Mol. Nutr. Food Res. 2013, 57, 21952205 2195 DOI 10.1002/mnfr.201300168RESEARCH ARTICLESelenium alters miRNA prole in an intestinal cell line:Evidence that miR-185 regulates expression of GPX2and SEPSH2Anabel Maciel-Dominguez1,2, Daniel Swan1, Dianne Ford1,2and John Hesketh1,21Institute for Cell and Molecular Biosciences, Newcastle University, Newcastle upon Tyne, UK2Human Nutrition Research Centre, Newcastle University, Newcastle upon Tyne, UKScope: Intake of the essential micronutrient selenium (Se) has health implications. This workaddressedwhethersomeeffectsofSeongeneexpressionareexertedthroughmicroRNAs(miRNA).Methods and results: Human colon adenocarcinoma cells (Caco-2) were grown in Se-decientor Se-adequate medium for 72 h. RNA was extracted and subjected to analysis of 737 miRNAusing microarray technology. One hundred and forty-ve miRNAwere found to be expressed inCaco-2 cells. Twelve miRNA showed altered expression after Se depletion: miR-625, miR-492,miR-373*, miR-22, miR-5325p, miR-106b, miR-30b, miR-185, miR-203, miR1308, miR-285p,miR-10b. These changes were validated by quantitative real-time PCR (RT-qPCR). Transcrip-tomic analysis showed that Se depletionaltered expressionof 50 genes including selenoproteinsGPX1, SELW, GPX3, SEPN1, SELK, SEPSH2 and GPX4. Pathway analysis identied arachi-donic acid metabolism, glutathione metabolism, oxidative stress, positive acute phase responseproteins and respiration of mitochondria as Se-sensitive pathways. Bioinformatic analysis iden-tied 13 transcripts as targets for the Se-sensitive miRNA; three were predicted to be recognisedby miR-185. Silencing of miR-185 increased GPX2 and SEPSH2 expression.Conclusions: We propose that miR-185 plays a role in up-regulation of GPX2 and SEPHS2expression. In the case of SEPHS2 this may contribute to maintaining selenoprotein synthesis.The data indicate that micronutrient supply can regulate the cell miRNA expression prole.Keywords:Epithelial cells / Micronutrient / MicroRNA / Selenoprotein / TranscriptomicsReceived: March 7, 2013Revised: April 25, 2013Accepted: May 23, 2013

Additional supporting information may be found in the online version of this article atthe publishers web-site1 IntroductionSub-optimal micronutrient intake has been proposed to af-fecthealthduringageing[1].Selenium(Se),zincandvita-Correspondence: Dr. John Hesketh, Institute for Cell and Molecu-lar Biosciences, Newcastle University, Newcastle-upon-Tyne, NE24HH, UKE-mail: j.e.hesketh@ncl.ac.ukFax: +44-1912227424Abbreviations: Caco-2, human colon adenocarcinoma cells;Ct, cyclethresholdvalue; FDR, falsediscoveryrate; GAPDH,glyceraldehyde 3-phosphate dehydrogenase; IPA, ingenuitypathwayanalysis; miRNA, microRNA; MRE, microRNArecog-nitionelement; NMD, nonsense-mediateddecay; RISC, RNA-induced silencing complex; RT-qPCR, quantitative real-timePCR; Se, selenium; Sec-tRNA, selenocysteine tRNAmin Dhaveimportantmetabolicrolesthatarereectedintheireffectsoncell growthandproliferationandintheirsub-optimal intake impacting on human health [24]. Tran-scriptomic analysis has shown that for each of these micronu-trients changes in intake lead to altered gene expression inseveral biochemical pathways and networks [58]. However,despite these known effects and the proposed role of microR-NAs (miRNA) as regulators of gene networks [9], there is littleinformation available on whether epigenetic mechanisms in-volving regulation through miRNA play a role in metabolicregulation by micronutrients. However, recently zinc deple-tion was reported to alter blood miRNAlevels [10] and miR-22has been proposed to be a target of 1,25-dihydroxyvitaminD3 in colon cancer cells [11]. The present work investigatedwhetherchangesinSesupplyaffectedmiRNAexpressionin human colon adenocarcinoma cells (Caco-2), an intestinalC2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com2196 A. Maciel-Dominguez et al. Mol. Nutr. Food Res. 2013, 57, 21952205cell line. The focus onSe and intestinal cells arose fromobser-vations that Se intake affects colorectal adenoma and cancerrisk (see [1214]) and the reported importance of Se and Se-containingselenoproteinsininammatorymechanismsinthe gut [1517].Se is incorporated into selenoproteins as the amino-acidselenocysteine using a specic tRNA (Sec-tRNA), a structurewithin the 3

untranslated region (3

UTR) of the mRNA (theSECIS)andaUGAcodon[18].Theselenoproteinsincludethe glutathione peroxidases (GPx1, GPx2 and GPx4), whichprotect cellsfromreactiveoxygenspecies, thethioredoxinreductases, which function in redox control, selenoproteinsNandS, whichhaverolesintheendoplasmicreticulumand selenoprotein P, which is a Se-transport protein [18, 19].Changes in Se intake alter the levels and activity of seleno-proteins, andcriticallytheeffectsdifferbetweendifferentselenoproteinsanddifferenttissues.Thisresultsinahier-archy of selenoproteins in which some are more sensitive tochanges in intake than others [18, 19]. The mechanistic basisforthishierarchyisunclearbutbothdifferencesin3

