research article differential gene expression in high-...

Research ArticleDifferential Gene Expression in High- and Low-ActiveInbred Mice

Michelle Dawes12 Trudy Moore-Harrison3 Alicia T Hamilton3 Tyrone Ceaser3

Kelli J Kochan4 Penny K Riggs4 and J Timothy Lightfoot123

1 Department of Health and Kinesiology Texas AampM University College Station TX 77843 USA2 Sydney and JL Huffines Institute for Sports Medicine and Human Performance Texas AampM UniversityCollege Station TX 77843 USA

3Department of Kinesiology University of North Carolina Charlotte Charlotte NC 28223 USA4Department of Animal Science Texas AampM University College Station TX 77843 USA

Correspondence should be addressed to Michelle Dawes dawesmichlkntamuedu

Received 15 October 2013 Accepted 15 December 2013 Published 16 January 2014

Academic Editor Jaakko Kaprio

Copyright copy 2014 Michelle Dawes et alThis is an open access article distributed under the Creative CommonsAttribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Numerous candidate genes have been suggested in the recent literature with proposed roles in regulation of voluntary physicalactivity with little evidence of these genesrsquo functional roles This study compared the haplotype structure and expression profilein skeletal muscle and brain of inherently high- (C57LJ) and low- (C3HHeJ) active mice Expression of nine candidate genes[Actn2 Actn3 Casq1 Drd2 LeprMc4rMstn Papss2 and Glut4 (aka Slc2a4)] was evaluated via RT-qPCR SNPs were observedin regions of Actn2 Casq1 Drd2 Lepr and Papss2 however no SNPs were located in coding sequences or associated with anyknown regulatory sequences In mice exposed to a running wheel Casq1 (119875 = 00003) and Mstn (119875 = 0002) transcript levels inthe soleus were higher in the low-active mice However when these genes were evaluated in naıve animals differential expressionwas not observed demonstrating a training effect Among naıve mice no genes in either tissue exhibited differential expressionbetween strains Considering that no obvious SNPmechanismswere determined or differential expressionwas observed our resultsindicate that genomic structural variation or gene expression data alone is not adequate to establish any of these genesrsquo candidacyor causality in relation to regulation of physical activity

1 Introduction

The benefits of physical activity on health and disease havebeen demonstrated convincingly [1] Despite this evidencephysical activity continues to decline in humans [2 3] withdata suggesting that less than 5 of adults completemoderateactivity on a regular basis and 25 of adults are not active atall during their leisure time Physical inactivity is a risk factorfor many health outcomes such as cardiovascular diseasediabetes some forms of cancer and obesity [4]

Studies of both human and animal models stronglysuggest that genetic factors play a role in physical activity withlittle common environmental effect [5ndash15] Heritability ofphysical activity has been observed to widely range from 20to 92 in humans and mice depending on the heritabilityindex used the activity measurement employed the sex and

age of the subject and species among other factors Whilecopious evidence exists that genetics are associated with thedetermination of physical activity levels little direct evidencesupports involvement of specific genetic mechanisms inactivity regulation

Recently several putative candidate genes have beenproposed to play roles in physical activity however there hasbeen no definite consensus about what constitutes ldquosufficientevidencerdquo to define a candidate gene Traditional experimen-tal approaches most often have used the single criterion offunctional relevance as the standard for candidate gene dec-laration [16] DiPetrillo et al [17] suggested that a candidategene can be declaredwhen a potential candidate gene exhibitsat least three lines of evidence as to its involvement in thephenotype of study which includes location within a knownQTL differences in gene expression the aforementioned

Hindawi Publishing CorporationBioMed Research InternationalVolume 2014 Article ID 361048 9 pageshttpdxdoiorg1011552014361048

2 BioMed Research International

ldquofunctional relevancerdquo andor alteration in the phenotypewith manipulation of the gene An example of this approachcan be seen with two candidate genes for physical activitymdashdopamine receptor 1 (Drd1) and nescient helix loop helix 2(Nhlh2)mdashwhich have shown functional relevance to activity[18 19] interval-specific haplotype differences in animalsexhibiting differential phenotypes [20] localization withinidentified activity single-effect and epistatic QTL [10 12 21ndash24] expression differences between high- and low-activeanimals [25] andor a change in phenotype with genemanipulation [18 26] However unlike Drd1 and Nhlh2 themajority of potential candidate genes suggested to be associ-atedwith physical activity have little evidence to support theircandidacy [27]

Therefore the purpose of this study was to examine theinterval-specific haplotype structure and gene expression ofthe nine previously suggested [10ndash12 27] but weakly sup-ported candidate genes in high- (C57LJ) and low-active(C3HHeJ) mice in both central brain (nucleus accumbens)and peripheral musculoskeletal (soleus) tissues with the goalof adding additional lines of evidence to support these genesas candidates for future causal activity regulation studies

2 Methods

21 Overall Procedures Based on the available literature ninegenes with direct or indirect association (through functionalrelevance or GWAS) to physical activity were investig-ated actinin 2 (Actn2 [27]) actinin 3 (Actn3 [27]) calse-questrin 1 (Casq1 [28]) dopamine receptor 2 (Drd2 [29])leptin receptor (Lepr [13 30]) melanocortin 4 receptor(Mc4r [31]) myostatin (Mstn [27]) 31015840-phosphoadenosine51015840-phosphosulfate synthase 2 (Papss2 [32]) and glucosetransporter 4 (Glut4mdashaka Slc2a4 [33]) Two methods wereused to investigate these genes Initially published databaseswere interrogated to identify regional haplotype differencesindicating potential genomic variation in the candidate genebetween the high- and low-active mouse strains SecondmRNA expression was measured in both naıve and runningwheel-exposed mice of both strains

