INTRODUCTION — Stem cells are those cells that have the capability of self-renewal and differentiation. First identified in the hematopoietic system, they are likely to be present in many other tissues. Stem cells have altered the care of individuals with hematologic, oncologic, dermatologic, ophthalmologic, and orthopedic conditions. The range of possible applications of stem cells to medicine extends beyond the conception of stem cells as replacement parts (table 1).
The evolving role of stem cells in clinical medicine is developing along at least three lines:
This topic will review the biology of stem cells and present an overview of their use in clinical and research settings, as well as ethical considerations that have generated much controversy.
Hematopoietic stem cells are discussed separately. (See "Overview of hematopoiesis and stem cell function" and "Sources of hematopoietic stem cells".) Hematopoietic stem cell transplantation for specific clinical conditions is discussed in multiple related topics.
TYPES OF STEM CELLS — All stem cells share two cardinal features: they are capable of self-renewal and they can differentiate. Self-renewal is the ability of cells to proliferate without the loss of differentiation potential and without undergoing senescence (biologic aging). Self-renewal does not imply that each cell division results in two exact replicas of the stem cell; daughter cells may be either stem cells or more differentiated cells. Indeed, stem cells are hypothesized to be able to divide symmetrically (in which both daughter cells are either stem cells or differentiated cells) or asymmetrically (yielding both a stem cell and a more differentiated cell) [1,2] (figure 1).
The potency of a stem cell is defined by the types of more differentiated cells that the stem cell can make (table 2). Stem cells can be either totipotent, pluripotent, multipotent, or unipotent.
SOURCES OF STEM CELLS — Classically, the potency of cells has been thought to be tied to the time of embryonic development of the organism. That is, cells that arise from the first few cell divisions following fertilization of the egg are generally the only cells that have totipotency. Pluripotent cells were thought to be limited to cells derived from either the inner cell mass of the blastocyst (a pre-implantation stage of development occurring approximately 7 to 10 days after fertilization in the human) or nascent germ cells in the embryo. Cells cultured and cell lines established from these structures are called embryonic stem (ES) cells and embryonic germ cells respectively. It is now known that pluripotent cells can arise from other cells types as well. (See 'Induced pluripotent stem cells (iPS)' below.)
Once the primitive streak forms in embryonic development (day 10 to 14 post fertilization in the human), it is thought that most stem cells are restricted to be either multipotent or unipotent. These have often been called 'adult' stem cells. If they are derived from tissue other than the germ cells, they may be called 'somatic' stem cells. Such cells would include cells derived from the cord blood, sometimes mistakenly considered equivalent to ES cells in the popular press.
Adult stem cells are thought to be present in most, but not all, tissues and to persist throughout life (figure 2). They are thought to provide the basis for tissue maintenance and response to injury. This is particularly true for tissues where there is high cell turnover, such as the blood, skin, and intestine, where stem cells have been clearly defined experimentally [3-5]. There is also clear evidence for a resident stem cell population for some tissues where the rate of cell turnover is lower (such as muscle and brain) [6,7].
However, for other tissues, including the islet cells of the pancreas, it is not clear whether adult stem cells exist and it is possible that stem cells are a durable source of replacement cells in some tissues but not in others. Division of mature cells may be the basis for the replacement after loss of some cells in other tissues. The implications of this are that mature cells are thought to only have a limited number of cell divisions, as opposed to stem cells. Therefore in settings of extensive injury, a tissue that depends upon residual mature cells to replace the injured cells may be more limited in its regenerative capacity than tissues that can rely on resident stem cells. This is one hypothesis for the failure of repair of pancreatic islet damage in type I diabetes mellitus, whereas extensive cell damage to the blood with cancer chemotherapy can result in complete restoration of tissue function.
