Through the years, scientists have defined stem cells in many ways. The consensus definition would encompass three main principles. First, a stem cell must be capable of self-renewal i.e. undergoing symmetric or asymmetric divisions through which the stem cell population is maintained. Second, a single cell must be capable of multilineage differentiation. The third principle is in vivo functional reconstitution of a given tissue. (Verfaillie et al., 2002)
A fertilized egg is capable not only of forming cells of the ectoderm, endoderm, and mesoderm layer, and germ cells, but also the supporting extraembryonic tissues required for the survival of the developing embryo. Therefore, these cells are at the top of the stem cell hierarchy and termed “totipotent”. (Verfaillie et al., 2002) Embryonic stem (ES) cells and embryonic germ (EG) cells, isolated from the inner cell mass of the blastocyst or from primordial germ cells of an early embryo, give rise to ectoderm, endoderm, and mesoderm layers and germ cells but not extra-embryonic tissues, and are therefore termed “pluripotent”. Stem cells isolated from various adult organs can self-renew and differentiate into multiple tissue specific cell types. These stem cells are termed “multipotent stem cells”. Committed cells generally have limited or no self-renewal ability and differentiate into only one defined cell type and are dubbed “progenitor cells” or “precursor cells”. (Verfaillie et al., 2002)
The ES cell is the quintessential pluripotent stem cell as it fulfills all criteria. Embryonic stem (ES) cells are pluripotent stem cells that can be propagated indefinitely in an undifferentiated state. ES cells differentiate to all cell lineages in vivo and also differentiate into many cell types in vitro. ES cells have been isolated from humans, however their use in research as well as in clinical practice has been hampered by ethical and technical considerations (Frankel, 2000). As embryonic stem cells readily form teratomas (pluripotent tumors), it will be critical to develop a method that ensures that all ES cells differentiate and none are left pluripotent.
As has been known for decades, stem cells exist for most tissues, including hematopoietic, neural, gastrointestinal, epidermal, hepatic and mesenchymal stem cells. (Verfaillie et al., 2002) Compared with ES cells, tissue-specific stem cells have less self-renewal and proliferative ability, and are not pluripotent. Only recently has it been shown that tissue-specific stem cells could not only differentiate into cells of the tissue of origin but into other lineages. For example, following transplantation of donor bone marrow (BM) or enriched hematopoietic stem cells (HSC) into allogeneic recipients, skeletal myoblasts, cardiac myoblasts, endothelium, hepatic and biliary duct epithelium, lung, gut and skin epithelia, and neuroectodermal cells of donor origin have been detected (Verfaillie et al., 2002).
Adult Stem Cells-Plasticity
Any discussion on stem cells would be incomplete without a full discussion on stem cell plasticity and the present controversy in the stem cell field. Traditionally, adult stem cells have been viewed as committed to a particular cell fate. For example, hematopoietic stem cells (HSC) were viewed to only contribute to lineages that are part of the hematopoietic system i.e. RBC’s and WBC’s and not unrelated tissues, such as hepatocytes or neurons. (Verfaillie et al., 2002) Many studies question this belief or dogma by demonstrating that cells from a given tissue might differentiate into cells of a different tissue. (LaBarge and Blau, 2002)
If true this would suggest that understanding that postnatal stem cells give rise to only cells of the tissue of origin may not be correct. HSC besides giving rise to blood cells, may also give rise to hepatocytes. NSC may not only give rise to nerve cells but also to early hematopoietic precursors. This ability of a tissue-specific stem cell to acquire the fate of a cell type different from the original tissue has been termed adult stem cell plasticity, although no consensus exists to what the exact definition should be. (Verfaillie et al., 2002)
However, the idea is almost a century old. In the late 19th and early 20th century it was recognized that there are epithelial changes in tissues in response to different stresses (Cotran, 1999, pp. 31-38). These changes in which one adult cell type is replaced by another cell type was termed metaplasia. An example includes the change from columnar epithelium to squamous epithelium in the respiratory tract of smokers in response to chronic irritation caused by smoking (Cotran, 1999, p. 36). Another example is the change from squamous epithelium to columnar epithelium due to gastric reflux that occurs in Barrett’s esophagus (Cotran, 1999). The possible mechanisms for this plasticity will be described later.