22 Method One Interval-Specific Haplotype ComparisonsHaplotypes of the nine candidate genes were comparedwithin and between the high- and low-active mouse strainsto identify potential genetic structural differences that couldcontribute to phenotypic variation Initial haplotype analysiswas conducted using the dense single nucleotide polymor-phism (SNP) map from Perlegen Inc (Mountain ViewCA) (asymp83 million SNPs) The Perlegen database utilizedsequence data from 55 inbred strains of mice to pre-dict haplotypes using pairwise comparisons between mousestrains The specific chromosomal location of each tar-get gene was determined using the NCBI GENE database(httpwwwncbinlmnihgovgene) and then inserted intothe haplotype block viewer [20] The haplotype viewer pro-vided a binary determination of whether the haplotype andany SNPs if present were similar or dissimilar between thestrains Following the recent dismantling of the Perlegenonline mouse haplotype viewer the haplotype data were

subsequently reverified using the Mouse Phylogeny Viewer(httpmsubcsbiouncedu [34])

23 Method Two Gene Expression Determination We hadpreviously identified C3HHeJ inbred mice as low-active andC57LJ inbred mice as high-active [10] with the C57LJ micerunning on average 271 farther on a daily basis thanthe low-active C3HHeJ mice At eight weeks of age fourC57LJ and four C3HHeJ mice (Jackson Labs Bar HarborME) were housed individually in cages with a 450mmcircumference solid surface running wheel (Ware Manu-facturing Phoenix AZ) interfaced with a magnetic sensorand computer odometer (Sigma Sport BC600 St CharlesIL) that counted revolutions of the running wheel and totaltime the mouse ran Each cage computer was calibrated (asper manufacturerrsquos instructions) for the circumference ofthe cage wheel allowing for measurement of distance (km)and time (min) the animals ran on thewheel with subsequentcalculation of speed (mmin) After one week of adaptationto the wheel the activity of each mouse was monitoredevery 24 hours beginning at 63 days of age (9 weeks) forseven consecutive days Each day the wheels were checkedto insure that they turned freely These methods have beenvalidated for repeatability [35] Subsequently due to concernsthat wheel exposure would cause training-induced geneexpression changes a separate group of high-active andlow-active mice (119899 = 12 3 D and 3 C of each strain)were housed with locked (ie nonturning) wheels from 8to 10 weeks of age Mice of respective activity groups werehoused in the same room of the university vivarium with12 h lightdark cycles (see discussion) with temperature andhumidity maintained at 19ndash21∘C and 50ndash60 respectivelyFood (Harland Tekland 8604 Rodent DietMadisonWI) andwater were provided ad libitum Mice were weighed on aweekly basis At 10 weeks of age the mice were anesthetizedwith 2ndash4 isofluorine for body composition testing andsubsequently euthanized The nucleus accumbens and thesoleus muscle were harvested and flash frozen in liquidnitrogen and then stored at minus80∘C for later analysis Bodycomposition was analyzed in the naıve animals prior totissue harvesting using the Lunar Piximus DEXA (dual-energy X-ray absorptiometry) instrument (Fitchberg WI)All procedures were approved by the University of NorthCarolina Charlotte and Texas AampM University InstitutionalAnimal Care and Use Committees

Target gene transcript expression was measured by quan-titative real-time polymerase chain reaction (RT-qPCR) asreported previously with minor modifications [25] TotalRNA was isolated from nucleus accumbens and soleus tissueusing the Qiagen RNeasy mini kit (Qiagen Valencia CA)Immediately following the elution step DNA was removedwith a DNA-free kit (Ambion Austin TX) RNA was quan-tified using a NanoDrop 1000 spectrophotometer (ThermoScientific Waltham MA) and in naıve animals qualityof RNA was determined by an Agilent 2100 Bioanalyzer(Santa Clara CA) RNA samples with RIN quality valuesgt75 were included in RT-qPCR assays RNA was reversetranscribed using iScript Reverse Transcription Supermixfor RT-PCR (Bio-Rad Laboratories Hercules CA) Then

BioMed Research International 3

RT-qPCR was conducted using SsoFast Probes Supermixwith ROX (Bio-Rad Laboratories) along with predesignedPrimeTime RT-qPCR Assays (Integrated DNA Technologies(IDT) Coralville IA) and 2 120583L cDNA to detect the transcriptsequence of interest All reactions were run in duplicate RT-qPCR reactions were run on an Applied Biosystems 7900HTFast Real-Time PCR System (Carlsbad CA) A fivefold RNAdilution series was utilized to determine efficiency of eachqPCR assay Amplification data were analyzed with SequenceDetection Software v 222 (Applied Biosystems) Expressionwas normalized to an endogenous control (18S ribosomalRNA (RN18S IDT)) using methods described by Pfaffl [36]A gene expression ratio was calculated that is positivelyrelated to expression level and takes the efficiency of eachassay into consideration Briefly gene expression ratio (GER)= target gene efficiency(CT target referenceminusCT target gene)controlgene efficiency(CT control referenceminusCT control gene) The referencevalue (calibrator) used for a given gene was the average Ctof all samples (in both strains) for that gene Efficiency wascalculated using the slope of the standard curve (10(minus1slope))

Actn2Casq1Glut4Lepr andMstn expression levels weremeasured in both the nucleus accumbens (central) and soleus(peripheral) tissue of animals exposed to running wheelsBased on evidence in the literature and results from thewheel-exposed animals expressions of Actn3 Actn2 Casq1Glut4 Lepr andMstn were assayed in the soleus of the naıveanimals while Drd2 Mc4r Papss2 and Lepr were measuredin the nucleus accumbens of the naıve animals

24 Statistics Gene expression data were checked for nor-mality using a two-sided 119865 test (JMP 100 SAS InstituteCary NC) If the expression ratio was not normal (119875 lt005) the expression data were analyzed by Chi-squarenonparametric approaches Normally distributed expressionratios were compared by a pooled 119905-test (if variances wereequal) or Studentrsquos 119905-test (if variances were not equal) Alphavalues were set a priori at 005 In all analyses expressionvalues that were greater than 25 standard deviations awayfrom the mean were considered outliers and eliminated fromthe dataset If differential expression was observed data weresubsequently analyzed for sex differences

3 Results

No difference in body weight was observed between strainsfor mice exposed to the running wheel (237 g plusmn 26 g C57LJversus 233 g plusmn 36 g C3HHeJ mean plusmn SD 119875 = 093) or naıveanimals (240 g plusmn 26 g C57LJ versus 244 g plusmn 21 g C3HHeJ119875 = 043) In naıve animals percent body fat was not differentbetween the strains (127 plusmn 16 C57LJ versus 145 plusmn 18C3HHeJ 119875 = 009)

