Stem Cells

Molecular signature of adult bone marrow-purified very small embryonic-like stem cells supports their developmental epiblast/germ line origin


We postulated that Oct4+SSEA-1+Sca-1+LinCD45 very small embryonic-like stem cells (VSELs) isolated from adult bone marrow (BM) could be a reserve population for tissue-committed stem cells. The aim of this study was to elucidate the developmental origin of these cells. We report that during embryogenesis, VSELs are enriched in embryonic day (E)12.5 murine fetal livers (FLs) and subsequently follow the developmental route of hematopoietic stem cells (H)SCs to colonize BM. Molecular analysis of purified VSELs revealed that both FL-derived VSELs and their adult BM-derived counterparts express: (i) several epiblast/primordial germ cell (PGC) markers; (ii) migrating PGC-like epigenetic reprogramming profiles of Oct4, Nanog and Stella loci; as well as (iii) a unique pattern of genomic imprinting. Thus, these data suggest that VSELs may originate from epiblast/migrating PGC-like cells and, in spite of the expression of pluripotent stem cell markers, changes in the epigenetic signature of imprinted genes keep these cells quiescent in adult tissues and prevent them from teratoma formation.


The rapidly developing field of regenerative medicine is searching for safe and therapeutically efficient sources of pluripotent stem cells (PSCs). By definition, PSCs should: (i) give rise to cells from all three germ layers; (ii) complete blastocyst development; and (iii) form teratomas after inoculation into experimental animals.1 Unfortunately, in contrast to immortalized embryonic (E)SC lines2 or inducible (i)PSCs,3 these last two criteria have not been obtained thus far with any potential PSC candidates isolated from adult tissues. There are two potential explanations for this discrepancy. The first is that PSCs isolated from adult tissues are not fully pluripotent; the second is that there are some physiological mechanisms involved in keeping these cells quiescent in adult tissues to preclude their unleashed proliferation and risk of teratoma formation.

Recently, our group isolated a population of Sca-1+LinCD45 very small embryonic-like stem cells (VSELs) from adult murine bone marrow (BM) and murine fetal livers (FLs) that express several morphological (for example, relatively large nuclei containing euchromatin) and molecular (for example, expression of SSEA-1, Oct4, Nanog and Rex1) markers characteristic for ESCs.4, 5 The true expression of Oct4 and Nanog in BM-derived VSELs (BM-VSELs) was recently confirmed by showing transcriptionally active chromatin structures of Oct4 and Nanog promoters.6 In the case of BM-VSELs, we also described a mechanism based on parent-of-origin-specific reprogramming of genomic imprinting that keeps VSELs quiescent in a dormant state in tissues.6 We also hypothesized that the same mechanism prevents VSELs from blastocyst complementation and teratoma formation,6 but this requires further experimental evidence. Genomic imprinting is an epigenetic mechanism that has a crucial role in early embryogenesis.7 The proper pattern of genomic imprints is maintained in all adult somatic cells. However, genomic imprinting is reprogrammed early in embryogenesis in migrating Oct4+ primordial germ cells (PGCs) to prevent them from unleashed proliferation.8 Thus, the expression of germ line markers (Oct4 and SSEA-1) and modulation of somatic imprints suggest a potential developmental similarity between VSELs and germ line-derived PGCs.

Around embryonic day (E)7.25 in mice, some proximal epiblast-residing PSCs are specified to PGCs and egress from the epiblast into extra-embryonic tissues.9 Subsequently, through the primitive streak, they return to the embryo proper and migrate to genital ridges, in which they ultimately give rise to gametes. Interestingly, it was also postulated that PGCs could give rise to developmentally early hematopoietic (H)SCs.10 To support this notion, the first primitive HSCs seem in the extra-embryonic tissues in yolk sac blood islands at a time when proximal epiblast-specified PGCs enter the extra-embryonic mesoderm.11 Furthermore, the appearance of definitive HSCs in the aorta-gonad-mesonephros region in the embryo proper corresponds timewise to migration of PGCs to the genital ridges through the aorta-gonad-mesonephros. Therefore, this spatial and temporal developmental overlap between PGCs and HSCs suggests a developmental relationship between them.12 Recently, we also postulated that FL-derived VSELs (FL-VSELs) follow the developmental migratory pathway of HSCs.5 However, the relationship between FL- and BM-residing VSELs and epiblast (Epi)SCs and PGCs has not been determined.

Thus, the aim of this study was to elucidate the developmental origin of VSELs and study their potential relationship to epiblast-derived PGCs. To address this, we isolated VSELs from adult BM and embryonic tissues, studied their molecular signature, and evaluated their developmental potential in vitro. Our data support the hypothesis that a population of migratory PSCs is specified early during embryogenesis in the proximal epiblast that gives rise to PGCs and related VSELs. The VSELs are subsequently deposited in developing organs as precursors for tissue-committed (TC)SCs; for example, BM-deposited VSELs may differentiate into HSCs.

Materials and methods

Animals and preparation of embryo, FL and BM cells for fluorescence-activated cell sorting

This study was performed in accordance with the guidelines of the Animal Care and Use Committee of the University of Louisville School of Medicine and with the Guide for the Care and Use of Laboratory Animals (Department of Health and Human Services, Publication No. NIH 86–23). Murine FLs, FL-depleted embryo (w/o FL) tissues (E11.5–E14.5), and BM mononuclear cells (BMMNCs; 4–5 weeks) were isolated from pathogen-free C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME, USA). MNCs from each tissue without Ficoll–Hypaque separation were lysed in BD lysing buffer (BD Biosciences, San Jose, CA, USA) for 15 min in room temperature and washed twice in phosphate-buffered saline. FLs from 10 to 15 fetuses were combined in each experiment. Tissues were mechanically fragmented and released cells were washed and filtered through a 40 FL-depleted embryo (w/o FL) tissues were mechanically fragmented and dissociated using collagenases I–IV (Worthington Biochemical Corp., Lakewood, NJ, USA). VSELs (Sca-1+LinCD45) and HSCs (Sca-1+LinCD45+) were isolated from each tissue by multiparameter live cell sorting (FACSVantage SE or MoFlo), as previously described.4

Expansion and VSEL-derived sphere formation culture

Freshly sorted Sca-1+LinCD45 (VSELs) from FL or FL-depleted embryo (w/o FL) cells were cultured over a C2C12 murine myoblast feeder layer seeded on a 22 mm glass bottom plate (World Precision Instruments, Sarasota, FL, USA). Cells were cultured in medium containing a low percentage of serum (Dulbecco's modified Eagle's medium with 2% fetal bovine serum, Molecular Probes, Invitrogen, Carlsbad, CA, USA) without any supplementing recombinant growth factors. VSEL-derived sphere formation was estimated after 9 days of culture by counting and Oct4 staining.

