Louisiana State University LSU Digital Commons Faculty Publications Department of Biological Sciences 1-1-2018 Intergenic and intronic DNA hypomethylated regions as putative regulators of imprinted domains Arundhati Bakshi Louisiana State University Corey L. Bretz Louisiana State University Terri L. Cain Louisiana State University Joomyeong Kim Louisiana State University Follow this and additional works at: https://digitalcommons.lsu.edu/biosci_pubs Recommended Citation Bakshi, A., Bretz, C., Cain, T., & Kim, J. (2018). Intergenic and intronic DNA hypomethylated regions as putative regulators of imprinted domains. Epigenomics, 10 (4), 445-461. https://doi.org/10.2217/epi-2017-0125 This Article is brought to you for free and open access by the Department of Biological Sciences at LSU Digital Commons. It has been accepted for inclusion in Faculty Publications by an authorized administrator of LSU Digital Commons. For more information, please contact ir@lsu.edu. Research ArticleFor reprint orders, please contact: reprints@futuremedicine.comIntergenic and intronic DNA hypomethylatedregions as putative regulators of imprinteddomainsArundhati Bakshi1, Corey L Bretz1, Terri L Cain1 & Joomyeong Kim*,11Department of Biological Sciences, Louisiana State University, Baton Rouge, LA 70803, USA* Author for correspondence: Tel.: +1 225 578 7692; Fax: +1 225 578 2597; jkim@lsu.eduAim: To investigate the regulatory potential of intergenic/intronic hypomethylated regions (iHMRs) withinimprinted domains. Materials & methods: Based on the preliminary results of the histone modificationand conservation profiles, we conducted reporter assays on the Peg3 and H19 domain iHMRs. The invitro results were confirmed by the in vivo deletion of Peg3-iHMR designed to test its function in thePeg3 imprinted domain. Results & conclusion: Initial bioinformatic analyses suggested that some iHMRsmay be noncanonical enhancers for imprinted genes. Consistent with this, Peg3- and H19-iHMRs showedcontext-dependent promoter and enhancer activity. Further, deletion of Peg3-iHMR resulted in allele- andsex-specific misregulation of several imprinted genes within the domain. Taken together, these resultssuggest that some iHMRs may function as domain-wide regulators for the associated imprinted domains.Graphical abstract:Sex- and allele-specific regulation byiHMR in Peg3 domain (ECR18)iHMRs withsimilarprofiles mayconstitute anovel class ofimprinteddomainregulatorsNochangeFemaleMaleNochangePeg3↓Usp29↓Zim1↓Usp29↓DELpDELm(cell type-dependent)Yes~79%(tissue-dependent)(cell type-dependent)Yes~87%(tissue-dependent)[Peg3-ECR18]iHMR location:Histone modifications:Conservation:(mouse vs. human)(MEF cells)CTCF occupancy:RNA Pol II occupancy:Promoter/enhancer activity:(in vitro)CRISPR/Cas9 deletion:(in vivo)Peg3 Usp29 H19 Igf2[H19-HMR]HypomethylatedH3K4me1H3K27acH3K4me1/K27acH3K4me3/K27acCTCF enrichmentPol II enrichmentFirst draft submitted: 2 October 2017; Accepted for publication: 4 December 2017; Published online:23 March 2018Keywords: DNA methylation • iHMR • imprinted domains • imprinted gene regulation • imprinted genes • imprint-ing • intergenic hypomethylated regionGenomic imprinting is a unique epigenetic regulatory mechanism in placental mammals by which a subset ofgenes are expressed in a parental allele-specific fashion. These genes generally share roles in embryonic and neonatalgrowth regulation as well as brain development and behavior [1]. A fraction of imprinted genes also shows signs oftumor suppressor activity and susceptibility to epigenetic changes in many cancer states [2,3]. Only a few hundredEpigenomics (2018) 10(4), 445–461 ISSN 1750-1911 44510.2217/epi-2017-0125 C© 2018 Future Medicine LtdResearch Article Bakshi, Bretz, Cain & Kimgenes are subjected to genomic imprinting in the mammalian genome, and they are typically found clustered inrelatively conserved domains of about 500 kilobases (kb) to two megabases (Mb) in length [1,4]. Imprinted geneswithin a domain are known to be co-regulated, based on the juxtaposition of reciprocally imprinted genes, as wellas their frequently shared spatial and temporal expression patterns [1,5,6]. This co-regulation is mediated throughregulatory elements called imprinting control regions (ICR). ICRs obtain DNA methylation during gametogenesisand maintain the allele-specific DNA methylation in somatic cells throughout the lifetime of eutherians. ICRsare also responsible for the imprinting and transcriptional regulation of entire imprinted domains [1,7]. However,besides ICRs, not much is currently known regarding the cis-regulatory elements that impact imprinted geneexpression, although other enhancer-like regions have been predicted to direct the spatial and temporal expressionpatterns of imprinted genes [5,7–10].In this postgenomic era, many potential regulatory elements have been revealed through high-throughputsequencing. One such class of elements is the intergenic/intronic hypomethylated region (iHMR), which are oftensmaller, more tissue-specific and better indicative of gene activity than hypomethylated promoter regions. TheiHMRs have been characterized to resemble insulator, promoter, enhancer or bivalent elements [11]. In general,insulator-like elements are transcriptionally silent and bound by CTCF [11], with potential roles in chromatindomain organization [12]. Promoter- and enhancer-like iHMRs, on the other hand, are transcriptionally activeelements bound by various transcription factors and RNA polymerase II. The enhancer-like iHMRs, however,are differentiated by their tissue-specific hypomethylation, as opposed to the constitutive hypomethylation patternobserved at many promoter- and insulator-like elements [11]. A fair share of each of the classes of iHMRs is foundwithin imprinted domains, where they may act as enhancers, alternative promoters as well as chromatin architectureregulators. The functions of all of the aforementioned categories of cis-regulatory elements have been previouslylinked with the regulation of imprinting mechanism and/or imprinted gene expression patterns [13–17]. Moreover,many of these elements also exhibit epigenetic aberrations in cancers where imprinted genes tend to be affected [2,3].Therefore, identifying the role of iHMRs in imprinted domains could lead to a further understanding of howimprinted genes are regulated in normal as well as clinically relevant states such as in cancer.A few iHMRs have been previously characterized as evolutionarily conserved elements in the Peg3, H19 and Dlk1domains. All three iHMRs are constitutively hypomethylated, and two of them bind CTCF. Yet, these iHMRs appearmore enhancer-like than insulator-like in terms of their epigenetic and functional states. The list of these elementsincludes ECR18 (evolutionary conserved region 18) within the Peg3 domain [8,9], H19-CCD (central conserveddomain) within the H19/Igf2 domain [18] and Dlk1-CE4 (conserved element 4) in the Gtl2/Dlk1 domain [10].Given the unusual co-regulation patterns of imprinted genes, we hypothesized that other such noncanonicalelements may also exist within imprinted domains that are involved in properly orchestrating gene expressionfrom opposite parental alleles. Thus, here we have surveyed five additional iHMRs within imprinted domainsthat share hypomethylation patterns across multiple tissues, similar to the three previously described noncanonicalenhancer-like iHMRs. Analysis of the eight total elements suggest that many of these iHMRs may be transcriptionalregulators within imprinted domains, despite their nontraditional presentation. Moreover, CRISPR/Cas9-baseddeletion of the intronic HMR, ECR18, affected multiple genes in the Peg3 domain, suggesting that some iHMRsmay even be regulators impacting the transcription of the entire imprinted domains.Materials & methodsBioinformatic analyses of iHMRs in imprinted domainsDNA methylation and histone modifications were visualized for eight intergenic/intronic hypomethylated regionsselected for this study using the publicly available datasets on UCSC Genome Browser (Supplementary Figure1) [19,20]. Based on placental mammal conservation data (PhyloP track) from UCSC Genome Browser, the mostconserved 100–500 bp segment of the iHMR was selected for further analyses (Supplementary Material 1) [21].Data