Differential haplotypes were exhibited across the entiretranscribed region of Actn2 Casq1 Drd2 Lepr and Papss2as reflected by a number of SNPs in each gene (Table 1) Thestrains exhibited similar haplotype patterns for Mstn Glut4Mc4r and Actn3 (ie no differential SNPs) Casq1 (119875 =00003) and Mstn (119875 = 0002) transcript expression in thesoleus was found to be different between the high-and low-active mice exposed to a running wheel (Table 2 Figure 1)

while there were no differences observed inActn2 (119875 = 055)Glut4 (119875 = 020) or Lepr (119875 = 085) However when thesegenes were evaluated in the soleus between strains of naıveanimals differences in expression of Casq1 and Mstn werenot observed (119875 = 040 and 119875 = 027 resp) No differentialexpression was observed in any of the genes evaluated in thenucleus accumbens (Actn2 119875 = 013 Casq1 119875 = 064 Glut4119875 = 058 Lepr 119875 = 072Mstn 119875 = 037 Table 2) in animalsexposed to the running wheels

In naıve animals gene expression results indicated nodifferential expression between high- and low-active animalsfor any of the genes in the soleus (Actn2 119875 = 058 Actn3119875 = 058 Casq1 119875 = 040 Glut4 119875 = 022 Lepr 119875 = 082Mstn 119875 = 027 Table 2) No difference was seen betweenstrains in Drd2 (119875 = 006) Lepr (119875 = 018)Mc4r (119875 = 008)or Papss2 (119875 = 040) in the nucleus accumbens (Table 2)Gene expression differences between sexes were not observedin either strain

4 Discussion

As an extension of quantitative genetic approaches that havebeen used to investigate the genetic control of physicalactivity several genes have been suggested to be associ-ated with activity regulation with little or no supportingphysiological evidence for their involvement This studyrsquospurpose was to investigate whether nine putative candidategenes had interval-specific haplotype structure variabilityand were actually expressed differentially between high- andlow-active mice Although differential gene expression is notthe only determinant of whether a gene is a candidate geneit is one line of evidence suggesting that a gene may beinvolved in regulation of a particular phenotype We foundthat prior exposure to a running wheel in and of itselfcaused changes in gene expression demonstrating a trainingeffect Thus as our goal was to investigate innate differencesin gene expression between strains with varying activitylevels expression was subsequently measured in naıve miceInterestingly although these strains ofmice have distinctivelydiverse levels of activity none of the genes evaluated weredifferentially expressed between naıve high- and low-activemice in the nucleus accumbens or soleus The majority ofgenes evaluated in this study were chosen from genome-wide association studies utilizing genomic DNA which doesnot correspond to transcript levels Therefore differentialexpression between phenotypes should not necessarily beexpected from genotype association studies alone While notruling these genes out as potential regulators of physicalactivity our data provides evidence that differences in activityare not due to variability in transcript abundance in thismodel Likewise given that there are no SNPs locatedin protein-coding regions for any of the genes evaluatedgenomic variability between the strains in these genes doesnot account for phenotypic differences between strainsThuswhile association and functional relevance provide two linesof evidence we suggest that further functional validationof these genes is necessary possibly including investiga-tion of post-transcriptional modification and differences in


Table 1 SNP variation between high- and low-active mice

Gene Sequence accession no Chromosome no SNP position Nucleotide change RegionOn chromosome In gene C3HHeJ C57LJ

Actn2 NC 000079 1

12269977 551 T C intron 112277437 8011 G A intron 112282994 13568 C T intron 112290237 20811 G A intron 112299638 30212 A T intron 212300030 30604 A C intron 212301365 31939 G A intron 412336213 66787 C T intron 1812336314 66888 G A intron 18

Casq1 NC 000067 1 172213506 3612 T G intron 3172213681 3787 A G intron 3

Drd2 NC 000075 9

49344757 4095 G A 51015840 UTR49353453 12791 A G 51015840 UTR49367545 26883 T C 51015840 UTR49372928 32266 G A 51015840 UTR49372964 32302 C T 51015840 UTR49373405 32743 A T 51015840 UTR49376079 35417 A G 51015840 UTR49376164 35502 T C 51015840 UTR49396183 55521 G A intron 1

Lepr NC 000070 4 101802391 84984 C T intron 17101802709 85302 T G intron 17

Papss2 NC 000085 19

32607198 11483 T C intron 132607669 11954 G A intron 132613744 18029 C T intron 132626984 31269 C T intron 132633562 37847 T C intron 132649103 53388 A C intron 732653353 57638 G A intron 832665695 69980 C A 31015840 UTR

Haplotype and SNP data were obtained from the Mouse Phylogeny Viewer (httpmsubcsbiouncedu) C57LJ are high-active mice and C3HHeJ are low-active mice

regulatory mechanisms as additional lines of support for thegenersquos candidacy in relation to any phenotype regulation

It has been well established that genetic background is asignificant regulator of daily physical activity in both humansand mice with little input from common environmentalinfluences [5ndash15 23 37ndash39] In spite of themounting evidenceconfirming genetic control of physical activity little is knownabout the actual regulatory mechanisms including the iden-tity of the responsible genes Identification of potential candi-date genes has been primarily through speculated functionalrelevance andor location within an identified quantitativetrait locus with little or no functional validation Moreoften than not further examination of potential candidategenes has indicated that use of QTL locationperceivedphysiological relevance results in a large number of falsepositive quantitative trait genes (QTG) Indeed the earlypromise of discovering QTG from QTL has had limitedsuccess with some authors reporting less than a 1 successrate in finding QTG in QTL [16] Flint et al [16] also suggest

that candidate genes derived frommost QTL studies accountfor very small phenotypic effects Therefore the small effectsof putative candidate genes associated with QTL combinedwith sequence variance and position of the QTL relative tothe coding region of the gene make determining the actualcausative gene and function using traditional quantitativegenetic approaches extremely difficult

