Catherine M. Verfaillie, M.D.
Division of Hematology,
Department of Medicine, and Stem Cell Institute,
University of Minnesota

 

INTRODUCTION

In this paper, we want to provide updated information regarding a rare cell population, we have named, multipotent adult progenitor cells or MAPC. In 2001-2002, we published a series of papers demonstrating that while attempting to select and culture mesenchymal stem cells (MSC) from human and subsequently mouse and rat bone marrow (BM), we accidentally identified a rare population of cells that has characteristics unlike most adult somatic stem cells in that they appear to proliferate without senescence, and have pluripotent differentiation ability in vitro and in vivo ^1,2 .

Phenotype of Bone Marrow MAPC: MAPC can be cultured from human, mouse and rat bone marrow (BM). Unlike MSC, MAPC do not express major histocompatibiliy (MHC)- class I antigens, do not express, or express only low levels of, the CD44 antigen, and are CD105 (also endoglin, or SH2) negative ^1,2 . Unlike hematopoietic stem cells (HSC), MAPC do not express CD45, CD34, and cKit ^1,2 , but like HSC, MAPC express Thy1, AC133 (human MAPC) and Sca1 (mouse) albeit at low levels ^1,2 .In the mouse, MAPC express low levels of stage specific embryonic antigen (SSEA)-1, and express low levels of the transcription factors Oct4 and Rex1, known to be important for maintaining embryonic stem (ES) cells undifferentiated ³ and to be down-regulated when ES cells undergo somatic cell commitment and differentiation ² .

MAPC can also be isolated from other tissues, and other species:

We also showed that MAPC can be cultured from mouse brain and mouse muscle ⁴ . Of note, the differentiation potential and expressed gene profile of MAPC derived from the different tissues appears to be highly similar. These studies used whole brain and muscle tissue as the initiating cell population, therefore containing more than neural cells and muscle cells, respectively. The implications of this will be discussed below. Studies are ongoing to determine if cultivation of MAPC from other organs is possible, and whether culture of MAPC, like ES cells, is mouse-strain dependent.  Initial studies suggest that a population of MAPC-like cells can also be cultured from bone marrow from cynomologous monkeys (unpublished observations)(studies done by our collaborator Felipe Prosper, University of Navarra, Pamplona, Spain) and from bone marrow of dogs (unpublished observations)(studies done at the University of Minnesota).

Non-senescent nature of MAPC:

Unlike most adult somatic stem cells, MAPC proliferate without obvious signs of senescence, and have active telomerase. In humans, the length of MAPC telomeres is 3-5kB longer than in neutrophils and lymphocytes, and telomere length is not different when MAPC are derived from young or old donors ¹ . This suggests that MAPC are derived from a population of cells that either has active telomerase in vivo, or that is highly quiescent in vivo, and therefore has not yet incurred telomere shortening in vivo. In human MAPC cultures we have not yet seen cytogenetic abnormalities. As human MAPC are however undergoing symmetrical cell divisions, it remains possible that despite lack of gross cytogentic changes, minor mutations accumulate over time. We are therefore planning to use comparative genomic hybridization to address the question at what time genetic abnormalities occur, if they do. Initial results from gene array analysis suggest that MAPC, like ES cells, have a large number of DNA repair genes expressed (unpublished observations), which may protect them from more frequent genetic abnormalities in view of the fact that they undergo multiple sequential symmetrical cell divisions.

However, several subpopulations of mouse MAPC, and to a lesser extent rat MAPC, have become aneuploid, even though additional subpopulations thawed subsequently were cytogenetically normal. Aneuploidy is seen more frequently once mouse (and rat) MAPC have been expanded for >60-70 population doublings and following repeated cryopreservtions and thawing episodes. This characteristic of mouse MAPC is not dissimilar from other mouse cell populations, including mouse ES cells.