Ex vivo differentiation of VSELs into hematopoietic cells in co-cultures with OP9 cells

Freshly sorted Sca-1+LinCD45 (VSELs) and Sca-1+LinCD45+ (HSCs) from FLs or FL-depleted embryos (w/o FL) were plated over OP9 cells in α modified Eagle's medium with 20% fetal bovine serum (Molecular Probes, Invitrogen) for 5 days and subsequently trypsinized and plated in methylocellulose-based medium (StemCell Tech, Vancouver, BC, Canada) supplemented with murine stem cell growth factor, interleukin-3 and granulocyte-macrophage colony-stimulating factor. Colony-forming units were estimated after 15 days of culture by counting and the property of HSCs was confirmed by surface CD45 (30-F11, phycoerythrin-conjugated, rat immunoglobulin (Ig)G2b; BD Biosciences) staining.

Cell culture

Murine ESC-D3 cells (ATCC, Manassas, VA, USA) were grown in Dulbecco's modified Eagle's medium (GIBCO, Invitrogen, Carlsbad, CA, USA) containing 4 mM L-glutamine, 1.5 g/l sodium bicarbonate, 4.5 g/l glucose, 0.1 mM β-mercaptoethanol (Sigma, St Louis, MO, USA), 15% heat-inactivated fetal bovine serum (GIBCO), 100 IU/ml penicillin, 100 μg/ml streptomycin (Invitrogen) and 5 ng/ml of recombinant mouse leukemia-inhibitory factor (Chemicon-Millipore, Temecula, CA, USA) without a feeder layer. Embryoid body formation was performed by the hanging drop method as previously described.6

Reverse transcriptase- and quantitative real-time polymerase chain reaction (RQ-PCR)

Total RNA from various cells (approximately 20 000 cells) was isolated using the RNeasy Mini Kit (Qiagen Inc., Valencia, CA, USA) with removal of genomic DNA using the DNA-free Kit (Applied Biosystems, Foster City, CA, USA). Isolated messenger (m)RNA was reverse-transcribed with Taqman Reverse Transcription Reagents (Applied Biosystems) according to the manufacturer's instructions. Reverse transcriptase-PCR was performed using Amplitaq Gold (Applied Biosystems) at 1 cycle of 8 min at 95 °C, 2 cycles of 2 min at 95 °C, 1 min at 62 °C and 1 min at 72 °C, subsequent 38 cycles of 30 sec at 95 °C, 1 min at 62 °C, 1 min at 72 °C and 1 cycle of 10 min at 72 °C by using sequence-specific primers. Quantitative measurement of target transcript expression was performed by RQ-PCR using an ABI Prism 7500 Sequence Detection System (Applied Biosystems). Complementary DNA from indicated cells was amplified using SYBR Green PCR Master Mix (Applied Biosystems) and specific primers. All primers were designed with Primer Express software (Applied Biosystems), as at least one primer included an exon–intron boundary. The threshold cycle (Ct) was determined and relative quantification of the expression level of target genes was performed with the 2ΔΔCt method, using that of β2-microglobulin (β2 mg) as an endogenous control gene and that of MNCs as a calibrator. All primers used in reverse transcriptase- and RQ-PCR are available on request.

Bisulfite sequencing

Exactly 100 ng of genomic DNA prepared from indicated cells were used in bisulfite modification using the EpiTect Bisulfite Kit (Qiagen Inc.) according to the manufacturer's instructions. The promoters of stemness (Oct4 and Nanog), germ line genes (Stella, Mvh, Dazl and Sycp3), repetitive sequences (LINE1 and IAP) and differentially methylated regions (DMRs) of imprinted genes were amplified by nested PCR using bisulfite-treated genomic DNA and specific primers as previously described.6 The methylation pattern in DMRs was analyzed using CpGViewer software (University of Leeds, UK). All experiments were conducted with three independent isolations of all cell populations and two independent PCRs of each isolated cell. All primers used in bisulfite sequencing are available on request.

Chromatin immunoprecipitation (ChIP)

ChIP analysis was performed using THP-1 cells as a source of carrier chromatin as previously described.6 The ChIP assay was performed using the Magna ChIP G kit (Upstate-Millipore, Billerica, MA, USA) according to the manufacturer's instructions. In all, 5 × 106 THP-1 cells were mixed with 2 × 104 freshly isolated VSELs, HSCs, MNCs, ESC-D3s and embryoid body-derived cells in culture media. After fixing with 1% formaldehyde and applying sonication, IP was performed using 3 μg of ChIP grade antibodies against H3Ac (Upstate-Millipore), H3K9me2 (Abcam, Cambridge, MA, USA), H3K4me3 (Abcam), H3K27me3 (Upstate-Millipore) or rabbit IgG control antibodies (Sigma). The bound and unbound sheared cross-linked chromatins were used as regular PCR or quantitative PCR reactions. Regular PCR for ChIP products was performed as previously described.6 To quantify the enrichment of each histone modification, the copy number of bound or unbound PCR products was measured by the absolute quantification method using an ABI Prism 7500 Sequence Detection System (Applied Biosystems). The enrichment of each histone modification was calculated as the ratio of amplicon amounts from bound to unbound fractions; fold differences are shown as mean±s.d. from at least four independent experiments. All PCR products for these ChIP primers were subsequently sequenced to rule out the possibility of amplification of pseudogenes or nonspecific sequences (data not shown). The previously published primer set was used in ChIP experiments for the Oct4, Nanog6 and Stella13 promoters.