For example De Moor et al [32] found novel SNPs inthe Papss2 gene region related to activity levels in humanssuggesting Papss2 was associated with leisure time exercisebehaviorPapss2produces a sulfonation enzyme thatmodifiesmacronutrients and exogenous compounds and is expressedin many tissues including skeletal muscle and brain [32]In our mouse model however we found no differences inPapss2 expression between our high- and low-active mice inthe nucleus accumbens a region of the brain that has beensuggested as a primary site of activity regulation [25 40]Papss2was not expressed at observable amounts in the soleusof our mice using the methods employed although this may


18

16

14

12

10

08

06

04

02

00

Expr

essio

n ra

tio (A

U)

High-active wheelLow-active wheel

High-active naiveLow-active naive

Casq1

lowast

(a)

25

20

15

10

05

00

Expr

essio

n ra

tio (A

U)

Mstn

lowast



(b)

Figure 1 Expression ofCasq1 (a) andMstn (b) in soleusmuscle In both panels comparisonsmade between strains within activity state (high-active versus low-active) lowastSignificantly different fromwheel-exposed high-active mice (119875 lt 005) Values are mean plusmn SD AU arbitrary units

Table 2 Gene expression ratios

Gene Tissue Expression ratio (AU)119875 value

C57LJ (high-active) C3HHeJ (low-active)

Wheel exposed

Actn2 sol 11 plusmn 032 095 plusmn 022 055Casq1 sol 085 plusmn 004 117 plusmn 004 00003lowast

Glut4 sol 094 plusmn 007 108 plusmn 013 02Lepr sol 10 plusmn 027 105 plusmn 026 085Mstn sol 059 plusmn 005 17 plusmn 014 0002lowast

Actn2 NA 014 plusmn 012 044 plusmn 053 013Casq1 NA 016 plusmn 014 012 plusmn 005 064Glut4 NA 017 plusmn 010 02 plusmn 07 058Lepr NA 177 plusmn 247 133 plusmn 073 072Mstn NA 012 plusmn 007 022 plusmn 022 037

Naıve

Actn2 sol 127 plusmn 096 133 plusmn 11 058Actn3 sol 125 plusmn 094 135 plusmn 157 058Casq1 sol 105 plusmn 054 079 plusmn 040 04Glut4 sol 122 plusmn 056 089 plusmn 019 022Lepr sol 094 plusmn 066 103 plusmn 040 082Mstn sol 098 plusmn 058 138 plusmn 055 027Drd2 NA 169 plusmn 085 080 plusmn 047 006Lepr NA 035 plusmn 009 026 plusmn 010 018Mc4r NA 088 plusmn 019 121 plusmn 036 008Papss2 NA 117 plusmn 033 097 plusmn 048 04

Gene expression ratio was calculated fromPfaffl [36] Values are described asmeanplusmn SD ldquosolrdquo indicates soleus ldquoNArdquo nucleus accumbens lowastIndicates differentialexpression of gene between strains


have been due to the small quantities of RNA available touse in the reverse transcription reaction Interestingly all ofthe SNPs found in De Moorrsquos work were located in intron1 of Papss2 Likewise we found five SNPs in intron 1 ofPapss2 between our strains of mice (Table 1) however aBLAST comparison of the human and mouse gene sequenceshows that none of the SNPs identified seem to matchbetween species Intronic SNPs are spliced out of the mRNAtherefore not affecting sequence of the mature transcriptIntronic sequence variance would only impact transcriptlevels through alteration of miRNA sequences or by locationin the promoter region None of these modes of regulationare currently presented in the literature for Papss2 As DNAsequence variation does not have a causal relationship withtranscript abundance we should not be surprised that ourresults differ from those of De Moorrsquos et al [32]

Unraveling the regulatory mechanisms of voluntaryactivity is further complicated by a variety of genetic mecha-nisms contributing to transcriptional regulation Thereforeit is critical that potential candidate genes be examinedthoroughly before they become entrenched in the literatureas ldquocausativerdquo of a phenotype With only 2 of the humangenome actually coding for proteins it is not surprisingthat mechanisms other than structural gene variation con-tribute to differences in phenotype Regulatory regions ofnoncoding sequences may be contributing to regulation ofvoluntary physical activity through a variety of mechanisms(eg miRNA siRNA and ribosomal binding proteins [41])These regulatory mechanisms have not been fully character-ized and may be contributing to activity regulation as wehave previously suggested [10] We have shown that othergenetic mechanisms such as epistasis (gene interactions)and pleiotropy (one gene has multiple effects) can affectphysical activity regulation [23 42 43] Glut4 was selectedas a putative candidate gene for inherent physical activityregulation based on QTL association [21 23 27] and fromfunctional relevance [33] Glut4 functions to move glucoseacross the plasma membrane of cells is found in skeletalmuscle and is induced by insulin or exercise [33] Tsao andcolleges [33] observed that mice with Glut4 overexpressionran four times further than controls Glut4 was found tobe close to the ldquomini-musclerdquo gene region [21] as well asnear a QTL exhibiting significant epistasis for distance run[23] Considering these previous physical and functionalexperiments of the role of Glut4 in physical activity weexpected to see differential expression between our inherentlyhigh- and low-active strains of mice However like Papss2we observed no differences in expression It is possible thatGlut4 may function through epistasis with other genes thusdifferential expression of Glut4 itself would not be detected

Considering the multitude of mechanisms contributingto gene regulation it is not unreasonable that the onlydifferential gene expression observed between strains in thisstudy was due to a training effect It is well known that avariety of perturbations can influence gene expression suchas repeated exercise bouts altering transcript levels in skeletalmuscle and brain tissue [44 45]While we had not previouslyshown alteration in brain gene expression after runningwheelactivity [25] it is not surprising that even aminimal exposure

to wheel running (seven days) produced changes in someof the skeletal muscle genes measured (Mstn and Casq1)The literature is ambiguous for Mstn expression changes inskeletal muscle with endurance exercise training showingvariable results depending on species training mode andtime elapsed after exercise session amongst other factors [4647] Casq1 protein levels have been shown to decrease in thesoleus with endurance training by Kinnunen and Manttari[48] which is comparable to the gene expression results seenin our high-active mice These observations highlight theneed to use naıve animals when investigating inherent geneexpression differences

