UBC Faculty Research and Publications

miRNA and miRNA target genes in copy number variations occurring in individuals with intellectual disability Qiao, Ying; Badduke, Chansonette; Mercier, Eloi; Lewis, Suzanne M; Pavlidis, Paul; Rajcan-Separovic, Evica Aug 10, 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12864_2013_Article_5259.pdf [ 275.47kB ]
JSON: 52383-1.0221344.json
JSON-LD: 52383-1.0221344-ld.json
RDF/XML (Pretty): 52383-1.0221344-rdf.xml
RDF/JSON: 52383-1.0221344-rdf.json
Turtle: 52383-1.0221344-turtle.txt
N-Triples: 52383-1.0221344-rdf-ntriples.txt
Original Record: 52383-1.0221344-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessmiRNA and miRNA target genes in copy numbervariations occurring in individuals withintellectual disabilityYing Qiao1,2†, Chansonette Badduke1†, Eloi Mercier3, Suzanne ME Lewis2, Paul Pavlidis3and Evica Rajcan-Separovic1*AbstractBackground: MicroRNAs (miRNAs) are a family of short, non-coding RNAs modulating expression of human proteincoding genes (miRNA target genes). Their dysfunction is associated with many human diseases, includingneurodevelopmental disorders. It has been recently shown that genomic copy number variations (CNVs) can causeaberrant expression of integral miRNAs and their target genes, and contribute to intellectual disability (ID).Results: To better understand the CNV-miRNA relationship in ID, we investigated the prevalence and function ofmiRNAs and miRNA target genes in five groups of CNVs. Three groups of CNVs were from 213 probands with ID(24 de novo CNVs, 46 familial and 216 common CNVs), one group of CNVs was from a cohort of 32 cognitivelynormal subjects (67 CNVs) and one group of CNVs represented 40 ID related syndromic regions listed in DECIPHER(30 CNVs) which served as positive controls for CNVs causing or predisposing to ID. Our results show that 1). Thenumber of miRNAs is significantly higher in de novo or DECIPHER CNVs than in familial or common CNV subgroups(P < 0.01). 2). miRNAs with brain related functions are more prevalent in de novo CNV groups compared to commonCNV groups. 3). More miRNA target genes are found in de novo, familial and DECIPHER CNVs than in the commonCNV subgroup (P < 0.05). 4). The MAPK signaling cascade is found to be enriched among the miRNA target genesfrom de novo and DECIPHER CNV subgroups.Conclusions: Our findings reveal an increase in miRNA and miRNA target gene content in de novo versus commonCNVs in subjects with ID. Their expression profile and participation in pathways support a possible role of miRNAcopy number change in cognition and/or CNV-mediated developmental delay. Systematic analysis of expression/function of miRNAs in addition to coding genes integral to CNVs could uncover new causes of ID.Keywords: Micro RNA (miRNA), Copy number variants (CNVs), Copy number variant regions (CNVRs), Intellectualdisabilities (ID), Functional pathwaysBackgroundMicroRNAs (miRNAs) are an abundant class of short,non-coding, endogenous RNAs that regulate gene expres-sion at the post-transcriptional level [1,2]. The maturemiRNA is a ~20-23 nucleotides long single stranded se-quence, which derives from primary transcript miRNA(pri-miRNAs) [3]. One miRNA can bind to hundreds oftarget genes at the 3’-UTR of mRNAs (miRNA targetgenes), and a single miRNA target gene can be targeted bymultiple miRNAs [4,5]. Based on bioinformatic predic-tions, up to 90% of human genes are believed to be regu-lated by miRNA [5]. MiRNAs have primarily beendemonstrated to mediate translational repression or targetgene degradation and silencing. Recently, it has beenshown that they can also upregulate gene expression bytargeting gene regulatory sequences [6].Genetic polymorphisms in miRNA or their targets canadd to the complexity of miRNA regulation and func-tion. Single nucleotide polymorphisms (SNPs) in miRNA* Correspondence: eseparovic@cw.bc.ca†Equal contributors1Department of Pathology and Lab Medicine, BC Child and Family ResearchInstitute, University of British Columbia (UBC), Vancouver, BC V5Z 4H4,CanadaFull list of author information is available at the end of the article© 2013 Qiao et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.Qiao et al. BMC Genomics 2013, 14:544http://www.biomedcentral.com/1471-2164/14/544binding sequences have indeed been shown to affectmiRNA-mediated gene regulation and alter the expres-sion of target genes [7]. A low SNP density at miRNAgenes exists compared to the reference human genome,suggesting a negative selection for miRNA with SNPs[8-10]. CNVs, as a major class of genomic variations, alsohave an effect on miRNA, as demonstrated by under-representation of miRNA in highly polymorphic CNVscompared to the reference genome [10]. In contrast, thenumber of miRNA target genes in polymorphic CNVs ishigher than in non-CNV regions, suggesting that genes in-tegral to polymorphic CNVs are more likely to be regu-lated by miRNAs, in order to counteract their expressionchanges due to copy number variability of the region inwhich they reside [9]. Multiple cancer studies show thatmiRNAs integral to CNVs demonstrate gain or loss at thegenomic level, and are associated with expression changesfor ~10% -20% miRNAs [11,12].In addition to cancer, miRNAs are involved in other hu-man diseases, for example, cardiovascular diseases andneurological/neurodevelopmental disorders [13,14]. Evi-dence for the role of miRNA in these diseases is based onidentifying mutations or differential expression of specificmiRNAs and/or their global expression [15-17], and the ef-fect of genomic copy number change on miRNA functionis largely unknown. In rare instances association of CNVswith miRNA expression was studied