Stringent culture conditions required for maintenance of the undifferentiated state of MAPC:

Culture of MAPC is, however, technically demanding. Major factors that play a role in successful maintenance of MAPC include cell density, CO₂ concentration and pH of the medium, lot of fetal calf serum that is used, and even the type of culture plastic that is used. Control of cell density appears to be species specific: mouse, rat and perhaps cynomologous monkey MAPC need to be maintained at densities between 500 and 1,000 cells / cm2, whereas human and perhaps dog MAPC need to be maintained between 1,500 and 3,000 cells/ cm² . The reason why MAPC tend to differentiate to the default MSC lineage when maintained at higher densities is not known. However, for MAPC to have clinical relevance, this will need to be overcome. Gene array and proteomics studies are ongoing to identify the contact and / or soluble factors that may be responsible for causing differentiation when MAPC are maintained at higher densities. These very demanding technical skills can however be "exported" from the University of Minnesota as, after training at the University of Minnesota, investigators at the University of Tokai, Japan (manuscript submitted) and investigators at the University of Gent, Belgium have successfully isolated MAPC from human bone marrow, and investigators at the University of Navarra, Spain, have successfully isolated MAPC from rat bone marrow.

IN VITRO DIFFERENTIATION POTENTIAL OF MAPC:

We published last year that human, mouse and rat MAPC can be successfully differentiated into typical mesenchymal lineage cells, including osteoblasts, chondroblasts, adipocytes and skeletal myoblasts ¹ . In addition, human, mouse and rat MAPC can be induced to differentiate into cells with morphological, phenotypic and functional characteristics of endothelial cells ⁵ , and morphological, phenotypic and functional characteristics of hepatocytes ⁶ .

Neuroectodermal differentiation

Since then, we have also been able to induce differentiation of MAPC from mouse bone marrow into cells with morphological, phenotypic and functional characteristics of neuroectodermal cells⁷ . Differentiation of MAPC to cells with neuroectodermal characteristics occurred by initial culture in the presence of basic fibroblast growth factor (bFGF) as the sole cytokine, followed by culture with FGF-8b and sonic hedgehog (SHH), and then brain derived neurotrophic factor (BDNF) ^8,9 . Differentiation using these sequential cytokine stimuli was associated with activation of transcription factors known to be important in neural commitment in vivo and differentiation from NSC and mES cells in vitro. Cells staining positive for astrocyte, oligodendrocyte and neuronal markers were detected. Neuron-like cells became polarized, and as has been described in most studies in which ES cells or NSC were differentiated in vitro to a mid-brain neuroectodermal fate using FGF8 and SHH, approximately 25% of cells stained positive for dopaminergic markers, 25% for serotonergic markers, and 50% for GABA-ergic markers. Subsequent addition of astrocytes induced further maturation and prolonged survival of the MAPC-derived neuron-like cells, which now also acquired electrophysiological characteristics consistent with neurons, namely voltage gated sodium channels and synaptic potentials ^10,11 .

Muscle differentiation:

In addition, we now have convincing evidence that MAPC can differentiate into cells with phenotypic as well as functional characteristics of smooth muscle cells (manuscript in preparation). Interestingly, the lineage that continues to be elusive is cardiac myoblasts, despite the fact that mouse MAPC injected in the blastocyst contribute to the cardiac muscle² . Although a number of in vitro differentiation conditions induce expression of Nkx2.5, GATA4, and myosin heavy chain mRNA and proteins^12-14 , we have been unable to induce differentiation of MAPC to cells with the typical functional characteristic of cardiac myoblasts, i.e. spontaneous rhythmic contractions or beating, a differentiation path that is almost a default differentiation pathway for mouse ES cells.  The reason for the lack of functional cardiac myoblast properties is currently unknown.

Another important cell lineage that has not yet been generated is insulin-producing cells, even though initial studies suggest that differentiation to cells expressing at least early pancreatic and endocrine pancreas transcription factors can be obtained.