More recently, reports on stem cell plasticity have brought much excitement within the lay and scientific communities (Verfaillie et al., 2002). In addition, they have also generated great skepticism. This is largely because the concept of stem cell plasticity conflicts with the established dogma of stem cell hierarchy and its role in developmental biology which has widely believed that cell fate and lineage restriction was determined during gastrulation and subsequent organ morphogenesis. (Verfaillie et al., 2002)
However, if correct, the potential clinical benefit of the ability of a postnatal stem cell to change fate is immeasurable. It is therefore critical to rigorously define stem cell plasticity such that it can be carefully tested. Moreover, studies have shown that the unexpected lineage differentiation was derived from a single cell that was also shown to differentiate into its traditional and expected lineages (Verfaillie et al., 2002). In the few studies that do, the frequency of this plasticity was very low. In addition, the criteria used to determine the cell type of the unexpected lineage differentiation generally included only phenotypic and morphologic characteristics, but rarely functional characteristics.
Proof of differentiation has depended on demonstration of co-expression of markers in the transplanted stem cells, such as GFP, (3-gal, the Y chromosome, or a particular antigen on the differentiated cell type, an approach that can be fraught with technical problems. Most importantly, most plasticity studies await independent confirmation from other researchers. When taken together, these problems result from the fact that no clear criteria or standards have been established to demonstrate or refute stem cell plasticity. (Holden and Vogel, 2002)
Mechanisms of Plasticity
There are several possible explanations for the perceived stem cell plasticity: 1) multiple tissue-specific stem cells are present in each organ; 2) fusion of the donor cell with the host cells; 3) stem cells are capable of dedifferentiation and differentiation into another cell type; 4) multipotent or pluripotent stem cells actually exist in adults. Present data support all four explanations with examples in nature. (Holden and Vogel, 2002)
The first mechanism, namely that stem cells for a given tissue may reside in unrelated tissues has now been demonstrated in several studies. It has been long established that HSC exit the BM and either specifically home to or are resident in various different organs. This appears to be the case for skeletal muscle as it has been shown that HSC can be isolated from skeletal muscle. Several experiments have shown that sex-mismatched bone marrow transplants in human or rodent results in the appearance of a small number of donor derived cells with the phenotype of hepatocytes suggesting transdifferentiation of HSC into hepatocytes. (Holden and Vogel, 2002)
However, at least three studies showed that liver progenitors may be present in the bone marrow. Avital et al. (2001) showed that in the rat, a population of hepatocyte progenitors characterized as Thy-1 positive and Beta2-microglobulin negative may exist. These cells can be induced to express mature hepatocyte markers and produce urea when cultured in vitro. Likewise, Fiegel et al. (2003) showed that in cultures of human bone marrow, cells with hepatocyte markers can be found, even though they did not examine functional activity of such hepatocyte-like cells. Therefore, these studies suggest the possibility that the BM contains hepatic progenitors. Consequently, when transplanted, these may be the cells that contribute to the host liver. In both instances, the apparent lineage switch would then not be caused by transdifferentiation of a single stem cell but rather caused by the presence of multiple stem cells, thus giving the perception of plasticity.
Most studies suggesting plasticity have not proven that a single cell can reconstitute the hematopoiesis and a second non-hematopoietic lineage, which has generated skepticism. Many studies tried to address clonal origin of differentiated progeny using cloning rings. This approach is not full-proof, as cells are very motile in culture and therefore one cannot fully demonstrate that single cells give rise to multiple lineages. Other studies have relied on better and more reliable methods such as single-cell sorting, or retroviral marking strategies to demonstrate single cell derivation of multiple lineage differentiation.