in subject with cogni-tive delay. For example in Down syndrome an extra copyof chromosome 21 was associated with upregulation ofseveral miRNAs from this chromosome, while their targetsare downregulated [18]. In addition, a deletion of 1p21.3containing a miRNA MIR137 has recently been reportedby Willemsen et al. [19], and resulted in downregulation ofthe miRNA, and upregulation of three targets in subjectswith ID and congenital abnormalities. Considering the roleof MIR137 and its targets in brain function (synapse mat-uration, morphogenesis of young neurons, and axongrowth), their dysfunction is considered causative ofthe patients’ phenotype [19]. Finally, microdeletion of amiRNA cluster MIR17HG on chromosome 13 is associ-ated with Feingold syndrome which includes microceph-aly, skeletal abnormalities and variable levels of ID [20].The role of this miRNA in the human phenotype is furthersupported by a mouse knock-out model [20].Although the above examples demonstrate the relevanceof copy number change on miRNA function and its role inhuman disease, these studies are focusing on one type ofCNV exclusively; for example, on common CNVs fromnormal populations [8-10], lump of CNVs from single dis-ease cohorts such as cancer [11,12] or autism [21], or a fewindividual pathogenic CNVs associated with neuro-developmental disorders [19,20]. Considering that ~ 70% ofmiRNAs are expressed in brain [22], and function inneurodevelopment, neurotransmission, synaptic plasticity,neurite outgrowth and dysregulation [18,23], we aimed tocharacterize and understand the total presence andfunctions of miRNAs in CNVs detected in idiopathic IDin comparison to neurotypical controls. Five differentCNV subgroups were studied for their miRNA andmiRNA target gene contents: (i) de novo, (ii) familial,(iii) common CNVs detected in subjects with idiopathicID, (iv) common CNVs in cognitively normal subjects,and (v) CNVs associated with syndromic ID selected fromDECIPHER (Database of Chromosomal Imbalance andPhenotype in Humans using Ensembl Resources) database(http://decipher.sanger.ac.uk/). CNVs from DECIPHERwere considered as positive controls. Our study is uniquein serving as the first report comparing the miRNA con-tent and function in different CNV subgroups from IDsubjects and cognitively typical controls.ResultsCNV detection and sub-classificationInitially, we identified 24 de novo, 46 familial and 216common CNVs from 213 cases with idiopathic ID, 67common CNVs from 32 cognitively normal subjects, and30 CNVs collected from 40 ID-related syndromic regionsin DECIPHER database representing CNVs known tocause or predispose to developmental delay. All the CNVswere identified experimentally except the CNVs from DE-CIPHER which were retrieved from the DECIPHER data-base. Overlapping CNVs from the same subgroups weremerged into CNVRs (CNV Regions). The final set in-cluded 22 de novo, 46 familial, 30 DECIPHER CNVRs,and respectively 210 and 61 common CNVRs from casesversus controls (Table 1).The genomic coverage (Build 36hg18) for all CNVRs and the miRNAs present in eachCNVR are provided in Additional file 1: Table S1-S5.Characteristics of miRNA in different CNV subgroupsNumber of miRNAs in different CNV groupsThe number and loci of miRNAs in different CNV sub-groups were obtained using Galaxy (https://main.g2.bx.psu.edu/) and miRBase (http://www.mirbase.org/). To avoidbias caused by varying CNV size, the number of miRNA ineach CNV region was weighted by the size of the CNV, andexpressed as miRNA/Mb. The weighted median numberof miRNAs/Mb in each CNV subgroup is shown inTable 1 and was compared among different pairs ofCNV subgroups using the Wilcoxon rank-sum test. Wefound the median number of miRNAs/Mb is signifi-cantly higher in de novo and DECIPHER CNVs than fa-milial or common CNVs from cases, and commonCNVs from controls (P < 0.01) (Table 1 and Figure 1).However, the average miRNA/Mb between the commonand pathogenic subgroups is comparable which is likelydue to the presence of few very small common CNVs withQiao et al. BMC Genomics 2013, 14:544 Page 2 of 10http://www.biomedcentral.com/1471-2164/14/544high miRNA density (for example, 8 miRNAs within onecommon CNV at 16p13 (16,322,499-16,682,499).Genomic coverage of miRNAs in different CNV groupsThe genomic coverage of miRNA and protein codinggenes (expressed as their number per CNV group pertheir total number in the whole genome) tended to behigher for de novo and DECIPHER CNVs, in compari-son to the coverage of miRNAs and protein codinggenes of randomly sampled sections from the referencegenome of the same size as each of the CNV subgroups.This increase in comparison to the reference genome wasthe highest for miRNAs in de novo CNVs, but was not sig-nificant (p > 0.05, Wilcoxon signed-rank test) (Figure 2).The de novo and DECIPHER CNV coverage of proteincoding genes was also slightly higher than expected bychance but not significant. In contrast, the protein codinggenes covered a significantly smaller fraction of the com-mon CNVs than expected by chance, i.e. when referencegenome of the same size as CNVs was used (P < 0.01)(Figure 2B), in keeping with their suspected benign nature.For the familial CNVs, the observed-to-expected miRNAand protein coding gene coverage was comparable.Functional characteristics of miRNA from different CNVgroupsExpression of miRNAs from CNVsThe expression profiles of miRNAs in different CNVsubgroups were obtained using two web-tools, themimiRNA database (http://mimirna.centenary.org.au/mep/formulaire.html) [24] and microRNA.org-Targets(http://www.microrna.org/microrna/home.do) [25]. Bothweb-tools identified a small number of miRNAs (20 ~ 30%of miRNAs