ORIGINAL RESEARCH

Embryonic stem cell-related miRNAsare involved in differentiation ofpluripotent cells originating from thegerm lineAthanasios Zovoilis1,2,*,†, Angeliki Pantazi1,†, Lukasz Smorag1,†,Lennart Opitz3, Gabriela Salinas Riester3, Marieke Wolf4,Ulrich Zechner4, Anna Holubowska1, Colin L. Stewart5, andWolfgang Engel1

1Institute of Human Genetics, University of Goettingen, 37073 Goettingen, Germany 2ENI Goettingen, University of Goettingen, GrisebachStr. 5, D-37077 Goettingen, Germany 3DNA Microarray Facility, University of Goettingen, 37073 Goettingen, Germany 4Institute of HumanGenetics, University of Mainz, 55131 Mainz, Germany 5Institute of Medical Biology, Singapore 138648, Singapore

*Correspondence address. Tel: +49-551-3910380; Email: azovoil@gwdg.de

Submitted on February 15, 2010; resubmitted on June 3, 2010; accepted on June 15, 2010

abstract: Cells originating from the germ cell lineage retain the remarkable property under special culture conditions to give rise tocells with embryonic stem cell (ESC) properties, such as the multipotent adult germline stem cells (maGSCs) derived from adult mouse testis.To get an insight into the mechanisms that control pluripotency and differentiation in these cells, we studied how differences observed duringin vitro differentiation between ESCs and maGSCs are associated with differences at the level of microRNAs (miRNAs). In this work, weprovide for a first time a connection between germ cell origin of maGSCs and their specific miRNA expression profile. We found thatmaGSCs express higher levels of germ cell markers characteristic for primordial germ cells (PGCs) and spermatogonia compared withESCs. Retained expression of miR-290 cluster has been previously reported in maGSCs during differentiation and it was associated withhigher Oct-4 levels. Here, we show that this property is also shared by another pluripotent cell line originating from the germ line, theembryonic germ cells. In addition, we provide proof that the specific miRNA expression profile of maGSCs has an impact on their differ-entiation potential. Low levels of miR-302 in maGSCs during the first 10 days of leukaemia inhibitory factor deprivation are shown to benecessary for the maintenance of high levels of early germ cell markers.

Key words: embryonic stem cells / multipotent adult germline stem cells / microRNAs, miR-290, miR-302 / differentiation / primordialgerm cell marker, Dppa-3

IntroductionDuring recent years a number of publications reported the remarkableproperty of the germ cell lineage to give rise to cells with embryonicstem cell (ESC) properties after culture under special conditions(Matsui et al., 1992; Kanatsu-Shinohara et al., 2004; Guan et al.,2006). Thus, although ESCs still remain the gold standard of pluripo-tency, the repertoire of pluripotent stem cells in mouse has beenextended to include EGCs derived from primordial germ cells(PGCs), multipotent germline stem cells (mGSCs) from neonataltestis, multipotent adult germline stem cells (maGSCs) from adulttestis and gPS cells (germline-derived pluripotent stem cells) from

adult testis (Matsui et al., 1992; Resnick et al., 1992; Kanatsu-Shinoharaet al., 2004; Guan et al., 2006; Seandel et al., 2007; Izadyar et al., 2008;Kanatsu-Shinohara et al., 2008; Ko et al., 2009). Similar results havebeen recently obtained also for human cells (Conrad et al., 2008;Kossack et al., 2009). Despite the direct implications of these findingsfor regenerative medicine, the exact mechanisms that control pluripo-tency and differentiation in these cells are still poorly understood.

A possible regulatory mechanism includes miRNAs, a class of smallnon-coding RNAs that regulate gene expression at the post-transcriptional level. Regulation of gene expression by miRNAs is rea-lized by the formation of complex regulatory networks, in which eachmiRNA can target many different mRNAs, and conversely, several

† These authors contributed equally to this work.

& The Author 2010. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.For Permissions, please email: journals.permissions@oxfordjournals.org

Molecular Human Reproduction, Vol.16, No.11 pp. 793–803, 2010

Advanced Access publication on June 21, 2010 doi:10.1093/molehr/gaq053

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

different miRNAs can cooperatively control a single mRNA target (Heand Hannon, 2004; Kim, 2005; Esquela-Kerscher and Slack, 2006).Surprisingly, even for the well studied ESCs, only little is knownabout the role of miRNAs in the transcriptional network and themechanisms controlling pluripotency and differentiation in thesecells. Recently, members of three miRNA clusters have been associ-ated with pluripotency in ESCs: miR-290, miR-302 and miR-17-92clusters. The first two miRNA clusters are exclusively expressed inundifferentiated pluripotent stem cells as well as during first stagesof in vitro differentiation and embryonic development (Houbaviyet al., 2003, 2005; Suh et al., 2004; Strauss et al. 2006; Tang et al.,2007; Chen et al., 2007; Zovoilis et al., 2008; Mendell, 2008; Foshayand Gallicano, 2009).

Until now, maGSCs is the only type of pluripotent stem cells orig-inating from the germ line which has been tested at the miRNA level.miRNA expression analysis revealed that these cells express alsomiR-290 and miR-302 clusters and miRNA expression levels weresimilar between undifferentiated ESCs and maGSCs (Zovoilis et al.,2008). However, during in vitro differentiation expression profiles ofmiR-290 and miR-302 clusters were found to differ between thetwo cell types. miR-290 cluster expression was found to decreasemuch later in maGSCs compared with ESCs, in which miR-290cluster members are strongly down-regulated within 5 days of differ-entiation. In addition, miR-302 cluster is strongly up-regulated inESCs during first stages of in vitro differentiation, although it increasesmuch more slowly in maGSCs.