In vitro differentiation of MAPC as model system for gene discovery:

A last comment regarding in vitro differentiation of MAPC is that, in contrast to differentiation of ES cells in vitro, the final differentiated cell product derived from MAPC is commonly >70-80% pure. This should allow using these in vitro differentiation models for gene and drug discovery. For instance, in a recently published study¹⁵ we compared the expressed gene profile in human MAPC induced to differentiate to osteoblasts and chondroblasts, two closely related cell lineages. We could demonstrate that although a large number of genes are co-regulated when MAPC differentiate to these two lineages, specificity in differentiation can readily be detected. For instance a number of known and yet to be fully characterized transcription factor mRNAs were differentially expressed during the initial phases of differentiation. Studies are ongoing to further define the role of these genes in lineage specific differentiation. These studies exemplify however the power of this model system to study lineage specific differentiation in vitro.

DEGREE OF PLURIPOTENCY OF MAPC:

We have shown that transfer of 10-12 mouse MAPC into mouse blastocysts results in the generation of chimeric mice. When 10-12 MAPC, expanded for 50-55 population doublings, were injected approximately 80% of offspring were chimeric, with the degree of chimerisms varying between 1-40%⁴ . Cells found in different organs acquire phenotypic characteristics of the tissue. For instance MAPC derived cells detected in the brain of chimeric animals differentiate appropriately into region specific neurons, as well as astrocytes and oligodendrocytes¹⁶ . More recent studies using MAPC from later population doublings have shown that the frequency of chimerism decreases when MAPC are maintained for longer time in culture, even though animals with chimerism of more than 70% could be obtained (unpublished observations). These studies indicate that like ES cells, MAPC can give rise to most if not all somatic cell types of the mouse. Whether MAPC can do this without help of other cells in the inner cell mass, i.e. can generate a mouse by tetraploid complementation¹⁷ , is not yet known. Also not yet known is whether MAPC contribute to the germ line when injected in the blastocyst.

POST-NATAL CONTRIBUTION TO TISSUES:

Neither human nor mouse MAPC injected into the muscles of severe combined immunodeficient (SCID) mice have led to the development of teratomas (unpublished observations). Likewise, we have not yet detected donor-derived tumor formation following IV injection of human or mouse MAPC in NOD-SCID animals. However, when mouse undifferentiated MAPC are administered IV to NOD-SCID mice, engraftment in the hematopoietic system as well as epithelia of gut, liver and lung is seen² .  Preliminary studies using human MAPC suggest that a similar pattern of engraftment may occur, even though the level of contribution to blood, liver, gut and lung is lower (unpublished observations). Noteworthy is the fact that neither mouse nor human MAPC appear to contribute to other tissues when injected IV, except to endothelium (see below). Although PCR analysis for human DNA in human - mouse transplants or for b-galactosidase in mouse-mouse transplants yielded positive signals in many tissues, we believe that this is mainly due to contaminating blood cells. When tissues were carefully examined for tissue specific differentiated MAPC progeny, we could not detect MAPC-progeny in brain, skeletal muscle, cardiac muscle, skin or kidneys. Lack of engraftment in brain, skeletal and cardiac muscle may be due to the fact that transplants were done in non-injured animals, where the blood brain barrier is intact, and where little or no cell turnover is expected in muscle. More difficult to explain is the absence of MAPC-derived progeny in skin, possibly the organ with the greatest cell turnover. Studies are ongoing to trace the homing behavior of MAPC following infusion in non-injured animals and injured animals, which may shed light on these observations.

In vivo differentiation into skeletal muscle:

Muguruma et al have also shown that undifferentiated human MAPC injected in the muscle of non-obese diabetic (NOD)-SCID mice differentiate into cells that stain positive for muscle transcription factors and muscle cytoskeletal proteins (manuscript submitted). Similar results were seen in Minnesota. We also found that pre-treatment of human MAPC with 5-azacytidine, required to induce muscle differentiation in vitro, enhanced the degree of engraftment of human cells in mouse muscle, suggesting that pre-differentiation of MAPC may under certain circumstances enhance the level of engraftment (unpublished observations).