from the MultiZ alignment tool in UCSC Genome Browser [22] were harnessed to retrieve orthologoussequence for each iHMR for all available placental mammals. Additionally, the BLAT program was used to identifyorthologous iHMR sequences that had not been detected through the MultiZ alignment. Each of the sequenceswere then aligned against the reference mouse sequence using blastn, and the percentage identity (ID) recordedfor all alignments ≥100 bp in length. Next, the alignments were manually surveyed for conserved DNA motifs,based on chromatin immunoprecipitation (ChIP)-seq datasets from various mouse and human tissues, availableon UCSC Genome Browser [23–26]. Additionally, for five highly conserved iHMR, orthologous sequences from fivespecies – mouse, human, cat, horse and cow – were analyzed using DCODE to identify conserved transcription446 Epigenomics (2018) 10(4) future science groupCharacterizing iHMR in imprinted domains Research ArticlePeg3-ECR18Bone marrowCerebellumColonCortexHeartIntestineKidneyLiverLungOlfactory bulbPancreasPlacentaSkinSpleenStomachThymusUterusESCMouse ESC and NP, Stadler 2011Gnas-HMRGnas-HMREpigenetic memory at embryonic enhancers, Hon et al 2013GnasNespasChr2: 174,120,000 174,140,000 174,160,000Chr7: 6,700,000 6,750,000 6,800,000 6,850,000Peg3 Usp29Ndn-HMR67,150,000 69,500,000 69,530,000 69,560,000Chr7:SnrpnSnurfNdn MageI2 Mkrn3Figure 1. Intergenic/intronic hypomethylated regions in imprinted domains. (A) The various intronic hypomethylated regions of theGnas domain are depicted by dark blue bars [19,20]. The iHMR selected for this study is marked by a light blue bar and labeled as‘Gnas-HMR’. Light blue bars also mark the other iHMRs selected for this study: (B) Peg3-ECR18 (Peg3 domain); (C) Ndn-HMR (Ndndomain); (D) H19-HMR (H19 domain). (E) The four iHMRs selected from the Dlk1 domain are also marked and labeled as shown for theGnas domain in (A).HMR: Hypomethylated regions; ICR: Imprinting control region; iHMR: Intergenic/intronic HMR.factor-binding motifs [27]. The phylogenetic tree in Figure 4 was downloaded in Newick format from UCSCGenome Browser and visualized using the PlyloPng program.Combined Bisulfite Restriction AssayGenomic DNA was isolated using phenol–chloroform (Invitrogen, Waltham, MA, USA) extraction, followed byethanol precipitation, from 14.5-days post coitum (dpc) whole embryo, 17.5-dpc embryo head and the followingadult tissues: brain, tail, kidney, spleen, lung and liver. Approximately 1 μg of genomic DNA was subjected to thesodium bisulfite treatment using the EZ DNA Methylation™ kit (Zymo Research, Irvine, CA, USA). Next, 1 μl(∼20 ng) of the bisulfite-converted DNA was amplified using several primer sets (Supplementary Material 2). ThePCR products were then digested with appropriate restriction enzymes to monitor the methylation level at the CpGfuture science group www.futuremedicine.com 447Research Article Bakshi, Bretz, Cain & KimBone marrowCerebellumColonCortexHeartIntestineKidneyLiverLungOlfactory bulbPancreasPlacentaSkinSpleenStomachThymusUterusESCMouse ESC and NP, Stadler 2011Dlk1-HMR1 HMR2 HMR3 HMR4Dlk1-HMR1 to -HMR4Epigenetic memory at embryonic enhancers, Hon et al 2013H19Meg3Dlk1Chr12: 110,650,000 110,700,000 110,750,000H19-HMRChr7: 149,770,000 149,800,000 149,800,000Igf2HMRiHMR characterized in this studyICRFigure 1. Intergenic/intronic hypomethylated regions in imprinted domains (cont.). (A) The various intronic hypomethylated regions ofthe Gnas domain are depicted by dark blue bars [19,20]. The iHMR selected for this study is marked by a light blue bar and labeled as‘Gnas-HMR’. Light blue bars also mark the other iHMRs selected for this study: (B) Peg3-ECR18 (Peg3 domain); (C) Ndn-HMR (Ndndomain); (D) H19-HMR (H19 domain). (E) The four iHMRs selected from the Dlk1 domain are also marked and labeled as shown for theGnas domain in (A).HMR: Hypomethylated regions; ICR: Imprinting control region; iHMR: Intergenic/intronic HMR.sites within their recognition sequences. In order to compare differences in DNA methylation between normaland tumor tissues, each sample was tested at least three-times (starting with the bisulfite-conversion process). TheImageJ program was subsequently used to quantify the relative band densities derived from the methylated andunmethylated DNA fractions in each sample.Chromatin immunoprecipitationChromatin prepared from mouse embryonic fibroblast (MEF) cells [28] was sonicated and divided into threefractions. The first fraction was saved as input control. The second fraction was precleared with protein A/GPLUS-agarose beads (Santa Cruz Biotechnology, Dallas, TX, USA), and treated with 5 μg of either anti-Pol II(N-20: sc-899X, Santa Cruz Biotechnology) or anti-CTCF antibody (catalog number 07–729, Millipore Cor-poration, Darmstadt, Germany). The third fraction was precleared, but without any antibody, thus serving asthe no-antibody (negative) control. The antibody/protein/DNA complex was precipitated using protein A/GPLUS-agarose beads. Subsequently, all three fractions were decrosslinked, and the DNA were recovered by phenol–chloroform (Invitrogen) extraction followed by ethanol precipitation.One microliter DNA from all three fractions was then amplified using primers (Supplementary Material 2)that were designed to flank the predicted transcription factor-binding motif and quantified by ChIP-qPCR448 Epigenomics (2018) 10(4) future science groupCharacterizing iHMR in imprinted domains Research ArticleGnas Peg3 Ndn H19HMR ECR18 HMR HMR HMR1 HMR2 HMR3 HMR4Gnas Peg3 Ndn H19HMR ECR18 HMR HMR HMR1 HMR2 HMR3 HMR4TissuesDlk1Bone marrowTissuesDlk1OlfactorybulbColonCerebellumCortexSpleenStomachCortexES-Bruce4 ES-E14Heart (E14.5)Limb (E14.5)Liver (E14.5)LungLiverPlacentaSkinIntestineKidneyLungMEFMELOlfactory bulbPlacentaHeartThymusUterusPancreasESSmall intestineSpleenTestisThymusBrain (E14.5)Bone marrowH3K4me1 H3K27ac H3K4me1/K27ac H3K4me3/K27acHypomethylatedFigure 2. Epigenetic profiling of intergenic/intronic hypomethylated region in imprinted domains. (A) Hypomethylation status of theiHMRs in the respective tissue is shown in blue. Although some iHMRs are constitutively hypomethylated, others are hypomethylated atvarious levels of tissue specificity. (B) Histone modification at iHMRs. Most of the iHMRs are marked by active histone marks (dark blue –H3K4me1; gold – H3K27ac; green – H3K4me1 and H3K27ac; magenta – H3K4me3 and H3K27ac) in a tissue-specific manner.ES: Embryonic stem; HMR: Hypomethylated regions; iHMR: Intergenic/intronic HMR.MouseMouse HumanUU U MBrainTailKidneySpleenLungLiverThymusE M LThymiclymphomaE14.5E17.5Br-mat-NBr-mat-TBr-CLu-mat-NLu-mat-TLu-CUMUMU UMU U UMUMU UUMUMUUUMUUMUUM MUUMUMMMUUMHypermethylated HypomethylatedNo changeM M UMGnas-HMRNdn-HMRH19-HMRDlk1-HMR1Dlk1-HMR2Dlk1-HMR3Dlk1-HMR4TaqαIHpyCH4IVHpyCH4IVBstUIHpy188IIITaqαITaqαIHinfITaqαIHpyCH4IIIBstUIHpyCH4IVFokIFokIFigure 3. DNA methylation of imprinted domain intergenic/intronic hypomethylated region in mice and humans.Methylation level of the iHMRs were analyzed with COBRA in: (A) five adult mouse tissues; (B) two embryonic stages;(C) within murine thymic lymphoma [3]; (D) at the orthologous region of the iHMRs in humans. Boxes underneathCOBRA images indicate hypermethylation (red), hypomethylation (green) and no significant change (gray) in tumortissues relative to the normal control. The band representing the methylated fraction of the DNA is marked with a red‘M’ and the unmethylated fraction with a blue ‘U’.Br-C: Breast cancer (unmatched); Br-mat-N: Matched breast normal; Br-mat-T: Matched breasttumor; COBRA: Combined Bisulfite Restriction Assay; E: Early stage; iHMR: Intergenic/intronic hypomethylatedregion; L: Late stage; Lu-C: Lung cancer (unmatched); Lu-mat-N: Matched lung normal; Lu-mat-T: Matched lungtumor; M: Middle stage.future science group www.futuremedicine.com 449Research Article Bakshi, Bretz, Cain & KimSlothArmadilloManateeTenrecRock hyraxElephantShrewHedgehogMegabatMicrobatHorsePandaDogCatCowSheepDolphinAlpacaPigTree shrewBushbabyMouse lemurTarsierSquirrel monkeyMarmosetBaboonRhesusEEGibbonOrangutanGorillaChimpHumanPikaRabbitSquirrelGuinea pigNaked mole ratKangaroo ratRatMouseEEE/CE/CE/CE/CEE/CE/CE/CE/CEE/CE/CE/CE/CE/CE/CE/CE/CEE/CE/CE/CE/CE/CEEE/CEE/CE/CE/CE/CHMR(133 bp)GnasCCCE/Ce/Ce/Ce/Ce/Ce/CE/CE/CE/CE/CE/CE/CE/CCE/CE/CE/CCE/Ce/CE/CE/CHMR(155 bp)eeeee/Ce/Ce/Ceee/CCe/Ce/CeCe/Ce/Ce/Ce/Ce/Ce/Ce/Cee/ceeeCCHMR2(398 bp)Dlk1H19eEEEEEEEEEeEEEEEEEEEEEECR18(395 bp)EEeEEEeEEEEEEEEEEEEEEEEEEEEEHMR1(199 bp)EEEEEeEEEEEeEEEEEEEEEEEEEEEEEEEEEEEHMR3(132 bp)eeeeeeeeeeeeeeeeeeeHMR4(270 bp)Peg3CHMR(114 bp)NdnEeCE-box (conserved)E-box (non-conserved)CTCF (conserved)Reference species≥90% ID≥80 to <90% ID≥70 to <80% ID<70% IDN/A or insufficientalignment length(Min alignment length ≥100 bp)Figure 4. Conservation profile of intergenic/intronic hypomethylated regions among placental mammals. The grayscale profile (darkerthe gray, higher the percent identity as compared with the reference [mouse]) shows that the iHMRs are relatively well conserved overall,and harbor some conserved DNA-binding motifs, such as CTCF (‘C’) and E-box (‘E’/‘e’). Upper case ‘E’ indicates that the actual sequencewithin the motifs are conserved in the respective species as compared with mouse. Lower case ‘e’ indicates the presence of a differentversion of the DNA-binding motif.iHMR: Intergenic/intronic hypomethylated region.