Immunocytochemistry was performed as previously described4 for the following proteins: Oct4 (clone 9E3.2, mouse monoclonal IgG1, Millipore); Nanog (W-18, goat polyclonal IgG, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA); SSEA-1 (clone MC-480, mouse monoclonal IgM, Millipore); Stella (M-150, rabbit polyclonal IgG, Santa Cruz Biotechnology Inc.); Mvh (Abcam); and Blimp1 (3H2-E8, mouse monoclonal IgG1, Novus Biologicals, Littleton, CO, USA).

Statistical analysis

All data in quantitative ChIP and gene expression analyses were analyzed using one factor analysis of variance with Bonferroni's Multiple Comparison Test. We used the Instat1.14 program (GraphPad Software, La Jolla, CA, USA) and statistical significance was defined as P<0.05 or P<0.01.


VSELs share some markers with epiblast and epiblast-specified PGCs

The dynamic change of gene expression and epigenetic profiles in early embryogenesis is essential for proper development of SCs. We hypothesized that VSELs could be epiblast-derived precursors of TCSCs1 and, as those first focused on expression in adult BM-VSELs, of genes that are characteristic for EpiSCs.13, 14 Figure 1 and Supplementary Figure 1 show the expression levels of mRNA for epiblast-related genes in Sca-1+LinCD45 VSELs and Sca-1+LinCD45+ HSCs purified by fluorescence-activated cell sorting as well as BMMNCs and ESC-D3 cells from the established ESC cell line. We noticed that VSELs express Oct4, Nanog, Sox2 and Klf4, which all support their pluripotent character. The expression levels of transcripts of Oct4 and Nanog in VSELs were around 50 and 20%, respectively, compared with ESC-D3 cells (Figure 1a). In contrast, VSELs expressed a similar level of Sox2 transcript and 3.5 times more Klf4 as compared with ESC-D3 cells.

Figure 1

Adult BM-VSELs share some markers with epiblast-like PSCs. RQ-PCR analysis of stemness genes (Oct4, Nanog, Sox2 and Klf4; a) and EpiSC markers (Gbx2, Fgf5, Nodal and Rex1; b). It is noteworthy that VSELs express a considerable amount of all these genes, which is specific to early embryonic PSCs. The relative expression level is represented as the fold difference to the value of BMMNCs and shown as the mean±s.d. from at least four independent experiments on different samples of murine adult BM-isolated VSELs, HSCs, BMMNCs and ESC-D3. *P<0.05, **P<0.01 compared with BMMNCs.

Next, we examined the expression levels of Gbx2, Fgf5 and Nodal. As reported, these are upregulated in EpiSCs, but expressed at lower levels in ESCs isolated from the inner cell mass of pre-implantation blastocysts. In contrast, the level of Rex1/Zfp42 transcripts is highly expressed in inner cell mass cells.13, 14 Again, we found that VSELs highly express Gbx2, Fgf5 and Nodal, but express less Rex1/Zfp42 transcript as compared with ESC-D3s (Figure 1b). Thus, our RQ-PCR analysis generally supports the notion that VSELs are more differentiated than inner cell mass-derived ESCs and share several markers with more differentiated EpiSCs.

After gastrulation, epiblast-derived somatic cells lose the expression of pluripotency core factors (for example, Oct4, Nanog and Sox2); however, proximal epiblast-derived PGCs continue to express them to keep their pluripotency.8 Previously, we reported that murine Oct4+ BM-VSELs show some reprogramming of genomic imprinting, which is specific for PGCs.6 Thus, we hypothesize that VSELs could be developmentally related to epiblast-derived PGCs. The specification of PGCs is tightly regulated by the sequential expression of germ line master regulators such as Fragilis, Blimp1 and Stella in response to signals delivered from extra-embryonic tissues by bone morphogenetic proteins.15, 16 As reported, the Stella promoter is first methylated during development of inner cell mass blastocyst ESCs to post-implantation embryo EpiSCs and, subsequently, is again demethylated to be expressed during the specification of PGCs.13, 17 Thus, to analyze the relationship between VSELs and epiblast-derived PGCs, we used bisulfite sequencing and examined the DNA methylation status of this promoter in VSELs, HSCs, BMMNCs, ESC-D3s, and cells isolated from ESC-D3-derived early stage (1-day) embryoid bodies. We found that VSELs show partial demethylation of this locus (Figure 2a). Next, we focused on the expression of genes involved in the germ line specification of the epiblast (for example, Stella, Prdm14, Fragilis, Blimp1, Nanos3 and Dnd1).16 By using RQ-PCR, we noticed that VSELs highly expressed all the genes involved in germ line specification from the epiblast (Figure 2b). Subsequently, we confirmed the expression of Stella, Blimp1 and Mvh in purified VSELs at the protein level by immunostaining (Figure 2c). It is noteworthy that Oct4 and Stella proteins were not detected in BM-purified HSCs (Supplementary Figure 2).