Contribution to endothelium in vivo:

When endothelial cells generated from human MAPC by incubation in vitro with vascular endothelial growth factor (VEGF)⁵ were infused in animals in which a tumor had been implanted underneath the skin, we detected enhanced tumor growth and found that up to 30% of the tumor vasculature was derived from the human endothelial cells. Likewise, wounds in the ears of these animals as a result of ear tagging contained human endothelial cells. One of the animals developed a host-tumor, an occurrence seen frequently in aging NOD-SCID mice. We detected contribution of MAPC-derived endothelium to tumor vessels² .  Likewise, one of the NOD-SCID mice that received human MAPC developed a host thymic lymphoma. Human MAPC, like mouse MAPC, appeared to differentiate into endothelial cells that contribute to tumor angiogenesis.

Engraftment of MAPC in stroke model:

In yet another in vivo study¹⁸ we evaluated the effect of human MAPC in a rat stroke model. Cortical brain ischemia was produced in male rats by permanently ligating the right middle cerebral artery distal to the striatal branch. Animals were placed on cyclosporine-A and 2 weeks later, 2x10⁵ human MAPC were injected around the infarct zone. As controls, animals received normal saline or MAPC conditioned medium. Limb placement test and tactile stimulation test were blindly assessed 1 week before brain ischemia, 1 day before transplantation, and at 2 and 6 weeks after grafting. The limb placement test included eight subtests described by Johansson and coworkers¹⁹ . In a tactile stimulation test²⁰ , a small piece of adhesive tape was rapidly applied to the radial aspect of each forepaw. The rats were then returned to their home cages, and the order of the tape removal (i.e., left versus right) was recorded. Three to five trials were conducted on each test day. Each trial was terminated when the tapes were removed from both forepaws or after 3 min. Animals were subsequently sacrificed to determine the fate of the human cells injected in the brain.  After 2 and 6 weeks, animals that received human MAPC scored statistically significantly better in the limb placement test as well as tactile stimulation test compared with animals that received only cyclosporine-A (CSA), or were injected with normal saline or MAPC conditioned medium. The level of recuperation of motor and sensory function was 80% of animals without stroke. When the brain was examined for the presence and differentiation of human MAPC to neuroectodermal cells, we found that human MAPC were present, but remained rather immature. Therefore, we cannot attribute the motor and sensory improvement to region specific differentiation to neuronal cells and integration of neurons derived from MAPC in the host brain. Rather the improvement must be caused by trophic effects emanated by the human MAPC to either improve vascularization of the ischemic area, to support survival of the remaining endogenous neurons, or to recruit neuronal progenitors from the host brain. These possibilities are currently being evaluated.

POSSIBLE MECHANISMS UNDERLYING THE PHENOMENON OF MULTIPOTENT ADULT PROGENITOR CELLS:

Currently we do not fully understand the mechanism(s) underlying the culture selection of MAPC. We have definitive data to demonstrate that the pluripotency of MAPC is not due to co-culture of several stem cells.

Pluripotency cannot be attributed to multiple stem cells:

First, using retroviral marking studies we have definitive proof that a single cell can differentiate in vitro to cells of mesoderm, both mesenchymal and non-mesenchymal, neuroectoderm and hepatocyte-like cells, and this for human^{1, 6} , mouse and rat MAPC ^{2, 6} . Second, we have shown that a single mouse MAPC is sufficient for generation of chimeric animals² . Indeed, we published that 1/3 animals born from blastocysts in which a single MAPC was injected were chimeric with chimerism degrees varying between 1 and 45%. This rules therefore out that the pluripotent nature of these cells is due to co-existence in culture of multiple somatic stem cells.