(QuantStudio 6, Invitrogen). The Ct values for each amplicon derived from the no-antibody control (Neg)and immunoprecipitated DNA (Ab) were normalized against those from the input.Reporter assaysA promoterless luciferase reporter construct [9] was modified by cloning Peg3-ECR18 (chr7:6866573–6866967)or a 1649 bp region of the H19-HMR (chr7:149796116–49797764) in both forward and reverse orientationsupstream of the luc2 gene. Enhancer activity for Peg3-ECR18 [9] and H19-HMR was tested by cloning the regions(as described for the promoter assay) upstream of the main promoter from each of their imprinted domains (Peg3and H19, respectively) in the luc2 reporter construct. HEK 293T, HeLa, Neuro2A and 3T3 cells were grown in aT-75 flask with Dulbecco’s modified Eagle medium Plus GlutaMAX medium (DMEM; Fisher Scientific, Pittsburg,PA, USA) containing 10% fetal bovine serum (FBS; Fisher Scientific) and 1% antibiotic-antimycotic (A-A;Corning, Tewksbury, MA, USA). Cells were co-transfected with 2 μg of a luciferase reporter construct in a 6-wellplate along with 2 μg of a β-geo reporter construct to monitor transfection efficiency. Transfections for HEK293T, HeLa and Neuro2A cells were performed with 10 μl Lipofectamine 2000 (Invitrogen). Transfection medium(DMEM without FBS and A-A) was replaced with fresh culture medium (DMEM with 10% FBS, 1% A-A) 5.5 hpost-transfection. The 3T3 cells were transfected using 6 μl GenJet in vitro DNA Transfection Reagent for 3T3Cells (SignaGen Laboratories, Rockville, MD, USA), according to the manufacturer’s protocol. Three independenttransfections were performed for each set of reporter constructs in all four cell lines. The cells were subsequentlyharvested 48 h post-transfection in 250 μl of the reporter lysis buffer (0.1 M Tris-Cl, pH 7.8, 0.1% NP-40). Dueto very low number of 3T3 cells, the three replicate transfections were combined when harvesting the cells (in total,250 μl reporter lysis buffer) for the luciferase and β-galactosidase assays. The luciferase assays were conducted usinga commercial reagent according to manufacturer’s protocol (Promega, Fitchburg, WI, USA), and the readings werenormalized against β-galactosidase activity. The graphs in Figure 6 show one representative transfection experiment450 Epigenomics (2018) 10(4) future science groupCharacterizing iHMR in imprinted domains Research Article0CTCFPoI II24681012% inputCTCFGnas-HMR****Pol214CTCFPeg3-ECR18***** ********** * *Pol2 CTCFNdn-HMRPol2 CTCFH19-HMRPol2 CTCFDlk1-HMR1Pol2 CTCFDlk1-HMR2Pol2 CTCFDlk1-HMR3Pol2 CTCFDlk1-HMR4Pol2CTCF Pol2NegAbFigure 5. CTCF and RNA polymerase II binding at imprinted domain intergenic/intronic hypomethylated regions.Top: enrichment (Ab) for CTCF (dark green) and Pol II (dark blue) is shown as a percentage of the input DNA,compared with the no-antibody control (Neg; CTCF, light green; Pol II, light blue). Error bars show standard deviationbetween technical replicates. Bottom: significant enrichment of antibody (Ab) over control (Neg) is shown as a filledbox for CTCF (green) and Pol II (blue). All normalized enrichment <1% of input DNA (gray bar) was ignored from theanalysis.*p ≤ 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; *****p < 0.00001 (Student’s t-test).N.D.: Not detected.for HEK 293T, HeLa and Neuro2A cells and the combined readings of the three independent transfections for3T3 cells.Generation of the ECR18 alleleA CRISPR/Cas9-based scheme was used to target a 395-bp genomic interval (mm9, chr7: 6,866,573–6,866,967) termed ECR18 for deletion. In brief, 200 fertilized eggs were isolated from time-matedC57BL/6J mice. The isolated eggs were injected with two single-stranded guide RNAs (5′ guide: 5′–TGATCGATCATCACGATCACGG–3′ and 3′ guide: 5′–TCCACAAGCACTACTCCTCACGG–3′) along withCas9 mRNA. The injected eggs were then implanted into the uteruses of six pseudo-pregnant females, whichderived 39 live pups. The pups were screened for CRISPR-mediated deletion by PCR genotyping using the fol-lowing primers: P1, 5′– GCAATCTTCTCCCCCGACTC–3′; P2, 5′–AGATCACATTTCCCAGGGGC–3′; P3,5′–ACACCCGGAGCTATGAATGC–3′. PCR genotyping identified one female mouse with CRISPR-mediateddeletion of ECR18. The genotyping PCR products were then cloned and sequenced to confirm the actual dele-tion of ECR18. The sequenced PCR products were aligned to the mouse genome using the UCSC genomebrowser, which revealed that the CRISPR-based deletion scheme removed a 273-bp region of ECR18 (mm9, chr7:6,866,588–6,866,861). The founder female mouse containing the deletion was bred to a C57BL/6J male mouseto establish the ECR18 mouse line. F1 progeny were bred to transmit the ECR18 allele through both thematernal and paternal germ lines.Gene expression analysesTotal RNA was extracted from the mouse heads of 1-day-old neonates using the Trizol Reagent (Invitrogen).Approximately 3.5 μg total RNA was converted to cDNA using random primers and M-MuLV Reverse Tran-scriptase (NEB, Ipswich, MA, USA). Next, 1 μl of 0.5× (diluted) cDNA was used for quantitative real-time PCR(QuantStudio 6, Invitrogen) using SYBR-green (Bio-Rad, Hercules, CA, USA). Three biological replicates for eachof the sexes were processed for wild-type (WT) and paternal or maternal deletion of Peg3-ECR18. Expression levelsof two internal controls, β-actin and Gapdh, were measured alongside those of Peg3, Usp29 and Zim1 in neonateheads. Ct values for all genes (three technical replicates for each animal) were independently normalized, usingfuture science group www.futuremedicine.com 451Research Article Bakshi, Bretz, Cain & Kim0123Peg3-ECR18Relativeluciferase activity 4 *************01020ECR18 + Peg3-promoter30*****0123H19-HMRRelativeluciferase activity4*********0HEK 293T510H19-HMR + H19-promoter2015HeLaNeuro2ANIH3T3HEK 293THeLaNeuro2ANIH3T3******HMR LucHMRLuc HMRLucLucFigure 6. Transcriptional regulation by Peg3-ECR18 and H19-HMR. Two constructs were used to test promoteractivity of the iHMR, in forward (dark colors in [B & C]) and reverse (light colors in [B & C]) orientations (A).Peg3-ECR18 (B) and H19-HMR (C) showed promoter activity in HEK293T (orange) and 3T3 (green) cells, but not inHeLa (gold) or Neuro2A (blue) cells. Luciferase activity from two constructs (D), without (hatched bars in [E & F]) andwith (solid bars in [E & F]) the iHMR cloned upstream of luc2 that was driven by the corresponding promoter in eachdomain. Peg3-ECR18 showed enhancer activity in HEK 293T, HeLa and NIH 3T3 cells (E), whereas H19-HMR (F)upregulated H19 in HEK293T but downregulated H19 in 3T3 cells. All luciferase activity is shown relative to that ofthe promoterless luc2 (dotted line at 1). Error bars show standard deviation between technical replicates.