in de novo or familial CNV groups, and ~10%of miRNAs in common CNV groups) with published ex-pression profiles. Although the low number of miRNAwith expression/function data does not allow definitiveconclusions, we note that the miRNAs expressed in brainrelated tissues and/or having experimental evidence of in-volvement in nervous system dysfunction (based on man-ual search of PubMed) seem to be more prevalent in thede novo, familial and DECIPHER CNV group (~50% of allde novo, familial and DECIPHER CNV miRNAs with ex-pression data) than in the common CNVs tested (0-25% ofall common CNV miRNAs with expression data) (Table 2).A list of miRNAs with available expression data and brainor nervous system related function is shown in Table 3. Incomparison to the common CNVs from patients, the twofold increase in the number of brain function relatedmiRNA in familial CNV group suggests that some genesin familial CNVs could contribute to the disease process.There were no brain function-related miRNAs in commonCNVs from cognitively typical controls (Table 2).Functional and pathways enrichment analysis of miRNAtargetsUsing the web-tool WebGestalt, we searched for pathwayenrichment of miRNA target genes from different CNVsubgroups compared to the reference genome. We foundthat approximately 8-11% of genes from de novo CNVs,DECIPHER CNVs or familial CNVs were targeted bymiRNAs in comparison to 0-1% of genes from commonCNV subgroups (P < 0.01; two-sided Fisher’s exact test)(Table 4). KEGG pathway enrichment analysis wasperformed using WebGestalt for identification of pathwaysTable 1 Comparison of the number of miRNAs in different CNV subgroupsCNV category No.ofCNVsNo. ofCNVRsSize oftotalCNVRs(Mb)Averagesize ofCNVRs(Mb)miRNATotal No. (#) Average miRNA/Mb #/CNVR Weighted median No.De novo 24 22 70.5 3.21 84 1.20 3.8 0.6aFamilial 46 46 29.5 3.34 14 0.47 0.3 0bCommon from cases 216 210 105.2 0.64 67 0.64 0.3 0bCommon from control cohort 67 61 34.1 0.56 30 0.88 0.5 0bPathogenic from DECIPHER 35 30 100.1 0.50 90 0.90 3.0 0.8aa,b:Significant difference was found between data a and b (P < 0.05) (a Wilcoxon signed-rank test).Common from controlsFamilialDenovoFamilialDenovoCommonfromcontrolsDecipherDecipherCommonfromcasesCommon from casesFigure 1 Comparison of weighted median number of miRNAs/Mbin different CNV subgroups.Qiao et al. BMC Genomics 2013, 14:544 Page 3 of 10http://www.biomedcentral.com/1471-2164/14/544enriched for the miRNA target genes in different CNVgroups. The top 10 pathways are listed in Table 4 andcontained nervous system-related pathways (Axon guid-ance and Neurotrophin signaling, respectively) for onlyde novo and DECIPHER CNVs (Table 4). We found thattargets from both de novo and DECIPHER CNVs par-ticipate in MAPK signaling pathways more thanexpected by chance. Ubiquitin-mediated proteolysis wasthe only pathway found to be enriched for miRNA tar-get genes in the familial CNV subgroup. There was noapparent pathway enrichment for the miRNA targetgenes from our common CNV subgroup. The miRNAstargeting the genes from the top 10 pathways for eachof our CNV subgroups are also listed in Table 4.DiscussionMiRNA-mediated gene regulation is complex and gen-omic variations, such as CNVs and SNPs, which canmodulate miRNA expression, add to this complexity.The miRNA-CNV relationship has been rarely studiedand predominantly involved polymorphic CNVs foundin control cohorts [9,10].Our study is novel in its comparison of miRNA contentin different classes of CNVs detected in a cohort of sub-jects with ID relative to cognitively typical subjects. Wefound a significant increase in the number of miRNAs inde novo and DECIPHER CNVs versus common CNVs(weighted median 0.6 and 0.8 versus 0.0, P < 0.05). Inaddition, the miRNAs in de novo CNVs were more likelyto have expression in brain-related tissues or cell linescompared to the miRNAs from common CNV groups.Our collective findings suggest that miRNAs from denovo and putatively pathogenic CNVs could contributeProtein coding gene coverageCNVs in our datasetsReference genome with similar length of CNVsA. B.Mature miRNA coverageFractionofCNV-covered0. 2 3 4 5FractionofCNV-covered0. 2 3 4 5Figure 2 Genomic coverage of miRNA (A) and protein coding genes (B) in different CNV subgroups. The fraction was defined as observednumber of miRNA (or protein coding genes) in each CNV group divided by the total miRNAs (or total number of protein coding genes) inhuman genome. The miRNA or gene fraction in CNVs was compared to the miRNA or gene fraction in the reference genome which wasgenerated by extracting random DNA fragments with similar length to the respective CNVs 1000 times from human genome. * Indicates p < 0.05(a Wilcoxon signed-rank test). 1: De novo CNVs; 2: DECIPHER CNVs; 3: Familial CNVs; 4: Common CNVs from controls; 5: Common CNVs from cases.Table 2 Expression/Function of miRNAs in different CNVgroupsCNV category No. ofmiRNANo. of miRNA withavailable expressiondata*Brain expression/function relatedmiRNA**De novo 84 18 10 (55%)Familial 14 4 2 (50%)Commonfrom casecohort67 4 1 (25%)Commonfrom controlcohort30 3 0PathogenicfromDECIPHER90 24 11 (46%)Notes: *indicates data collected from websites: http://www.microrna.org/microrna/ and mimiRNa (http://mimirna.centenary.org.au/mep/mir.htm).