A second possible explanation for plasticity is that fusion of the transplanted cells with a host cell of a different lineage may occur. This would lead to the transfer of the cell contents, including proteins, DNA, and RNA from the transplanted cells to the host cell. This idea is decades old and has been studied since the early 20th century, and was then known as the heterokaryon technique. (Holden and Vogel, 2002) For example, myoblast fusion with fibroblasts induces expression of muscle proteins in the fibroblasts. This indicates that the cell cytoplasm contains factors, which induce specification, and differentiation, which is not surprising. The cloning of “Dolly” and “Copy-cat” are clear examples of this. Nuclear cloning involves the transfer of a nucleus from a somatic cell into an oocyte. It is widely known that some factors in the cytoplasm are involved in the dedifferentiation of the somatic nucleus though the specific factors and mechanisms are still not known. (Holden and Vogel, 2002)
Studies clearly demonstrated that though rare (~1/100,000-1/1,000,000), coculture of adult cells with embryonic stem cells leads to cell fusion. (Wang et al., 2003) When cocultured with ES cells, BM cells or NSC acquired ES characteristics and appeared to have transdifferentiated. On closer examination however, karyotyping and cell marking clearly demonstrated fusion. In both studies this phenomenon required a strong selectable pressure. This in vitro study suggests that apparent lineage switch may be caused by donor-host cell fusion. This phenomenon would likely be more prevalent when strong selectable pressure exists in vivo, such as in acute organ failure or tissue death. This phenomenon may also be more likely in organs that normally exhibit polyploidy such as muscle, hepatocytes, and cerebellar purkinje cells. (Wang et al., 2003)
That fusion may occur in vivo was first demonstrated by Vassilopoulos et al. (2003). Wang et al. (2003) has showed that the rescue of FAH mice with bone marrow derived cells may not be the result of the transdifferentiation of HSC to hepatocytes but the result of fusion of HSC or their hematopoietic progeny with hepatocytes. The transfer of genetic material from the normal HSC to the hepatocyte with the genetic defect resulted in hepatocytes that were able to produce the missing enzyme and consequently rescue the mice. Camargo et al. (2004) confirmed these results demonstrating that the fusogenic cell is most likely from the myelomonocytic fraction and not directly from HSC. However, their results are confounded by a recent report that demonstrated that their Cre/Lox based strategy labels both HSC and the myelomonocytic cells raising the possibility again that HSC may account for some of the fusogenic events seen in their models. (Holden and Vogel, 2002)
In another study it was shown that bone marrow derived cells contribute to adult mouse Purkinje neurons through cell fusion (Weimann, 2004). The percentage of purkinje cells that has fused increased linearly over 1.5 years suggesting a normal physiologic process. Even more surprising it appears that cell fusion at least in the case of Purkinje cells is stable (over the course of this study) resulted in reprogramming of the donor nuclei which exhibit dispersed chromatin and were shown to activate a purkinje specific transgene, L7-GFP. This is notable as stable heterokaryons (products of cell fusion without subsequent chromosome loss) have only been observed in artificial conditions in vitro (Blau et al., 1983). Moreover, cell fusion in vivo had only been observed in situations of extreme selective pressure (Vassilopoulos et al., 2003).
Studies have shown that donor bone marrow or even peripheral blood can contribute to donor derived cells in regenerating skeletal muscle. Studies used whole bone marrow as a transplant source (LaBarge and Blau, 2002). The study by Corbel et. al (2003) was the first to use a purified donor HSC (cKit+Lin-Sca1+) population for transplantation in order to determine their myogenic potential in vitro and in vivo. Camargo et al. (2003) performed a group of experiments to examine whether contribution was due to fusion. In injured skeletal muscle, 0.013% to 0.076% was derived from donor HSC while no HSC contribution was shown in uninjured skeletal muscle. Experiments using LysM-Cre/ROSAflox/STOP mice showed that HSC contribution to skeletal muscle was through fusion of a cell from a myeloid lineage with a skeletal muscle cell and not through a muscle stem cell. (Camargo et al., 2003)
Echoing these results, Nygren et al. (2004) suggest that bone marrow derived hematopoietic cells generate cardiomyocytes at a low frequency through cell fusion without any evidence of transdifferentiation. Four weeks after cardiac infarcts, 0.0065% of all cardiomyocytes were donor derived. In contrast, Wagers et al. (2002) was unable to show any contribution of HSC to cardiomyocytes following transplant into lethally irradiated animals or with animals in parabiosis. Using mice with the construct, ?-myosin heavy chain promoter driving expression of LacZ or EGFP, Murry et al. (2004) also was unable to show that HSC differentiate into cardiomyocytes in myocardial infarcts in vivo.
A third explanation is that cells can undergo dedifferentiation or transdifferentiation. A somatic cell from many mammalian species can be reprogrammed to dedifferentiate into pluripotent cells. There are many examples of this in nature such as in drosophila, in planaria, and in newt. For example, experiments with newts have shown that during limb regeneration, postmitotic cells undergo “dedifferentiation” at the site of injury, and form a cap-like structure known as a blastema (Simon et al., 1995). These “dedifferentiated” cells now can enter mitosis and are capable of reforming the limb. Extracts from blastema cells were capable of “dedifferentiating” fully formed newt myotubes and murine myotubes (McGann, 2001). When the extract was removed, these “dedifferentiated” cells expressed markers consistent with cells undergoing adipogenesis, chrondogenesis, osteogenesis, and myogenesis.
It has been suggested that the nuclear protein msx1 may play a role in blastema formation and dedifferentiation process (Simon et al., 1995). In developing mouse limbs, the expression or lack of expression of msx1 demarcates the boundary between differentiated and undifferentiated cells (Simon et al., 1995). Overexpression of msx1 in myotubes derived from C2C12 (a mouse skeletal muscle cell line) resulted in loss of expression of skeletal muscle markers and transcription factors and the regression of myotubes into smaller multinucleated myotubules or mononucleated myoblasts. It has not been examined whether these pathways examined in vitro or in vivo in Newt and mice plays a role in higher mammals.