We hypothesized that there is a connection between the abovedifferences in miRNA levels and the origin of maGSCs from germcells as well as their differentiation potential. To this end, we firsttested whether maGSCs express higher levels of markers representa-tive of their germ cell origin compared with ESCs. Then, we ques-tioned whether expressing high miR-290 cluster levels duringdifferentiation is also a property shared by another pluripotent cellline originating from germ cells, the EGCs, and whether these levelsare associated with high Oct-4 levels also in these cells. Finally, weinvestigated the question of whether maGSC cultures would expresshigher levels of germ cell markers during differentiation than ESCsdue to their germ cell origin and whether differences in miRNAlevels could be involved.

Materials and Methods

Cell cultureDerivation and culture of maGSC and ESC lines have been described pre-viously (Cheng et al., 2004; Guan et al., 2006). With the exception ofFig. 1a and b, ESCs, maGSCs and EGCs used in this study had the samegenetic background, that of the 129/Sv mouse line, so that impact of differ-ent strain backgrounds on the results could be eliminated. In brief, the ESCline (ESCs) is the well-characterized ES R1 cell line derived from the 129/Svmouse line (Nagy et al., 1993). The maGSC line (maGSCs) is the maGSC129SV line used and characterized previously (Zovoilis et al., 2008;Zechner et al., 2009). The EGC line (EGCs) is the EG-1 line used in thework of Sharova et al. (2007), which also originates from the 129/Svmouse line (Stewart et al., 1994). To maintain cells in undifferentiatedstate, cells were cultured as reported previously (Zovoilis et al., 2008).Prior to any RNA extraction and subsequent analysis, cells (two replicatesper line) were cultured for at least 2 days in the absence of feeder layer

(FL). For the differentiation studies cells were seeded on a 0.1% gelatine-coated 6-well culture plates (5 × 105 cells/well) without leukaemia inhibi-tory factor (LIF). Differentiation experiments were repeated twice.

miRNA transfectionFor increasing miRNA levels, cells were transfected using HiPerFect trans-fection reagent (Qiagen, Hilden, Germany) according to manufacturer’slong-term transfection protocol. In brief, at Day 0 (beginning of cultureunder differentiation conditions), cells were transferred (5 × 105 cells/well) to 0.1% gelatine-coated 6-well plates. Subsequently, transfectioncomplexes containing 18 ml of the transfection reagent and the miRNAprecursor were added drop-wise onto the cells. Complexes were lefton the cells until they reached 60–80% confluence (after 2 days), thensub-cultured again and re-transfected as described above. This procedurewas repeated until Day 10 or until analysis of the cells. Pre-miR miRNAprecursor molecules (Ambion, Austin, USA) for miR-302 a, b and d,were used in a final concentration of 5 nM each. The negative controls(neg) were transfected with the respective Pre-miR control, a randomsequence Pre-miR molecule validated not to produce effects on knownmiRNA function and cells were cultured under exactly the same con-ditions for comparison. Pre-miR controls were used in a final concen-tration of 15 nM. miRNA precursor molecules used in this study arelisted in detail in Supplementary, Table S1). Under these transfection con-ditions no cell toxicity effects were observed. miRNA transfection exper-iments were repeated twice.

Real-time PCRTotal RNA (including miRNAs) was isolated using the miRNeasy mini Kit(Qiagen). Conversion into cDNA and Real-time PCR detection ofmiRNAs were carried out using the miScript Reverse Transcription Kitand the miScript SYBR-Green PCR Kit (Qiagen), respectively, on an ABIPrism 7900HT Sequence Detection System (Applied Biosystems, Darm-stadt, Germany). Optimized miRNA-specific primers for each miRNA aswell as for the endogenous control RNU6B (miScript Primer Assays,Qiagen) are listed in Supplementary, Table S2. PCR specificity waschecked as described previously (Zovoilis et al., 2008). For Real-time quan-titative RT–PCR (qRT–PCR), the QuantiTect SYBR-Green PCR MasterMix(Qiagen) was used with gene-specific primers (listed in Supplementary,Table S3) for the following genes: Dppa-3, Stra-8, Oct-4, C-kit, Dazl andBrachyury. All quantities were further normalized to values of RNU6Band Sdha for miRNA or mRNA quantification, respectively.

Bioinformatics’ approaches andstatistical analysisFor computational prediction of miRNA targets, we used the TargetScanweb platform (www.targetscan.com /release 5.1), whereas for statisticalanalysis we used the STATISTICA software package for performing aone-way ANOVA (followed by Fischer LSD’s multiple comparison test)or t-test. Data are expressed as mean+ SD. P , 0.05 was consideredstatistically significant. For functional annotation of miR-302 clustertargets, we used the data mining environment provided by the DAVIDplatform (Huang et al., 2009). The functional annotation module wasapplied for gene ontology terms in PANTHER database using an EASEscore of 0.001, a minimum number of two counts and a Benjaminivalue ≤0.01. For functional annotation clustering medium classificationstringency was used.

DNA methylation analysisGenomic DNA was isolated from maGSC and ESC lines using standardprotocols. Bisulfite treatment of genomic DNA isolated was performed

794 Zovoilis et al.

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

using the EpiTect bisulfite kit (Qiagen, Hilden, Germany). The methylationstatus of the promoters of Oct4, miR290 and miR302 was analysed bybisulfite pyrosequencing. Bisulfite pyrosequencing was performed on aPSQTM96MA Pyrosequencing System (Biotage, Uppsala, Sweden) withthe PyroGold SQA reagent kit (Biotage). Pyro Q-CpG software(Biotage) was used for pyrosequencing data analysis. Primers that wereused and genomic locations of the regions that were tested are listed inSupplementary, Table S4.

