Bone marrow-derived very small embryonic-like stem cells: Their developmental origin and biological significance


  • M. Kucia,

    1. Stem Cell Biology Program at James Graham Brown Cancer Center, University of Louisville, Louisville, Kentucky
    Search for more papers by this author
  • W. Wu,

    1. Stem Cell Biology Program at James Graham Brown Cancer Center, University of Louisville, Louisville, Kentucky
    Search for more papers by this author
  • M.Z. Ratajczak

    Corresponding author
    1. Stem Cell Biology Program at James Graham Brown Cancer Center, University of Louisville, Louisville, Kentucky
    • Stem Cell Institute, James Graham Brown Cancer Center, University of Louisville, 580 S. Preston Street, Baxter II, Room 119E, Louisville, KY 40202
    Search for more papers by this author


Data from our and other laboratories provide evidence that bone marrow (BM) contains a population of stem cells that expresses early developmental markers such as (1) stage-specific embryonic antigen (SSEA) and (2) transcription factors Oct-4 and Nanog. These are the markers characteristic for embryonic stem cells, epiblast stem cells, and primordial germ cells (PGC). The presence of these stem cells in adult BM supports the concept that this organ contains some population of pluripotent stem cells that is deposited in embryogenesis during early gastrulation. We hypothesize that these cells could be direct descendants of the germ lineage that, to pass genes on to the next generations, has to create soma and, thus, becomes a “mother lineage” for all somatic cell lineages present in the adult body. Germ potential is established after conception in totipotent zygotes and retained in blastomeres of morula, cells from the inner cell mass of blastocyst, epiblast, and population of PGC. We will present a concept that SSEA+ Oct-4+ Nanog+ cells identified in BM could be descendants of epiblast cells as well as some rare migrating astray PGC. Developmental Dynamics 236:3309–3320, 2007. © 2007 Wiley-Liss, Inc.


Bone marrow (BM) was for many years primarily envisioned as the “home organ” of hematopoietic stem cells (HSC). However, recent research indicates that BM in addition to HSC also contains a heterogeneous population of nonhematopoietic stem cells. It is hypothesized that the presence of these various populations of stem cells in the BM is a result of the “developmental migration” of stem cells during ontogenesis and the presence of the permissive environment that attracts these cells to the BM tissue. HSC and other nonhematopoietic stem cells are actively chemoattracted by factors secreted by BM stroma cells and osteoblasts (e.g., stromal derived factor-1 [SDF-1] and hepatocyte growth factor [HGF]) and colonize marrow by the end of the second and the beginning of the third trimester of gestation (Kmiecik et al.,1992; Ma et al.,1998; Nagasawa,2000; Taichman et al.,2001). Accumulating evidence suggests that these nonhematopoietic stem cells residing in the BM play some role in the homeostasis/turnover of peripheral tissues and if needed could be released/mobilized from the BM into circulation during tissue injury and stress, thus facilitating the regeneration of damaged organs (LaBarge and Blau,2002; Kale et al.,2003; Kollet et al.,2003; Abbott et al.,2004; Kucia et al.,2004,2006c; Wojakowski et al.,2004; Long et al.,2005; Gomperts et al.,2006). It is also possible that, in steady state conditions, they shuttle at very low level between BM and stem cell niches in peripheral organs/tissues.

In this study, we will review data from our laboratory that demonstrate that BM contains a population of SSEA+ Oct-4+ Nanog+ stem cells that express markers of pluripotent stem cells (PSC; Kucia et al.,2006b). We named these cells very small embryonic-like (VSEL) stem cells. We hypothesize that they are deposited early in the development in the marrow tissue and are descendants of epiblast-derived stem cells (EPSC) and perhaps some primordial germ cells (PGC; Kucia et al.,2006a; Ratajczak et al.,2007). Generally, VSEL (1) could be a dormant quiescent population of PSC that resides in BM, (2) a population of stem cells that actively contributes to long-term hematopoiesis and turnover of other tissue specific (monopotent) stem cells that are located in peripheral niches, and (3) after being mobilized into peripheral blood for example during stress situation/organ injury, these cells may contribute to tissue organ regeneration. On the other hand, it is very likely that this population of very primitive cells, if exposed to mutagens, may give rise to cancer (e.g., teratomas, germinal tumors, and pediatric sarcomas; Kucia et al.,2006d).

Because VSEL show tropism to stromal cells and undergo emperipolesis in cocultures with bone marrow-derived fibroblasts (Kucia et al.,2005), we postulate that they could be coisolated with the adherent fractions of BM cells that were described as mesenchymal stem cells (MSC; Prockop,1997), multipotent adult progenitor cells (MAPC; Jiang et al.,2002), or marrow-isolated adult multilineage inducible (MIAMI) cells (D′Ippolito et al.,2004).