Induced pluripotent stem cells (iPS) — The concept of the close relationship between stem cell potential and stage of development dramatically changed in 2006 . In a remarkable set of experiments, Shinya Yamanaka and his colleagues took genes that were expressed in pluripotent ES cells, but not generally in mature cells, and introduced them into mature cells. They did so in a manner such that the genes would now be “ectopically” expressed, ie, expressed in a cell type where the gene is normally not expressed. A small number of the mature cells reverted back to a highly immature cell state that resembled an ES cell. This process, now called reprogramming, induced a pluripotent state in a previously differentiated cell type (figure 3). These cells are therefore called induced pluripotent cells (iPS).
The ability to induce pluripotency has changed the landscape of stem cell biology along several lines.
While under active development, such approaches remain far from clinical application.
CLINICAL APPLICATIONS: CELL REPLACEMENT — The paradigm for stem cells as a means of replacing injured or diseased tissue was first explored in response to the threat of nuclear warfare after World War II. Research on overcoming the effects of radiation injury led to the first demonstration of stem cells, whose existence had been hypothesized since the early twentieth century. In 1963, researchers in Toronto demonstrated that a bone marrow-derived cell could replace all the blood elements and rescue an otherwise lethally-irradiated animal by simple infusion of donor cells into the blood .
This demonstration quickly led to clinical testing and application. Over the ensuing twenty years, hematopoietic stem cell transplantation has become a standard means of treating individuals with hematologic malignancies as well as clinical and acquired bone marrow failure states, including radiation injury. In 1975, it was reported that cultured cells from the skin could result in the generation of large numbers of cells sufficient to provide an autologous cutaneous barrier in patients with severe burns . Thus both un-manipulated stem cells from bone marrow and manipulated stem cells from skin have tremendous clinical utility in otherwise fatal settings.
In addition to treatment for hematologic disease and burns, stem cells are now used for bone grafting in orthopedics and corneal generation from limbal stem cells in ophthalmology. The presence of stem cells in other tissue types has led to the preclinical testing of stem cell therapies for other disorders, including those involving muscle and nerve tissue.
Preclinical examples using human ES cells — Human embryonic stem cells have been successfully differentiated in vitro into multiple cell types for therapeutic uses, including oligodendrocytes, pancreatic cells, cardiomyocytes, motor and dopaminergic neurons, and hematopoietic precursor cells . The therapeutic potential for ES cell-derived somatic cells has been demonstrated in animal models of retinal blindness, Parkinson's disease, spinal cord injury, myocardial infarction, and type I diabetes mellitus.
Retinal disease — Human ES cell-derived retinal photoreceptors have been used to improve visual functions in blind mice . Following intraocular injection, retinal cells derived from human ES migrated into the appropriate retinal layers and expressed markers of differentiated rod and cone photoreceptor cells. Subretinal transplantation of the cells into the subretinal space of mice modeling Leber's congenital amaurosis restored light responses to the animals. (See "Retinitis pigmentosa: Treatment", section on 'Retinal cell transplantation'.)
Parkinson disease — A highly enriched population of midbrain neural stem cells was derived from mouse ES cells . The dopamine neurons generated by these stem cells showed electrophysiologic and behavioral properties expected of neurons from the midbrain. These ES-derived neurons survived after being transplanted into rats with symptoms of Parkinson disease, showed appropriate electrophysiological properties, and ameliorated pathologic movements in the animals. Subsequently, functional improvement was seen with use of human ES cell-derived dopaminergic neurons after transplantation in a rat model of Parkinsonism .
Spinal cord injury — Transplantation of human ES-derived oligodendrocyte progenitor cells into adult rats, seven days after the induction of a spinal cord injury, was shown to enhance remyelination and promote improved motor function . Transplanted cells survived, redistributed over short distances, and differentiated into oligodendrocytes, demonstrating their therapeutic potential after recent spinal cord injury.