Results

maGSCs express high levels of germ cellmarkers characteristic for PGCs andspermatogoniaIn contrast to ESCs, pluripotent cells from testis do not originate frompluripotent stem cells like those of the inner cell mass of the blastocyst

but from unipotent germ cells of testis (Kanatsu-Shinohara et al., 2008;Ko et al., 2009). In this respect, they resemble EGCs that also originatefrom the germline, after reprogramming of PGCs to an ESC-like state.We raised the question whether maGSCs show expression of earlygerm cell markers after reprogramming to the ESC-like state and towhat extent compared with ESCs. We tested two key germ cellmarkers: Dppa-3, a widely used PGC marker expressed early duringgerm cell development, and Stra-8, which is highly expressed in sper-matogonia (Oulad-Abdelghani et al., 1996; Lacham-Kaplan, 2004;Yabuta et al., 2006; Mark et al., 2008). In addition, we used mousetestis 7 days p.p. as well as adult testis as control. maGSC cultureswere consistently found to express both markers in higher levelsthan cultures of ESCs originating from the same mouse background(Fig. 1a and b). As expected, expression of Dppa-3, which is anearly expressed germ cell marker, was very low in adult testis(Fig. 1a) in contrast to the spermatogonia marker Stra-8, which washighly expressed in adult testis. Moreover, neonatal testis, which has

Figure 1 Expression levels of the PGC marker Dppa-3 and the spermatogonia marker Stra-8 in undifferentiated ESCs, maGSCs and EGCs. Cellswere cultured under standard conditions in presence of LIF and FLs (+LIF/+FL). Expression in neonatal (7 days p.p.) and adult testis was also testedas a control. Levels have been determined by qRT–PCR and quantities were normalized to the endogenous control (Sdha). (a, b) Comparison ofDppa-3 and Stra-8 levels between ESCs and maGSCs originating from three different mouse strains. Asterisk depicts statistical significance for thedifference between maGSCs of each strain and the respective ESCs from the same strain. (c, d) Comparison of Dppa-3 and Stra-8 levelsbetween ESCs and EGCs originating from the 129/Sv strain. Asterisk depicts statistical significance for the difference in Dppa-3 levels betweenmaGSCs and EGCs.

Differentiation of pluripotent cells 795

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

a high ratio of spermatogonia to more mature germ cell types (Tege-lenbosch and de Rooij, 1993), showed higher expression of Stra-8compared with adult testis (Fig. 1b).

Next, we tested whether such high expression levels of Dppa-3 andStra-8 in maGSC cultures are also observed in cultures of anotherpluripotent cell type that originates from germ cells, EGCs. EGC cul-tures express Stra-8 in levels similar to those of maGSCs (Fig. 1d).Consistent with their PGC origin, expression levels of Dppa-3 inEGCs were more than 8-fold higher than in maGSCs (Fig. 1c).These results suggest that even after reprogramming maGSCs andEGCs retain expression of key germ cell markers.

Pluripotent cells originating from germ cellsretain expression of pluripotency-relatedmiRNAs and Oct-4 during differentiationmiR-290 cluster is down-regulated in maGSCs later than in ESCs duringin vitro differentiation (deprivation of FLs and LIF; Zovoilis et al., 2008).Since EGCs have similarities to maGSCs and retain a similar expression

pattern of early germ cell markers after reprogramming, we testedwhether maintenance of high miR-290 cluster levels during in vitro differ-entiation is another common trait of these cells. To this end, we usedqRT–PCR to profile expression levels of members of the miR-290cluster in EGC cultures upon deprivation of FL and LIF for 21 days, incomparison to ESCs and maGSCs (Fig. 2a). In addition, we testedrepresentative members of another cluster, the miR-17-92 cluster,which as discussed later, has been also recently connected to ESCand PGC proliferation (Hayashi et al., 2008; Fig. 2b).

Surprisingly, like maGSCs, EGCs retained high expression levels ofmiRNA members of both clusters upon deprivation of LIF. Althoughexpression profiles of both cell types were definitely not the same,the similarity of their profiles compared with that of ESCs was striking:ESCs demonstrate a very early decrease in miRNA levels during thedifferentiation period that was tested, whereas maGSCs and EGCsdo not. Importantly, these differences in miRNA expression patternsof maGSCs and EGCs compared with ESCs are observed also whenother in vitro differentiation strategies are followed (Supplementary,Fig. S1).

Figure 2 miRNA expression profiling by qRT–PCR of members of miR-290 and miR-17-92 clusters in ESCs and maGSCs during in vitro differen-tiation. (a, b) Intensities depict expression levels relative to those of undifferentiated cells (Day 0) and were developed using Matrix2png/version 1.0.6as described previously (Pavlidis and Noble, 2003). Quantities were normalized to the endogenous control (RNU6B) and calibrated with levels ofundifferentiated cells that depict intensity values of 1. Each column depicts expression at a different time point during differentiation and differentrows represent the different miRNAs tested. (c) Schematic representation of the differentiation strategy that was followed. +LIF/+FL culture in pres-ence of FLs and LIF, refers to undifferentiated cells, before beginning of differentiation. 2LIF/2FL, culture in gelatine-coated flasks in absence of FLsand LIF that promotes differentiation.