Cell fusion is not likely explanation:

A second possibility for the greater degree of differentiation potential would be that cells undergo fusion and acquire via this mechanism greater pluripotency. Fusion has been shown to be responsible for apparent ES characteristics of marrow and neural stem cells ^21,22 that had been cocultured with ES cells in vitro, and more recently for the apparent lineage switch of bone marrow cells to hepatocytes when hematopoietic cells were infused in animals with hereditary tyrosinemia due to lack of the fumarylacetoacetate hydroxylase (FAH) gene²³ [Wang et al, Nature 2003]. In the former two studies, the majority of genes expressed in the marrow or neural cell that fused with ES cells were silenced, and the majority of the genes expressed in ES cell were persistently expressed. Likewise for the bone marrow-hepatocyte fusion, the majority of genes expressed normally in hematopoioetic cells (except the FAH gene) were silenced whereas genes expressed in hepatocytes predominated. Finally, the cells generated were in general tetraploid or aneuploid.

We do not believe that this phenomenon underlies the observation that MAPC are pluripotent. Cultivation and differentiation in vitro (in general, except the final differentiation step for neuroectoderm) does not require that MAPC are co-cultured with other cells, making the likelihood that MAPC are the result of fusion very low. Smith et al suggested in a recent commentary that MAPC could be caused by fusion of multiple cell types early on during culture leading to reprogramming of the genetic information and pluripotency [REF]. However, we have no evidence that MAPC are tetraploid or aneuploid early during culture, making this possibility less likely. Nevertheless, studies are ongoing to rule this out. The in vivo studies were not set up to fully be capable of ruling out this possibility.

However, a number of findings suggest that fusion may not likely be the cause for the engraftment seen postnatally, nor the chimerism in the blastocyst injection experiment. The frequency of the fusion event described for the ES-BM, ES-NSC, and HSC-hepatocyte fusion was in general very low, i.e. 1/100,000 cells. Expansion of such fused cells could only be detected when drug selection was applied in the in vitro systems, and withdrawal of NTBC (2-(2-nitro-4-trifluoro-methylbenzoyl)-1,3-cyclohexanedione) in the FAH mouse model was used to select for cells expressing the FAH gene. The percent engraftment seen in our post-natal transplant models was in the range of 1% - 9%. The frequency of chimerism seen in blastocyst injection studies ranged between 33% and 80% when 1 and 1 and 10-12 MAPC were injected, respectively. These frequencies are significantly higher than what has been described for fusion events with ES cells in vitro, and in the HSC-hepatocyte fusion studies in vivo.

Furthermore, in contrast to what was described in the papers indicating that fusion may be responsible for apparent plasticity, all in vivo studies done with MAPC were done without selectable pressure, mainly in non-injured animals. Therefore, it is less likely that the pluripotent behavior of MAPC in vivo is due to fusion between the MAPC and the tissues where they engraft / contribute to. However, specific studies are currently being designed to formally rule this out.

Primitive ES-like cells that persist vs. de-differentiation:

Currently, we do not have proof that MAPC exist as such in vivo. Until we have positive selectable markers for MAPC, this question will be difficult to answer. If the cell exists in vivo, one might hypothesize that it is derived for instance from primordial germ cells that migrated aberrantly to tissues outside the gonads during development. It is, however, also possible that removal of certain (stem) cells from their in vivo environment results in "reprogramming" of the cell to acquire greater pluripotency. The studies on human MAPC suggest that such a cell that might undergo a degree of reprogramming is likely a protected (stem) cell in vivo, as telomere length of MAPC from younger and older donors is similar, and significantly longer than what is found in hematopoietic cells from the same donor. The fact that MAPC can be isolated from multiple tissues might argue that stem cells from each tissue might be able to be reprogrammed. However, as was indicated above, the studies in which different organs were used as the initiating cell population for generation of MAPC did not purify tissue specific cells or stem cells. Therefore, an alternative explanation is that the same cells isolated from bone marrow that can give rise to MAPC in culture might circulate, and be collected from other organs. However, we have until now been unsuccessful in isolating MAPC from blood or from umbilical cord blood, arguing against this phenomenon. Finally, cells selected from the different organs could be the same cells resident in multiple organs, such as MSC that are present in different locations, or cells associated with tissues present in all organs such as for instance blood vessels. Studies are ongoing to determine which of these many possibilities is correct.