Myocardial infarction — Several studies have demonstrated that transplantation of human ES-derived cardiomyocytes improves contractile function of the infarcted mouse heart. Using a combination of pro-survival factors that limited cardiomyocyte death after transplantation, one group demonstrated decreased rates of heart failure and partial remuscularization (as evidenced by systolic thickening of the infarction wall) for animals treated with human ES-derived cardiomyocytes in comparison to control animals treated with non-cardiac human ES derivatives . Similar results in another study  highlight the potential of human ES for myocardial cell therapy strategies. Another report, however, found that the beneficial effects were transient, stressing the importance of long-term follow-up in the assessment of therapeutic benefit in animal models . (See "Genetic and cellular therapy in heart failure and myocardial infarction", section on 'Hematopoietic stem cell therapy' and "New therapies for angina pectoris", section on 'Stem cell therapy'.)
Diabetes mellitus type I — There are currently no reproducible methods for the efficient direct differentiation of human ES cells to insulin-producing beta cells, although the generation of beta cells in vitro by selecting for the expression of beta cell genes or varying growth conditions has been reported . However, the physiologic ability of human ES-derived pancreatic endoderm to differentiate in vivo into glucose-responsive insulin-secreting cells in immunodeficient mice with streptozotocin-induced diabetes, ameliorating hyperglycemia in these mice, has been demonstrated [20,21].
Preclinical studies using iPS cells — The recognition that a cell taken from an individual can be induced to become pluripotent (a cell type capable of forming any other cell type in that individual's body) provides unprecedented opportunities for regenerative medicine. Easily accessible patient cell types, such as skin fibroblasts or blood cells, can be reprogrammed to iPS. These pluripotent cells are envisioned to then be differentiated into mature cells that may be deficient in diseases such as islet cells for type I diabetes, or into tissue-specific adult (stem) cells. The cells can theoretically be used for tissue regeneration in the same patient (figure 4).
In such a way, iPS cell technology could overcome two important obstacles associated with human ES cells: immune rejection after transplantation and ethical concerns regarding the use of human embryos. The approach would be particularly powerful in monogenic diseases, where ‘patient-specific’ iPS can be generated, the diseased gene in those cells potentially corrected, and the gene-corrected cells transplanted to restore organ function. This field is rapidly evolving, although there remains a great deal of research needed to make such a schema viable in the clinical setting.
Many of the studies to evaluate iPS cells as a source of cell replacement are being conducted in rodents. These studies permit testing of basic principles in a physiologic setting and are therefore useful, particularly where murine models of human diseases have been generated. Two such contexts are Parkinson disease  and hemophilia A . In both of these settings, iPS cells were generated and used to differentiate cells that could then be transplanted for therapy.
Other technologies developed for gene therapy have also been used to test whether iPS cells can be useful to create gene corrected cells for transplantation. One proof-of-principle study used a mouse model for sickle cell disease where the hemoglobin genes of the mouse contained the mutations known to cause red cell sickling in human sickle cell anemia. Investigators generated iPS cells from skin cells of the mouse and introduced a normal hemoglobin gene allele to replace the sickle hemoglobin gene (this can be done by gene targeting technology used to engineer mouse strains for research) . They then generated blood stem cells from the gene corrected iPS cells. Transplantation of these cells into sickle cell anemic mice improved the disease phenotype. This approach is a model for how iPS may be used as cell therapy in the setting of genetic disorders. If this could be applied to humans, correcting genetic disorders in iPS cells might then yield a cell type for transplantation.
This approach was extended to humans, using fibroblasts from Fanconi anemia patients that were corrected for the diseased Fanconi gene . The normal (non-diseased) Fanconi gene was transferred into the fibroblast cells using a virus specifically engineered for transferring genes into cells. The gene-corrected cells were reprogrammed to pluripotency to generate patient-specific iPS cells. The iPS cells now carrying a normal Fanconi gene were tested for their ability to become blood cells in tissue culture outside the body. The cells with the normal Fanconi gene were able to generate blood cell types in a manner similar to cells from normal volunteers. This study indicated that skin cells from patients with an abnormal gene (as in the setting of Fanconi anemia) can be used to generate gene-corrected iPS cells that might serve as a source of cells for transplantation. Therefore, we are edging closer to demonstrating the feasibility of using iPS cells as therapy.