796 Zovoilis et al.

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

Because high expression levels of miR-290 cluster during in vitrodifferentiation have been previously associated with high Oct-4expression levels (Zovoilis et al., 2008; Zovoilis et al., 2009), wealso tested Oct-4 expression profiles in EGCs during differentiationin comparison to maGSCs and ESCs (Fig. 3). Like maGSCs, and incontrast to ESCs, Oct-4 levels in EGC cultures did not decreaseupon deprivation of LIF after 10 days of in vitro differentiation.

maGSCs express high levels of germ cellmarkers during differentiationSince some properties of maGSCs associate with their germ cell origin,we hypothesized that this origin may also favour their differentiationtowards germ cells upon deprivation of LIF. Indeed, cultures ofmaGSCs were found to express during differentiation higher levelsof Stra-8 and other later expressed germ cell markers, like C-kit andDazl (Fig. 4a) compared with ESC cultures. The in vitro differentiation

environment employed here differs from that observed in vivo.However, it was reasonable to assume that increased levels of germcell markers in spontaneously differentiating maGSC cultures mayresemble the germ cell differentiation programme observed in vivo.Indeed, differentiating maGSC cultures expressed significantly higherlevels of Dppa-3 than ESC cultures (Fig. 4b), which in vivo is indicativeto commitment towards germ line and proceeds expression of C-kitand Dazl. It should be noted that expression levels of germ cellmarkers tested here refer to heterogeneous cultures undergoing spon-taneous differentiation and are not indicative of expression in thesingle cell level.

Low miR-302 expression levels indifferentiating maGSCs correlate withincreased Dppa-3 levelsWe questioned whether the specific miRNA expression profile of dif-ferentiating maGSCs can explain the observed differences in germ cellmarkers. Recently, over expression of members of miR-290 clusterwas found to reduce Dppa-3 expression in ESCs (Zovoilis et al.,2009). However, differentiating maGSCs retain stable miR-290levels; thus, it is rather unlikely that this could be a reason for theobserved differences in Dppa-3 levels in differentiating maGSCs.Another pluripotency-related cluster, miR-302, is predicted to sharesame mRNA targets with some miR-290 cluster members (www.targetscan.org/Release 5.1) and could have a similar effect asmiR-290 cluster. Thus, we investigated whether increased Dppa-3levels in differentiating maGSC cultures correlate with levels ofmiR-302 cluster in these cells. Indeed, miR-302 cluster memberswere found to be expressed in lower levels in differentiatingmaGSCs than in differentiating ESCs (Fig. 4c).

Increased miR-302 cluster levels result indown-regulation of Dppa-3 and Stra-8To confirm the association between miR-302 cluster and Dppa-3 wetested whether artificially increased levels of miR-302 cluster in differ-entiating maGSC cultures are able to inhibit Dppa-3 up-regulation.

Figure 5a summarizes the strategy that was followed. Pre-miRNAprecursor molecules (small, chemically modified double-strandedRNA molecules designed to mimic endogenous mature miRNAs)for miR-302 a, b, d were introduced into maGSCs cultured underdifferentiation conditions (2LIF/2FL) using conventional siRNAtransfection methods. To monitor the efficiency of our treatmentmiRNA levels in treated cells were compared with those of negativecontrol. In all cases successful treatment was confirmed by miRNAqRT–PCR (Supplementary, Fig. S2).

As shown in Fig. 5c, consistent with our hypothesis, increased levelsof miR-302 cluster members prevented Dppa-3 up-regulation inmaGSC cultures and resulted in decreased expression after 10 daysdifferentiation in vitro. Using the same transfection strategy we con-firmed this finding using another early germ cell marker, Stra-8 (Sup-plementary, Fig. S3) and also for ESC cultures. Although cultures ofESCs express already high levels of the miR-302 cluster members,further increase at their levels led to even lower Dppa-3 levels (Sup-plementary, Fig. S4).

We also tested the impact of high miR-302 expression levels onOct-4 expression profile in both cell types. Increased miR-302 levels

Figure 3 Expression levels of the pluripotency marker Oct-4during 10 days in vitro differentiation of ESCs, maGSCs and EGCs.Cells were cultured in absence of LIF and FLs (2LIF/2FL). Levelswere determined by qRT–PCR and quantities were normalized tothe endogenous control (Sdha) and calibrated with expressionlevels of cells before differentiation (d 0). d, day of differentiation.Asterisk depicts statistical significance at day 10 compared with undif-ferentiated cells (d 0).

Differentiation of pluripotent cells 797

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

did not affect Oct-4 expression levels in either ESCs or maGSCs(Fig. 5b).

Functional annotation of miR-302 clusterpredicted targetsTo get an insight into possible ways through which miR-302 clustermay control differentiation, we conducted bioinformatics analysisregarding genes predicted to be targeted by miR-302 clustermembers. To this end, we employed the Targetscan platform formouse which makes use of an algorithm identifying mRNAs with con-served complementarity to the seed (nucleotides 2–7) of the miRNA(Lewis et al., 2005). Targetscan yielded 417 conserved targets, with atotal of 447 conserved sites and 133 poorly conserved sites for themiR-302 cluster seed. Genes with only poorly conserved sites werenot tested further, in order to reduce the false positive rate. Allgenes that were included in further analysis are listed in Supplemen-tary, Table S5. Using the data mining environment provided by theDAVID platform (Huang et al., 2009), we applied the respective func-tional annotation module to statistically highlight the most enrichedbiological annotations of miR-302 predicted targets regardingcommon biological processes and molecular functions. The exact par-ameter values used are listed in Materials and Methods section.

As shown in Fig. 6 the majority of enriched annotations include pro-cesses connected with mRNA transcription as well as regulation oftranscription. Consistent with this, the most predominant molecularfunctions in which miR-302 targets are significantly enriched refer totranscription factor activity and DNA-protein interactions.