Figure 2

The expression and epigenetic status of Stella locus in VSELs. (a) Bisulfite sequencing results of DNA methylation of the Stella promoter in murine adult BM-isolated VSELs, HSCs, BMMNCs, ESC-D3, and cells from ESC-D3-derived 1-day (1d) embryoid bodies (EBs). Methylated and unmethylated CpG sites are shown in filled and open circles, respectively. The numbers under the bisulfite sequencing profiles indicate the percentage of methylated CpG sites. (b) RQ-PCR analysis of Stella, Blimp1, Prdm14 and Nanos3 in indicated cells. The relative expression level is represented as the fold difference to the value of BMMNCs and shown as the mean±s.d.; n=4; *P<0.05, **P<0.01 compared with BMMNCs. (c) Immunostaining of Oct4, Stella, Blimp1 and Mvh protein in VSELs. Nucleus was stained with 4,6-diamidino-2-phenylindole (DAPI) blue. BF, bright field. Regular ChIP (d) or quantitative (q)ChIP (e, f) analyses for the Stella promoter for enrichment of transcriptionally active (H3Ac and H3K4me3; e) or repressive (H3K9me2 and H3K27me3; f) histone modifications in the indicated cells. The enrichment of each histone code is represented as fold differences to the value of ESC-D3s and shown as the mean±s.d.; n=4; *P<0.05, **P<0.01 compared with BMMNCs.

To confirm expression of Stella in VSELs (Figures 2b and c), we evaluated histone modifications of this locus by using a modified ChIP assay. Our ChIP results show that the Stella promoter in VSELs shows transcriptionally active histone modifications (acetylated-histone3 (H3Ac) and trimethylated-lysine-4 of histone3 (H3K4me3)) and was less enriched for transcriptionally repressive histone markers (dimethylated-lysine-9 of histone3 (H3K9me2) and trimethylated-lysine-27 of histone3 (H3K27me3); Figures 2d, e, f, and Supplementary Figure 3). Thus, our results show that VSELs express: (i) genes and (ii) show a Stella promoter chromatin structure that is characteristic for germ line specification.

VSELs share some markers with migratory PGCs, but not post-migratory PGCs

PGCs that complete specification in extraembryonic tissues migrate back into the embryo proper to colonize the genital ridge. During migration, PGCs initiate epigenetic reprogramming such as the erasure of genomic imprinting, DNA demethylation and X chromosome reactivation.8, 15, 16, 17, 18 Such epigenetic reprogramming could have an important role in tuning the timely expression of genes involved in germ cell development such as Stella, Blimp1, Mvh, Dazl and Sycp3.8 For example, during PGC specification, the Stella promoter becomes demethylated, which reactivates its transcriptional activity in early PGCs, and the DNA methyltransferase 1 (Dnmt1) mutant embryos show premature expression of post-migratory markers (for example, Sycp3 and Dazl).17, 18 Overall, epigenetic reprogramming in PGCs is a dynamic process initiated early during their migration (E8.5) and completed after colonizing genital ridges.19 As a result of these epigenetic changes, early specialized, migratory and post-migratory PGCs show distinguished gene expression and epigenetic signatures.8, 18, 19, 20, 21 When we focused on expression of genes specific to various PGC developmental stages in addition to genes involved in PGC specification (Figures 2b and c), VSELs highly express Dppa2, Dppa4 and Mvh, which characterize late migratory PGCs (E10.5–11.5)18, 20 (Figure 3a).

Figure 3

Similar gene expression and epigenetic profiles between VSELs and migratory PGCs. (a) RQ-PCR analysis of Dppa2, Dppa4 and Mvh. Bisulfite sequencing results of DNA methylation of the Mvh promoter (b) and LINE1-5′LTR (c) in the indicated cells. The percentage of methylated CpG sites was shown under the bisulfite sequencing results. RQ-PCR analysis of post-migratory PGC markers (Sycp3 and Dazl, d) and LINE1 (e). The relative expression level is represented as the fold difference to the value of BMMNCs and shown as the mean±s.d.; n=4; *P<0.05, **P<0.01 compared with BMMNCs.

Thus, to further confirm any similarity between VSELs and migratory PGCs, we checked DNA methylation for repetitive DNA sequences (LINE1, IAP) and promoters characterizing post-migratory PGC-expressed genes (Mvh, Dazl and Sycp3). DNA demethylation of these genes is initiated in late migratory PGCs (E10.5) and completed after arrival to the genital ridge (>E13.5).18, 21 Thus, late migratory PGCs (E10.5–E11.5) show partial DNA demethlylation levels of IAP (74%), LINE1 (65%), Mvh and Sycp3 and dynamic demethylation is first detectable in a population of post-migratory PGCs (E13.5).18, 21 In our bisulfite sequencing results, VSELs, similarly to previously reported data on late migratory PGCs,18, 21 show partial demethylation of the Mvh promoter (80%, Figure 3b), Sycp3 (82.1%, Supplementary Figure 4a) and LINE1-5′LTR (65%, Figure 3c). In contrast, the Dazl promoter and IAP sequence in VSELs were fully methylated (Supplementary Figure 4b and c). As expected, while other adult cells (HSCs and BMMNCs) showed the complete methylation of all post-migratory PGC-related gene promoters and repetitive sequences, cells from early embryoid bodies showed the lowest level of methylation.

Finally, we evaluated expression of Sycp3, Dazl and LINE1 genes that are highly expressed in post-migratory PGCs.18, 21 As shown in Figures 3d and e, VSELs do not express significant levels of these post-migratory PGC markers. Thus, our results suggest that VSELs deposited into murine BM show some similarities in gene expression and epigenetic signatures to epiblast-derived migratory PGCs (E10.5–E11.5).

Molecular characterization of VSELs isolated from second trimester embryos supports their epiblast/migratory PGC origin

After completing the previously described evaluations, we were interested in the molecular signature of VSELs isolated from murine embryos. As VSELs that reside in adult BM were identified as small Sca-1+CD45Lin cells4 and Sca-1 antigen is expressed in E11.5 developing embryos,11 we used E11.5–E14.5 embryos in our experiments. VSELs were purified separately from FL and FL-depleted embryonic tissues.

