7, 1199–1203.

Datta, S., Patterson, E.H., Vincitore, M., Tonkiss, J., Morgane, P.J.,Galler, J.R., 2000. Prenatal protein malnourished rats showchanges in sleep/wake behavior as adults. J. Sleep Res. 9,71–79.

Davidson, A.J., Castanon-Cervantes, O., Stephan, F.K., 2004. Dailyoscillations in liver function: diurnal vs circadian rhythmicity.Liver Int. 24, 179–186.

Dibner, C., Schibler, U., Albrecht, U., 2010. The mammaliancircadian timing system: organization and coordination ofcentral and peripheral clocks. Annu. Rev. Physiol. 72, 517–549.

Duran, P., Cintra, L., Galler, J.R., Tonkiss, J., 2005. Prenatal proteinmalnutrition induces a phase shift advance of thespontaneous locomotor rhythm and alters the rest/activityratio in adult rats. Nutr. Neurosci. 8, 167–172.

Elahi, M.M., Cagampang, F.R., Mukhtar, D., Anthony, F.W., Ohri, S.K.,Hanson, M.A., 2009. Long-term maternal high-fat feeding fromweaning through pregnancy and lactation predisposes offspringto hypertension, raised plasma lipids and fatty liver in mice. Br.J. Nutr. 102, 514–519.

Hruz, T., et al., 2011. RefGenes: identification of reliable andcondition specific reference genes for RT-qPCR datanormalization. BMC Genomics 12, 156.

Hsiao, L.L., et al., 2001. A compendium of gene expression innormal human tissues. Physiol. Genomics 7, 97–104.

Hvid, H., Ekstrom, C.T., Vienberg, S., Oleksiewicz, M.B.,Klopfleisch, R., 2011. Identification of stable and oestrus cycle-independent housekeeping genes in the rat mammary glandand other tissues. Vet. J. 190, 103–108.

Jabs, E.W., Thomas, P.J., Bernstein, M., Coss, C., Ferreira, G.C.,Pedersen, P.L., 1994. Chromosomal localization of genesrequired for the terminal steps of oxidative metabolism: alphaand gamma subunits of ATP synthase and the phosphatecarrier. Hum. Genet. 93, 600–602.

Mahoney, M.M., 2010. Shift work, jet lag, and femalereproduction. Int. J. Endocrinol. 2010, 813764.

McCall, M.N., Almudevar, A., 2012. Affymetrix GeneChipmicroarray preprocessing for multivariate analyses. Brief.Bioinform. 13, 536–546.

Meller, M., Vadachkoria, S., Luthy, D.A., Williams, M.A., 2005.Evaluation of housekeeping genes in placental comparativeexpression studies. Placenta 26, 601–607.

Mohawk, J.A., Green, C.B., Takahashi, J.S., 2012. Central andperipheral circadian clocks in mammals. Annu. Rev. Neurosci.35, 445–462.

Murthi, P., Fitzpatrick, E., Borg, A.J., Donath, S., Brennecke, S.P.,Kalionis, B., 2008. GAPDH, 18S rRNA and YWHAZ are suitable

endogenous reference genes for relative gene expressionstudies in placental tissues from human idiopathic fetalgrowth restriction. Placenta 29, 798–801.

Panda, S., Hogenesch, J.B., 2004. It’s all in the timing: many clocks,many outputs. J. Biol. Rhythms 19, 374–387.

Peirson, S.N., Butler, J.N., Duffield, G.E., Takher, S., Sharma, P.,Foster, R.G., 2006. Comparison of clock gene expression inSCN, retina, heart, and liver of mice. Biochem. Biophys. Res.Commun. 351, 800–807.

Radonic, A., Thulke, S., Mackay, I.M., Landt, O., Siegert, W.,Nitsche, A., 2004. Guideline to reference gene selection forquantitative real-time PCR. Biochem. Biophys. Res. Commun.313, 856–862.

Rana, S., Mahmood, S., 2010. Circadian rhythm and its role inmalignancy. J. Circadian Rhythms 8, 3.

Roenneberg, T., Kumar, C.J., Merrow, M., 2007. The humancircadian clock entrains to sun time. Curr. Biol. 17, R44–R45.

Rosbash, M., 2009. The implications of multiple circadian clockorigins. PLoS Biol. 7, e62.

Rosenwasser, A.M., 2010. Circadian clock genes: non-circadianroles in sleep, addiction, and psychiatric disorders?. Neurosci.Biobehav. Rev. 34, 1249–1255.

Sadek, K.H., Cagampang, F.R., Bruce, K.D., Macklon, N., Cheong, Y.,2012a. Variation in stability of housekeeping genes in healthyand adhesion-related mesothelium. Fertil. Steril. 98, 1023–1027.

Sadek, K.H., Cagampang, F.R., Bruce, K.D., Shreeve, N.,Macklon, N., Cheong, Y., 2012b. Variation in stability ofhousekeeping genes in endometrium of healthy and polycysticovarian syndrome women. Hum. Reprod. 27, 251–256.

Sellayah, D., Sek, K., Anthony, F.W., Hanson, M.A., Cagampang, F.R., 2008. Sensitivity of housekeeping genes in thehypothalamus to mismatch in diets between pre- andpostnatal periods in mice. Neurosci. Lett. 447, 54–57.

Sirover, M.A., 2005. New nuclear functions of the glycolyticprotein, glyceraldehyde-3-phosphate dehydrogenase, inmammalian cells. J. Cell Biochem. 95, 45–52.

Suzuki, T., Higgins, P.J., Crawford, D.R., 2000. Control selection forRNA quantitation. Biotechniques 29, 332–337.

Takeda, N., Maemura, K., 2010. Circadian clock and vasculardisease. Hypertens. Res. 33, 645–651.

Tanic, N., Perovic, M., Mladenovic, A., Ruzdijic, S., Kanazir, S.,2007. Effects of aging, dietary restriction and glucocorticoidtreatment on housekeeping gene expression in rat cortex andhippocampus-evaluation by real time RT-PCR. J. Mol. Neurosci.32, 38–46.

Vandesompele, J., De, P.K., Pattyn, F., Poppe, B., Van, R.N., De, P.A.,Speleman, F., 2002. Accurate normalization of real-timequantitative RT-PCR data by geometric averaging of multipleinternal control genes. Genome Biol. 3 (research0034.1-research0034.11).

Warrington, J.A., Nair, A., Mahadevappa, M., Tsyganskaya, M.,2000. Comparison of human adult and fetal expression andidentification of 535 housekeeping/maintenance genes.Physiol. Genomics 2, 143–147.

Young, M.E., Razeghi, P., Taegtmeyer, H., 2001. Clock genes in theheart: characterization and attenuation with hypertrophy.Circ. Res. 88, 1142–1150.

Zhong, H., Simons, J.W., 1999. Direct comparison of GAPDH, beta-actin, cyclophilin, and 28S rRNA as internal standards forquantifying RNA levels under hypoxia. Biochem. Biophys. Res.Commun. 259, 523–526.