*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; *****p < 0.00001 (Student’s t-test).iHMR: Intergenic/intronic hypomethylated region.the 2Ct method [29], against both internal controls to determine the accuracy of fold changes observed. Relativeexpression levels of Peg3, Usp29 and Zim1 were pooled for all replicates (technical and biological) of a particular sexand genotype; statistical significance of expression level differences between WT and mutant was established usingthe Mann–Whitney–Wilcoxon test (p < 0.05; RStudio). All mouse experiments were performed in accordancewith NIH guidelines for care and use of animals and also approved by the Louisiana State University InstitutionalAnimal Care and Use Committee, protocol #16-060.ResultsIdentification of iHMR as potential enhancers for imprinted domainsWhole-genome bisulfite sequencing has revealed several iHMRs that do not correspond to a known promoter orenhancer element. We visualized these regions located within five imprinted domains with the UCSC GenomeBrowser showing DNA methylation profiles of adult mouse tissues, placenta [20] and embryonic stem (ES) cells(Figure 1) [19]. We detected the largest number of iHMR [23] in the approximately 50 kb intronic region spanningGnas and Nespas (Gnas domain), 22 in the ∼200 kb intron of Usp29 (Peg3 domain) and 21 in the ∼100 kbintergenic region upstream and downstream of Dlk1 (Gtl2/Dlk1 domain). Five iHMRs were noted in the ∼50 kbintergenic spaces between Ndn, Magel2 and Mkrn3 (Ndn domain), and eight in the ∼40 kb region betweenH19 and Igf2 (H19 domain) (Supplementary Figure 1). A large majority of the iHMR showed tissue-specifichypomethylation patterns; however, a few of them showed hypomethylation in a majority of the tissues studied.Although the former category has been previously shown to contain enhancer-like characteristics [11], the latterhas remained largely uncharacterized. Therefore, here we largely focused on those iHMR, which showed sharedhypomethylation patterns across multiple mouse tissues (Gnas-HMR, as an example in Figure 1A).452 Epigenomics (2018) 10(4) future science groupCharacterizing iHMR in imprinted domains Research ArticleEight representative iHMRs were selected from the following imprinted domains – Gnas, Peg3, Ndn, H19 andDlk1 – and named based on the imprinted domain they were located in. The genomic location of each iHMR, alongwith the nomenclature, is shown in Figure 1A–E. Since the Peg3-HMR was initially identified as ECR18, we retainedthe established terminology. According to whole-genome bisulfite sequencing (Figure 2A), four iHMRs (Peg3-ECR18, Ndn-HMR, H19-HMR and Dlk1-HMR2) were hypomethylated in all adult tissues as well as placenta.The remaining four iHMRs (Gnas-HMR, Dlk1-HMR1, -HMR3 and -HMR4) were hypomethylated in a majorityof the tissues. Three of the iHMRs selected showed a constitutive hypomethylation pattern (hypomethylated in EScells as well as all adult tissues): Ndn-HMR, H19-HMR and Dlk1-HMR2 (Figure 2A).In order to understand the potential functional significance of these iHMRs, we surveyed the histone modi-fications at these regions [23]. The iHMRs were largely marked by H3K4me1 (monomethylation of lysine 4 onhistone H3) and/or H3K27ac (acetylation of lysine 27 on histone H3) (Figure 2B), indicating a potential enhancerfunction for these regions [30]. We noted a unique case of H3K4 trimethylation on Peg3-ECR18 in the embryonicheart tissue, suggesting tissue-specific promoter activity for this region. The greatest number of active histone markswas observed in E14.5 tissues (heart, limb, brain and liver), placenta and adult olfactory bulb. This is consistentwith the fact that imprinted genes are mainly expressed during the embryonic and neonatal stages [7]. Therefore, theoverall histone profile suggested that the iHMRs may be tissue- and stage-specific enhancers for imprinted genesin their respective domains.DNA methylation status of iHMR in normal & tumor tissuesIn order to validate the hypomethylation status of the iHMRs, we performed independent COmbined BisulfiteRestriction Assay (COBRA) [31] analyses using the DNA from the six adult tissues representing the three germ layers.According to the results (Figure 3A & Supplementary Figure 2A), five iHMRs appeared completely unmethylatedin most adult tissues (Peg3-ECR18, Ndn-HMR, Gnas-HMR, Dlk1-HMR1 and -HMR3), whereas the remaindershowed partial methylation patterns. The iHMRs also showed some tissue-specific methylation. Despite beinggenerally hypomethylated, Ndn-HMR (Figure 3A) and Peg-ECR18 (Supplementary Figure 2A) showed minor levelsof methylation in the brain; H19-HMR, Dlk1-HMR2 and -HMR3, on the other hand, showed higher methylationlevels in the liver, compared with the other tissues (Figure 3A). In order to test whether the hypomethylation statusof the loci in the adult tissues is set up earlier during development, we performed COBRA analyses on embryonictissues at 14.5 and 17.5 dpc. The iHMR loci appeared to have either none or very low levels of DNA methylation atboth embryonic stages (Figure 3B & Supplementary Figure 2B); thus indicating that the hypomethylation status atthe iHMRs is likely established during embryonic development and maintained in the adult tissues. In sum, this setof DNA methylation analyses largely agreed with the next-generation sequencing (NGS)-based results (Figure 2A),and confirmed the hypomethylation of the eight imprinted domain iHMRs selected for this study. Furthermore,the COBRA data revealed subtle tissue-specific differences in methylation levels at the iHMRs, and a tendency forgreater methylation at some loci in the brain and liver.Previously, it had been suggested that potential enhancer regions in imprinted domains (such as Peg3-ECR18)may be epigenetically unstable in mouse thymic lymphoma [3]. Thus, we tested the methylation levels of theremaining iHMRs at three stages of murine thymic lymphoma driven by a KrasG12D mutation [3]. Besides Peg3-ECR18 (Supplementary Figure 2C), five other iHMRs also showed aberrant DNA methylation levels in thethymic neoplasm (Figure 3C). For the most part, the iHMRs tended toward hypermethylation in the cancer tissue;however, two iHMRs, H19-HMR and Dlk1-HMR1, which were partially methylated in the normal thymus,became hypomethylated in the lymphoma. Given that the open chromatin at hypomethylated regions wouldbe susceptible to global epigenetic changes in cancer, it is interesting to note that Ndn-HMR and Gnas-HMRshowed no change and remained completely unmethylated in all tumor samples (Figure 3C). This suggests thatthe epigenetic instability at imprinted domain iHMRs may not be simply an outcome of global changes to theepigenome, but rather a functional response to the carcinogenesis process.Imprinted genes have also been implicated in various human cancers [2]. In particular, various epigenetic changeshave also been noted at the associated ICRs and ECRs, including at the human ortholog of Peg3-ECR18 [2].Thus, we assessed the methylation levels of the human orthologous regions of the iHMRs in six human samples:two pairs of breast and lung tumors along with their matched normal tissues and two unmatched samples ofbreast and lung cancer. The orthologous regions for the iHMRs in humans appeared largely hypomethylated inthe normal human breast and lung tissue, albeit with partial methylation at all loci except hNdn-HMR, hH19-HMR and hDlk1-HMR2. Besides ECR18 (Supplementary Figure 2D), five other iHMRs showed aberrant DNAfuture science group www.futuremedicine.com 453Research Article Bakshi, Bretz, Cain & Kimmethylation in at least one of the human tumor samples (Figure 3D). The hGnas-HMR locus appeared the mostsensitive, responding to all four breast and lung tumor samples. At least three loci, hGnas-HMR, hDlk1-HMR1and -HMR3, were deemed capable of responding specifically to carcinogenesis in the breast and/or lung, since theyshowed aberrant methylation levels in at least one of the matched tumors. Two loci, hNdn-HMR and hH19-HMR,showed no change in DNA methylation in any of the tumors, remaining completely unmethylated in all normaland tumor samples (Figure 3D). Taken together, the seemingly targeted, nonglobal nature of epigenetic changesobserved at the human orthologs of the iHMRs, suggest that these elements may be functionally responsive tocancer states in humans as well as in mice.Conserved putative regulatory motifs found within iHMRIn order to further characterize potential functions of the iHMRs, we assessed the overall sequence conservationlevels at the hypomethylated loci in 40 placental mammals. Since many of the iHMRs spanned several kb, the mostconserved 100–500 bp segment was selected based on the conservation data available from the UCSC GenomeBrowser [21]. The mouse iHMR sequence (Supplementary Material 1) was used