From our gene expression results we can conclude thatdifferences in inherent variation in activity levels are not dueto differences in transcript abundance of the genes investi-gated Additionally we propose that expression differencesseen inMstn andCasq1 in the wheel-exposed animals did notarise through genomic structural differences Our interstrainhaplotype results indicated that five of the nine genes (Actn2Casq1 Drd2 Lepr and Papss2) contained SNPs althoughnone of the SNPs were located in coding regions Drd2contains 51015840UTRSNPs however no known regulatory regionswere found at these locations None of the SNPs determinedin this study were found to have obvious mechanisms ofvariation

There are limitations that warrant consideration in thisstudy beginning with the tissues assessed the number ofstrains evaluated and the inability to compare betweenwheel treatments Only slow-twitch oxidative muscle fiberwas evaluated in this study without consideration of fast-twitch fibers Previous studies [10] have shown that theaverage daily duration of activity in the high-active C57LJmice was lengthy (3511 plusmn 616minsday) suggesting thatthe slow-twitch fibers would be the primary locomotormuscles used however the genes we evaluated might beexpressed differently in fast-twitch fibers For instance whileKinnunen et al [48] foundCasq1 protein levels to decrease insoleus fibers Casq1 was increased in fast-twitch EDL musclewith endurance training Additionally while the nucleusaccumbens was removed with the utmost care [26] it ispossible that surrounding portions of hypothalamus weredissected along with the nucleus accumbens leading tovariability in expression levels We do not expect this to bethe case however as variability of the expression ratios of thenaıve animals (as to not account for any variability causedby training adaptation) is consistent between genes in thisstudy Gene expression variability in the nucleus accumbenswas also similar to that seen by Knab et al [25] Thereforeas the nucleus accumbens is considered the central rewardcenter and a potential site of activity regulation [25 40]we believe that our results reflect true differences in geneexpression Furthermore it should be noted that only twostrains of mice were evaluated in this study It is possiblethat the mechanisms controlling activity in these two strainsare specific to only those strains While there are no directdata regarding this point studies from our lab have shownthat physical activity-QTL derived using two strain intercrossmethods (ie positional cloning approaches) [12] differsfrom physical activity-QTL derived using multiple strain


genome-wide association approaches [10]Thus it is possiblethat the potential candidate genes we examined in this studymight be expressed differentially in other strains Finally itis worth mentioning that gene expression comparisons werenot made between wheel-exposed and naıve animals dueto variability between these groups Given that differencesbetween the wheel-running animals and the naıve animalswere not our primary hypothesis as well as the fact that theanimals were housed at different locations (wheel-exposedanimals were housed at UNC-Charlotte while naıve animalswere housed at Texas AampM) which has been known to causedifferent phenotypic responses [49] gene expression com-parisons between wheel-exposed mice and naıve mice maypossibly lead to an inaccurate depiction of the physiologicaldifferences between these groups

In conclusion results showed augmented gene expressionof Casq1 and Mstn in the soleus of low-active mice thatwere exposed to a running wheel In addition we foundthat exposure to a running wheel resulted in differences intranscript abundance in and of itself implying a trainingeffect and highlighting the need to measure gene expressionin naıve mice when studying naıve genetic regulation Noneof the nine suggested activity-related candidate genes weredifferentially expressed between inherently high- and low-active mice in soleus or nucleus accumbens Five genes havegenomic structural differences (Actn2Casq1Drd2 Lepr andPapss2) however no SNPs were found in coding regionsnor were any associations made between any 31015840 UTR SNPsand known miRNA targets Thus the SNPs we found donot indicate an obvious mechanism of variation As theunderstanding of genetic regulation continues to mature it isclear that considering genomic structural variation solely assuggested by association studies is not adequate to establisha genersquos candidacy for a regulatory role and that informationregarding transcriptional expression transcriptional regula-tory mechanisms and proteomic data is needed to establishsolid genetic candidates for further causal investigations

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Authorsrsquo Contribution

Michelle Dawes performed assays analyzed results anddrafted the paper and figures Trudy Moore-Harrisondesigned real-time experiments conducted real-time assaysand wroteedited the first version of the paper Alicia THamilton designed experiments conducted qPCR assaysand edited the paper Tyrone Ceaser helped design andimplement haplotype comparison analyses and edited thepaper Kelli J Kochan assisted with assays Kelli J Kochanand Penny K Riggs assisted with results J Timothy Lightfootwas primarily responsible for conception design of researchanalysis of results project financial support development andpaper preparation

Acknowledgments

The project described was supported by a National Institutesof Health Grant (NIAMS AR050085mdashJ T Lightfoot) and bya Texas AampM start-up fund (Lightfoot) as well as studentsupport from theWhole SystemsGenomics Initiative at TexasAampM The authors would like to thank the Vivarium staffsof both Texas AampM and UNC Charlotte Additionally theauthors wish to thank David Ferguson Emily Schmitt andAnalisa Jimenez for assistance with data collection

References

[1] C K Roberts and R J Barnard ldquoEffects of exercise and diet onchronic diseaserdquo Journal of Applied Physiology vol 98 no 1 pp3ndash30 2005

[2] B R Belcher D Berrigan K W Dodd B A Emken C-PChou andD Spruijt-Metz ldquoPhysical activity inUS youth effectof raceethnicity age gender and weight statusrdquo Medicine andScience in Sports and Exercise vol 42 no 12 pp 2211ndash2221 2010

[3] R P Troiano D Berrigan K W Dodd L C Masse TTilert and M Mcdowell ldquoPhysical activity in the United Statesmeasured by accelerometerrdquoMedicine and Science in Sports andExercise vol 40 no 1 pp 181ndash188 2008

[4] T Rankinen and C Bouchard ldquoInvited commentary physicalactivity mortality and geneticsrdquo American Journal of Epidemi-ology vol 166 no 3 pp 260ndash262 2007

[5] M F W Festing ldquoWheel activity in 26 strains of mouserdquoLaboratory Animals vol 11 no 4 pp 257ndash258 1977

[6] A M Joosen M Gielen R Vlietinck and K R WesterterpldquoGenetic analysis of physical activity in twinsrdquoAmerican Journalof Clinical Nutrition vol 82 no 6 pp 1253ndash1259 2005