As with their BM-derived counterparts, we noticed that Sca-1+LinCD45 FL-VSELs express a similar pattern of pluripotent and epiblast/germ line genes at the mRNA and protein levels (Figure 4a and Supplementary Figure 5). Accordingly, the Oct4 gene was highly expressed in FL-VSELs at the mRNA (Figure 4a) and protein levels (Figure 5b). More importantly and similarly to BM-VSELs, the Oct4 promoter was hypomethylated (Figure 4b). The open transcriptionally active structure of Oct4 promoter in FL-VSELs was subsequently confirmed by a high ratio of H3Ac/H3K9me2 histone codes (Figure 4c). In addition, FL-VSELs also express Nanog and Stella (Figures 4a and 5b), which was further supported by: (i) the hypomethylation status of promoters and (ii) the H3Ac/H3K9me2 ratio characteristic for open, transcriptionally active chromatin (Supplementary Figure 6). It is noteworthy that we also noticed that FL-VSELs, as compared with BM-VSELs, show the more embryonic-like DNA methylation status of the Oct4, Nanog and Stella promoters (Supplementary Figure 6 and 7). On the basis of PGCs only showing open chromatin structures and expression of these stemness genes after gastrulation, we found that FL-VSELs show a PGC-like molecular signature similar to BM-VSELs.

Figure 4

The expression and epigenetic profiles of FL-VSELs. (a) RQ-PCR analysis of pluripotency (Oct4 and Nanog), EpiSCs (Gbx2) and germ line (Stella) markers in the indicated cells isolated from murine FLs (E11.5–E12.5) by fluorescence-activated cell sorting (FACS). The relative expression level is represented as the fold difference to the value of FL-MNCs and shown as the mean±s.d.; n=4; *P<0.05, **P<0.01 compared with FL-MNCs. (b) Bisulfite sequencing results of DNA methylation of the Oct4 promoter in indicated cells. (c) Quantitative (q)ChIP analyses for the Oct4 promoter in indicated cells to evaluate its association with H3Ac and H3K9me2 histone codes. Fold differences are shown as the mean±s.d.; n=4; *P<0.05, **P<0.01 compared with FL-MNCs. (d) Bisulfite sequencing profiles of DNA methylation of paternally methylated Igf2-H19 and Rasgrf1 DMRs in the indicated cells. The percentage of methylated CpG sites was shown by using bisulfite modification and sequencing results.

Figure 5

The functional assays for VSEL populations from FLs and other embryonic tissues. (a) RQ-PCR analysis of Oct4, Nanog, Sox2 and Stella in VSEL populations from different embryonic days (E11.5–E14.5), FLs, and embryonic tissue without FLs (w/o FL). VSELs were isolated by fluorescence-activated cell sorting (FACS) as (Sca-1+LinCD45: SL-CD45) together with the HSC fraction (Sca-1+LinCD45+: SL-CD45+) by FACS. The relative expression level for each gene is represented as the percentage of that for the housekeeping gene, β2 mg, and shown as the mean±s.d.; n=4; *P<0.05, **P<0.01 compared with MNCs for each tissue. Immunostaining (b) and frequency (c) of cells expressing Oct4, Stella (red), Stella and SSEA-1 (green) proteins in VSEL populations (FACS isolated as Sca-1+LinCD45) from both E12.5 FLs and FL-depleted embryo tissues (w/o FL). Nucleus was stained with 4,6-diamidino-2-phenylindole (DAPI) blue. The Oct4 protein staining (d) and number of colonies (e) of embryoid body (EB)-like spheres formed by the C2C12 myoblast co-culture system using a VSEL population (SL-CD45) prepared from E11.5 FLs. ND, not determined. (f) Growth of hematopoietic colonies in secondary methylocellulose culture from cobblestone area cells formed from co-culture over OP9 feeder cells using VSEL (SL-CD45) or HSC (SL-CD45+) populations prepared from E13.5 FLs. The property of HSCs was confirmed by surface CD45 staining. BF, bright field.

Next, we examined the DNA methylation status of selected imprinted genes in FL-VSELs. Again, similarly to PGCs,8, 22 FL-VSELs show the hypomethylation of DMRs for both paternally methylated (insulin-like growth factor-2 (Igf2)-H19 and Rasgrf1) and maternally methylated (Kcnq1 and Igf2R) loci (Figure 4d and Supplementary Figure 8). The degree of reprogramming for genomic imprinting was different on each locus and paternally DMRs tended to be more sensitive to the DNA demethylation. This status of genomic imprints in FL-VSELs does not correspond exactly to that observed in BM-VSELs, which show the parent-of-origin-specific methylation pattern exhibited by erasure of paternally methylated loci, but hypermethylation of maternally methylated ones (Figure 4d and Supplementary Figure 8). This would indicate that VSELs deposited in the adult BM microenvironment later in development increase DNA methylation of maternal imprints, which increases their quiescence. Thus, our gene expression and epigenetic studies show that VSELs sharing several migratory PGC-like markers colonize FLs early in embryogenesis.

To shed more light on the distribution of VSELs in FLs versus other embryonic tissues, we compared E11.5–E14.5 FL-derived Sca-1+LinCD45 VSELs with Sca-1+LinCD45 cells isolated from FL-depleted embryos (embryo w/o FL). We noticed that E12.5 FL-derived Sca-1+LinCD45 cells express Oct4, Nanog, Sox2 and Stella at the highest levels (Figure 5a). This points to the temporal link between the emergence of VSELs in FLs and the migration of PGCs to genital ridges through the aorta-gonad-mesonephros region.23 Furthermore, Sca-1+LinCD45 cells isolated from E12.5 embryonic tissue without FLs (w/o FL) were also found to highly express Nanog, Sox2 and Stella transcripts. However, in contrast to FL-VSELs, the Oct4 transcript was expressed at a much lower level (Figure 5a), indicating that these cells could represent other SC populations (for example, post-migratory PGCs, neural crest-derived SCs, or some other primitive TCSCs). The expression of these developmentally early genes was subsequently confirmed in FLs and FL-depleted embryos by cytochemical staining (Figure 5b). In agreement with our RQ-PCR results, the frequency of Oct4+Stella+ VSELs was much higher in FLs than other embryonic tissues (Figure 5c).