as a reference to harness theorthologous sequences in the other mammals, using the MultiZ alignment tool [22] and BLAT. The sequenceswere then aligned against the mouse sequence, using the blastn program. The sequence identity (% ID) for eachalignment ≥100 bp was recorded, and has been represented as a grayscale in Figure 4. Except for Ndn-HMR andDlk1-HMR4, the iHMRs appeared conserved in a large number of mammalian species (Figure 4). Among theconserved iHMRs, H19-HMR and Dlk1-HMR3 showed highest level of sequence identity in humans (∼88%)and Dlk1-HMR2 showed the lowest (∼68%). The percentage ID between mouse and human for the remainingthree conserved iHMRs ranged from approximately 73–79% (Supplementary Figure 3A). Overall, six out of eightimprinted domain iHMRs showed approximately 70–90% sequence identity between mice and humans, indicatinghigh degrees of sequence conservation at these regions.Next, we overlapped publicly available ENCODE ChIP-seq data to identify any DNA-binding factors sharedamong the iHMRs. We found enrichment for CTCF at Gnas-HMR, Ndn-HMR, H19-HMR and Dlk1-HMR2in a majority of cell types, suggesting that they may be constitutive CTCF-binding sites [12]. Examination of theChIP-seq data from the human ENCODE project also revealed CTCF, SMC3 and RAD 21 (cohesin complexfactors) enrichment at the orthologous regions in humans. Furthermore, the identical CTCF-binding motif sharedby hGnas-HMR and mGnas-HMR was found in 26 of the 40 (65%) placental mammals (Supplementary Figure3B). Similarly, the CTCF-binding motifs at H19-HMR and Dlk1-HMR2 were also conserved, with approximately63 and 45% of the placental mammals, respectively, also harboring the identical sequence (‘C’ in Figure 4). Wefurther tested potential binding of CTCF to the iHMR through an independent series of ChIP experiments usingMEF cells (Figure 5). The results confirmed the binding at the four loci, whereas the remaining iHMRs showed nosignificant enrichment (Figure 5). Together, these data suggest that CTCF may be the most obvious DNA-bindingprotein that is involved in some unknown functions of the iHMR.We also used the following bioinformatic approach to identify additional transcription factor motifs. We selectedfive mammal sequences for each iHMR representing five mammalian orders, which were then used for predictingpotential binding motifs for known DNA-binding proteins with the DCODE program [27]. One of the motifsconserved across multiple iHMRs turns out to be the E-box (enhancer-box); all except Ndn-HMR displayed someform of the generic motif (CANNTG; ‘E’/’e’ in Figure 4 & Supplementary Figure 3). Therefore, we conductedmanual inspection of the 40 mammalian sequences to detect its presence within each iHMR. The E-box wasparticularly conserved at Gnas-HMR and Dlk1-HMR3. The E-box motif is detected at these loci in approximately85% of the species surveyed, with most species harboring the identical sequence as mouse (‘E’ in Figure 4 &Supplementary Figure 3B). In some species, however, the exact sequence of the E-box varied from that in mouse,but a generic motif for the E-box was still present (‘e’ in Figure 4). Interestingly, despite the poor overall conservationstatus of Dlk1-HMR4, several placental mammals nonetheless harbor some form of the E-box motif in that region.Consistent with a conserved E-box, we discovered putative-binding sites for E2A, MYOD and AP4 at Peg3-ECR18 [8], Dlk1-HMR1, and Gnas-HMR (Table 1 & Supplementary Material 3), which are known to bind tothe E-box motif. Overall, these surveys identify the E-box motif as another frequent motif within the imprinteddomain iHMR.Besides CTCF and E-box motifs, we also discovered other conserved motifs within the iHMRs (Table 1 &Supplementary Material 3). Both Dlk1-HMR1 and -HMR2 showed putative binding domains for GATA-bindingfactors, which are known to be over-represented at enhancer elements [32,33]. Putative motifs for hormone receptors454 Epigenomics (2018) 10(4) future science groupCharacterizing iHMR in imprinted domains Research ArticleTable 1. Conserved putative transcription factor-binding motifs in imprinted domain intergenic/intronic hypomethylatedregion.Locus Conserved putative transcription factor motifsGnas-HMR YY1, E12, MYOD, TAL1Peg3-ECR18† MYOD, AP4, E2A, PITX2H19-HMR ATF1, ATF3, ATF6, ATF2/CREBP1, T3R, CREB, CREBP1CJUNDlk1-HMR1 E2A, MYOD, AP4, MYOGENIN, MEF2, MEF2A/RSRFC4, GATA1, GATA2, GATA3, NF1Dlk1-HMR2 GATA1, GATA2, GATA3, GATA6, WT1Dlk1-HMR3 AR, PR, GR, POU3F2, POU1F1, PAX6, PAX8, HNF1, IPF1/PDX1, LHX3, IRF1, NKX2–5A detailed summary of the results from DCODE analysis is provided as Supplementary Material 3.†From reference [8] (not part of DCODE analysis in this study).(e.g., AR, GR) were found on Dlk1-HMR3, along with motifs for neuronal proteins such as POU1F1, PAX6 andPAX8. Motifs for sex hormone binding in a potential regulatory region for an imprinted gene is intriguing, sincemany of them show a sexually dimorphic expression pattern [34]. Moreover, the putative-binding motifs for theneuronal proteins are also interesting, given that Dlk1 is functionally important for neurogenesis [35,36]. Finally,H19-HMR showed various cAMP response element-associated factors (e.g., CREB, ATF1 and ATF3), along withputative-binding sites for T3R and the enhancer-binding protein JUN (an AP1 transcription factor subunit).Altogether, the iHMRs showed several conserved putative binding domains for enhancer-associated transcriptionfactors [32]. Furthermore, many of the transcription factors, whose putative DNA-binding motifs are conserved atthe iHMRs, appear to be functionally relevant for the regulation of imprinted genes in the domain. These findings,along with their epigenetic profile (Figure 2), support the hypothesis that some of these iHMRs may be potentialenhancer elements within imprinted domains.Transcriptional regulation by Peg3-ECR18 & H19-HMRTo further test whether the iHMRs act as regulatory elements in imprinted domains, we conducted ChIP-qPCRfor RNA polymerase II (Pol II) in MEF cells (Figure 5). According to the results, five iHMRs were significantlyenriched for Pol II: Peg3-ECR18, H19-HMR, Dlk1-HMR2, HMR3 and HMR4. Of those, we could successfullyclone two loci, H19-HMR (also bound by CTCF) and Peg3-ECR18 (not bound by CTCF), in order to conductreporter assays in vitro.Enhancer elements have been previously shown to harbor transcriptional potential [37,38]. Therefore, we firstconducted a series of promoter assays for the two iHMRs, using a set of constructs designed to compare promoterless-luc2 basal expression with the two iHMRs cloned upstream of luc2 in both orientations (Figure 6A). Accordingto the results, both Peg3-ECR18 and H19-HMR showed the ability to drive luciferase expression in HEK 293Tand 3T3 cells. Although ECR18 appeared to have promoter functionality in an orientation-independent mannerin both cell lines, H19-HMR was only able to drive transcription in the forward orientation in HEK293T cells(Figure 6B & C). However, neither of the iHMR showed any significant promoter activity in HeLa or Neuro2Acells in either orientation. Thus, the promoter assays concluded that the two iHMRs possess the ability to drivetranscription, albeit in a tissue-specific manner. Moreover, the orientation-independent capacity of Peg3-ECR18and H19-HMR to initiate transcription in HEK293T and/or 3T3 cells may be reminiscent of the bidirectionaltranscription observed at many enhancers [37,38].Next, we directly tested the potential enhancer/repressor activity of the loci by comparing luciferase expressionwithout and with the iHMRs cloned upstream of luc2, driven by the main promoter from each of the imprinteddomains (Figure 6D). The results suggested that ECR18 and H19-HMR may possess regulatory capacity, but actin a context-dependent manner. We observed significant upregulation of Peg3 in HEK 293T, HeLa and 3T3 cellswith the upstream ECR18 (Figure 6E). H19-HMR, on the other hand, influenced H19 expression levels differentlybased on the cell line. In HEK293T, the upstream HMR acted as a transcriptional enhancer for H19, whereas ithad a repressive effect in 3T3 cells (Figure 6F). We were unable to determine the regulatory potential of H19-HMR in Neuro2A cells because of the lack of H19 promoter activity in this cell line. Overall, the