**indicate manual PubMed search for each miRNA.Qiao et al. BMC Genomics 2013, 14:544 Page 4 of 10http://www.biomedcentral.com/1471-2164/14/544to ID etiopathogenesis, in addition to coding genes inte-gral to CNVs.Similar to the increase in the number of miRNAs inde novo and DECIPHER CNVs, the number of miRNAtarget genes integral to these CNV types was also higherin comparison to familial or common CNVs. Lowernumbers of miRNAs and miRNA target genes in com-mon CNVs compared to the de novo or pathogenicCNVs suggests that they participate in processes thatcan tolerate functional variation. Previous studies haveshown that targets for miRNAs from polymorphicCNVs tend to participate in environment-orientatedprocesses including stimulus responses and immune re-sponses, while the targets from non-CNV regions of thegenome are enriched for fundamental biological pro-cesses such as maintenance of chromatin, chromosomesegregation and nucleic acid processes [26]. In ourstudy, genes from common CNVs that are targets formiRNAs do not show enrichment in any pathway whilethose from familial CNVs show enrichment in ubiquitinmediated proteolysis. In contrast, brain function relatedpathways such as axon guidance, Huntington’s disease,neutrophin signaling pathway, are found to be amongthe top 10 pathways enriched for targets from de novoand pathogenic CNVs. Targets from these two CNVgroups also showed enrichment in the MAPK (mitogen-activated protein kinase) signaling pathway. The MAPKsignaling cascade is involved in a wide variety of cellularprocesses and has been recently reported to be involvedin ID pathogenesis [27-29]. Seven genes from de novoand DECIPHER CNVs which are targets for miRNAswere found to be involved in the MAPK pathway(MAPK9, MAPT, MEF2C, MKNK2, PAK2, RPS6KA1,and TAOK2). Three of them are known to be related toTable 3 MiRNAs with expression and/or function related to brain or nervous system#Chr Start (bp) End (bp) miRNAIDCNVfeatureExpression* Functional relevance** Reference1 1103258 1103279 hsa-miR-200a/bde novo hsa-pancreatic islets, hsa-breast adenocarcinoma,HT29, breast malignanttumorolfactory neurogenesis; neuronaldifferentiation of neural stem/progenitorcellsPMID:18184563;PMID: 229934451 1104435 1104456 hsa-miR-429de novo cancer-related non-brain cancer related; neuroprotectiveeffect in in vitro ischemiaPMID: 21684154;PMID: 205769535 87962684 87962705 hsa-miR-9de novo brain, astrocytoma,neuroblastomaUpregulation in HD (Huntington’s disease);involved in spinal motor neuron disease;increased in fmr1/fxr2 knock-out micePMID: 19118166;PMID: 20616011PMID:219572335 179442361 179442382 hsa-miR-340de novo neuroblastoma cancer-related including neuroblastoma PMID: 227970597 99691233 99691253 hsa-miR-25de novo neuroblastoma neural stem cells differentiation PMID: 213861327 99691438 99691460 hsa-miR-93de novo cervix-Hela neural stem cells differentiation PMID: 213861327 99691625 99691646 hsa-miR-106bde novo Neurobl-SHSY5Y_IFN, AML-THP1, kidney-embryo-HEK2neuronal differentiation; pathogenesis ofAlzheimer’s diseasesPMID: 21386132;PMID: 2070903019 4770712 4770734 hsa-miR-7de novo neuroblastoma Repression of alpha-synuclein accumaulationin Parkinson’s disease; brain cancersPMID: 20106983;PMID: 2191268120 61151558 61151579 hsa-miR-1de novo heart, thyroid, Ewing-sarcomamuscle development; modulating neuriteoutgrowthPMID: 22365735;PMID: 2117074520 61809866 61809887 hsa-miR-124de novo hsa-hippocamp-adult, hsa-celebellum-adultneuronal differentiation andincorporation of neural-specific exons;increased in fmr1/fxr2 knock-out micePMID: 17679093PMID:219572331 94312436 94312455 hsa-miR-760familial hsa-Neurobl-SHSY5Y, hsa-Hodgkin-KMH2, hsa-midbrain-adultnon-brain cancer related PMID: 2297020915 45725296 45725317 hsa-miR-147bfamilial hsa-fibrobl-CMV, hsa-medullobl-DADY, hsa-DLBCL-DLBL3, dendritic cellrectal cancer-specific PMID: 228505662 32757280 32757298 hsa-miR-558common-caseovary and ovary-relatedcancercancer-related including neuroblastoma PMID: 21498633Notes: *indicates data collected from websites: http://www.microrna.org/microrna/ and mimiRNa (http://mimirna.centenary.org.au/mep/mir.htm).**indicate manual PubMed search for each miRNA.Qiao et al. BMC Genomics 2013, 14:544 Page 5 of 10http://www.biomedcentral.com/1471-2164/14/544ID (MAPT [30], MEF2C [31,32], PAK2 [33]). Copynumber changes of these targets could affect theirregulation by miRNA. Target genes for CNV miRNAsdetected in subjects with autism were also found to beenriched in MAPK pathway [21].For familial CNVs we noted that although the numberof miRNAs is much lower than in de novo or DECIPHERpathogenic CNV groups, and not different than in com-mon CNVs, the number of miRNA target genes was sig-nificantly higher than in common CNVs. This mightsuggest potential phenotypic impact of these familialCNVs. In our cohort we had two familial CNVs, predis-posing to ID: 16p11.2 and 1q21.1. These CNVs are well-known for their heterogeneous phenotypes and familial orTable 4 Summary of miRNA target genes within the CNV subgroups and the miRNAs targeting the CNV genesCNVcategoryNo. of codinggenes inCNVsGenes in CNVstargeted bymiRNAPathways enrichedfor miRNA targetgenesmiRNA targeting the genes in CNVs in enrichment analysisDe novo fromID737 84 (11.5%)a MAPK signalingpathwayMIR-128A,MIR-128B,MIR-194,MIR-27A,MIR-27B,MIR-296,MIR-302A,MIR-302B,MIR-302C,MIR-302D,MIR-326,MIR-329,MIR-34A,MIR-34C,MIR-372,MIR-373,MIR-449,MIR-503,MIR-520A,MIR-520B,MIR-520C,MIR-520D,MIR-520E,MIR-526B,MIR-9,MIR-93LysosomeInsulin signalingpathwayAxon guidanceHuntington’s diseaseHedgehog signalingpathwayEndocytosisN-Glycan biosynthesisp53 signalingpathwayProgesterone-mediated oocytematurationPathogenicfrom Decipher838 87 (10.4%)a Neurotrophinsignaling pathwayMIR-196A,MIR-196B,MIR-24,MIR-320,MIR-506,MIR-493,MIR-25,MIR-32,MIR-92,MIR-363,MIR-367,MIR-512-5P,MIR-302C,MIR-30A-5P,MIR-30C,MIR-30D,MIR-30B,MIR-30E-5P,MIR-485-3PRenal cell carcinomaMAPK signalingpathwayPathways in cancerChronic myeloidleukemiaChemokine signalingpathwayLong-termpotentiationErbB signalingpathwayTGF-beta