CONCLUSION:

We believe that MAPC would have clinical relevance whether they exist in vivo, or are created in vitro. However, understanding the nature of the cell will have impact on how one would approach their clinical use. If they exist in vivo, it will be important to learn where they are located, and to determine whether their migration, expansion and differentiation in a tissue specific manner can be induced and controlled in vivo. If they are a culture creation, understanding the mechanism underlying the reprogramming event will be important as that might allow this phenomenon to happen on a more routine and controlled basis.

Either way, a long road lies ahead before MAPC might be applicable in clinical trials. Hurdles to be overcome include development of robust culture systems that will allow automatization. Like of other stem cells, including ES cells, we will need to determine in preclinical models whether undifferentiated vs. lineage committed vs. terminally differentiated cells should be used to treat a variety of disorders. If lineage committed or terminally differentiated cells will be needed, robust clinical scale differentiation cultures will need to be developed. Furthermore, studies will need to be performed to demonstrate whether potentially contaminating undifferentiated MAPC will interfere with engraftment, and / or differentiate inappropriately in vivo. Likewise, studies aimed at determining what level of HLA-mismatch will be tolerated in transplantations, whether tolerization via hematopoietic engraftment from MAPC will be required. As is also the case for other extensively cultured cells, we will need to further determine if prolonged expansion leads to genetic abnormalities in cells that might lead to malignancies when transplanted in vivo.

As a final remark, MAPC appear to have pluripotent potential both in vitro and in vivo. Furthermore, they appear to proliferate without obvious senescence when maintained under very stringently controlled culture conditions. Because of these reasons, some have argued that they might be a viable alternative to ES cells. However, at this stage of the research, I feel that such a conclusion is premature. Whether MAPC have equal longevity as ES cells, and have the ability to create all >200 cell types in the body is still not known. Moreover, there appear to be certain cell types that are more readily generated from ES cells compared with MAPC, such as for instance cardiac myoblasts, whereas it appears for instance more easy to generate hepatocyte like cells from MAPC than ES cells. Therefore, I continue to strongly believe that strict comparative studies between the two cell populations are needed to determine the true potential of the cells, and that the scientific insights gained from these studies should be used to determine which of the cells will be suitable for use in the clinical setting.

_________________

ENDNOTES/References

1.  Reyes M, Lund T, Lenvik T, Aguiar D, Koodie L, Verfaillie CM. Purification and ex vivo expansion of postnatal human marrow mesodermal progenitor cells. Blood. 2001;98:2615-2625

2.  Jiang Y, Jahagirdar B, Reyes M, Reinhardt RL, Schwartz RE, Chang H-C, Lenvik T, Lund T, Blackstad M, Du J, Aldrich  S, Lisberg A, Kaushal S, Largaespada DL, Verfaillie CM. Pluripotent nature of adult marrow derived mesenchymal stem cells. Nature. 2002;418:41-49

3.  Niwa H, Miyazaki J, Smith AG. Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells. Nat Genet. 2000;24:372-376

4.  Jiang Y, Vaessen B, Lenvik T, Blackstad M, Reyes M, Verfaillie CM. Multipotent progenitor cells can be isolated from post-natal murine bone marrow, muscle and brain. Exp Hematol. 2002;30(8):896-904

5.  Reyes M, Dudek A, Jahagirdar B, Koodie K, Marker PH, Verfaillie CM. Origin of endothelial progenitors in human post-natal bone marrow. J Clin Invest. 2002;109:337-346

6.  Schwartz RE, Reyes M, Koodie L, Jiang Y, Blackstad M, Lund T, Lenvik T, Johnson S, Hu W-S, Verfaillie CM. Multipotent adult progenitor cells from bone marrow differentiate into functional hepatocyte-like cells. J Clin Invest. 2002;109(10):1291-1302