[7] J M Kaprio M Koskenvuo and S Sarna ldquoCigarette smokinguse of alcohol and leisure-time activity among same-sexed adultmale twinsrdquo in Twin Research 3 Epidemiological and ClinicalStudies L Gedda P Parisi and W E Nance Eds Progress inClinical and Biological Research pp 37ndash46 Alan R Liss NewYork NY USA 1981

[8] D S Lauderdale R Fabsitz J M Meyer P Sholinsky VRamakrishnan and J Goldberg ldquoFamilial determinants ofmoderate and intense physical activity a twin studyrdquo Medicineand Science in Sports and Exercise vol 29 no 8 pp 1062ndash10681997

[9] I Lerman B C Harrison K Freeman et al ldquoGenetic variabilityin forced and voluntary endurance exercise performance inseven inbred mouse strainsrdquo Journal of Applied Physiology vol92 no 6 pp 2245ndash2255 2002

[10] J T Lightfoot L Leamy D Pomp et al ldquoStrain screen andhaplotype association mapping of wheel running in inbredmouse strainsrdquo Journal of Applied Physiology vol 109 no 3 pp623ndash634 2010

[11] J T Lightfoot M J Turner M Daves A Vordermark and S RKleeberger ldquoGenetic influence on daily wheel running activitylevelrdquo Physiological Genomics vol 19 pp 270ndash276 2004

[12] J T Lightfoot M J Turner D Pomp S R Kleeberger and LJ Leamy ldquoQuantitative trait loci for physical activity traits inmicerdquo Physiological Genomics vol 32 no 3 pp 401ndash408 2008

[13] J H Stubbe D I Boomsma J M Vink et al ldquoGeneticinfluences on exercise participation in 37051 twin pairs fromseven countriesrdquo Plos One vol 1 no 1 article e22 2006


[14] J H Stubbe D I Boomsma and E J C De Geus ldquoSportsparticipation during adolescence a shift from environmental togenetic factorsrdquoMedicine and Science in Sports and Exercise vol37 no 4 pp 563ndash570 2005

[15] J G Swallow T Garland Jr P A Carter W-Z Zhan and GC Sieck ldquoEffects of voluntary activity and genetic selection onaerobic capacity in house mice (Mus domesticus)rdquo Journal ofApplied Physiology vol 84 no 1 pp 69ndash76 1998

[16] J Flint W Valdar S Shifman and R Mott ldquoStrategies formapping and cloning quantitative trait genes in rodentsrdquoNatureReviews Genetics vol 6 no 4 pp 271ndash286 2005

[17] K DiPetrillo X Wang I M Stylianou and B Paigen ldquoBioin-formatics toolbox for narrowing rodent quantitative trait locirdquoTrends in Genetics vol 21 no 12 pp 683ndash692 2005

[18] D J Good C A Coyle and D L Fox ldquoNhlh2 a basic helix-loop-helix transcription factor controlling physical activityrdquoExercise and Sport Sciences Reviews vol 36 no 4 pp 187ndash1922008

[19] J S Rhodes and T Garland Jr ldquoDifferential sensitivity toacute administration of Ritalin apormorphine SCH 23390 butnot raclopride in mice selectively bred for hyperactive wheel-running behaviorrdquo Psychopharmacology vol 167 no 3 pp 242ndash250 2003

[20] K A Frazer E Eskin H M Kang et al ldquoA sequence-basedvariation map of 827 million SNPs in inbred mouse strainsrdquoNature vol 448 no 7157 pp 1050ndash1053 2007

[21] J Hartmann T Garland R M Hannon S A Kelly G Munozand D Pomp ldquoFine mapping of ldquomini-musclerdquo a recessivemutation causing reduced hindlimb muscle mass in micerdquoJournal of Heredity vol 99 no 6 pp 679ndash687 2008

[22] S A Kelly D L Nehrenberg J L Peirce et al ldquoGeneticarchitecture of voluntary exercise in an advanced intercross lineof micerdquo Physiological Genomics vol 42 no 2 pp 190ndash2002010

[23] L J Leamy D Pomp and J T Lightfoot ldquoAn epistatic geneticbasis for physical activity traits inmicerdquo Journal of Heredity vol99 no 6 pp 639ndash646 2008

[24] D L Nehrenberg S Wang R M Hannon T Garland and DPomp ldquoQTLunderlying voluntary exercise inmice interactionswith the ldquomini musclerdquo locus and sexrdquo Journal of Heredity vol101 no 1 pp 42ndash53 2009

[25] A M Knab R S Bowen A T Hamilton A A Gulledge andJ T Lightfoot ldquoAltered dopaminergic profiles implications forthe regulation of voluntary physical activityrdquo Behavioural BrainResearch vol 204 no 1 pp 147ndash152 2009

[26] A M Knab R S Bowen A T Hamilton and J T LightfootldquoPharmacological manipulation of the dopaminergic systemaffects wheel-running activity in differentially active micerdquoJournal of Biological Regulators and Homeostatic Agents vol 26no 1 pp 119ndash129 2012

[27] J T Lightfoot ldquoCan you be born a couch potato The genomicregulation of physical activityrdquo in Exercise Genomics L SPescatello and S M Roth Eds Molecular and TranslationalMedicine pp 45ndash72 Humana Press New York NY USA 2011

[28] R M Murphy N T Larkins J P Mollica N A Beard and GD Lamb ldquoCalsequestrin content and SERCAdetermine normaland maximal Ca2+ storage levels in sarcoplasmic reticulum offastmdashand slow-twitch fibres of ratrdquo Journal of Physiology vol587 no 2 pp 443ndash460 2009

[29] R L Simonen T Rankinen L Perusse et al ldquoA dopamine D2receptor gene polymorphism and physical activity in two family

studiesrdquo Physiology and Behavior vol 78 no 4-5 pp 751ndash7572003

[30] L Richert T Chevalley D Manen J-P Bonjour R Rizzoli andS Ferrari ldquoBone mass in prepubertal boys is associated witha Gln223Arg amino acid substitution in the leptin receptorrdquoJournal of Clinical Endocrinology and Metabolism vol 92 no11 pp 4380ndash4386 2007

[31] G Cai S A Cole N Butte et al ldquoA quantitative trait locuson chromosome 18q for physical activity and dietary intake inHispanic childrenrdquo Obesity vol 14 no 9 pp 1596ndash1604 2006