BM bone marrow CB cord blood EG embryonic germ cells ESC embryonic stem cells EPSC epiblast stem cells HGF/SF hepatocyte growth factor/scatter factor HSC hematopoietic stem cells ICM inner cell mass of blastocyst MAPC multipotent adult progenitor cells MIAMI marrow-isolated adult multilineage inducible cells mPB mobilized peripheral blood MSC mesenchymal stem cells PGC primordial germ cells PSC pluripotent stem cells SDF 1 stromal-derived factor 1 SSEA stage-specific embryonic antigen USSC unrestricted somatic stem cells VSEL very small embryonic-like stem cells


Several investigators in the past few years have demonstrated that BM-derived cells can contribute to the regeneration of nonhematopoietic organs and tissues (Jackson et al.,2001; Orlic et al.,2001; Corbel et al.,2003; Kale et al.,2003). These observations were mainly explained by the hypothesis that HSC are “plastic” and, thus, could trans-dedifferentiate into stem cells committed for various organs and tissues (Mezey et al.,2000; Corti et al.,2002b). This phenomenon to occur requires (1) a parallel switch of commitment for HSC in the compartment of stem cells or (2) alternatively, as postulated, a step back in the differentiation process from compartment of HSC with their dedifferentation into multipotent (one germ layer-committed) or even pluripotent (three germ layer-committed) stem cells.

This hypothetical possibility that HSC are plastic and able to trans-dedifferentiate raised much hope that HSC isolated from BM, mobilized peripheral blood (mPB), or cord blood (CB) could become a universal source of stem cells for tissue/organ repair. This excitement was bolstered at that time by several reports that demonstrated the remarkable regenerative potential of “HSC” in animal models, for example, after heart infarct (Orlic et al.,2001), stroke (Hess et al.,2004), spinal cord injury (Corti et al.,2002a), and liver damage (Petersen et al.,1999).

However, these first exciting and promising reports on the role of BM stem cells in the repair of damaged organs has become with time controversial (McKinney-Freeman et al.,2002; Orkin and Zon,2002; Wagers et al.,2002). Further experiments with highly purified populations of HSC showed them not to be effective in regenerating damaged heart (Murry et al.,2004) or brain (Castro et al.,2002). In response to these unexpected results, the scientific community became polarized in its view of the concept of stem cell plasticity.

These obvious discrepancies in published results could be explained by differences in the tissue injury models used and/or problems in detection of tissue chimerism. However, several other possibilities have been proposed to explain these discrepancies (Table 1). First, it is possible that, for trans-dedifferentiation of HSC to occur, the appropriate tissue damage models are required, which are able to create the permissive “pro-plastic” environment enriched in factors needed to promote this process. Hypothetically, such a “permissive” environment could induce epigenetic changes in HSC and, thus, force them to change lineage commitment (Morshead et al.,2002; Pomerantz and Blau,2004). Second, it has been postulated that some plasticity data could be explained simply by the phenomenon of cell fusion (Terada et al.,2002; Ying et al.,2002; Alvarez-Dolado et al.,2003; Vassilopoulos et al.,2003; Nygren et al.,2004). Accordingly, the donor-derived cells observed in damaged tissues that express nonhematopoietic markers could be in fact heterokaryons, the result of the fusion of BM-derived stem cells with somatic host cells in the damaged organs. Both cell fusion and epigenetic changes, however, are extremely rare and randomly occurring events that certainly could not fully account for all of the positive trans-dedifferentiation data published. Furthermore, fusion as a major contributor to the observed donor-derived chimerism has been excluded in several recently published studies (Harris et al.,2004; Wurmser et al.,2004; Brittan et al.,2005).

Table 1. Alternative Explanations of the Phenomenon of Trans-dedifferentiation or Plasticity of HSC and Their Potential Contribution to Organ Regenerationa
  • a

    BM, bone marrow; HSC, hematopoietic stem cells; PSC, pluripotent stem cells.

Epigenetic changesFactors present in the environment of damaged organs induce epigenetic changes in genes that restrict pluripotency of HSC (involvement of changes in DNA methylation, acetylation of histones). More evidence needed that it is a robust and reproducible phenomenon.
Cell fusionThe relatively rare phenomenon by which infused HSC may fuse with cells in damaged tissues and form heterokaryons. Heterokaryons created this way express markers of both donor and recipient cells (pseudochimerism).
Paracrine stimulationHSC are a source of different trophic and angiopoietic factors that may promote tissue/organ repair.
Microvesicle-dependent transfer of moleculesSome of the plasticity data could be explained by a transient modification of cell phenotype by the transfer of receptors, proteins, and mRNA between HSC and damaged cells by membrane-derived microvesicles.
Heterologous population of stem cells in BMIn addition to HSC, BM contains other stem cell populations. Regeneration could be explained by the presence of endothelial progenitors that promote neovasculogenesis and also by the presence of other stem cells, including PSC. This could also explain the loss of contribution of BM cells to organ regeneration with use of highly purified populations of HSC.