Challenges in the clinical use of stem cells as replacement — The use of stem cells to replace abnormal or missing cells is conceptually compelling. The successful use of stems cells for bone marrow and skin therapies, and the identification of stem or progenitor populations in a number of tissue types, offers great promise for expanded clinical applications. There is the additional potential that human pluripotent cells may be a source of virtually any cell type. However, several complexities must be addressed before widespread application becomes feasible.
The first issue is how transplanted cells will integrate into surrounding tissue to achieve a physiologically beneficial effect. This is of particular relevance where coordination of complex networks of cells is essential, such as in the heart and brain where aberrant circuits can result in serious adverse events. This challenge will require extensive further testing. One encouraging observation is that some cells appear to have an inherent capacity to incorporate into existing structures. An example is that, in an animal model, human endothelial cells derived from ES cells were able to organize into tubular structures and integrate into the host vasculature when inserted as dispersed cells into a host tissue .
A second concern is the potential for transplanted cells to form tumors. This is of particular importance when using pluripotent cells, since these are characterized by the ability to form teratomas (neoplastic tumors containing cells corresponding to all three embryonic layers) in animal models [8,26]. Thus, the differentiation state of transplanted cells will need to be defined with high precision to avoid delivery of residual pluripotent cells that may differentiate aberrantly in vivo. Evidence also suggests that this issue is not restricted to pluripotent cells. In one case, a child was given cultured fetal brain cells intrathecally and subsequently developed multiple CNS tumors of donor origin . The culture process may enable the outgrowth of genetically abnormal cell types that could be of potential danger. The specifics of what types of testing will be required for pluripotent or other cells to assure genetic integrity of transplanted cells is an active area of consideration by the field.
The issue of malignant potential of cells is of greatest concern in iPS cells. The methods of greatest efficiency for reprogramming cells are currently retrovirus or lentivirus-based, and therefore run the risk of mutagenesis by virtue of viral integration into the host genome. In addition, some of the genes used to induce reprogramming have known oncogenic potential (eg, c-Myc) . Progress in reducing the number of gene products needed for reprogramming, and in the use of either non-integrating viruses or small molecules to supplant retrovirus-based reprogramming, is ongoing [29-33]. These developments may mitigate the concerns about insertional mutagenesis, but will not entirely assuage concern for altered growth control of modified cells, particularly those with pluripotency.
A third concern is the ability to direct the differentiation state of the cells to be used. Generating the proper cell type from pluripotent cells remains a significant challenge for some cell types. Protocols have now been devised to create some neural cell types of clear clinical importance . However, for other tissues, such as the blood, the cell types created most closely resemble embryonic blood cells and are not capable of engrafting the bone marrow without further and undesirable genetic manipulation .
Achieving the right stage of differentiation is another consideration in development of the stem cell derived cell therapies. It may be most desirable to generate progenitors, rather than fully mature terminally differentiated cells in some tissues so that the replaced cells do not quickly senesce and die.
Finally, the uniformity and consistency of the cell product may be difficult to achieve across different donor sources of cells.
These hurdles are not likely to be prohibitive, but each will require considerable effort and attention.
OTHER CLINICAL APPLICATIONS
Disease modifiers — Beyond use of stem cells for cell replacement, the ability of certain stem cells to alter disease without engrafting has been most extensively explored as a means of modifying cellular response to injury or aberrant immune activity. Such non-engrafting stem cells are hypothesized to provide complex signals, affecting disease outcome without directly replacing injured cells.
The identification of a population of cells in the bone marrow that could form a number of mesenchymal cell populations ex vivo led to the concept of a mesenchymal stem cell [36,37]. This population of cells is defined primarily by its ability to form colonies in tissue culture, and by the ability of single cells derived from the colonies to differentiate into osteoblasts, adipocytes or chondrocytes in vitro. Identification of these cells in the body, and understanding of how they function in vivo, remains controversial.