signalingpathwayAdherens junctionFamilial fromID188 15 (8.1%)a Ubiquitin mediatedproteolysisMIR-193A,MIR-193B,MIR-495,MIR-302C,MIR-198Commonfrom ID casecohort454 0b N/A 0Commonfrom controlcohort266 3 (1.1%)b N/A MIR-526BNotes: Significant difference was found between data a and data b (P < 0.05) (a two sided Fisher’s exact test). N/A: not available.Qiao et al. BMC Genomics 2013, 14:544 Page 6 of 10http://www.biomedcentral.com/1471-2164/14/544de novo occurrence. The 1q21.1 CNV contains a singlemiRNA of unknown role (miR-5087), while the 16p11.2paternal duplication covers 2 miRNAs: miR-3680-3p andmiR-3680-5p and the de novo 16p11.2 deletion has nomiRNA content (Additional file 1: Table S2). Conse-quences of the variability in copy number and function ofmiRNA integral to these CNVs is yet unknown. Under-standing the role of miRNAs and miRNA targets in famil-ial CNVs is of interest since this type of CNV represents asignificant and clinically relevant interpretational challengethat could be guided by their miRNA features.The ultimate proof that miRNAs influence the patho-genicity of genomic changes will come from an empiricalconfirmation of their copy number and expression change,and influence on the expression of their targets. The limi-tation of our own and other studies is due to their depend-ence on accurate prediction of miRNA targets and the factthat miRNA numbers increase dramatically between dif-ferent versions of miRBase database. Therefore, compari-son between different studies in terms of miRNA contentin CNVs is challenging. Furthermore, some miRNAs maystill represent false positive discoveries [10]. Unfortunately,we do not have miRNA expression data from our CNVsas RNA was not available. However, a recent study byGarcia-Orti et al. [34], demonstrated 10% of 259 studiedmiRNAs from regions of gain and loss detected in AMLhad significant change in expression concordant with thetype of copy number change.A more recently published study highlighted the sig-nificance of specific CNV-miRNAs and their targets inautism [21]. It assessed the content and function of 378autism-associated CNVs, and detected 71 miRNAs. FivemiRNAS were previously reported in ASD and 3 wereknown to have neuronal function. In our study, amongthe 84 miRNAs in the de novo CNVs, 3 were foundto be associated with neurodegenerative disorders(miRNA-7, miRNA-9, miRNA-106b) [35-37] and 1 withID (miRNA-9) [37].Our analysis of miRNAs and miRNA targets related toCNVs is a first attempt to evaluate their role in a patientcohort manifesting idiopathic intellectual disabilities. Add-itional studies of cohorts of subjects with ID would benefitfurther evaluation of the apparent increase in miRNA con-tent in de novo CNVs demonstrated in this study. In clin-ical practice, the interpretation of miRNAs that occur inpatient CNVs is frequently challenging, particularly if theyare the only genes that are included in the CNV. With glo-bal investigations of miRNAs in subjects with ID, followedby their expression analysis, our understanding of the roleof miRNA in ID pathogenesis will be further improved.ConclusionsOur findings support a possible role of miRNA copynumber change in cognition and/or CNV-mediateddevelopmental delay based on increased number ofmiRNA and miRNA target genes in de novo versus com-mon CNVs in subjects with ID as well as their expres-sion profile and participation in pathways. Systematicanalysis of expression/function of miRNAs in additionto coding genes integral to CNVs could uncover newcauses of ID.MethodsSubjects213 subjects with idiopathic ID were recruited for arrayCGH analysis by clinical geneticists across Canada.Selected individual or groups of cases were reportedpreviously [38-42]. A cohort of 32 cognitively normalsubjects had array testing as internal control samples.All of the subjects were tested by either Agilent 105 KOligo array (227 subjects) or NimbleGen array (18 sub-jects). Forty syndromic genomic regions were selectedfrom the DECIPHER database (http://decipher.sanger.ac.uk/) [43] and represented patients with neurode-velopmental delay associated with CNV findings ofestablished potential for pathogenicity. The use of theDNA from these patients in our cohort was approved bythe Committee for Ethical Review of Research involvingHuman Subjects, University of British Columbia. All sub-jects gave written informed consent for participation inthe study and anonymized data were used for CNV/miRNA analysis.Array CGHGenomic DNA was extracted from peripheral blood usingPUREGENE DNA Isolation Kits (Gentra, Minneapolis,MN). A pool of normal male or female control DNAs(Promega, Madison, WI) was used as reference DNAmatching the sex of the proband samples.Agilent 105 K array analysis was performed accordingto the protocol provided by the company (version 4.0,June 2006, Agilent Technologies, CA, USA) [44]. Fea-ture Extraction software (version, Agilent Technolo-gies) rendered image analysis using the manufacturer’srecommended settings (CGH_v4_95) and human genomeassembly hg18. The minimum absolute average of log2 ra-tio was 0.25.Higher-resolution 385 K oligonucleotide genome arrayCGH was performed courtesy of NimbleGen. Array log2ratio > ±0.2 was used for a segmentation (region).For both the Agilent and Nimblegen array platforms, ≥3consecutive probes were required for a significant CNVcall. CNVs that overlapped in genomic coverage were con-sidered to represent the same CNV loci.Types of CNVsAll detected CNVs were grouped into 3 subgroups (denovo, familial and common CNVs) based on the criteriaQiao et al. BMC Genomics 2013, 14:544 Page 7 of 10http://www.biomedcentral.com/1471-2164/14/544described previously [45]. Briefly, CNVs