Another possible explanation of some of the benefits observed in organ/tissue regeneration after infusion of BM cells is the result of paracrine effects. It is well known that BM-derived cells are a source of several trophic cytokines and growth factors and that these factors if released from these cells could promote tissue repair and vascularization (Rafii and Lyden,2003). Furthermore, it has been recently shown that cells may transiently modify the phenotype of neighboring cells by transferring surface receptors, intracellular proteins, and mRNA in mechanisms that involve exchange of cell-membrane–derived microvesicles (Ratajczak et al.,2006a,b). Shedding of membrane-derived microvesicles is a physiological phenomenon that accompanies cell growth and cell activation, for example, hypoxia or oxidative injury (Beaudoin and Grondin,1991; VanWijk et al.,2003; Morel et al.,2004; Hugel et al.,2005). Thus, microvesicle-mediated exchange of receptors, proteins, and mRNA between infused BM cells and cells in damaged organs could temporarily modify the phenotype of cells in the damaged tissues. An open question is whether microvesicles can also exchange some of the reporter-gene markers used to detect tissue chimerism (e.g., green fluorescent protein or β-galactosidase).

Surprisingly, during all of these deliberations concerning stem cell plasticity (Table 1) and the potential contribution of BM-derived cells to organ regeneration, the concept that BM may contain heterogeneous populations of stem cells was not taken into careful consideration (Orkin and Zon,2002; Ratajczak et al.,2004). We postulate that regeneration studies demonstrating a contribution of donor-derived BM, mPB, or CB cells to nonhematopoietic tissues without addressing this possibility (by including the appropriate controls) has led to misleading interpretations. It is reasonable to assume that the presence of nonhematopoietic population of stem cells in BM, mPB, or CB would be considered before experimental evidence is interpreted as plasticity or trans-dedifferentiation of HSC (Kucia et al.,2005).

Hence, the presence of nonhematopoietic stem cells in BM rather than “trans-dedifferentiation” of HSC, could explain some of the positive results of tissue/organ regeneration as witnessed by several investigators using BM-derived cells (Petersen et al.,1999; Lagasse et al.,2000; Orlic et al.,2001). To support this, when highly purified HSC were used for regeneration experiments, nonhematopoietic stem cells were likely excluded from these cell preparations and plasticity was not observed. Thus, in the current state of knowledge, the phenomenon of trans-dedifferentiation of HSC and their contribution to regeneration of damaged tissues remains questionable.


PSC is defined as stem cells that gives rise to the stem cells from all three germ layers (meso-, ecto-, and endoderm) both in vitro in appropriate culture conditions as well as in vivo after injection into the developing blastocyst. PSC are derived from embryos from (1) blastomeres of morula, (2) inner cell mass of the blastocyst (ICM), and (3) EPSC. They could also be derived as immortalized cell lines in in vitro cultures from (1) ICM cells (embryonic stem cells [ESC]) or from (2) embryonic germ cells (EG) that derive in vitro cultures form population of PGC.

As mentioned above, recent research indicates that BM in addition to HSC also contains a heterogeneous population of nonhematopoietic stem cells and some of these cells were proposed to be PSC. These cells have been variously described in the literature as (1) MSC (Prockop,1997; Peister et al.,2004), (2) MAPC (Jiang et al.,2002), (3) MIAMI cells (D′Ippolito et al.,2004), and (4) USSC (Kogler et al.,2004). Recently, we isolated from BM a population of VSEL stem cells (Kucia et al.,2006b). It is likely that in many cases similar or overlapping populations of primitive stem cells in the BM were detected using different experimental strategies and, hence, were assigned different names. Although these cells were demonstrated in in vitro cultures to differentiate into cells from all three germ layers (meso-, ecto-, and endoderm), the developmental contribution of these cells to multiple organs and tissues in vivo after injection into the developing blastocyst is still missing. A possible explanation for this could be, as we hypothesize, an erasure of somatic imprint on PSC deposited in the developing embryo during organ formation, a mechanism that prevents parthenogenesis and teratoma formation in mammals. We will discuss this possibility further on in this review.


From the developmental and evolutionary point of view, the main goal of the multicellular organism is to pass genes to the next generations and this process is orchestrated by the appropriate interplay between germ and somatic cell lines (Weismann,1885; Kucia et al.,2006a). The germ line carries the genome (nuclear and mitochondrial DNA) from one generation to the next generation and is the only cell lineage that retains true developmental totipotency. In this context, we can envision that all somatic cell lines are descendants from the germ line and help germ cells to accomplish this mission effectively (Fig. 1). We will discuss this hypothesis below.

Figure 1.