It is possible that this process is similar to the cloning mechanism described earlier. In both cases differentiated cells adopt a more immature and less differentiated state. In nuclear cloning this has been shown to result from epigenetic changes; i.e. decreased DNA methylation and histone acetylation (Rideout et al., 2001). A similar process may partially explain the dedifferentiation and redifferentiation of myotubules. The exact mechanism involved in dedifferentiation and redifferentiation is not yet known but closer examination of these pathways may demonstrate a key role for these processes in stem cell plasticity.
The fourth explanation is the persistence of a truly multipotent or pluripotent stem cells into postnatal life. Pluripotent stem cells are exemplified by embryonic stem cells. Embryonic stem cells have been characterized based on cell surface markers including stage specific embryonic antigens (SSEA 1-4), expression of the transcription factors Oct4 and Rex1. Oct4 is a transcription factor expressed in the pregastrulation embryo, the inner cell mass, germ cells, and in embryonic carcinoma cells. When ES cells differentiate, Oct4 is downregulated. Oct4 expression is required for maintenance of the undifferentiated phenotype of ES cells and plays a major role in early embryogenesis. Another property of ES cells is the expression of telomerase, which prevents shortening of telomeres and thus allows ES cells to undergo virtually unlimited cell divisions. As ES cells are the quintessential pluripotent stem cells, some or all of these characteristics should be present. In lower animal species such as in planaria or starfish, reproduction (planaria) and regeneration (planaria and starfish) is mediated through the existence of pluripotent stem cells that can regenerate all tissues. Plasticity may be explained by the persistence of some pluripotent life beyond embryogenesis into adult life. There is ample evidence for the existence of cells with greater potency than previously appreciated; SKPs, MAPC, MSC, NSC, mesangioblast and epidermal stem cells. (Odelberg et al., 2000) However, it is important to remember that most of the evidence for these cells’ “plasticity” is in vitro which could result in a nonphysiologic process such as dedifferentiation.
Potential Uses of Adult Stem Cells
Often in the controversy of adult stem cells and stem cell plasticity, the potential opportunities for therapeutic purposes is lost. Millions of people suffer from diseases that could benefit from a cellular therapy. Tissues and organs that actively regenerate themselves from stem cells have been the best targets for cellular therapy. This maxim explains the present successes in cellular therapy. For example, hematopoietic stem cells have been used clinically to reestablish the hematopoietic system following radiation and/or chemotherapy for over 30 years. More recently, keratinocyte stem cells are being used as a source of artificial skin and the use of corneal and neural stem cells are now being evaluated in a number of studies.
This principle underlies both the success and failure of cellular therapy. Diseases resulting from a intrinsic stem (i.e. the problem lies in the stem cells themselves) would be an ideal target for cellular therapy. This explains the success of bone marrow transplant in diseases resulting in bone marrow failure such as Fanconi anemia. In contrast diseases that results from an extrinsic stem cell failure (i.e. the problem lies outside the stem cells) or results from remodeling of the stem cell environment may be a more intractable problem. For instance transplantation of hepatic stem cells in the setting of hepatitis-C-mediated liver cirrhosis may have little positive impact.
Despite these concerns it is important to examine the use of stem cells and mature cells in cellular therapy. Our appreciation for the possible capabilities of adult cells is not being realized. As early as the 1980’s several groups reported on whether BM or peripheral blood cells contribute to tissues outside the hematopoietic system. Results from these and present studies were inconclusive and mixed. For instance, many studies strongly argue that stromal cells were of donor origin while others asserted the opposite. However, more recent reports such as that by Horwitz et al. (2001) indicate that bone marrow transplantation may ameliorate symptoms of osteogenesis imperfecta such as bone brittleness and fractures. It is thought that this results from MSC engraftment and donor derived osteoblast formation, although characterization of the originating cell or what the mechanism is remains unclear.
The notion that HSC may for instance engraft and differentiate into satellite and skeletal muscle cells is exciting as it opens the possibility for treatment for diseases such as muscular dystrophy. Clinical usefulness will however require that the degree of “plasticity” is higher than the commonly reported 0.1-1%. Nevertheless, if the current published studies are correct, they would serve as the proof of principle experiment for a technology and method still in its infancy. Initial experiments with bone marrow transplant both in animals yielded poor results. Only after better appreciation for the overall mechanisms were the success rates increased. Over the next few years it will be important that the field addresses these fundamental questions and criticisms, and to develop the “plasticity” of stem cells individually to applicable approaches.