7.  Jiang Y, Henderson D, Blackstadt M, Chen A, Miller FF, Verfaillie CM. Neuroectodermal differentiation from mouse multipotent adult progenitor cells. Proc. Natl. Acad. Sci U S A. 2003;In Press

8.  Ling Z, Potter E, Lipton J, Carvey P. Differentiation of mesencephalic progenitor cells into dopaminergic neurons by cytokines. Exp Neurol. 1998;149:411-423

9.  Lee SH, Lumelsky N, Studer L, Auerbach JM, McKay RD. Efficient generation of midbrain and hindbrain neurons from mouse embryonic stem cells. Nat Biotechnol. 2000;18:675-679

10.  Wagner J, Akerud P, Castro DS, Holm PC, Canals JM, Snyder EY, Perlmann T, Arenas E. Induction of a midbrain dopaminergic phenotype in Nurr1-overexpressing neural stem cells by type 1 astrocytes. Nat Biotech. 1999;17:653-659

11.  Song H, Stevens CF, Gage FH. Astroglia induce neurogenesis from adult neural stem cells. Nature. 2002;417:39-44

12.  Tanaka M, Chen Z, Bartunkova S, Yamasaki N, Izumo S. The cardiac homeobox gene Csx/Nkx2.5 lies genetically upstream of multiple genes essential for heart development. Development. 1999;126:1269-1280

13.  Laverriere AC, MacNeill C, Mueller C, Poelmann RE, Burch JB, Evans T. GATA-4/5/6, a subfamily of three transcription factors transcribed in developing heart and gut. J Biol Chem. 1994;269:23177-23184

14.  Klug MG, Soonpaa MH, Koh GY, Field LJ. Genetically selected cardiomyocytes from differentiating embronic stem cells form stable intracardiac grafts. J Clin Invest. 1996;98:216-224

15.  Qi H, Aguiar DJ, Williams SM, La Pean A, Pan W, Verfaillie CM. Identification of genes responsible for osteoblast differentiation from human mesodermal progenitor cells. Proc. Natl Acad. Sci U S A. 2003 Mar18;100(6):3305-3310

16.  Keene CD, Ortiz-Gonzalez XR, Jiang Y, Largaespada DA, Verfaillie CM, Low WC. Neural differentiation and incorporation of bone marrow-derived multipotent adult progenitor cells after single cell transplantation into blastocyst stage mouse embryos. Cell Transplant. 2003;12(3):201-213

17.  Wang ZQ, Kiefer F, Urbanek P, Wagner EF. Generation of completely embryonic stem cell-derived mutant mice using tetraploid blastocyst injection. Mech Dev. 1997;62:137-145

18.  Zhao L-R, Duan W-M, Reyes M, Keene CD, Verfaillie CM, Low WC. Human bone marrow stem cells exhibit neural phenotypes and ameliorate neurological deficits after grafting into the ischemic brain of rats. Exp Neurol. 2002;174:11-20

19.  Ohlsson AL, Johansson BB. Environment influences functional outcome of cerebral infarction in rats. Stroke. 1995;4:644-649

20.  Netto CA, Hodges JD, Sinden JD, LePeillet E, Kershaw T, Schallert T, Whishaw IQ. Bilateral cutaneous stimulation of the somatosensory system in hemidecorticate rats. Behav. Neurosci. 1984;98:518-540

21.  Ying QY, Nichols J, Evans EP, Smith AG. Changing potency by spontaneous fusion. Nature. 2002;416:545-548

22.  Terada N, Hamazaki T, Oka M, Hoki M, Mastalerz DM, Nakano Y, Meyer EM, More Ll, Petersen BE, Scott EW. Bone marrow cells adopt the phenotype of  other cells by spontaneous cell fusion. Nature. 2002;416:542-545

23.  Lagasse E, Connors H, Al-Dhalimy M, Reitsma M, Dohse M, Osborne L, Wang X, Finegold M, Weissman IL, Grompe M. Purified hematopoietic stem cells can differentiate into hepatocytes in vivo.  Nat Med. 2000;6:1229-1234