Interestingly, enriched biological annotations included in 15% ofcases developmental processes, in 5% cell cycle control and in 7% pro-liferation, which is in agreement with previous reports about the roleof miR-302 cluster in cyclin D1 regulation (Card et al., 2008). In orderto narrow our search for processes directly associated with germ celldevelopment, we applied the functional annotation clustering moduleand searched for groups of annotated targets that are connected withthis process. Supplementary, Fig. S5 depicts the resulting internalrelationships among the clustered germ cell development terms andmiR-302 predicted target genes.

miRNA promoter methylation dynamics indifferentiating ESCs and maGSCsFinally, we wanted to test whether differences between ESCs andmaGSCs in expression profiles of Oct-4 as well as miR-290 andmiR-302 clusters during differentiation are connected with the DNAmethylation status of the promoters of these genes (Fig. 7). Ingeneral, the tested promoters showed a strong hypomethylation inthe tested undifferentiated and differentiating ESCs and maGSCs.DNA methylation of Oct-4 and miR-290 promoters in maGSCs,although marginally higher than in undifferentiated ESCs, remainedunchanged or even decreased during differentiation, in contrast toESCs that increased methylation. These data suggest that, after 10days of differentiation, ESCs have already initiated the process of pro-moter silencing of pluripotency-related genes like Oct-4 and miR-290cluster. Although maGSCs depict a basal DNA methylation marginallyhigher than ESCs, do not seem to have initiated this process, indicating

Figure 4 Differences in expression levels of germ cell markers between ESC and maGSC cultures after in vitro differentiation associate with differ-ences in miR-302 cluster expression. Cells were cultured in absence of LIF and FLs (2LIF/2FL) for 10 days. Levels were determined by qRT–PCRand quantities were normalized to the endogenous control (Sdha or RNU6B for miRNAs) and calibrated with expression levels of testis (a, b) andmaGSCs (c), respectively. Asterisk depicts statistical significance for the comparison between ESCs and maGSCs. (a) Expression levels for the sper-matogonia marker Stra-8 and the premeiotic germ cell markers C-kit and Dazl. (b) Expression levels of the PGC marker Dppa-3. (c) Expression levelsof members of miR-302 cluster.

798 Zovoilis et al.

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

their need to retain miR-290 and Oct-4 expression. For the samereason, miR-302 cluster, which is up-regulated during first stages ofESC differentiation, did not show any increase of promoter DNAmethylation, in contrast to miR-290 cluster and Oct-4.

DiscussionIn this work, we provide for a first time a connection between germcell origin of maGSCs and their specific miRNA expression profile.Retained expression of miR-290 cluster has been previously reportedin maGSCs during differentiation and it has been associated withhigher Oct-4 levels. Here, we show that this property is also sharedby EGCs. In addition, in this work, we provide proof that the specificmiRNA expression profile of maGSCs has an impact on their differen-tiation potential. Low levels of miR-302 observed in cultures ofmaGSCs during the first 10 days of LIF deprivation were shown tobe associated with high Dppa-3 and Stra-8 levels. However, it islikely that other mechanisms may also contribute to this phenotype.Of course, under the cell culture system employed here cells (ESCs,maGSCs or EGCs) do not differentiate exclusively towards a specificlineage or with the same strict hierarchical pattern that blastocyst

cells do in vivo. Thus, the data presented here only indicate mRNAlevels in a complex population of cells. However, this systemremains invaluable for detecting general trends in differentiationpotential of each of the different cell lines used here.

The ability of undifferentiated maGSCs to retain high expression ofearly germ cell markers after reprogramming to the ESC-like state is ofparticular interest. Two recent reports show that pluripotent cellsoriginating from the germ line differ from ESCs in DNA methylationof imprinted genes (Ko et al., 2009; Zechner et al., 2009). In thecontext of these reports our finding suggests that the process ofreprogramming towards the ESC-like state does not completelyerase all properties of the cells from which they originate. This demon-strates one of the few differences observed until now between undif-ferentiated maGSCs and ESCs, which otherwise show similarexpression patterns in global gene expression array profiling (Meyeret al., unpublished, personal communication), western blotting forkey pluripotency markers (Guan et al., 2006; Zovoilis et al., 2008)as well as whole microRNAome expression profiles (Supplementary,Fig. S6).

Although ESCs and maGSCs do not differ significantly under cultureconditions that inhibit differentiation, during in vitro differentiation,

Figure 5 Schematic representation of the miRNA transfection strategy followed in this study and expression levels of Oct-4 and Dppa-3 after trans-fection. (a) Members of the miR-302 cluster were transfected (pos) during 10 days differentiation in the absence of LIF and FLs (2LIF/2FL). At thesame time cells cultured under exactly the same conditions were transfected with the respective Pre-miR controls (neg). (b) qRT–PCR for the plur-ipotency marker Oct-4 during transfection in differentiating ESCs and maGSCs (2LIF/2FL). Quantities were normalized to the endogenous control(Sdha) and calibrated with levels of undifferentiated (untreated) cells (+LIF/+FL). Asterisk indicates statistical significance for the comparison betweendifferentiated and undifferentiated cells. No statistically significant difference was observed between ‘pos’ and ‘neg’ in either cell type. (c) qRT–PCR forthe PGC marker Dppa-3 during transfection in differentiating maGSCs (2LIF/2FL). For normalization and calibration refer to subfigure (b). Oneasterisk indicates statistical significance for the comparison between ‘pos’ and ‘neg’. Two asterisks indicate statistical significance for the comparisonbetween differentiated (pos or neg) and undifferentiated cells (+LIF/+FL).

Differentiation of pluripotent cells 799

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

Oct-4 expression is maintained for a longer time and the levels ofdifferentiation markers increase more slowly in maGSC cultures thanin those of ESCs (Zovoilis et al., 2008). These remarkable propertiesof maGSCs can be partially explained by their specific expressionprofile of miR-290 cluster during differentiation. The miR-290 clusterhas been shown to maintain and induce pluripotency in multipleways, by controlling Wnt signalling, de novo DNA methylation aswell as cell cycle regulators (Sinkkonen et al., 2008; Wang et al.,2008; Zovoilis et al., 2009). Particularly, ESCs with artificiallyinduced abnormal high levels of miR-290 cluster have been shownto resemble maGSCs’ delayed differentiation under the same differen-tiation conditions used in the current study (Zovoilis et al., 2009).Members of miR-17-92 cluster, although not exclusively expressedin pluripotent cells, have been also shown to control proliferation ofESCs (Foshay and Gallicano, 2009) and could also contribute to theabove phenotype.