[32] M H M De Moor Y-J Liu D I Boomsma et al ldquoGenome-wide association study of exercise behavior in Dutch andAmerican adultsrdquo Medicine and Science in Sports and Exercisevol 41 no 10 pp 1887ndash1895 2009

[33] T-S Tsao J Li K S Chang et al ldquoMetabolic adaptations inskeletal muscle overexpressing GLUT4 effects on muscle andphysical activityrdquoThe FASEB Journal vol 15 no 6 pp 958ndash9692001

[34] H Yang J R Wang J P Didion et al ldquoSubspecific origin andhaplotype diversity in the laboratory mouserdquo Nature Geneticsvol 43 no 7 pp 648ndash655 2011

[35] A M Knab R S Bowen T Moore-Harrison A T HamiltonM J Turner and J T Lightfoot ldquoRepeatability of exercisebehaviors in micerdquo Physiology and Behavior vol 98 no 4 pp433ndash440 2009

[36] M W Pfaffl ldquoA new mathematical model for relative quantifi-cation in real-time RT-PCRrdquo Nucleic Acids research vol 29 no9 article e45 2001

[37] S Carlsson T Andersson P Lichtenstein K Michaelsson andA Ahlbom ldquoGenetic effects on physical activity results fromthe Swedish twin registryrdquo Medicine and Science in Sports andExercise vol 38 no 8 pp 1396ndash1401 2006

[38] L Perusse A Tremblay C Leblanc and C Bouchard ldquoGeneticand environmental influences on level of habitual physicalactivity and exercise participationrdquo American Journal of Epi-demiology vol 129 no 5 pp 1012ndash1022 1989

[39] J G Swallow P A Carter and T Garland Jr ldquoArtificial selectionfor increased wheel-running behavior in house micerdquo BehaviorGenetics vol 28 no 3 pp 227ndash237 1998

[40] A M Knab and J T Lightfoot ldquoDoes the difference betweenphysically active and couch potato lie in the dopamine systemrdquoInternational Journal of Biological Sciences vol 6 no 2 pp 133ndash150 2010

[41] E Szostak and F Gebauer ldquoTranslational control by 31015840-UTR-binding proteinsrdquo Briefings in Functional Genomics vol 12 no1 pp 58ndash65 2012

[42] L J Leamy D Pomp and J T Lightfoot ldquoGenetic variation inthe pleiotropic association between physical activity and bodyweight inmicerdquoGenetics Selection Evolution vol 41 no 1 article41 2009

[43] L J Leamy D Pomp and J T Lightfoot ldquoEpistatic interactionsof genes influence within-individual variation of physical activ-ity traits in micerdquo Genetica vol 139 no 6 pp 813ndash821 2011

[44] A Allen and C Messier ldquoPlastic changes in the astrocyteGLUT1 glucose transporter and beta-tubulin microtubule pro-tein following voluntary exercise in micerdquo Behavioural BrainResearch vol 240 pp 95ndash102 2013

[45] V G Coffey and J A Hawley ldquoThe molecular bases of trainingadaptationrdquo Sports Medicine vol 37 no 9 pp 737ndash763 2007

[46] A Matsakas A Friedel T Hertrampf and P Diel ldquoShort-term endurance training results in a muscle-specific decrease


of myostatin mRNA content in the ratrdquo Acta PhysiologicaScandinavica vol 183 no 3 pp 299ndash307 2005

[47] T Schiffer S Geisler B Sperlich and H K Struder ldquoMSTNmRNA after varying exercise modalities in humansrdquo Interna-tional Journal of SportsMedicine vol 32 no 9 pp 683ndash687 2011

[48] S Kinnunen and S Manttari ldquoSpecific effects of enduranceand sprint training on protein expression of calsequestrin andSERCA in mouse skeletal musclerdquo Journal of Muscle Researchand Cell Motility vol 33 no 22 pp 123ndash130 2012

[49] J C Crabbe DWahlsten and B C Dudek ldquoGenetics of mousebehavior interactions with laboratory environmentrdquo Sciencevol 284 no 5420 pp 1670ndash1672 1999

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Anatomy Research International

PeptidesInternational Journal of


Hindawi Publishing Corporation httpwwwhindawicom

International Journal of

Volume 2014

Zoology


Molecular Biology International

GenomicsInternational Journal of


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


BioinformaticsAdvances in

Marine BiologyJournal of



Signal TransductionJournal of


BioMed Research International

Evolutionary BiologyInternational Journal of



Biochemistry Research International

ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Genetics Research International


Advances in

Virolog y

Hindawi Publishing Corporationhttpwwwhindawicom

Nucleic AcidsJournal of

Volume 2014

Stem CellsInternational



Enzyme Research



Microbiology


ldquofunctional relevancerdquo andor alteration in the phenotypewith manipulation of the gene An example of this approachcan be seen with two candidate genes for physical activitymdashdopamine receptor 1 (Drd1) and nescient helix loop helix 2(Nhlh2)mdashwhich have shown functional relevance to activity[18 19] interval-specific haplotype differences in animalsexhibiting differential phenotypes [20] localization withinidentified activity single-effect and epistatic QTL [10 12 21ndash24] expression differences between high- and low-activeanimals [25] andor a change in phenotype with genemanipulation [18 26] However unlike Drd1 and Nhlh2 themajority of potential candidate genes suggested to be associ-atedwith physical activity have little evidence to support theircandidacy [27]

Therefore the purpose of this study was to examine theinterval-specific haplotype structure and gene expression ofthe nine previously suggested [10ndash12 27] but weakly sup-ported candidate genes in high- (C57LJ) and low-active(C3HHeJ) mice in both central brain (nucleus accumbens)and peripheral musculoskeletal (soleus) tissues with the goalof adding additional lines of evidence to support these genesas candidates for future causal activity regulation studies