The potential for these mesenchymal stem cells to form muscle led to exploration of their use in the setting of ischemic injury of the heart. While initial indications suggested that the cells may directly contribute to generating cells in the area of injury, it has subsequently been shown that the cells do not engraft . Rather, it is proposed that they provide complex signals, in a paracrine manner, that may alter the ability of the heart (and other tissues) to respond to ischemic injury [39,40]. This use of a stem cell population remains highly contested despite numerous clinical trials in various settings, in part because the mechanisms by which they work are not clear. The mesenchymal stem cell populations studied are highly variable and have been variably reported to secrete a number of factors including indoleamine 2, 3-dioxygenase, prostaglandins, interleukin-6, hepatocyte growth factor, inducible nitric oxide synthase, and tumor growth factor (TGF)-ß1 . Which of these might participate in particular clinical settings is difficult to discern.
It has been proposed that the mesenchymal stem cell population is capable of altering immune function and may therefore modulate injury responses as in ischemia, or mitigate the effects of immune mediated diseases. As an example, mesenchymal stem cells have been tested in clinical trials in humans in settings of graft versus host disease following allogeneic hematopoietic stem cell transplantation, and in inflammatory bowel disease. For each, there has been controversy regarding the impact of the cells and the mechanism by which the cells may be acting. (See "Treatment of acute graft-versus-host disease: Clinical trials", section on 'Mesenchymal stem cells' and "Investigational therapies in the medical management of Crohn's disease", section on 'Stem cell therapy'.)
While this use of stem cell populations as disease modifying cell platforms is an area of active investigation, the concept should not be considered proven at this time. The prospect that mesenchymal cell populations may have the ability to arrive at sites of injury and provide paracrine signals is an exciting and potentially important application of stem cell biology, but it is not yet a defined therapy.
Drugs targeting endogenous stem cells — In addition to cells being introduced into the patient as therapy, there is emerging recognition that stem cells resident in adult tissues may be targeted pharmacologically. In this context, the therapeutic intervention is drug-based with the goal of altering the function of endogenous stem cells. This approach may be less likely to engender the issues related to tissue integration that are concerns with exogenously applied cells.
Growing understanding of the signaling processes that induce activation of stem cells, or modify their differentiation, makes this approach one of growing interest.
Analogous to the use of erythropoietin to modify the activity of red cell progenitors, agents with similar effects on stem cells might generate reconstitution of a broader number of cells in the case of hematopoietic stem cells, and of other mature cell types in other tissue settings.
Some studies have identified agents capable of altering stem cell activity in settings of injury to the bone marrow or the bone [42,43]. One study has identified an agent (a derivative of prostaglandin E2) that is being tested for the ability to improve stem cell regeneration of hematopoiesis in the treatment of hematologic malignancies .
Whether other agents can be identified that can alter the ability of stem cells to regenerate cell lines after tissue injury is to be determined.
Disease models — Stem cells offer the possibility of creating in vitro disease models that may improve molecular understanding of disease and accelerate the development of new therapies. Human cells from affected individuals with diseases affecting the nervous system, for example, have been extremely difficult to obtain for in vitro analysis except through the use of stem cell technologies. With stem cell approaches, it is now possible to generate sufficiently large numbers of cells to be able to investigate molecules involved in the functional deficits associated with the disease and to use the cells to develop high through-put drug discovery strategies.
The development of iPS cells has led to a number of efforts to create in vitro disease models amenable to genetic and small molecule study. A major goal of such efforts is to identify compounds capable of altering the progression of cell events corresponding to disease progression. These efforts depend upon robust cell culture systems that resemble the in vivo context closely enough to be useful for drug evaluation. Fulfillment of this requirement still needs validation.
Two examples depict use of iPS cells of to create a disease model for further study of human disease pathogenesis and treatment.