To compare the developmental potential of Sca-1+LinCD45 cells from FLs and FL-depleted embryos, we used two different co-culture systems using myoblastic C2C12- and OP9-cell feeder layers to evaluate sphere formation and HSC differentiation of these cells, respectively. Accordingly, however, freshly isolated VSELs show little capacity to proliferate in vitro in co-cultures over C2C12. They are able to form VSEL-derived spheres, which contain primitive cells that regain proliferative potential and give rise to cells from all three germ layers in secondary tissue-specific culture conditions (data not shown). Thus, we routinely use this assay to evaluate the pluripotency of VSELs. As shown in Figures 5d and e, E11.5 FL-derived Sca-1+LinCD45 cells, in contrast to those derived from FL-depleted embryos, produced Oct4+ VSEL-derived spheres in co-cultures over C2C12 cells similarly to BM-VSELs, indicating their similar differentiation potential. In another in vitro assay based on the co-culture of VSELs over the OP9 feeder layer, we also evaluated their hematopoietic differentiation capacity.24 We noticed that FL-derived Sca-1+LinCD45 VSELs, similarly to BM-derived ones, formed cobblestone areas enriched for CD45+CD41+ cells (Figure 5f). After replating, these are able to grow colony-forming units for granulocytes and macrophages in secondary methylocellulose cultures. However, this phenomenon was not reproduced with Sca-1+LinCD45 cells isolated from FL-depleted embryos. Thus, FL-derived Sca-1+LinCD45 VSELs functionally resemble their BM-derived counterparts and our data support that the FL is a developmentally important niche during embryogenesis for VSELs, which follow the migratory route of HSCs to the BM.5


We recently postulated that a population of primitive Oct4+ VSELs is deposited early during embryonic development in developing organs as a potential reserve pool of precursors for TCSCs and that this population has an important role in tissue rejuvenation and regeneration. Data presented in this report support the concept that VSELs could in fact originate from highly migratory epiblast/germ line-like SCs, colonize FLs during embryogenesis and subsequently follow the migratory route of HSCs and colonize adult BM (Figure 6a). Owing to this unique developmental origin, VSELs show characteristic epigenetic reprogramming and gene expression in stemness-, germ line- and imprinted-genes (Figure 6b and Supplementary Figure 9) that maintain their pluripotency, but also prevent their unleashed proliferation and teratoma formation.

Figure 6

Proposed developmental origin and epigenetic profiles of VSEL population. (a) Plausible explanation for developmental origin of VSELs deposited in adult tissues. Migratory PGCs-like PSCs become specified in proximal epiblast and migrate first to extra-embryonic tissues (left panel). Some of them could potentially imitate primitive hematopoiesis in yolk sac blood islands (?). Subsequently, these cells, known as PGCs, return through the primitive streak into the developing embryo proper and migrate toward genital ridges. Some of these cells (VSELs?) go astray from the main migratory route and seed around the aorta in the aorta-gonad-mesonephros (AGM) region (middle panel). Further studies are needed to determine whether these cells may contribute to definitive HSCs. Finally, (right panel) we envision that these cells colonize FLs and BM by following the developmental route of HSCs and seed to adult organs through circulation. We cannot exclude that some VSELs emerge in other parts of the epiblast as precursors for TCSCs. A, aorta; AL, allantois; BM, bone marrow; DM, dorsal mesentry; ExEc, extra-embryonic ectoderm; ExEn, extra-embryonic endoderm; H, hindgut; G, genital ridge; M, mesonephros; PE, primitive ectoderm. (b) Gene expression profiles of ESC-D3s, PGCs, FL-VSELs, and BM-VSELs, -HSCs, and -MNCs. Heat-map analysis of Ct values from RQ-PCR experiments for the indicated cells was prepared using Heatmap Builder software (Stanford University, Stanford, CA, USA). The expression level is indicated by red (high expression) and green (low expression) colors. *The expression of the indicated genes in PGCs is represented by a symbol and is based on previously published data by other groups.8, 15, 16, 17, 18, 19, 20, 21, 22, 32, 33, 34 +(cyto) in Dnmt3b indicates transcript expressed and cytoplasmic localization.22 +/+ and +/− in the imprinted genes panel indicates the biallelic expression (+/+) and no expression (+/−) of imprinted genes after erasure of genomic imprinting.33, 34

Accumulating evidence also indicates that PGCs could somehow be related to HSCs, another population of highly migratory SCs. To support this hypothesis: (i) a tight spatio-temporal overlap exists between the migration route of PGCs and the developmental origin of HSCs, first in extra-embryonic tissues in yolk sac blood islands and then in the aorta-gonad-mesonephros region;12 (ii) PGCs isolated from murine embryos were described as being able to grow HSC colonies;10, 25 and (iii) robust hematopoietic differentiation was observed in classical germ tumors.26, 27, 28 Furthermore, we report here that E12.5 FL-derived Oct4+ VSELs, similarly to their BM-derived counterparts,29 may expand over OP9 cells into CD45+ hematopoietic progenitors, which form myeloid colonies in secondary methylocellulose cultures. Thus, the ability of both FL- and BM-VSELs to acquire hematopoietic potential is also somehow spatio-temporally related to the developmental migration of HSCs (Figure 6a). In the future, it will be important to evaluate the potential presence of VSELs in yolk sac blood islands and to determine whether VSELs are detectable in Ncx1−/− embryos that do not initiate a heart beat and thus lack definitive HSCs in embryonic tissues.30 Altogether, molecular analysis of purified VSELs supports the concept that FL-derived epiblast/PGC-like VSELs follow the developmental migratory pathway of HSCs to colonize BM and, perhaps through circulation, may spread to other organs as well (Figure 6a). However, we cannot exclude the possibility that some VSELs originate through a different mechanism in situ during specification of EpiSCs during gastrulation.