reporter assaysdemonstrated the cis-regulatory potential of Peg3-ECR18 and H19-HMR, and highlighted the context-dependentnature of their activity.future science group www.futuremedicine.com 455Research Article Bakshi, Bretz, Cain & Kim0.00.20.40.81.01.21.4MaleRelativeexpression levelZim11.60.6Usp29Peg3**** **** *0.00.20.40.81.01.21.4FemaleRelativeexpression levelZim11.60.6Usp29Peg30.00.20.40.81.01.21.4Relativeexpression levelZim11.60.6Usp29Peg3*** ****0.00.20.40.81.01.21.4Relativeexpression levelZim11.60.6Usp29Peg3WTDELpWTDELmFigure 7. In vivo regulatory function of Peg3-ECR18 in Peg3 imprinted domain. Relative expression levels of Peg3, Usp29 and Zim1(normalized to β-actin) for: (A & C) WT (dark blue) versus Peg3-ECR18 deletion on the paternal allele (light blue); and (B & D) WT (darkred) versus Peg3-ECR18 deletion on the maternal allele (light red). Deletion of Peg3-ECR18 affected multiple genes in the Peg3 domain inan allele- and sex-specific manner. Three biological replicates (individual bars) were tested for males (A & B) and females (C & D) of eachgenotype. All expression levels are shown relative to that of the three WT animals (dotted line at 1). Error bars show standard deviationbetween technical replicates for each biological replicate.*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; *****p < 0.00001.DELm: Maternal transmission of the deletion; DELp: Paternal transmission of the deletion; WT: Wild-type.Sex-specific regulation of the Peg3 domain by ECR18 in vivoPotential in vivo functions of the iHMR were further tested using a mutant mouse line, in which ECR18 wasdeleted with the CRISPR/Cas9 genome editing protocol. For this series of analyses, we derived two sets of mutantmice in the following manner. Male heterozygotes for the deletion of ECR18 were bred with female WT littermatesto derive the pups with the paternal transmission of the deletion (DELp), while the reciprocal cross yielded thepups with maternal transmission of the deletion (DELm). These two sets of pups were used to test the potentialfunctions of ECR18 in an allele-specific manner. Since ECR18 was previously predicted to be active in the neonatebrain [9], total RNA was isolated from neonate heads for subsequent cDNA synthesis. The expression levels ofthe three imprinted genes within the Peg3 domain, which are highly expressed in neonatal heads [9,39–42], weremeasured as a proxy for ECR18 activity (Figure 7).According to qRT-PCR analyses, deletion of ECR18 on the paternal allele in males resulted in significantdownregulation of the paternally expressed genes, Peg3 (∼50%) and Usp29 (∼60%) (Figure 7A). Downregulationof the maternally expressed Zim1 was inconsistent across the three DELp male biological replicates, indicatingthat it may be affected by extraneous factors, not related directly to the ECR18 deletion. However, the maternaldeletion of ECR18 resulted in consistent and significant downregulation of Zim1, by about 60%, in all three malebiological replicates (Figure 7B). In contrast, the paternally expressed genes were less affected in the DELm samples;Peg3 showed no difference in gene expression, and Usp29 was downregulated by approximately 30% relative toWT. These results strongly suggest that ECR18 functionally interacts with the promoters of Peg3/Usp29 and Zim1regulating their transcriptional activity in an allele-specific manner. However, since no effect of ECR18 deletion oneither allele was observed in females (Figure 7C & D), we concluded that the modulation of imprinted genes by456 Epigenomics (2018) 10(4) future science groupCharacterizing iHMR in imprinted domains Research ArticleECR18 is functional only in males. Taken together, the in vivo regulatory activity of Peg3-ECR18 suggest that theiHMR may be involved in domain-wide regulation of imprinted genes in an allele- and sex-specific mannerDiscussionRecent whole-genomic bisulfite sequencing [19,20] have revealed several intergenic hypomethylated regions, whosefunctional significance remain largely unknown. Here we have focused on a subset of intergenic and intronichypomethylated regions in imprinted domains (Figures 1 & 2A), which show hallmarks of regulatory elements. Amajority of them bear enhancer-associated histone marks in embryonic tissues known to express imprinted genes(Figure 2B) [39,40,43–47] and are occupied by RNA polymerase II in MEF cells (Figure 5). Furthermore, most of theiHMRs harbor various conserved putative transcription factor-binding motifs that are associated with enhancers(Figure 4 & Table 1), and half of them appear to be constitutive CTCF-binding sites. A series of functional assaysfurther suggested that these elements have transcriptional-regulatory potential (Figure 6) and may be domain-widemodulators of imprinted gene expression (Figure 7).Imprinted genes are typically clustered within domains, where reciprocally imprinted genes are frequently co-regulated. However, much of the mechanisms through which these genes are regulated remain uncharacterized.Data from the ENCODE project have yielded many useful hints, and a few intergenic regions within imprinteddomains, such as H19-CCD (overlapping with H19-HMR) and Dlk1-CE4 (overlapping with Dlk1-HMR2), havebeen shown to regulate Igf2 [18] and Dlk1 [10], respectively. Similarly, a survey on a number of ECRs in the Peg3domain concluded that they might harbor some regulatory potential as well. However, apart from ICRs, fewother classes of regulatory elements have been characterized that have domain-wide regulatory functions. Herewe show that the deletion of a conserved intronic HMR in the Peg3 domain, ECR18, resulted in allele-specificdownregulation of multiple genes in the domain (Figure 7). The results suggest that in neonate brain, ECR18modulates the expression levels of Peg3 and Usp29 on the paternal allele and Zim1 on the maternal allele in cis.Thus, this is the first indication that ECR18 may serve as a domain-wide enhancer element for the Peg3 domain.Given the similarities in the epigenetic profiles between ECR18 and many other iHMRs, we propose that otherintergenic hypomethylated regions may act as domain-wide regulators of imprinted genes as well. Thus, furtherstudies aimed at elucidating the roles of individual iHMRs are likely to yield important information regarding theco-regulation of imprinted genes within a domain.The various conserved putative transcription factor-binding sites at the iHMRs may offer a mechanistic cluetoward their regulatory potential. Of note, we found shared E-box-binding motifs across multiple imprinteddomain iHMRs (Figure 4), which is reminiscent of a similar finding across multiple ECRs in the Peg3 domain [8].A large number of imprinted genes are associated with growth regulation, myogenic and neuronal cell lineages andvarious mouse and human cancers [1–3,43,48]. A wide array of beta-Helix Loop Helix transcription factors that areimportant for the same, such as MYOD, NEUROD and MYC, are known to bind the E-box [49]. Thus, the sharedE-box motif may offer a clue regarding how multiple imprinted genes might be regulated in varying contexts.Many imprinted genes also share the unique feature that their expression levels are sexually dimorphic betweenthe two murine sexes [34]. Some iHMRs that show putative-binding motifs for sex hormone receptors (Table 1 &Supplementary Material 3) may mediate such sex-specific expression patterns of imprinted genes. In that respect,it is interesting to speculate the contribution of unknown sex-specific factors that mediate ECR18 activity onlyin males, but not in females. Furthermore, many of the CTCF-binding iHMRs could be involved in organizingparental allele-specific chromatin architecture, which may explain how nearby reciprocally imprinted genes areco-regulated, such as in the case of Peg3 and Zim1 (Figures 4 & 7) [6,50]. Moreover, the conserved CTCF-bindingiHMRs (Figure 4) may be the key to understanding how the imprinting status and co-regulation of genes haveremained conserved throughout all placental mammals. Many of localized transcription factor-binding sites mayalso help to understand associated imprinted gene function. For example, the conserved putative-binding sitesfor cAMP response elements at H19-HMR is particularly interesting given that Igf2r is known to be regulatedby