The cycle of life—from zygote to germ cells. From the developmental and evolutionary point of view, the germ line (shown in red) carries the genome (nuclear and mitochondrial DNA) from one generation to the next and all somatic cell lines bud out (gray color) during ontogenesis from the germ line to help germ cells accomplish this mission effectively. The germ potential is established in the fertilized oocyte (zygote), and subsequently retained in the morula, inner cell mass of blastocyst (ICM), epiblast stem cells (EPSC), primordial germ cells (PGC), and mature germ cells (oocytes and sperm). The first cells that bud out from the germ lineage are trophoectodermal cells that will give rise to the placenta. Subsequently during gastrulation, EPSC are a source of pluripotent stem cells for all three germ layers (meso-, ecto-, and endoderm) and PGC. We hypothesize that, at this stage, some EPSC could be deposited as Oct-4+ pluripotent stem cells in peripheral tissues/organs (red circles). Similarly, some migrating PGC could go astray from their major migratory route to the genital ridges and become deposited as well. Furthermore, we hypothesize that, similar to PGC, other EPSC deposited in the developing tissues undergo erasure of their somatic imprint (yellow arrows). This mechanism of erasure of methylation of somatic imprinted genes protects developing organism from the possibility of teratoma formation. However, it will affect some of the aspects of the pluripotentiality of these cells (e.g., potential of these cells to contribute to blastocyst development or teratoma formation after transplantation into immunodeficient mice).

To support this concept, we can envision a zygote, which derives directly during conception from the fusion of two germ cells (female oocyte and male sperm) as the most primitive totipotent germ stem cell able to form both embryo and extraembryonal tissues (placenta). In a zygote, haploid DNA derived from an oocyte is combined with haploid DNA of a male germ cell, sperm; thus, the zygote could be envisioned as a mother cell to the germ lineage, which “down the road” will give rise to (1) more differentiated cells from the germ lineage (to ensure transfer of genome to the next generation) and (2) other somatic lineages that will provide the soma to fulfill this mission. These somatic cells will derive or “bud out” from the germ lineage during embryogenesis. Figure 1 presents this concept showing “a circle of reproductive life,” which begins with the establishment of the most primitive germ line cell (zygote), which gives rise to somatic lineages (meso-, ecto-, and endoderm), and most important germ cells (oocytes or sperm), which ensure transfer of DNA to the next generation. In this context, the germ lineage, to pass genes to the progeny, must establish an adult organism that will provide a “vehicle” soma/body to fulfill this mission (Donovan,1998; Zwaka and Thomson,2005; Kucia et al.,2006a).

Figure 1 also shows that the germ potential (marked as red) of the zygote is retained in the (1) first blastomeres, (2) cells that are present in the center of a developing morula and, subsequently, in the (3) cells from the ICM of a blastocyst. At this time of development, however, some level of specification already occurs and the trophoectoderm buds out from the germ line (Fig. 1). Trophoectoderm will form the placenta and the remaining part of the blastocyst; ICM will give rise to the epiblast.

EPSC as demonstrated experimentally are also pluripotent and retain germ lineage potential as well (Evans and Kaufman,1981). Shortly before the epiblast is about to give rise to all three germ layers (ectoderm, mesoderm, and endoderm), the first morphologically identifiable precursors of PGC in mice become specified ∼ 6.0–6.5 days post coitum (dpc) in the proximal part of the epiblast (McLaren,1992,2003). Thus, precursors of PGC are the first population of stem cells that is specified in the embryo at the beginning of gastrulation. PGC in mice subsequently move for a short period of time first to the basis of alantois, which is located in the extraembryonic mesoderm and then migrate into the embryo proper toward the genital ridges—where they will undergo developmental differentiation to oocytes or sperm respectively (Molyneaux and Wylie,2004; Fig. 1). Therefore, we envision one individual life as one link in the chain of the consecutive events aimed at the transfer of the genome by germ cells from one generation to the next. We hypothesize that the somatic cell lines that bud from the germ line merely play a supportive role.


As mentioned above, a PSC is a cell that in vitro at the single cell level can give rise to the cells from all three germ layers (meso-, ecto-, and endoderm); however, the most valuable evidence for pluripotentiality of a stem cell is its contribution to the development of multiple organs and tissues in vivo after injection into the developing blastocyst. This finding had been demonstrated very well in the case of ESC isolated from embryos or established in vitro ESC lines (Martin,1981; Thomson et al.,1995; Amit et al.,2000). However, it is very difficult to get such evidence in a reproducible manner for any type of putative PSC isolated from adult tissues including our VSEL. One of the potential mechanisms responsible for this finding could be an erasure of somatic imprint in adult cells, and we will discuss this explanation later on in this review.

As mentioned above EPSC are pluripotent and express SSEA-1 (mice), SSEA-3/4 (human), Oct-4, and Nanog. They will give rise to all three germ layers (ecto-, meso-, and endoderm), including PGC. Thus, the epiblast, through the process of gastrulation, is the source of all stem cells, for all the germ layers and stem cells in these layers will give rise to all of the tissues and organs in the embryo. Thus, EPSC could be envisioned as a founder population of PSC for multipotent stem cells for ecto-, endo-, or mesoderm that give rise to unipotent stem cells that will develop given cell lineages.