completelyoverlapping with variants reported in at least two stud-ies in the Database of Genomic Variants (DGV) (http://projects.tcag.ca/variation/) or in our internal controlswere considered common CNVs; CNVs that overlappedpartially (<50%) or did not overlap with CNVs reportedin the DGV or our internal controls were called rareCNVs including de novo (not detected in proband’s par-ents) and familial CNVs (inherited from either of par-ents). All unique CNVs (de novo and familial) wereconfirmed by an independent secondary method (suchas FISH method) and only single copy gains or losseswere identified.Bioinformatics analysisThe complete list of miRNAs in the whole genome wasdownloaded from miRBase v19 (http://www.mirbase.org/) [46], containing >2000 mature miRNAs in human.The miRNAs in different CNV subgroups were obtainedby using intersecting tool in Galaxy (https://main.g2.bx.psu.edu/) [47].The expression profiles of miRNAs were obtained byusing mimiRNA database (http://mimirna.centenary.org.au/mep/formulaire.html) [24] and microRNA.org - Targetsand Expression (http://www.microrna.org/microrna/home.do) [25]. Both web-tools provide experimentally derivedmiRNA expression data after inputting a specificmiRNA name.The genomic locations of 19,905 human protein cod-ing genes (PCG) were extracted from the MISO database(http://genes.mit.edu/burgelab/miso/index.html). Genecontent of different CNV subgroups was obtained byconsidering genes within or covering CNV regions.WebGestalt2 (WEB-based GEne SeT AnaLysis ToolkitV2) (http://bioinfo.vanderbilt.edu/webgestalt/) is a publiclyavailable web-tool for functional enrichment analysis ofgene sets using a web-based integrated data mining system[48]. Using hypergeometric test, the top 10 groups of pro-tein coding genes from each CNV subgroup that are targetsfor miRNA were generated by this tool and ranked by adjP(p value adjusted by multiple test adjustment). Amongthese miRNA target genes, only those with adjP < 0.05, i.e.distributed in CNVs more often than expected whencompared to reference genome, were selected forKyoto Encyclopedia of Genes and Genomes (KEGG)pathway enrichment analysis. Similarly, the top tenKEGG pathways enriched for miRNA target gene weregenerated by the same tool for each CNV subgroup.Only the top ten enriched pathways with adjP < 0.05were selected and compared between different CNVsubgroups. The WebGestalt2 tool was also used toidentify miRNAs that target the target genes for eachCNV group.Statistical analysisAll statistical analyses were performed in R 2.12MiRNA and protein coding gene (PCG) coverage relative torandom distribution: for each CNV/CNVR subgroup wegenerated a set of random CNV/CNVRs with similarlength distribution and total genome coverage. The num-ber of miRNA regions and PCG regions affected by theserandom CNV/CNVRs was assessed. This operation wasrepeated 1000 times for each of the 5 CNV subgroups,generating a series of values that served as a measurementof the miRNA and PCG coverage of CNV/CNVRs undera random distribution. We performed a two-tailed test tocompute the p value of the actual miRNA and PCG cover-age of the 5 CNV/CNVR subsets. The p values were cal-culated as the number of values falling above (or below)the observed miRNA and gene coverage of the actualCNVRs datasets.Comparison of miRNA and PCG content: we computedthe number of miRNA and PCG in the individual CNVsrelative to their size i.e. the number of miRNA and PCGper Mb. We compared the subgroups to each other usinga Wilcoxon signed-rank test to assess whether a group hasa significantly different miRNA and PCG density.In order to determine the significance of the numbermiRNA related to brain function, we compared the frac-tion of brain-function related miRNA in the pool of avail-able CNV-miRNA between each pair of CNV subgroups.The statistical p value was computed using a two-sidedFisher’s exact test. Similarly, we performed two-sidedFisher’s exact tests to compare the fraction of genestargeted by miRNA in the different CNV subgroups usingthe total number of genes encompassed by the CNVs.Additional fileAdditional file 1: Genomic coordinates of CNV/CNVRs and miRNAsused in this study.AbbreviationsCNVs: Copy number variations; CNVRs: Copy number variant regions;DECIPHER: Database of chromosomal imbalance and phenotype in humansusing ensembl resources; ID: Intellectual disability; KEGG: Kyoto encyclopediaof genes and genomes; MAPK: Mitogen-activated protein kinase;MAPK9: Mitogen-activated protein kinase 9; MAPT: Microtubule-associatedprotein tau; MEF2C: Myocyte enhancer factor 2C; miRNA: MicroRNA;MKNK2: MAP kinase interacting serine/threonine kinase 2; PAK2: P21 protein(Cdc42/Rac)-activated kinase 2; PCG: Protein coding gene;pri-miRNAs: Primary transcript miRNA; RPS6KA1: Ribosomal protein S6 kinase;90 kDa: Polypeptide 1; SNPs: Single nucleotide polymorphisms; TAOK2: TAOkinase 2; WebGestalt2: WEB-based GEne SeT AnaLysis Toolkit V2.Competing interestsWe declare no conflict of interest in our manuscript titled as “miRNA andmiRNA target genes in copy number variations occurring in individuals withintellectual disability”.Authors’ contributionsYQ performed data acquisition, data analysis, and drafting of manuscript; CBinitiated the project and analyzed the data; EM performed statistical andQiao et al. BMC Genomics 2013, 14:544 Page 8 of 10http://www.biomedcentral.com/1471-2164/14/544bioinformatics analyses; SL recruited clinical cases and reviewed themanuscript; PP supervised statistical and bioinformatics analyses andreviewed the manuscript; ES supervised and designed the study, helped withdata interpretation, and critically revised the manuscript. All