2 Methods

21 Overall Procedures Based on the available literature ninegenes with direct or indirect association (through functionalrelevance or GWAS) to physical activity were investig-ated actinin 2 (Actn2 [27]) actinin 3 (Actn3 [27]) calse-questrin 1 (Casq1 [28]) dopamine receptor 2 (Drd2 [29])leptin receptor (Lepr [13 30]) melanocortin 4 receptor(Mc4r [31]) myostatin (Mstn [27]) 31015840-phosphoadenosine51015840-phosphosulfate synthase 2 (Papss2 [32]) and glucosetransporter 4 (Glut4mdashaka Slc2a4 [33]) Two methods wereused to investigate these genes Initially published databaseswere interrogated to identify regional haplotype differencesindicating potential genomic variation in the candidate genebetween the high- and low-active mouse strains SecondmRNA expression was measured in both naıve and runningwheel-exposed mice of both strains

22 Method One Interval-Specific Haplotype ComparisonsHaplotypes of the nine candidate genes were comparedwithin and between the high- and low-active mouse strainsto identify potential genetic structural differences that couldcontribute to phenotypic variation Initial haplotype analysiswas conducted using the dense single nucleotide polymor-phism (SNP) map from Perlegen Inc (Mountain ViewCA) (asymp83 million SNPs) The Perlegen database utilizedsequence data from 55 inbred strains of mice to pre-dict haplotypes using pairwise comparisons between mousestrains The specific chromosomal location of each tar-get gene was determined using the NCBI GENE database(httpwwwncbinlmnihgovgene) and then inserted intothe haplotype block viewer [20] The haplotype viewer pro-vided a binary determination of whether the haplotype andany SNPs if present were similar or dissimilar between thestrains Following the recent dismantling of the Perlegenonline mouse haplotype viewer the haplotype data were

subsequently reverified using the Mouse Phylogeny Viewer(httpmsubcsbiouncedu [34])

23 Method Two Gene Expression Determination We hadpreviously identified C3HHeJ inbred mice as low-active andC57LJ inbred mice as high-active [10] with the C57LJ micerunning on average 271 farther on a daily basis thanthe low-active C3HHeJ mice At eight weeks of age fourC57LJ and four C3HHeJ mice (Jackson Labs Bar HarborME) were housed individually in cages with a 450mmcircumference solid surface running wheel (Ware Manu-facturing Phoenix AZ) interfaced with a magnetic sensorand computer odometer (Sigma Sport BC600 St CharlesIL) that counted revolutions of the running wheel and totaltime the mouse ran Each cage computer was calibrated (asper manufacturerrsquos instructions) for the circumference ofthe cage wheel allowing for measurement of distance (km)and time (min) the animals ran on thewheel with subsequentcalculation of speed (mmin) After one week of adaptationto the wheel the activity of each mouse was monitoredevery 24 hours beginning at 63 days of age (9 weeks) forseven consecutive days Each day the wheels were checkedto insure that they turned freely These methods have beenvalidated for repeatability [35] Subsequently due to concernsthat wheel exposure would cause training-induced geneexpression changes a separate group of high-active andlow-active mice (119899 = 12 3 D and 3 C of each strain)were housed with locked (ie nonturning) wheels from 8to 10 weeks of age Mice of respective activity groups werehoused in the same room of the university vivarium with12 h lightdark cycles (see discussion) with temperature andhumidity maintained at 19ndash21∘C and 50ndash60 respectivelyFood (Harland Tekland 8604 Rodent DietMadisonWI) andwater were provided ad libitum Mice were weighed on aweekly basis At 10 weeks of age the mice were anesthetizedwith 2ndash4 isofluorine for body composition testing andsubsequently euthanized The nucleus accumbens and thesoleus muscle were harvested and flash frozen in liquidnitrogen and then stored at minus80∘C for later analysis Bodycomposition was analyzed in the naıve animals prior totissue harvesting using the Lunar Piximus DEXA (dual-energy X-ray absorptiometry) instrument (Fitchberg WI)All procedures were approved by the University of NorthCarolina Charlotte and Texas AampM University InstitutionalAnimal Care and Use Committees

Target gene transcript expression was measured by quan-titative real-time polymerase chain reaction (RT-qPCR) asreported previously with minor modifications [25] TotalRNA was isolated from nucleus accumbens and soleus tissueusing the Qiagen RNeasy mini kit (Qiagen Valencia CA)Immediately following the elution step DNA was removedwith a DNA-free kit (Ambion Austin TX) RNA was quan-tified using a NanoDrop 1000 spectrophotometer (ThermoScientific Waltham MA) and in naıve animals qualityof RNA was determined by an Agilent 2100 Bioanalyzer(Santa Clara CA) RNA samples with RIN quality valuesgt75 were included in RT-qPCR assays RNA was reversetranscribed using iScript Reverse Transcription Supermixfor RT-PCR (Bio-Rad Laboratories Hercules CA) Then





3 Results





4 Discussion





Actn2 NC 000079 1



Drd2 NC 000075 9



Papss2 NC 000085 19








18

16

14

12

10

08

06

04

02

00

Expr

essio

n ra

tio (A

U)



Casq1

lowast

(a)

25

20

15

10

05

00

Expr

essio

n ra

tio (A

U)

Mstn

lowast



(b)





Wheel exposed




Naıve

















Acknowledgments


References





























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology





3 Results





4 Discussion





Actn2 NC 000079 1



Drd2 NC 000075 9



Papss2 NC 000085 19








18

16

14

12

10

08

06

04

02

00

Expr

essio

n ra

tio (A

U)



Casq1

lowast

(a)

25

20

15

10

05

00

Expr

essio

n ra

tio (A

U)

Mstn

lowast



(b)





Wheel exposed




Naıve

















Acknowledgments


References





























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology




Actn2 NC 000079 1



Drd2 NC 000075 9



Papss2 NC 000085 19








18

16

14

12

10

08

06

04

02

00

Expr

essio

n ra

tio (A

U)



Casq1

lowast

(a)

25

20

15

10

05

00

Expr

essio

n ra

tio (A

U)

Mstn

lowast



(b)





Wheel exposed




Naıve

















Acknowledgments


References





























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


18

16

14

12

10

08

06

04

02

00

Expr

essio

n ra

tio (A

U)



Casq1

lowast

(a)

25

20

15

10

05

00

Expr

essio

n ra

tio (A

U)

Mstn

lowast



(b)





Wheel exposed




Naıve

















Acknowledgments


References





























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology















Acknowledgments


References





























































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology
















































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

research article differential gene expression in high-...

Documents