Our data are in strikingly consistent with findings of Hayashi et al.(2008) who tested miRNA expression patterns during PGC develop-ment and early spermatogenesis in vivo (Hayashi et al., 2008). In thatwork, miR-290 and miR-17-92 clusters as well as Oct-4 were foundto be constitutively highly expressed during germ cell development.

Although the differentiation environment in vitro differs significantlyfrom that in vivo, the similarities of the findings of Hayashi et al.(2008) to our findings are striking: spontaneous in vitro differentiationof maGSC cultures is characterized by higher levels of key PGC andgerm cell markers compared with ESCs and at the same timeexpression of the same miRNA clusters as in that work is maintainedat high levels. In addition, high levels of miR-290 cluster during differ-entiation of maGSCs are associated with remarkably high Oct-4 levels,a gene that in vivo continues to be expressed in stem cells of adulttestis (He et al., 2007). Moreover, not only maGSC cultures butalso EGC cultures were found to retain high levels of miR-290 and17/92 cluster members during first stages of in vitro differentiation inclose correlation with high Oct-4 levels. Thus, in the light of findingsof Hayashi et al., our data suggests that these miRNAs may play animportant role in the ability of stem cells of the germ line to maintainOct-4 expression during first stages of differentiation. This is furthersupported by our previous findings that maintaining artificially highlevels of miR-290 cluster in ESCs during differentiation preventstheir differentiation, and particularly it prevents expression of meso-derm markers (Zovoilis et al., 2009). The ability to suppress differen-tiation towards mesoderm is crucial for development of PGCs.

Figure 6 Functional annotation analysis of miR-302 cluster predicted targets. Gene ontology terms are based on PANTHER Database and refer toeither Biological Processes (Pie diagram) or Molecular Function (Table). Pie diagram depicts the percentage of biological processes found to beenriched among the identified by DAVID miR-302 target genes. Enrichment analysis was based on an EASE score 0.001 and resulted for all identifiedprocesses in a P ≤ 9.10E-04 and a Benjamini ≤ 1.30E-02. Table depicts top enriched terms for the same genes but concerning molecular function. Therespective P and Benjamini values are also presented. The exact parameter values used are listed in Materials and Methods section.

800 Zovoilis et al.

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

Previous works have suggested that the destiny of PGC-competentproximal epiblast cells is initially directed towards mesoderm lineageexhibiting expression of mesoderm genes like Brachyury. Thesegenes continue to increase in the neighbouring mesoderm somatictissue, but become repressed along with progression of specificationto PGCs (Hayashi et al., 2007). Interestingly, EGCs are alreadyknown not to preferentially differentiate in vitro towards cell types ofmesoderm lineage (Sharova et al., 2007), a fact that could be nowexplained by the high levels of miR-290 cluster that characterizesthese cells. Furthermore, as shown in Supplementary, Fig. S7, such aproperty applies also to maGSCs, in which Brachyury levels are signifi-cantly lower than in ESCs during differentiation, although for the ecto-derm marker Nestin such a difference is much less profound.

The inhibitory role of miR-290 cluster in maGSCs’ differentiationtowards different cell lineages could be ultimately confirmed by inhibit-ing these miRNAs in maGSCs. However, it was not possible toachieve such an effect in maGSCs and EGCs at non-toxic levels as

we had previously achieved it with ESCs (Zovoilis et al., 2009).Taking into account that these miRNAs consist of more than 70%of the total miRNA cell content in ESCs, these cells seem to beable to compensate for their loss during inhibition experiments at non-toxic levels. However, the importance of these miRNAs in germ celldevelopment is supported by a recently derived miR-290 deficientmouse line, which showed a role of these miRNAs in maintainingboth pluripotency and fertility. Particularly, viable miR-290 deficientmouse were found to be infertile (reviewed in Blakaj and Lin, 2008).

maGSC cultures have been shown in the current and a previous work(Zovoilis et al., 2008) to increase levels of differentiation markers for allthree germ layers later than ESC cultures. However, here we show thatthis does not apply to early germ cell markers that are significantlyup-regulated during differentiation in cultures of maGSCs. Forexample, as shown in Supplementary, Fig. S8, a flow cytometricapproach for the spermatogonia marker Stra-8, which quantified howmany of differentiating maGSCs actually activated Stra-8, revealed arange from 37 to 49% in all different mouse strains tested. Levels ofmiR-302 cluster, which share the same miRNA seed (and thustargets) with some miR-290 cluster members, could partially explainthis increase in germ cell markers in differentiating maGSCs. Duringdifferentiation, miR-302 cluster is up-regulated in maGSCs later thanin ESCs. Low levels of miR-302 cluster in maGSCs as compared withESCs are necessary for high Dppa-3 and Stra-8 levels in maGSCs,since increased miR-302 cluster levels were found to reduce Dppa-3and Stra-8 expression in maGSCs. Thus, this delay in up-regulation ofmiR-302 observed in maGSCs seems to be one of the underlying mech-anisms for the high early germ cell marker levels in maGSC culturesduring differentiation.

At the moment it remains unclear through which targets miR-302induces this effect. miR-302 cluster is predicted to target more than400 mRNAs. In addition, as noted above, miR-302 cluster shares atthe same time targets with many other miRNAs, so that applicationof massive parallel sequencing after HITS-CLIP is needed to elucidatecompletely the exact regulated networks. However, through func-tional annotation analysis of the predicted targets it was possible toshow here that miR-302 cluster acts more likely by influencing tran-scription through targeting of transcriptional factors. This is in agree-ment with a key role of miR-302 cluster during first stages ofdevelopment shown in the current work, since at that point the tran-scriptome is characterized by a dynamic switch on and off of genesrelated to differentiation or pluripotency, respectively. In addition, asshown in Supplementary, Fig. S5, predicted miR-302 cluster targetsinclude genes direct associated with germ cell development.