An additional application of stem cell biology to disease modeling is in the setting of cancer. In this context, the stem cell paradigm has been applied to consider cancer through the perspective of a tissue organized as a self-renewing subpopulation that generates diverse more differentiated daughter cells. The model hypothesizes that there is a definable population of malignant cells that can be demonstrated to have the two cardinal features of stem cells (self-renewal and differentiation capacity). These cells, like stem cells in normal tissues, would replace more mature cells whose lifespan is limited. They are therefore hypothesized to be the cells that enable persistence and perhaps metastasis of tumors.
The hypothesis has gained experimental support through the use of immunodeficient mice in whom human cancer cells can be engrafted [47,48]. A subset of tumor cells can be defined in these animals that is capable of engrafting the tumor, while other subsets cannot. Further, the tumor-initiating subset can be sequentially transplanted to initiate new tumors and can yield the full diversity of cells observed in the original malignancy, thereby fulfilling the experimental definition of stem cells. Human acute myelogenous leukemia, breast, colon, ovarian, pancreatic, head and neck cancer, and malignant glioma have been shown to have such a subpopulation of cells [47-53].
Whether all tumors have a stem-like cell is still controversial, and much of the definition depends upon engraftment in a mouse model, which is not necessarily a surrogate for function in humans. However, the model is beginning to have an impact on the way cancer is viewed and how oncologic drugs are developed. Studies are now investigating whether drugs affect the stem-like cells and not just tumor bulk, and whether agents can target distinctive features of the stem-like cells of cancer rather than the mixture of cells comprising a malignancy.
ETHICAL CONSIDERATIONS — The first derivation of human embryonic stem and embryonic germ cells in 1998 sparked enormous interest and controversy regarding pluripotent stem cells. The initial, and still most frequent, techniques used to derive ES cells disrupt the blastocyst from which the cells are derived, raising concern that an early human life form was being destroyed. The source of the blastocyst was generally discarded material from in vitro fertilization clinics. Nonetheless, for many the issue of using the material for research, even if that research was intended to assuage human suffering, was morally unacceptable. These controversies led to policy decisions among some nations that precluded ES cell research or, as in the case of the United States, federal funding to support such research.
The advent of reprogramming to generate iPS has provided an alternative source of pluripotent cells that has satisfied the concerns of many. It should be recognized that iPS have not yet been shown to be fully equivalent to ES cells, and ES cells still represent the best defined pluripotent cell population. However, iPS cells are increasingly the focus of research and appear to engender less ethical concern since they do not involve the use of blastocysts or embryonic tissue. The cells from animals do have the ability to contribute to a developing organism though, so other ethical concerns are not entirely mitigated.
An issue that is relevant for both pluripotent and other stem cells is how to define safety when proceeding to clinical trial. The area of cell therapies has been well explored in hematology, transfusion medicine, and hematopoietic stem cell transplant. Infusion of other cultured cell products has a well-defined safety profile for keratinocyte skin grafts and mesenchymal stem cell transplants. Several features are unique to the use of pluripotent cells, however:
Therefore, clinical experience will be necessary to define some of the risks in stem cell-based therapeutics, and careful design of clinical trials with a range of parameters specific to the field are in process.
Stem cells have enormous potential that is well recognized by the public, and are a source of hope for those dealing with otherwise untreatable medical problems. However, the field is in its infancy and progress is painstaking; for those desperate to get help, unproven therapies even outside of reputable centers or clinical trial are a highly tempting option. The potential for patient exploitation raises another extremely challenging area. There has been a profusion of centers in many parts of the world, often offering treatment without clear definition of what cells would be used, what the source of cells is, and what the full experience has been. Services are often offered for cash payment and with unclear provisions to monitor safety or respond to adverse events. Patients and providers should ask a range of specific questions about such therapies. The International Society for Stem Cell Research (ISSCR) has provided such suggested questions and other information that may be of use (www.isscr.org).
SUMMARY AND RECOMMENDATIONS