authors read andapproved the final manuscript.AcknowledgementsThis work was supported by funding from the Canadian Institutes forHealth Research (CIHR) (MOP 74502; PI: ERS), Canadian Foundation forInnovation-Leading Edge Fund (CFI-LEF; PI: MESL) and the BC KnowledgeDevelopment Fund (MOP 64217; PI: MESL), and Establishment Grantfunding from the Michael Smith Foundation for Health Research. PP wassupported by a career award from the Michael Smith Foundation for HealthResearch, a CIHR New Investigator award, the Canadian Foundation forInnovation, and the National Institutes of Health (GM076990). MESL and ERSare Career Scholars supported by the Michael Smith Foundation for HealthResearch. The authors appreciate the collaboration and support of theparticipating subjects and their families. This study makes use of datagenerated by the DECIPHER Consortium. A full list of centres whocontributed to the generation of the data is available from http://decipher.sanger.ac.uk and via email from decipher@sanger.ac.uk. Funding for theproject was provided by the Wellcome Trust.Author details1Department of Pathology and Lab Medicine, BC Child and Family ResearchInstitute, University of British Columbia (UBC), Vancouver, BC V5Z 4H4,Canada. 2Department of Medical Genetics, BC Child and Family ResearchInstitute, University of British Columbia (UBC), Vancouver, BC V6H 3N1,Canada. 3Centre for High-throughout Biology, 177 Michael SmithLaboratories, UBC, Vancouver, BC V6T 1Z4, Canada.Received: 23 March 2013 Accepted: 6 August 2013Published: 10 August 2013References1. Lee RC, Feinbaum RL, Ambros V: The C. elegans heterochronic gene lin-4encodes small RNAs with antisense complementarity to lin-14. Cell 1993,75:843–854.2. Wightman B, Ha I, Ruvkun G: Posttranscriptional regulation of theheterochronic gene lin-14 by lin-4 mediates temporal pattern formationin C. elegans. Cell 1993, 75:855–862.3. Kim VN: MicroRNA biogenesis: coordinated cropping and dicing. Nat RevMol Cell Biol 2005, 6:376–385.4. Lewis BP, Burge CB, Bartel DP: Conserved seed pairing, often flanked byadenosines, indicates that thousands of human genes are microRNAtargets. Cell 2005, 120:15–20.5. Miranda KC, Huynh T, Tay Y, Ang YS, Tam WL, Thomson AM, Lim B, Rigoutsos I:A pattern-based method for the identification of MicroRNA binding sitesand their corresponding heteroduplexes. Cell 2006, 126:1203–1217.6. Vasudevan S: Posttranscriptional upregulation by microRNAs.Wiley Interdiscip Reviews 2011, 3:311–330.7. Sethupathy P, Collins FS: MicroRNA target site polymorphisms and humandisease. Trends Genet 2008, 24:489–497.8. Duan S, Mi S, Zhang W, Dolan ME: Comprehensive analysis of the impactof SNPs and CNVs on human microRNAs and their regulatory genes.RNA Biol 2009, 6:412–425.9. Felekkis K, Voskarides K, Dweep H, Sticht C, Gretz N, Deltas C: Increased numberof microRNA target sites in genes encoded in CNV regions. Evidence for anevolutionary genomic interaction. Mol Biol Evol 2011, 28:2421–2424.10. Marcinkowska M, Szymanski M, Krzyzosiak WJ, Kozlowski P: Copy numbervariation of microRNA genes in the human genome. BMC Genomics 2011,12:183.11. Schiffman JD, Lorimer PD, Rodic V, Jahromi MS, Downie JM, Bayerl MG,Sanmann JN, Althof PA, Sanger WG, Barnette P, et al: Genome wide copynumber analysis of paediatric Burkitt lymphoma using formalin-fixedtissues reveals a subset with gain of chromosome 13q andcorresponding miRNA over expression. Brit J Haematol 2011, 155:477–486.12. Shim SM, Nam HY, Lee JE, Kim JW, Han BG, Jeon JP: MicroRNAs in humanlymphoblastoid cell lines. Crit Rev Eukar Gene 2012, 22:189–196.13. Bandiera S, Hatem E, Lyonnet S, Henrion-Caude A: microRNAs in diseases:from candidate to modifier genes. Clinical genetics 2010, 77:306–313.14. Esteller M: Non-coding RNAs in human disease. Nature reviews 2011,12:861–874.15. Ghahramani Seno MM, Hu P, Gwadry FG, Pinto D, Marshall CR, Casallo G,Scherer SW: Gene and miRNA expression profiles in autism spectrumdisorders. Brain research 2011, 1380:85–97.16. Provost P: MicroRNAs as a molecular basis for mental retardation,Alzheimer’s and prion diseases. Brain research 2010, 1338:58–66.17. Talebizadeh Z, Butler MG, Theodoro MF: Feasibility and relevance ofexamining lymphoblastoid cell lines to study role of microRNAs inautism. Autism Res 2008, 1:240–250.18. Bian S, Sun T: Functions of noncoding RNAs in neural development andneurological diseases. Mol Neurobiol 2011, 44:359–373.19. Willemsen MH, Valles A, Kirkels LA, Mastebroek M, Olde Loohuis N, Kos A,Wissink-Lindhout WM, de Brouwer AP, Nillesen WM, Pfundt R, et al:Chromosome 1p21.3 microdeletions comprising DPYD and MIR137 areassociated with intellectual disability. J Med Genet 2011, 48:810–818.20. de Pontual L, Yao E, Callier P, Faivre L, Drouin V, Cariou S, Van Haeringen A,Genevieve D, Goldenberg A, Oufadem M, et al: Germline deletion of themiR-17 approximately 92 cluster causes skeletal and growth defects inhumans. Nature genetics 2011, 43:1026–1030.21. Vaishnavi V, Manikandan M, Tiwary BK, Munirajan AK: Insights on thefunctional impact of MicroRNAs present in autism-associated copynumber variants. PloS one 2013, 8:e56781.22. Kosik KS: The neuronal microRNA system. Nat Rev Neurosci 2006, 7:911–920.23. Szulwach KE, Jin P, Alisch RS: Noncoding RNAs in mental retardation.Clinical genetics 2009, 75:209–219.24. Ritchie W, Flamant S, Rasko JE: mimiRNA: a microRNA expression profilerand classification resource designed to identify functional correlationsbetween microRNAs and their targets. Bioinformatics (Oxford, England)2010, 26:223–227.25. Betel D, Koppal A, Agius P, Sander C, Leslie C: Comprehensive modeling ofmicroRNA targets predicts functional non-conserved and