Although purified adult BM-VSELs share several markers characteristic for EpiSCs as well as migratory PGCs, there are some discrepancies between migratory PGCs and VSELs. First, migratory PGCs demethylate both paternally and maternally DMRs, although each imprinted locus shows a different sensitivity to demethylation.31 In contrast, VSELs deposited into adult BM show a different imprint pattern depending on the parental origin for DMRs (Figure 4d and Supplementary Figure 8). This could be related to differences in the expression level and sub-cellular location of de novo DNA methyltransferase 3a and 3b (Dnmt3a and Dnmt3b) and its associated protein (Dnmt3l; Supplementary Figure 1) between VSELs and PGCs.22, 32 Furthermore, VSELs express the same genes (Klf4, c-Myc, Stat3, Snai1 and Ecat1), which are highly expressed in ESCs but not in PGCs.32 Therefore, VSELs are similar to migratory PGCs. However, they still show some differences in gene expression, which could be explained by different modulatory effects of the microenvironments in niches in which they reside (genital ridge for PGCs and BM for VSELs).

In this study, we also show some differences between FL- and BM-VSELs. Accordingly, FL-VSELs show: (i) a more ESC-like methylation status and histone modifications in Oct4, Nanog and Stella promoters; and (ii) hypomethylation in both paternally and maternally DMRs, similar to PGCs (Figure 4). In contrast, BM-VSELs show the parent-of-origin-specific imprints (that is, the erasure of imprints for paternally DMRs and hypermethylation of maternally DMRs). It is noteworthy that imprinted genes affect fetal growth and proliferation in an antagonistic manner, depending on paternal origin of their expression. Generally, paternally expressed imprinted genes enhance fetal growth, whereas maternally expressed ones negatively affect such growth.7 Thus, the parent-of-origin-specific genomic imprinting reprogramming in BM-VSELs leads to upregulation of growth-repressive maternally expressed genes (H19, Igf2R and p57KIP2) and downregulation of growth-promoting paternally expressed imprinted ones (Igf2 and Rasgrf1).6 Taken together, as compared with BM-VSELs, FL-VSELs show the gene expression and epigenetic features that are more favorable for cell proliferation and maintenance of pluripotency (Figure 6b and Supplementary Figure 9). Thus, some intrinsic factor(s) and/or microenvironmental signals regulate epigenetic differences between FL-VSELs and BM-VSELs. Identification of such factors and signals will be crucial to optimize an ex vivo expansion protocol for adult BM-VSELs. One potential candidate that is crucial for VSEL expansion is probably Igf2, which is downregulated in VSELs because of erasure of the Igf2-H19 DMR.6

In conclusion, the proliferative and developmental potential of VSELs is tightly regulated by the epigenetic status of genes involved in pluripotency, germ line development and genomic imprinting. Identifying the intrinsic and extrinsic factor(s) that control their expression would be crucial for developing more powerful strategies to unleash the regenerative potential of these cells to use them efficiently in clinic. Furthermore, this study shows that VSELs are somehow related to the population of migratory PGCs that become specified early in the development of proximal epiblast. Thus, this part of the developing embryo, in addition to PGCs, could also be the potential origin of other related, highly migratory SCs (for example, VSELs, HSCs?).


  1. 1

    Ratajczak MZ, Machalinski B, Wojakowski W, Ratajczak J, Kucia M . A hypothesis for an embryonic origin of pluripotent Oct-4+ stem cells in adult bone marrow and other tissues. Leukemia 2007; 21: 860–867.

  2. 2

    Evans MJ, Kaufman MH . Establishment in culture of pluripotential cells from mouse embryos. Nature 1981; 292: 154–156.

  3. 3

    Takahashi K, Yamanaka S . Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 2006; 126: 663–676.

  4. 4

    Kucia M, Reca R, Campbell FR, Zuba-Surma E, Majka M, Ratajczak J et al. A population of very small embryonic-like (VSEL) CXCR4+SSEA-1+Oct-4+ stem cells identified in adult bone marrow. Leukemia 2006; 20: 857–869.

  5. 5

    Zuba-Surma EK, Kucia M, Rui L, Shin D-M, Wojakowski W, Ratajczak J et al. Fetal liver very small embryonic/epiblast like stem cells follow developmental migratory pathway of hematopoietic stem cells. Ann NY Acad Sci 2009; 1176: 205–218.

  6. 6

    Shin DM, Zuba-Surma EK, Wu W, Ratajczak J, Wysoczynski M, Ratajczak MZ et al. Novel epigenetic mechanisms that control pluripotency and quiescence of adult bone marrow-derived Oct4+ very small embryonic-like stem cells. Leukemia 2009; 23: 2042–2051.

  7. 7

    Reik W, Walter J . Genomic imprinting: parental influence on the genome. Nat Rev Genet 2001; 2: 21–32.

  8. 8

    Surani MA, Hayashi K, Hajkova P . Genetic and epigenetic regulators of pluripotency. Cell 2007; 128: 747–762.

  9. 9

    Ginsburg M, Snow MH, McLaren A . Primordial germ cells in the mouse embryo during gastrulation. Development 1990; 110: 521–528.

  10. 10

    Rich IN . Primordial germ cells are capable of producing cells of the hematopoietic system in vitro. Blood 1995; 86: 463–472.

  11. 11

    Mikkola HKA, Orkin SH . The journey of developing hematopoietic stem cells. Development 2006; 133: 3733–3744.

  12. 12

    De Miguel MP, Arnalich Montiel F, Lopez Iglesias P, Blazquez Martinez A, Nistal M . Epiblast-derived stem cells in embryonic and adult tissues. Int J Dev Biol 2009; 53: 1529–1540.

  13. 13

    Hayashi K, Lopes SMCdS, Tang F, Surani MA . Dynamic equilibrium and heterogeneity of mouse pluripotent stem cells with distinct functional and epigenetic states. Cell Stem Cell 2008; 3: 391–401.

  14. 14

    Brons IGM, Smithers LE, Trotter MWB, Rugg-Gunn P, Sun B, Chuva de Sousa Lopes SM et al. Derivation of pluripotent epiblast stem cells from mammalian embryos. Nature 2007; 448: 191–195.