CREB in rats [51]. Similarly, the putative-binding sites for PAX6 at Dlk1-HMR3 could shed light on how thenearby gene is regulated in the context of neurogenesis [36]. Therefore, future molecular and mutational analyses ofthe various motifs within the iHMRs could be critical for understanding the regulatory importance of these regionswithin imprinted domains.Detailed functional characterization of iHMRs as domain-wide gene regulators could offer important cluesregarding the role of imprinted genes in cancer as well. Recent studies have implicated epigenetic instabilityat ICRs, along with misregulation of imprinted genes, in the context of both mouse and human cancers [2,3].future science group www.futuremedicine.com 457Research Article Bakshi, Bretz, Cain & KimMethylation levels at putative enhancer regions (such as the iHMRs) also appeared sensitive to cancer states in bothspecies (Figure 3C & D) [2,3]. Currently, it is unknown whether the methylation status of these regulatory elementscould affect the expression level of imprinted genes; however, it is noteworthy that complete unmethylation at thePeg3-ICR leads to misexpression of several imprinted genes in the domain [52], many of which are also affectedby ECR18 deletion (Figure 7). Based on that we propose that the methylation level at some of the ECR18-likeiHMRs may have a domain-wide effect on gene regulation, when altered in the cancer environment. It is mostlikely that such an effect would be mediated through the activity of various transcription factors. In that respect, itis interesting to note that several transcription factors whose putative-binding sites exist on iHMRs, such as CTCF,MYOD, MYC and JUN, have been associated with various cancer states [53–56]. If the methylation status at iHMRsand the subsequent misregulation of imprinted genes are causally linked to tumor progression, then these regionscould potentially be developed as novel targets of cancer therapy. Therefore, further studies aimed at understandingthe functional aspects of iHMRs in imprinted domains could not only reveal how imprinted genes are regulatedunder normal circumstances, but the information could also have important applications in the context of cancerformation and treatment.Conclusion & future perspectiveIn this study, we have focused on a set of eight iHMRs, most of which are hypomethylated in a large number oftissues and bear CTCF-binding sites. Previous studies have categorized iHMRs, which show shared hypomethylationpatterns in multiple tissues largely as promoter regions and those which bind CTCF as insulators. However, a subsetof those regions, such Peg3-HMR and H19-HMR, appear to function as enhancers. These noncanonical enhanceriHMRs are generally well conserved and show histone marks and DNA-binding motifs, which are consistent withtheir enhancer function. We confirmed their ability to both initiate transcription as well as affect the transcriptionallevel of a nearby gene through reporter assays in multiple cell lines. Further, deletion of the Peg3-HMR in miceresulted in allele- and sex-specific downregulation of several genes in the Peg3 domain, suggesting that the iHMRmay be responsible for domain-wide regulation of imprinted genes. Based on these evidences, we propose theexistence of noncanonical enhancer iHMRs within imprinted domains. These iHMRs are likely to hold themechanistic key to the co-regulation of oppositely imprinted genes within a domain by allowing for allele-specificlooping interactions. Future studies aimed at mutating or deleting these regions could yield important informationfor imprinted domain regulation both in normal as well as disease states, such as cancer.Summary points• Most intergenic/intronic hypomethylated regions (iHMRs) can be categorized based on their methylation patternas follows: promoter, hypomethylated in multiple tissue types; enhancer, tissue-specific hypomethylation; orinsulator, CTCF-bound and hypomethylated in multiple tissues.• Here we assayed eight iHMRs in five imprinted domains which comprise a subset that show a combination ofthese patterns: hypomethylation in multiple tissues while also being bound by CTCF and harboring epigeneticsignatures of enhancer regions.• Since many imprinting regulators remain unknown and the finding that the iHMRs tend to be epigeneticallyunstable in multiple cancers, we further investigated their role in regulating imprinted genes, which arethemselves frequently misregulated during tumorigenesis.• Analysis of 40 mammalian sequences indicated that most iHMRs are well conserved and harbor conserved E-boxand CTCF-binding motifs.• Besides CTCF, chromatin immunoprecipitation assays also revealed RNA polymerase II occupancy at many of theiHMRs, suggesting their potential role as regulatory elements.• Two iHMRs, associated with the Peg3 and H19 imprinted domains, showed the context-dependent ability to drivelow-level transcription as well as regulate the transcriptional activity of Peg3 and H19, respectively.• In vivo deletion of the Peg3-HMR resulted in allele- and sex-specific downregulation of multiple imprinted geneswithin the Peg3 domain indicating the region may be a shared enhancer in the domain.• Based on this evidence, we propose that the noncanonical iHMRs may be involved in allele-specific chromatinlooping in imprinted domains, thereby mediating the co-regulation of oppositely imprinted genes.• Future mutational analysis of these regions is likely to reveal key information regarding imprinted generegulation, which may bear some clinical significance as well.458 Epigenomics (2018) 10(4) future science groupCharacterizing iHMR in imprinted domains Research ArticleSupplementary dataTo view the supplementary data that accompany this paper please visit the journal website at www.futuremedicine.com/doi/suppl/10.2217/epi-2017-0125AcknowledgementsThe authors thank H Kim for her assistance with the cell culture and reporter assay experiments, and C Pesson for his help withthe DNA methylation analyses of the human tumor panel. The authors also thank H Kim and W Frey for their thoughtful feedbackand discussion over the manuscript.Financial & competing interests disclosureThis work was supported by NIH (R01-GM066225 and R01-GM097074 to J Kim). The authors have no other relevant affiliationsor financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter ormaterials discussed in the manuscript apart from those disclosed.No writing assistance was utilized in the production of this manuscript.Ethical conduct of researchThe authors state that they have obtained appropriate institutional review board approval or have followed the principles outlinedin the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving humansubjects, informed consent has been obtained from the participants involved.ReferencesPapers of special note have been highlighted as: • of interest; •• of considerable interest1. Barlow DP, Bartolomei MS. Genomic imprinting in mammals. Cold Spring Harb. Perspect. Biol. 6(2), pii:a018382 (2014).•• An excellent review summarizing our current knowledge about genomic imprinting and future directions for the field.2. Kim J, Bretz CL, Lee S. Epigenetic instability of imprinted genes in human cancers. Nucleic Acids Res. 43(22), 10689–10699 (2015).3. Bretz CL, Langohr IM, Lee S, Kim J. Epigenetic instability at imprinting control regions in a Kras(G12D)-induced T-cell neoplasm.Epigenetics 10(12), 1111–1120 (2015).4. Kelsey G, Bartolomei MS. Imprinted genes ... and the number is? PLoS Genet. 8(3), e1002601 (2012).5. He H, Kim J. Regulation and function of the Peg3 imprinted domain. Genomics Inform. 12(3), 105–113 (2014).6. Su AI, Wiltshire T, Batalov S et al. A gene atlas of the mouse and human protein-encoding transcriptomes. Proc. Natl Acad. Sci.USA 101(16), 6062–6067 (2004).7. Barlow DP. Genomic imprinting: a mammalian epigenetic discovery model. Annu. Rev. Genet. 45, 379–403 (2011).8. Kim J, Ye A. Phylogenetic and epigenetic footprinting of the putative enhancers of the Peg3 domain. PLoS ONE 11(4), e0154216 (2016).• Bioinformatic analyses and MYOD-chromatin immunoprecipitation for ECR18 (and other ECRs) in Peg3 domain.9. Thiaville MM, Kim H, Frey WD, Kim J. Identification of an evolutionarily conserved cis-regulatory element controlling the Peg3imprinted domain. PLoS ONE 8(9), e75417 (2013).