We hypothesize that pluripotent EPSC are deposited during gastrulation in germ layers as stem cells initiating development of various organs (Fig. 1; Kucia et al.,2006a; Ratajczak et al.,2007). Furthermore, there is also some additional evidence that some epiblast-derived PGC themselves during migration through the embryo proper on their way to the genital ridges might go astray and seed to peripheral tissues (Jordan,1917; Upadhyay and Zamboni,1982; Francavilla and Zamboni,1985). In conclusion, it is very likely that the cells identified in adult tissues that express ICM/epiblast/PGC markers such as SSEA-1 (mice), SSEA-3/4 (human), Oct-4, and Nanog are populations of PSC that were deposited in these tissues early during gastrulation/embryogenesis and are derived mostly from epiblast stem cells and, in addition, from some rare migrating PGC that went astray from their major migratory route to the genital ridges.


Of interest, PGC in contrast to EPSC do not reveal real pluripotency in vitro and in vivo. Namely, in cultures freshly isolated from embryos, PGC proliferate for a few days only, and then disappear either because they differentiate or die (De Felici and McLaren,1983). Furthermore, whereas the nuclei of migrating PGC at 8.5–9.5 dpc can be successfully used as donors for nuclear transfer, nuclei from PGC at 11.5 dpc and later are incompetent to support full-term development (Yamazaki et al.,2003). This finding is somehow intriguing, taking in consideration that PGC are the population of stem cells that carries “developmental totipotency” for oocytes and sperm.

However, when PGC are cultured over murine embryonic fibroblasts and exposed ex vivo to three growth factors (kit ligand, leukemia inhibitory factor, and basic fibroblast growth factor), they continue to proliferate and form large colonies of EG, which similarly as ESC can be expanded indefinitely (Matsui et al.,1992; Shamblott et al.,1998; Turnpenny et al.,2003). EG had been derived from pre- and postmigratory as well as from migratory PGC in both mice and humans and are pluripotent (Matsui et al.,1992; Shamblott et al.,1998). Namely, EG in contrast to PGC fully contribute to blastocyst complementation giving rise in the developing embryo to all somatic lineages and germ cells.

To explain this phenomenon at the molecular level, it is known that the pluripotency of PGC nuclei depends on the methylation status of genomic imprinted genes (e.g., H19, Igf-2, Igf-2R, Snrpn; Mann,2001; Yamazaki et al.,2003). PGC until 9.5 dpc display a somatic imprint (paternal and maternal pattern of methylation) of H19, Igf-2, Igf-2R, Snrpn—that is crucial to maintain their pluripotency. A somatic type of imprint, however, is erased by demethylation, while these cells migrate toward the genital ridges ∼ 10.5 dpc (Lee et al.,2002). The erasing of the methylation (imprint) of H19, Igf-2, Igf-2R, and Snrpn in early PGC could be envisioned as one of the mechanisms that shuts down PGC developmental pluripotency and makes these cells resistant to potential parthenogenesis or formation of teratomas (Macchiarini and Ostertag,2004; Oosterhuis and Looijenga,2005). A proper somatic imprint is subsequently again reestablished in sperm and oocytes, so that a fertilized egg expresses a developmentally proper somatic imprint of these crucial genes.

That PGC-derived EG cells are pluripotent demonstrates that the re-establishment of a proper somatic imprint is possible. We hypothesize that perhaps a similar mechanism of somatic imprint erasure takes place during development not only in PGC but also in other EPSC-derived PSC (e.g., VSEL) that are deposited in the developing organs (Fig. 1, yellow mark). It is also likely that, similarly as PGC, these cells under certain circumstances (e.g., tissue/organ injury) may regain a proper somatic imprint. This may occur through epigenetic changes for example during tissue/organ damage.

To further explore this hypothesis, it is necessary to (1) assess methylation status of IGF2 and H19 genes in VSEL and (2) to evaluate developmental potential of VSEL-derived nuclei in nuclear transfer experiments. We postulate that the possible erasure of somatic imprint in VSEL may explain why these cells, in our hands if freshly isolated from the BM, failed in contrast to ES cells to form teratomas after transplantation into syngeneic recipients or to contribute to blastocyst development in blastocyst complementation assay. On the other hand, keeping in mind that a proper status of somatic imprint in these cells could be potentially reestablished (e.g., as seen during generation of EGC from PGC), it is likely that after appropriate treatment they could acquire again pluripotentiality in vivo.


Stem cells that express early embryonic stem cell markers, including Oct-4, Nanog, and SSEA-1, had been identified in BM in several potential multipotent/PSC isolated from the BM or CB such as VSELs (Kucia et al.,2006b,2007), MSC (Colter et al.,2001; Pochampally et al.,2004), MAPC (Jiang et al.,2002), MIAMI (D′Ippolito et al.,2004), USSC (Kogler et al.,2004), and precursors of oocytes/spermatogonia and precursors of shuttling between BM and heart cardiomyocytes (Johnson et al.,2005; Nayernia et al.,2006; Pallante et al.,2007). It is very likely that several investigators using different isolation strategies described the same populations of stem cells but gave them different names according to circumstance.