The most interesting conclusion that can be drawn from our findingsis that pluripotent cells have two miRNA clusters (miR-290 andmiR-302) that at least for some members share the same targetsbut are differentially expressed in different cell types during in vitrodifferentiation, as it is the case of ESCs and maGSCs. In this way,response of the different pluripotent cell types to different needs ordecisions (for example, differentiation towards germ cell lineage ornot) is co-ordinately orchestrated by the different expression patternsof miR-290 and miR-302 clusters in the different cell types. This isfurther supported by the different dynamics of DNA promotermethylation of these miRNA clusters that have been observed here,which confirmed that their expression is subjected to different tran-scriptional control in ESCs and maGSCs.

Figure 7 DNA methylation levels of (a) the Oct-4 5′ upstreamregion, (b) the miR-290 cluster pri-miRNA 5′ upstream region and(c) the miR-302 cluster pri-miRNA 5′ upstream region in undifferen-tiated (+LIF) and differentiating (2LIF) ESCs and maGSCs.

Differentiation of pluripotent cells 801

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

Supplementary dataSupplementary data are available at http://molehr.oxfordjournals.org/.

Authors’ rolesA.Z.: conception and design, collection and assembly of data, dataanalysis and interpretation, manuscript writing. A.P.: collection andassembly of data, data analysis, manuscript writing. L.S.: collectionand assembly of data, data analysis. L.O.: data analysis. G.S.R.,M.W., A.H.: collection and assembly of data. U.Z.: data analysis andinterpretation, manuscript writing, financial support. C.L.S.: provisionof study material, manuscript writing. W.E.: provision of studymaterial, manuscript writing, financial and administrative support.

AcknowledgementsWe would like to thank Sandra Meyer from the Institute of HumanGenetics of University of Goettingen for her support in the arrayexperiments.

FundingThis work was supported by the DFG SPP 1356 (EN 84/22-1; ZE442/4-1) and DFG FOR 1041 (EN 84/23-1; ZE 442/5-1).

ReferencesBlakaj A, Lin H. Piecing together the mosaic of early mammalian

development through microRNAs. J Biol Chem 2008;283:9505–9508.Card DA, Hebbar PB, Li L, Trotter KW, Komatsu Y, Mishina Y,

Archer TK. Oct4/Sox2-regulated miR-302 targets cyclin D1 in humanembryonic stem cells. Mol Cell Biol 2008;28:6426–6438.

Chen C, Ridzon D, Lee CT, Blake J, Sun Y, Strauss WM. Definingembryonic stem cell identity using differentiation-related microRNAsand their potential targets. Mamm Genome 2007;18:316–327.

Cheng J, Dutra A, Takesono A, Garrett-Beal L, Schwartzberg PL.Improved generation of C57BL/6J mouse embryonic stem cells in adefined serum-free media. Genesis 2004;39:100–104.

Conrad S, Renninger M, Hennenlotter J, Wiesner T, Just L, Bonin M,Aicher W, Buhring HJ, Mattheus U, Mack A et al. Generation ofpluripotent stem cells from adult human testis. Nature 2008;456:344–349.

Esquela-Kerscher A, Slack FJ. Oncomirs-microRNAs with a role in cancer.Nat Rev Cancer 2006;6:259–269.

Foshay KM, Gallicano GI. miR-17 family miRNAs are expressed duringearly mammalian development and regulate stem cell differentiation.Dev Biol 2009;326:431–443.

Guan K, Nayernia K, Maier LS, Wagner S, Dressel R, Lee JH, Nolte J,Wolf F, Li M, Engel W et al. Pluripotency of spermatogonial stemcells from adult mouse testis. Nature 2006;440:1199–1203.

Hayashi K, de Sousa Lopes SM, Surani MA. Germ cell specification in mice.Science 2007;316:394–396.

Hayashi K, Chuva de Sousa Lopes SM, Kaneda M, Tang F, Hajkova P,Lao K, O’Carroll D, Das PP, Tarakhovsky A, Miska EA et al.MicroRNA biogenesis is required for mouse primordial germ celldevelopment and spermatogenesis. PLoS ONE 2008;3:e1738.

He L, Hannon GJ. MicroRNAs: small RNAs with a big role in generegulation. Nat Rev Genet 2004;5:522–531.

He Z, Jiang J, Hofmann MC, Dym M. Gfra1 silencing in mousespermatogonial stem cells results in their differentiation via theinactivation of RET tyrosine kinase. Biol Reprod 2007;77:723–733.

Houbaviy H, Murray M, Sharp P. Embryonic stem cell-specific microRNAs.Dev Cell 2003;5:351–358.

Houbaviy H, Dennis L, Jaenisch R, Sharp PA. Characterization of a highlyvariable eutherian microRNA gene. RNA 2005;11:1245–1257.

Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysisof large gene lists using DAVID Bioinformatics Resources. Nature Protoc2009;4:44–57.

Izadyar F, Pau F, Marh J, Slepko N, Wang T, Gonzalez R, Ramos T,Howerton K, Sayre C, Silva F. Generation of multipotent cell linesfrom a distinct population of male germ line stem cells. Reproduction2008;135:771–784.

Kanatsu-Shinohara M, Inoue K, Lee J, Yoshimoto M, Ogonuki N, Miki H,Baba S, Kato T, Kazuki Y, Toyokuni S et al. Generation of pluripotentstem cells from neonatal mouse testis. Cell 2004;119:1001–1012.