UTRsequencesandmethylationofSec-tRNAarethoughttobeinvolved[2022].Seintakealsoaffectstheactivityofotherbiochemical pathways, presumably secondary effects due tometabolic changes as a result of altered selenoprotein expres-sion. Forexample, transcriptomicstudiesshowthat mod-erate Se deciency leads to alterations in inammatory, en-doplasmic reticulum unfolded protein response and proteinbiosynthetic pathways in the mouse colon [5]. In humans Sesupplementation changes protein biosynthetic pathways [23].However, the mechanisms by which Se intake inuences theexpression of these other genes are largely unknown.miRNAaresmall, 2030nt long, nonprotein-codingRNAs that can regulate gene expression through the forma-tion of an RNA-induced silencing complex (RISC) [24], whichcontrolsgenomefunctionthroughmRNAdegradationortranslational repression [25]. miRNA recognise complemen-tary miRNA recognition elements (MRE) throughout mRNAsequences, including the 3

and 5

UTRs. miRNA are exiblein their recognition of target sequences and this allows themto regulate groups of genes [9, 2426]. One mRNA may con-tain multiple MRE that recognise different miRNA resultinginthepotential forexpressiontoberegulatedbymultiplemiRNA. In addition, several mRNAs may contain MRE forone miRNA resulting in the potential for one miRNA to reg-ulate several genes within a biochemical pathway or network.For example, miR-19a has been suggested to regulate severalgenes with roles in apoptosis [26] and the mir-1792 family,which have MRE for different protein coding genes [27], rep-resent miRNAthat target a group of genes. Such observationsare consistent with an overall model of regulation at the levelof pathways.No information is available in the literature as to whetherchanges in Se intake and the subsequent metabolic changeslead to alterations in miRNA expression or changes in targetgene expression through miRNA. Two observations led us tohypothesise that changes in Se supply would lead to miRNAregulation of gene expression. First, transcriptomic analyseshave shownthat Se deciencyor supplementation leads tochangesinexpressionofmultiplegeneswithinapathway,e.g. the protein synthesis pathway/translation factors [5, 23].Second, protein binding to the 3

UTR, the region of mRNAsmost frequently reported to be a target for miRNA recogni-tion, is central to selenoprotein synthesis [18, 19]. Thereforethe aim of the present work was to use a transcriptomic ap-proach for the analysis of the response of both the miRNAprole and mRNA prole of an intestinal cell line to changesin Se supply.2 Materials and methods2.1 Cell culture and RNA analysisCaco-2 cells were grown at 37C in the presence of 5% CO2inDMEMwith4.5g/LglucoseandGlutamax, 10%fetal-calf serumand1%penicillin/streptomycin(100units/mLand100g/mL, respectively). Se-adequateor Se-decientmedium comprised DMEM (with 4.5 g/L glucose and Gluta-max) without added fetal-calf serum but containing 5 g/mLinsulin, 5 g/mLtransferrin, 1%penicillin/streptomycinandwithorwithout 7ng/mLsodiumselenite(40.5nM).Cells were grown in Se-decient or Se-adequate medium for72 h. Under all these culture conditions the Caco-2 cells wereundifferentiated. RNA was extracted by standard Trizol pro-cedure followed by passage through a glass bre membranespin column (PureLink RNA Mini Kit, Invitrogen, UK). RNAconcentrationandintegritywasassessedbyabsorbanceat260/280nmonaNanoDropR ND-1000andbyCEonanAgilent 2100 Bioanalyzer.2.2 miRNA microarray design and analysisAcustomised human genome V2 Agilent miRNAmicroarraywas designed using Agilents eArray Design Tool. The arrayincluded 737 distinct Homo sapiens miRNA and 10 negativecontrol miRNA. The miRNA accession numbers are stated inSupporting Information Table 1. The array targets included:194miRNAidentiedaspresent incolontissueorcoloncell lines [2632]; 292 miRNA identied by microrna.org, Mi-croCosm, GeneSet2MiRNA, miRWalkandIngenuitypath-way analysis (IPA) and bioinformatically predicted to targetthe genes within the protein biosynthetic pathways found tobeSe-sensitiveinahumansupplementationtrial[23];414miRNAtakenfromtheIlluminahumanv2chipplatformmiRNAarrayandPCRarraysforwhichmaturesequenceswere best characterised and some experimentally validated;lastly, 10 miRNA from Mus musculus as an internal negativecontrol. The miRNA expression proling of the customised8 15 K slide was assessed by microarray technology usingthehumangenomeV2Agilentplatform. Hybridisationof100 ng of total RNA was carried out as a service by WarwickC2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.comMol. Nutr. Food Res. 2013, 57, 21952205 2197University followingthemiRNAMicroarray SystemwithmiRNA Complete Labelling and Hyb Kit protocol. Spike-insolutions were added at the labelling and hybridisation stagewith the labelled RNAput through a further purication step.Array scan was performed using the Agilent G2562CA scan-ner. Differential miRNA expression analysis was performedusing t-tests in GeneSpring GX 11 (Agilent) and BenjaminiHochberg false discovery rate (FDR) multiple testing correc-tion where probes had a corrected p-value of