VSEL Stem Cells

The homogenous population of rare (∼0.01% of BM MNC) Sca-1+ lin CD45 cells was recently purified by fluorescence-activated cell sorting (FACS) from murine BM (Fig. 2) (Kucia et al.,2006b). They express (as determined by RQ-polymerase chain reaction and immunhistochemistry) SSEA-1, Oct-4, Nanog, and Rex-1 and Rif-1 telomerase protein, but do not express MHC-I and HLA-DR antigens and are CD90 CD105 CD29. Direct electron microscopical analysis revealed that these cells display several features typical for embryonic stem cells such as (1) a small size (2–4 μm in diameter), (2) a large nucleus surrounded by a narrow rim of cytoplasm, and (3) an open-type chromatin (euchromatin). Despite their small size, they posses diploid DNA and contain numerous mitochondria (Fig. 3A).

Figure 2.

Sort of very small embryonic-like stem cells (VSEL) from murine bone marrow mononuclear cells (BMMNC). Fluorescence-activated cell sorting of murine VSEL were sorted from BMMNC derived from 1-month-old mice. Cells were sorted by using a MoFlo sorter (DAKO) after immunostaining for murine Sca-1 and CD45 and lineage markers using specific antibodies conjugated with fluorochromes, PE-Cy5, APC-Cy7, and PE, respectively (BD Pharmingen). Cells were sorted based on size of cells estimated using beads (Flow Cytometry Size Calibration Kit, Invitrogen). A: Region R1 contains cells with a size between 2 and 6 μm. B: Dot-plots show cells from region R1. VSELs were obtained from a fraction of cells that were negative for CD45 and did not express lineage markers (region R12, B). CD45/Lin cells were separated based on Sca-1 expression (C). The percentage of Sca-1+ lin CD45 cells (0.01%) referees to the cells present in the total population of murine BMMNC. A representative scheme of VSEL sorting is shown.

Figure 3.

Transmission electron microscopy (TEM) of murine and human very small embryonic-like stem cells (VSEL). A: Murine VSEL stem cells are small and measure 2–4 μm in diameter. They possess a relatively large nucleus surrounded by a narrow rim of cytoplasm. At the ultrastructural level the narrow rim of cytoplasm possesses a few mitochondria, scattered ribosomes, small profiles of endoplasmatic reticulum, and a few vesicles. The nucleus is contained within a nuclear envelope with nuclear pores. Chromatin is loosely packed and consists of euchromatin. B: Human cord blood (CB) -VSEL are small and measure 3–5 μm in diameter. They possess a relatively large nucleus surrounded by a narrow rim of cytoplasm. At the ultrastructural level this narrow rim of cytoplasm possesses a few mitochondria, scattered ribosomes, small profiles of endoplasmatic reticulum, and a few vesicles. The nucleus is contained within a nuclear envelope with nuclear pores. Chromatin is loosely packed and consists of euchromatin. Representative TEM pictures for murine bone marrow- and human CB-derived VSEL are shown.

Of interest, ∼5–10% of purified VSELs if plated over a C2C12 murine myoblast cell feeder layer are able to form spheres that resemble embryoid bodies. Cells from these VSEL-derived spheres (VSEL-DS) are composed of immature cells with large nuclei containing euchromatin, and like purified VSELs are CXCR4+SSEA-1+Oct-4+. Furthermore, cells from VSEL-DS, after re-plating over C2C12 cells, may again (up to five to seven passages) grow new spheres or, if plated into cultures promoting tissue differentiation, expand into cells from all three germ-cell layers. Similar spheres were also formed by VSELs isolated from murine fetal liver, spleen, and thymus. Of interest, formation of VSEL-DS was associated with a young age in mice, and no VSEL-DS were observed in cells isolated from old mice (> 2 years). Because VSELs express several markers of the germ cell line (fetal-type alkaline phosphatase, Oct-4, SSEA-1, CXCR4, Mvh, Stella, Fragilis, Nobox, Hdac6), they are probably closely related to a population of EPSC and PGC.

VSELs are mobile and respond robustly to an SDF-1 gradient, adhere to fibronectin and fibrinogen, and may interact with BM-derived stromal fibroblasts (Kucia et al.,2005). Confocal microscopy and time laps studies revealed that these cells attach rapidly to, migrate beneath, and undergo emperipolesis in marrow-derived fibroblasts (Kucia et al.,2005). This robust interaction of VSELs with BM-derived fibroblasts has an important implication, namely as mentioned above isolated BM stromal cells may be contaminated by these tiny cells from the beginning. This observation may somehow explain the unexpected “plasticity” of marrow-derived cells with a fibroblastic morphology (e.g., MSC, MAPC, or MIAMI cells).