• Initial description of Peg3-ECR18 as a putative enhancer for Peg3 domain.10. Rogers ED, Ramalie JR, McMurray EN, Schmidt JV. Localizing transcriptional regulatory elements at the mouse Dlk1 locus. PLoSONE 7(5), e36483 (2012).• Initial description of the Dlk1-HMR2 locus and its transcriptional regulatory potential.11. Schlesinger F, Smith AD, Gingeras TR, Hannon GJ, Hodges E. De novo DNA demethylation and noncoding transcription define activeintergenic regulatory elements. Genome Res. 23(10), 1601–1614 (2013).•• Initial description of intergenic/intronic hypomethylated regions from genome-wide analyses.12. Li Y, Huang W, Niu L, Umbach DM, Covo S, Li L. Characterization of constitutive CTCF/cohesin loci: a possible role in establishingtopological domains in mammalian genomes. BMC Genomics 14, 553 (2013).13. Perera BP, Kim J. Alternative promoters of Peg3 with maternal specificity. Sci. Rep. 6, 24438 (2016).14. Luo Z, Lin C, Woodfin AR et al. Regulation of the imprinted Dlk1-Dio3 locus by allele-specific enhancer activity. Genes Dev. 30(1),92–101 (2016).15. Nordin M, Bergman D, Halje M, Engstrom W, Ward A. Epigenetic regulation of the Igf2/H19 gene cluster. Cell Prolif. 47(3), 189–199(2014).16. Court F, Baniol M, Hagege H et al. Long-range chromatin interactions at the mouse Igf2/H19 locus reveal a novel paternally expressedlong non-coding RNA. Nucleic Acids Res. 39(14), 5893–5906 (2011).future science group www.futuremedicine.com 459Research Article Bakshi, Bretz, Cain & Kim17. Nativio R, Sparago A, Ito Y, Weksberg R, Riccio A, Murrell A. Disruption of genomic neighbourhood at the imprinted IGF2-H19 locusin Beckwith–Wiedemann syndrome and Silver–Russell syndrome. Hum. Mol. Genet. 20(7), 1363–1374 (2011).18. Charalambous M, Menheniott TR, Bennett WR et al. An enhancer element at the Igf2/H19 locus drives gene expression in bothimprinted and non-imprinted tissues. Dev. Biol. 271(2), 488–497 (2004).• Initial description of the H19-HMR locus and its transcriptional regulatory potential.19. Stadler MB, Murr R, Burger L et al. DNA-binding factors shape the mouse methylome at distal regulatory regions. Nature 480(7378),490–495 (2011).20. Hon GC, Rajagopal N, Shen Y et al. Epigenetic memory at embryonic enhancers identified in DNA methylation maps from adultmouse tissues. Nat. Genet. 45(10), 1198–1206 (2013).21. Pollard KS, Hubisz MJ, Rosenbloom KR, Siepel A. Detection of nonneutral substitution rates on mammalian phylogenies. GenomeRes. 20(1), 110–121 (2010).22. Blanchette M, Kent WJ, Riemer C et al. Aligning multiple genomic sequences with the threaded blockset aligner. Genome Res. 14(4),708–715 (2004).23. Stamatoyannopoulos JA, Snyder M, Hardison R et al. An encyclopedia of mouse DNA elements (mouse ENCODE). GenomeBiol. 13(8), 418 (2012).24. Gerstein MB, Kundaje A, Hariharan M et al. Architecture of the human regulatory network derived from ENCODE data.Nature 489(7414), 91–100 (2012).25. Wang J, Zhuang J, Iyer S et al. Sequence features and chromatin structure around the genomic regions bound by 119 humantranscription factors. Genome Res. 22(9), 1798–1812 (2012).26. Wang J, Zhuang J, Iyer S et al. Factorbook.org: a Wiki-based database for transcription factor-binding data generated by the ENCODEconsortium. Nucleic Acids Res. 41, D171–D176 (2013).27. Ovcharenko I, Loots GG, Giardine BM et al. Mulan: multiple-sequence local alignment and visualization for studying function andevolution. Genome Res. 15(1), 184–194 (2005).28. Ye A, He H, Kim J. Paternally expressed Peg3 controls maternally expressed Zim1 as a trans factor. PLoS ONE 9(9), e108596 (2014).29. Winer J, Jung CK, Shackel I, Williams PM. Development and validation of real-time quantitative reverse transcriptase-polymerase chainreaction for monitoring gene expression in cardiac myocytes in vitro. Anal. Biochem. 270(1), 41–49 (1999).30. Creyghton MP, Cheng AW, Welstead GG et al. Histone H3K27ac separates active from poised enhancers and predicts developmentalstate. Proc. Natl Acad. Sci. USA 107(50), 21931–21936 (2010).31. Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 22(15), 2990–2997(1994).32. Andersson R, Gebhard C, Miguel-Escalada I et al. An atlas of active enhancers across human cell types and tissues. Nature 507(7493),455–461 (2014).33. Huang J, Liu X, Li D et al. Dynamic control of enhancer repertoires drives lineage and stage-specific transcription during hematopoiesis.Dev. Cell 36(1), 9–23 (2016).34. Faisal M, Kim H, Kim J. Sexual differences of imprinted genes’ expression levels. Gene 533(1), 434–438 (2014).35. Ferron SR, Charalambous M, Radford E et al. Postnatal loss of Dlk1 imprinting in stem cells and niche astrocytes regulates neurogenesis.Nature 475(7356), 381–385 (2011).36. Surmacz B, Noisa P, Risner-Janiczek JR et al. DLK1 promotes neurogenesis of human and mouse pluripotent stem cell-derived neuralprogenitors via modulating Notch and BMP signalling. Stem Cell Rev. 8(2), 459–471 (2012).37. Andersson R. Promoter or enhancer, what’s the difference? Deconstruction of established distinctions and presentation of a unifyingmodel. BioEssays 37(3), 314–323 (2015).38. Kim TK, Shiekhattar R. Architectural and functional commonalities between enhancers and promoters. Cell 162(5), 948–959 (2015).39. Kuroiwa Y, Kaneko-Ishino T, Kagitani F et al. Peg3 imprinted gene on proximal chromosome 7 encodes for a zinc finger protein. Nat.Genet. 12(2), 186–190 (1996).40. Kim J, Lu X, Stubbs L. Zim1, a maternally expressed mouse Kruppel-type zinc-finger gene located in proximal chromosome 7. Hum.Mol. Genet. 8(5), 847–854 (1999).41. Kim J, Bergmann A, Lucas S, Stone R, Stubbs L. Lineage-specific imprinting and evolution of the zinc-finger gene ZIM2.Genomics 84(1), 47–58 (2004).42. Kim J, Bergmann A, Wehri E, Lu X, Stubbs L. Imprinting and evolution of two Kruppel-type zinc-finger genes, ZIM3 and ZNF264,located in the PEG3/USP29 imprinted domain. Genomics 77(1–2), 91–98 (2001).43. Relaix F, Weng X, Marazzi G et al. Pw1, a novel zinc finger gene implicated in the myogenic and neuronal lineages. Dev. Biol. 177(2),383–396 (1996).44. Bartolomei MS, Zemel S, Tilghman SM. Parental imprinting of the mouse H19 gene. Nature 351(6322), 153–155 (1991).460 Epigenomics (2018) 10(4) future science groupCharacterizing iHMR in imprinted domains Research Article45. Dechiara TM, Robertson EJ, Efstratiadis A. Parental imprinting of the mouse insulin-like growth factor II gene. Cell 64(4), 849–859(1991).46. Yevtodiyenko A, Schmidt JV. Dlk1 expression marks developing endothelium and sites of branching morphogenesis in the mouseembryo and placenta. Dev. Dyn. 235(4), 1115–1123 (2006).47. Qu C, Jiang T, Li Y et al. Gene expression and IG-DMR hypomethylation of maternally expressed gene 3 in developing corticospinalneurons. Gene Expr. Patterns 13(1–2), 51–56 (2013).48. Perera BP, Kim J. Sex and tissue specificity of Peg3 promoters. PLoS ONE 11(10), e0164158 (2016).49. Jones S. An overview of the basic helix-loop-helix proteins. Genome Biol. 5(6), 226 (2004).50. Vietri Rudan M, Barrington C, Henderson S et al. Comparative Hi-C reveals that CTCF underlies evolution of chromosomal domainarchitecture. Cell Rep. 10(8), 1297–1309 (2015).51. Chen WK, Kuo WW, Hsieh DJ et al. CREB negatively regulates IGF2R gene expression and downstream pathways to inhibithypoxia-induced H9c2 cardiomyoblast cell death. Int. J. Mol. Sci. 16(11), 27921–27930 (2015).52. Bretz CL, Kim J. Transcription-driven DNA methylation setting on the mouse Peg3 locus. Epigenetics 12(11), 945–952 (2017).53. Kemp CJ, Moore JM, Moser R et al. CTCF haploinsufficiency destabilizes DNA methylation and predisposes to cancer. Cell Rep. 7(4),1020–1029 (2014).54. Cai C, Qin X, Wu Z et al. Inhibitory effect of MyoD on the proliferation of breast cancer cells. Oncol. Lett. 11(6), 3589–3596 (2016).55. Gabay M, Li Y, Felsher DW. MYC activation is a hallmark of cancer initiation and maintenance. Cold Spring Harb. Perspect. Med. 4(6),pii:a014241 (2014).56. Kappelmann M, Bosserhoff A, Kuphal S. AP-1/c-Jun transcription factors: regulation and function in malignant melanoma. Eur. J. CellBiol. 93(1–2), 76–81 (2014).future science group www.futuremedicine.com 461

status of IGF2R in

Abstract

Extracted data

status of IGF2R in

Abstract

Extracted data

Related items

Related items