Kanatsu-Shinohara M, Lee J, Inoue K, Ogonuki N, Miki H, Toyokuni S,ikawa M, Tomoyuki N, Ogura A, Shinohara T. Pluripotency of a singlespermatogonial stem cell in mice. Biol Reprod 2008;78:681–687.

Kim VN. MicroRNA biogenesis: coordinated cropping and dicing. Nat RevMol Cell Biol 2005;6:376–385.

Ko K, Tapia N, Wu G, Kim JB, Arauzo Bravo MJ, Sasse P, Glaser T,Ruau D, Han DW, Greber B et al. Induction of pluripotency in adultunipotent germline stem cells. Cell Stem Cell 2009;5:87–96.

Kossack N, Meneses J, Shefi S, Nguyen HN, Chavez S, Nicholas C,Gromoll J, Turek PJ, Reijo-Pera RA. Isolation and characterization ofpluripotent human spermatogonial stem cell-derived cells. Stem Cells2009;27:138–149.

Lacham-Kaplan O. In vivo and in vitro differentiation of male germ cells inmouse. Reproduction 2004;128:147–152.

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.

Mark M, Jacobs H, Oulad-Abdelghani M, Dennefeld C, Feret B, Vernet N,Codreanu CA, Chambon P, Ghyselinck NB. STRA8-deficientspermatocytes initiate, but fail to complete, meiosis and undergopremature chromosome condensation. J Cell Sci 2008;121:3233–3242.

Matsui Y, Zsebo K, Hogan BL. Derivation of pluripotential embryonic stem cellsfrom murine primordial germ cells in culture. Cell 1992;70:841–847.

Mendell JT. miRiad roles for the miR-17-92 cluster in development anddisease. Cell 2008;133:217–222.

Nagy A, Rossant J, Nagy R, Abramow-Newerly W, Roder JC. Derivationof completely cell culture-derived mice from early-passage embryonicstem cells. Proc Natl Acad Sci 1993;90:8424–8428.

Oulad-Abdelghani M et al. Characterization of a premeiotic germcell-specific cytoplasmic protein encoded by Stra8, a novel retinoicacid-responsive gene. J Cell Biol 1996;135:469–477.

Pavlidis P, Noble WS. Matrix2png: a utility for visualizing matrix data.Bioinformatics 2003;19:295–296.

Resnick JL, Bixler LS, Cheng L et al. Long-term proliferation of mouseprimordial germ cells in culture. Nature 1992;359:550–551.

Seandel M, James D, Shmelkov SV, Falciatori I, Kim J, Chavala S, Scherr DS,Zhang F, Torres R, Gale NW et al. Generation of functional multipotentadult stem cells from GPR125+ germline progenitors. Nature 2007;449:346–350.

Sharova LV, Sharov AA, Piao Y, Shaik N, Sullivan T, Stewart CL, Hogan BL,Ko MS. Global gene expression profiling reveals similarities anddifferences among mouse pluripotent stem cells of different originsand strains. Dev Biol 2007;307:446–459.

802 Zovoilis et al.

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from

Sinkkonen L, Hugenschmidt T, Berninger P, Gaidatzis D, Mohn F,Artus-Revel CG, Zavolan M, Svoboda P, Filipowicz W. MicroRNAscontrol de novo DNA methylation through regulation oftranscriptional repressors in mouse embryonic stem cells. Nat StructMol Biol 2008;15:259–267.

Stewart CL, Gadi I, Bhatt H. Stem cells from primordial germ cells canreenter the germ line. Dev Biol 1994;161:626–628.

Strauss W, Chen C, Lee CT, Ridzon D. Nonrestrictive developementalregulation of microRNA gene expression. Mamm Genome 2006;17:833–840.

Suh MR, Lee Y, Kim JY, Kim SK, Moon SH, Lee JY, Cha KY, Chung HM,Yoon HS, Moon SY et al. Human embryonic stem cells express aunique set of microRNAs. Dev Biol 2004;270:488–498.

Tang F, Kaneda M, O’Carroll D, Hajkova P, Barton SC, Sun YA, Lee C,Tarakhovsky A, Lao K, Surani MA. Maternal microRNAs are essentialfor mouse zygotic developement. Genes Dev 2007;21:644–648.

Tegelenbosch RA, de Rooij DG. A quantitative study of spermatogonialmultiplication and stem cell renewal in the C3H/101 F1 hybridmouse. Mutat Res 1993;290:193–200.

Wang Y, Baskerville S, Shenoy A, Babiarz JE, Baehner L, Blelloch R.Embryonic stem cell-specific microRNAs regulate the G1-Stransition and promote rapid proliferation. Nat Genet 2008;40:1478–1483.

Yabuta Y, Kurimoto K, Ohinata Y, Seki Y, Saitou M. Gene expressiondynamics during germ line specification in mice identified byquantitative single-cell gene expression profiling. Biol Reprod 2006;75:705–716.

Zechner U, Nolte J, Wolf M, Shirneshan K, Hajj NE, Weise D, Kaltwasser B,Zovoilis A, Haaf T, Engel W. Comparative methylation profiles andtelomerase biology of mouse multipotent adult germline stem cells andembryonic stem cells. Mol Hum Reprod 2009;15:345–353.

Zovoilis A, Nolte J, Drusenheimer N, Zechner U, Hada H, Guan K,Hasenfuss G, Nayernia K, Engel W. Multipotent adult germline stemcells and embryonic stem cells have similar microRNA profiles. MolHum Reprod 2008;14:521–529.

Zovoilis A, Smorag L, Pantazi A, Engel W. Members of the miR-290 clustermodulate in vitro differentiation of mouse embryonic stem cells.Differentiation 2009;78:69–78.

Differentiation of pluripotent cells 803

at Niedersaechsische Staats- u. U

niversitaetsbibliothek Goettingen on January 30, 2014

http://molehr.oxfordjournals.org/

Dow

nloaded from