Further Evidence for a Presence of Oct-4+ Cells in BM

To support a concept that VSEL could be in fact co-isolated with BM-derived adherent cells, there are several reports that Oct-4+ cells could be isolated from the BM. For example, a subpopulation of undifferentiated MSC expanded from BM adherent cells express embryonic stem cell markers such as Oct-4, Rex1, and Nanog (Lamoury et al.,2006). Furthermore, a cells similar to our VSEL population that SSEA-1+ Oct-4+ Nanog+ Sca-1+ lin CD45 were also recently isolated from the BM by another team with the suggestion that these cells could be associated with MSC (Anjos-Afonso and Bonnet,2006). Similarly, serum deprivation of human MSC cultures selects for expansion of so-called serum-deprived (SD) MSC (Pochampally et al.,2004), which express mRNA for embryonic markers, e.g., Oct-4.

Another population of BM-derived adherent cells, which was initially described as MAPC, was recently further enriched for small cells that are embryonic-like and express SSEA, Oct-4, and Nanog (pig) (Zeng et al.,2006b) and Oct-4 (human). Of interest, MAPC are the only population of BM-derived stem cells that, so far as is known, contribute to all three germ layers after injection into a developing blastocyst, indicating their pluripotency (Jiang et al.,2002). The contribution of MAPC to blastocyst development, however, requires confirmation by other, independent laboratories. Finally, MIAMI cells, which are isolated from human adult BM by culturing BM mononuclear cells in low oxygen tension conditions on fibronectin, express the embryonic stem cell markers Oct-4 and Rex-1 and differentiate into cells from multiple germ layers (D′Ippolito et al.,2004).

Recently, BM was also, somewhat unexpectedly, identified as a source of Oct-4+ oocyte-like (Johnson et al.,2005) and spermatogonia-like cells (Nayernia et al.,2006). This observation supports to some extent the concept that, during embryonic development, some of the stem cells from the germ lineage (Fig. 1) may go astray on their way to the genital ridges and home to the BM. To support this assumption, these cells express several PGC markers such as Oct-4, Mvh, Dazl, Stella, and Fragillis. At the same time, another independent group reported that BM cells may also be a source of Oct-4+ male germ cells (Nayernia et al.,2006). Finally, as recently demonstrated, BM cells may form culture aggregates of Oct3/4+ cells, which in in vitro cultures in the presence of platelet derived growth factor, may differentiate into cardiomyocytes (Pallante et al.,2007).

The potential relationship of all these Oct-4+ cells to Oct-4+ VSEL described by us is not clear at this point, although it is possible that these are overlapping populations of cells identified by slightly different isolation/expansion strategies.


Neonatal cord blood (CB) is an important source of nonhematopoietic stem cells. Generally, we can envision CB as neonatal peripheral blood mobilized by the stress related to delivery. Release of several cytokines and growth factors, as well as hypoxic conditions during labor, may mobilize neonatal marrow cells into circulation.

It is well known that CB-derived cells contribute to skeletal muscle (Gang et al.,2004; Brzoska et al.,2006), liver (Kakinuma et al.,2003; Newsome et al.,2003; Di Campli et al.,2004), neural tissue (Buzanska et al.,2002), and myocardium regeneration (Henning et al.,2004; Ma et al.,2005), and more importantly, recent multiorgan engraftment and differentiation has been achieved in goats after transplantation of human CB CD34+lin cells (Zeng et al.,2006a). Furthermore, several groups of investigators described the presence of Oct-4+, Nanog+, and SSEA-3/4+ stem cells in human CB and umbilical cord matrix (Carlin et al.,2006). They were enriched by different approaches using (1) immunomagnetic beads against CD133 (Baal et al.,2004); (2) expansion of a hematopoietic cell-depleted adherent population of CB cells in the presence of thrombopoietin, kit ligand, and flt3-ligand (McGuckin et al.,2005); or (3) isolation of mononuclear cells by Ficoll gradient centrifugation (Zhao et al.,2006).

Recently, we purified from human CB small cells resembling a population of murine BM-derived VSEL. These cells were purified and demonstrated at the single cell level by using a novel two-step isolation procedure, i.e., removal of erythrocytes by hypotonic lysis combined with multiparameter FACS sorting (Fig. 3B; Kucia et al.,2007). These CB-isolated VSEL (CB-VSEL) are very small (3–5 μm) and highly enriched in a population of CXCR4+AC133+CD34+lin CD45 CB mononuclear cells, possess relatively large nuclei containing unorganized euchromatin, express nuclear embryonic transcription factors Oct-4 and Nanog and surface embryonic antigen SSEA-4. Further studies are needed to determine whether human CB-isolated VSEL, similar to their murine BM-derived counterparts, are endowed with pluripotency. It is also not clear at this point what is the potential relationship of these cells to the so-called CB-derived USSC. USSC are very rare cells that are detectable in ∼40% of CB units at the frequency 1–11 cells/CB unit (Kogler et al.,2004). Unfortunately, the markers that are expressed on the founder cells for these colonies are not described. In in vitro cultures, unrestricted somatic stem cells differentiate into osteoblasts, chondroblasts, adipocytes, hematopoietic, and neural cells.