  15. 15

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

  16. 16

    Kurimoto K, Yabuta Y, Ohinata Y, Shigeta M, Yamanaka K, Saitou M . Complex genome-wide transcription dynamics orchestrated by Blimp1 for the specification of the germ cell lineage in mice. Genes Dev 2008; 22: 1617–1635.

  17. 17

    Hayashi K, Surani MA . Self-renewing epiblast stem cells exhibit continual delineation of germ cells with epigenetic reprogramming in vitro. Development 2009; 136: 3549–3556.

  18. 18

    Maatouk DM, Kellam LD, Mann MRW, Lei H, Li E, Bartolomei MS et al. DNA methylation is a primary mechanism for silencing postmigratory primordial germ cell genes in both germ cell and somatic cell lineages. Development 2006; 133: 3411–3418.

  19. 19

    Hajkova P, Ancelin K, Waldmann T, Lacoste N, Lange UC, Cesari F et al. Chromatin dynamics during epigenetic reprogramming in the mouse germ line. Nature 2008; 452: 877–881.

  20. 20

    Maldonado-Saldivia J, van den Bergen J, Krouskos M, Gilchrist M, Lee C, Li R et al. Dppa2 and Dppa4 are closely linked sap motif genes restricted to pluripotent cells and the germ line. Stem Cells 2007; 25: 19–28.

  21. 21

    Lane N, Dean W, Erhardt S, Hajkova P, Surani A, Walter J et al. Resistance of IAPs to methylation reprogramming may provide a mechanism for epigenetic inheritance in the mouse. Genesis 2003; 35: 88–93.

  22. 22

    Hajkova P, Erhardt S, Lane N, Haaf T, El-Maarri O, Reik W et al. Epigenetic reprogramming in mouse primordial germ cells. Mech Dev 2002; 117: 15–23.

  23. 23

    Molyneaux KA, Stallock J, Schaible K, Wylie C . Time-lapse analysis of living mouse germ cell migration. Dev Biol 2001; 240: 488–498.

  24. 24

    Vodyanik MA, Bork JA, Thomson JA, Slukvin II . Human embryonic stem cell-derived CD34+ cells: efficient production in the coculture with OP9 stromal cells and analysis of lymphohematopoietic potential. Blood 2005; 105: 617–626.

  25. 25

    Ohtaka T, Matsui Y, Obinata M . Hematopoietic development of primordial germ cell-derived mouse embryonic germ cells in culture. Biochem Biophys Res Commun 1999; 260: 475–482.

  26. 26

    Woodruff K, Wang N, May W, Adrone E, Denny C, Feig SA . The clonal nature of mediastinal germ cell tumors and acute myelogenous leukemia: A case report and review of the literature. Cancer Genet Cytogenet 1995; 79: 25–31.

  27. 27

    Saito A, Watanabe K, Kusakabe T, Abe M, Suzuki T . Mediastinal mature teratoma with coexistence of angiosarcoma, granulocytic sarcoma and a hematopoietic region in the tumor: a rare case of association between hematological malignancy and mediastinal germ cell tumor. Pathol Int 1998; 48: 749–753.

  28. 28

    Kritzenberger M, Wrobel K-H . Histochemical in situ identification of bovine embryonic blood cells reveals differences to the adult haematopoietic system and suggests a close relationship between haematopoietic stem cells and primordial germ cells. Histochem Cell Biol 2004; 121: 273–289.

  29. 29

    Zuba-Surma EK, Klich I, Wysoczynski M, Greco NJ, Laughlin MJ, Ratajczak MZ et al. In vitro and In vivo evidence that umbilical cord blood (UCB)-derived CD45-/SSEA-4+/OCT-4+/CD133+/CXCR4+/Lin- very small embryonic/epiblast like stem cells (VSELs) do not contain clonogenic hematopoietic progenitors but are highly enriched in more primitive stem cells—novel view on hierarchy of ucb stem cell compartment. Blood 2009; 114: 35.

  30. 30

    Lux CT, Yoshimoto M, McGrath K, Conway SJ, Palis J, Yoder MC . All primitive and definitive hematopoietic progenitor cells emerging before E10 in the mouse embryo are products of the yolk sac. Blood 2008; 111: 3435–3438.

  31. 31

    Shovlin TC, Durcova-Hills G, Surani A, McLaren A . Heterogeneity in imprinted methylation patterns of pluripotent embryonic germ cells derived from pre-migratory mouse germ cells. Dev Biol 2008; 313: 674–681.

  32. 32

    Yabuta Y, Kurimoto K, Ohinata Y, Seki Y, Saitou M . Gene expression dynamics during germline specification in mice identified by quantitative single-cell gene expression profiling. Biol Reprod 2006; 75: 705–716.

  33. 33

    Szabó PE, Mann JR . Biallelic expression of imprinted genes in the mouse germ line: implications for erasure, establishment, and mechanisms of genomic imprinting. Genes Dev 1995; 9: 1857–1868.

  34. 34

    Lee J, Inoue K, Ono R, Ogonuki N, Kohda T, Kaneko-Ishino T et al. Erasing genomic imprinting memory in mouse clone embryos produced from day 11.5 primordial germ cells. Development 2002; 129: 1807–1817.

Download references


This work was supported by NIH P20RR018733 from the National Center for Research Resources to MK and NIH R01 CA106281-01, NIH R01 DK074720, European Union structural funds, Innovative Economy Operational Program POIG.01.01.01-00-109/09-01 and the Henry M and Stella M Hoenig Endowment to MZR.

Author information

Correspondence to M Kucia or M Z Ratajczak.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Leukemia website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Shin, D., Liu, R., Klich, I. et al. Molecular signature of adult bone marrow-purified very small embryonic-like stem cells supports their developmental epiblast/germ line origin. Leukemia 24, 1450–1461 (2010).

Download citation


  • VSELs
  • epiblast
  • PGCs
  • CXCR4
  • epigenetics

Further reading