non-canonicalsites. Genome biology 2010, 11:R90.26. Wu X, Zhang D, Li G: Insights into the regulation of human CNV-miRNAsfrom the view of their target genes. BMC Genomics 2012, 13:707.27. Aoki Y, Niihori T, Narumi Y, Kure S, Matsubara Y: The RAS/MAPKsyndromes: novel roles of the RAS pathway in human genetic disorders.Hum Mutat 2008, 29:992–1006.28. Houge G, Rasmussen IH, Hovland R: Loss-of-function CNKSR2 mutation isa likely cause of non-syndromic X-linked intellectual disability.Mol Syndromol 2012, 2:60–63.29. McMillan EL, Kamps AL, Lake SS, Svendsen CN, Bhattacharyya A: Geneexpression changes in the MAPK pathway in both Fragile X and downsyndrome human neural progenitor cells. Am J Stem Cells 2012,1:154–162.30. Shaw-Smith C, Pittman AM, Willatt L, Martin H, Rickman L, Gribble S, Curley R,Cumming S, Dunn C, Kalaitzopoulos D, et al: Microdeletion encompassingMAPT at chromosome 17q21.3 is associated with developmental delay andlearning disability. Nat Genet 2006, 38:1032–1037.31. Nowakowska BA, Obersztyn E, Szymanska K, Bekiesinska-Figatowska M, Xia Z,Ricks CB, Bocian E, Stockton DW, Szczaluba K, Nawara M, et al: Severe mentalretardation, seizures, and hypotonia due to deletions of MEF2C. Am J MedGenet B Neuropsychiatr Genet 2010, 153B:1042–1051.32. Zweier M, Gregor A, Zweier C, Engels H, Sticht H, Wohlleber E, Bijlsma EK,Holder SE, Zenker M, Rossier E, et al: Mutations in MEF2C from the5q14.3q15 microdeletion syndrome region are a frequent cause ofsevere mental retardation and diminish MECP2 and CDKL5 expression.Hum Mutat 2010, 31:722–733.33. Quintero-Rivera F, Sharifi-Hannauer P, Martinez-Agosto JA: Autistic andpsychiatric findings associated with the 3q29 microdeletion syndrome:case report and review. Am J Med Genet 2010, 152A:2459–2467.34. Garcia-Orti L, Cristobal I, Cirauqui C, Guruceaga E, Marcotegui N, Calasanz MJ,Castello-Cros R, Odero MD: Integration of SNP and mRNA arrays withmicroRNA profiling reveals that MiR-370 is upregulated and targets NF1 inacute myeloid leukemia. PLoS One 2012, 7:e47717.35. Doxakis E: Post-transcriptional regulation of alpha-synuclein expressionby mir-7 and mir-153. J Biol Chem 2010, 285:12726–12734.36. Wang H, Liu J, Zong Y, Xu Y, Deng W, Zhu H, Liu Y, Ma C, Huang L, Zhang L,et al: miR-106b aberrantly expressed in a double transgenic mouse modelfor Alzheimer’s disease targets TGF-beta type II receptor.Brain Res 2010, 1357:166–174.Qiao et al. BMC Genomics 2013, 14:544 Page 9 of 10http://www.biomedcentral.com/1471-2164/14/54437. Xu XL, Zong R, Li Z, Biswas MH, Fang Z, Nelson DL, Gao FB: FXR1P but notFMRP regulates the levels of mammalian brain-specific microRNA-9 andmicroRNA-124. J Neurosci 2011, 31:13705–13709.38. Gibson WT, Harvard C, Qiao Y, Somerville MJ, Lewis ME, Rajcan-Separovic E:Phenotype-genotype characterization of alpha-thalassemia mentalretardation syndrome due to isolated monosomy of 16p13.3. Am J MedGenet 2008, 146A:225–232.39. Harvard C, Malenfant P, Koochek M, Creighton S, Mickelson EC, Holden JJ,Lewis ME, Rajcan-Separovic E: A variant Cri du Chat phenotype andautism spectrum disorder in a subject with de novo crypticmicrodeletions involving 5p15.2 and 3p24.3–25 detected using wholegenomic array CGH. Clin Genet 2005, 67:341–351.40. Tyson C, Harvard C, Locker R, Friedman JM, Langlois S, Lewis ME, Van Allen M,Somerville M, Arbour L, Clarke L, et al: Submicroscopic deletions andduplications in individuals with intellectual disability detectedby array-CGH. Am J Med Genet 2005, 139:173–185.41. Tyson C, McGillivray B, Chijiwa C, Rajcan-Separovic E: Elucidation of acryptic interstitial 7q31.3 deletion in a patient with a language disorderand mild mental retardation by array-CGH. Am J Med Genet 2004,129A:254–260.42. Rajcan-Separovic E, Harvard C, Liu X, McGillivray B, Hall JG, Qiao Y, Hurlburt J,Hildebrand J, Mickelson EC, Holden JJ, et al: Clinical and molecularcytogenetic characterisation of a newly recognised microdeletionsyndrome involving 2p15–16.1. J Med Genet 2007, 44:269–276.43. Firth HV, Richards SM, Bevan AP, Clayton S, Corpas M, Rajan D, Van Vooren S,Moreau Y, Pettett RM, Carter NP: DECIPHER: database of chromosomalimbalance and phenotype in humans using ensembl resources. Am J HumGenet 2009, 84:524–533.44. Fan YS, Jayakar P, Zhu H, Barbouth D, Sacharow S, Morales A, Carver V,Benke P, Mundy P, Elsas LJ: Detection of pathogenic gene copy numbervariations in patients with mental retardation by genomewideoligonucleotide array comparative genomic hybridization. Hum Mutat2007, 28:1124–1132.45. Qiao Y, Tyson C, Hrynchak M, Lopez-Rangel E, Hildebrand J, Martell S,Fawcett C, Kasmara L, Calli K, Harvard C, et al: Clinical application of 2.7Mcytogenetics array for CNV detection in subjects with idiopathic autismand/or intellectual disability. Clin Genet 2013, 83:145–154.46. Kozomara A, Griffiths-Jones S: miRBase: integrating microRNA annotationand deep-sequencing data. Nucleic acids research 2011, 39:D152–157.47. Goecks J, Nekrutenko A, Taylor J: Galaxy: a comprehensive approach forsupporting accessible, reproducible, and transparent computationalresearch in the life sciences. Genome Biol 2010, 11:R86.48. Zhang B, Kirov S, Snoddy J: WebGestalt: an integrated system forexploring gene sets in various biological contexts. Nucleic Acids Res 2005,33:W741–748.doi:10.1186/1471-2164-14-544Cite this article as: Qiao et al.: miRNA and miRNA target genes in copynumber variations occurring in individuals with intellectual disability.BMC Genomics 2013 14:544.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitQiao et al. BMC Genomics 2013, 14:544 Page 10 of 10http://www.biomedcentral.com/1471-2164/14/544


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items