A population of stem cells that express markers of ESC/epiblast/PGC cells was recently described in several nonhematopoietic organs, for example, in epidermis (Dyce et al.,2004,2006; Yu et al.,2006), bronchial epithelium (Ling et al.,2006), myocardium, pancreas, (Kruse et al.,2006; Danner et al.,2007), testis (Kanatsu-Shinohara et al.,2004; Guan et al.,2006), retina (Koso et al.,2006), and amniotic fluid (De Coppi et al.,2007). This finding supports a concept of developmental deposition of Oct-4+ EPSC in developing organs.


The “positive” data supporting plasticity of BM-derived stem cells can be re-interpreted by the fact that BM stem cells are heterogeneous and that BM tissue contains different types of stem cells, including perhaps a rare population of PSC (e.g., VSEL) (Fig. 4). The question remains, however, whether these PSC are only some developmental remnants from epiblast or perhaps neural crest (Pierret et al.,2006) or if they can continuously contribute in adult life to the renewal of other more committed adult stem cells, including HSC, which are the most numerous population of stem cells in BM. There are several answers to this provocative, but timely, question (Table 2), especially in view of the current widely performed clinical trails with BM-derived stem cells in cardiology and neurology.

Figure 4.

Versatile populations of stem cells in bone marrow (BM). BM is traditionally the home of self-renewing hematopoietic stem cells (HSC). BM also contains a small admixture of several nonhematopoietic stem cells such as endothelial progenitor cells (EPC), mesenchymal stem cells (MSC), and perhaps other tissue-committed (monopotent) stem cells. Recent research indicates that BM contains also pluripotent stem cells (e.g., very small embryonic-like stem cells [VSEL]); however, the open question is whether these cells reside in dormant state only or are functional and able to supply all other types of stem cell in BM, including HSC.

Table 2. Pitfalls Related to Potential Clinical Application of BM-Derived Nonhematopoietic Stem Cells (e.g., VSEL)
  1. aBM, bone marrow; VSEL, very small embryonic-like stem cells.

 Isolation of a sufficient number of these cells from BM
 Optimal delivery by systemic or local infusion
 Proper “repertoire” of homing signals in damaged organs
 Adequate homing to damaged organs
 Effective contribution to organ/tissue regeneration after homing to damaged organ
 Lack of efficient ex vivo expansion strategies of stem cells

First, there is the obvious problem of isolating a sufficient number of these primitive cells from the BM. The number of these cells among BM MNC is very low (e.g., VSEL represent 1 cell in 104–105 of BM MNC; Kucia et al.,2006b). Furthermore, there are some indications that these cells are enriched in the BM of young mammals and their number decreases with age. It is also likely that these nonhematopoietic stem cells if released from the BM, even if they are able to home to the areas of tissue/organ injury, play a role only in regeneration of minor tissue injuries. Heart infarct or stroke, on the other hand, would be considered to involve severe tissue damage beyond the capacity of these rare cells to effectively repair. Second, the allocation of these cells to the damaged areas depends on homing signals that may be inefficient in the presence of proteolytic enzymes released from leukocytes and macrophages associated with damaged tissue. For example, matrix metalloproteinases released from inflammatory cells degrade SDF-1 locally and, thus, perturb homing of CXCR4+ stem cells (Lataillade et al.,2000; Petit et al.,2002; Christopherson et al.,2004). Third, it is not clear whether the optimal delivery system of these cells is intravenous infusion or direct injection into damaged tissues. Fourth, in order to reveal the full regenerative potential, these cells have to be fully functional. We cannot exclude the possibility that VSEL, while residing/being “trapped” in the BM, are not fully functional but remain “locked” in a dormant state and need the appropriate activation signals by unidentified factors. Finally, a major limitation is ex vivo expansion of these cells, which will not be associated with their differentiation.

Humanity continually searches for the holy grail of an end to the suffering caused by illness, and a better quality of life in advancing years. There is no doubt that medical science is now looking to stem cells to provide some of milestones in this important quest. However, there is a crucial need for a reliable and noncontroversial source of stem cells. Adult BM stem cells could potentially provide a real therapeutic alternative to the controversial use of human ES cells and therapeutic cloning. Hence, while the ethical debate on the application of ES cells in therapy continues, the potential of BM-derived VSEL is ripe for exploration. Researchers must determine whether these cells could be efficiently used in the clinic or whether they are merely developmental remnants found in the BM that cannot be harnessed effectively for regeneration. The coming years will bring important answers to these questions.


M.Z.R. was funded by the NIH.