|Home | About | Journals | Submit | Contact Us | Français|
Stem cells, by definition, are capable of total or partial differentiation as well as self-renewal. The most primitive stem cells, in many people's opinion, are those derived from the blastocyst inner cell mass1. Such embryonic germ cells can be maintained in culture for a long time if an appropriate cytokine, leukaemia inhibitory factor, is present in the culture medium. They remain pluripotent; but, if the cytokine is removed, they can be modulated to produce various cell lines, from blood islands to neurons and epithelia2. These cells therefore meet the criteria of being able to divide symmetrically to expand their number and to divide asymmetrically in order to self-renew and give rise to a differentiated progeny3. However, there is no evidence that blastocyst cells self-renew in vivo and obviously they are not working over the entire lifespan of the organism, as stem cells should. Therefore they cannot be regarded as true stem cells, while among so-called germ-line stem cells (oocytes and sperm-producing cells) only male spermatogonial cells are present in mammals over the lifetime: oocytes are produced in a finite number near the time of birth. I confine this review to cells that can self-renew for the entire lifespan of the organism.
During embryonic development pluripotential stem cells become progressively more restricted, giving rise to stem cells with narrower commitments. Somatic stem cells appear relatively late in development: haemopoietic stem cells are seen first in the yolk sac, then in the paraortic region, later in fetal liver, finally in spleen and bone marrow. In irradiated embryos, repopulating stem cells tend to increase successively through the same sites (yolk sac, liver, marrow)5. Adult stem cells are often situated in ‘niches’ where they are available to start a differentiation pathway; the environment of such niches seems able to exert a critical influence on their biochemical and developmental potential. For instance, when adult haemopoietic stem cells from mouse bone marrow are injected into the inner mass of the mouse blastocyst, they express fetal rather than adult haemoglobin6.
Evidence for the presence of adult stem cells in the central nervous system dates back to the 1960s, when postnatal neurogenesis was found in rat and guinea-pig hippocampus7. Further research revealed neural stem cells in other regions such as the forebrain ventricular wall: in the subventricular zone, facing the lumen, a layer of proliferating subependymal cells may represent either a stem cell compartment or a stream of progenitor cells in transit8,9. It is therefore possible that such cells serve as an endogenous source for new neurons and glia in the adult mammalian forebrain. A potential for neuron generation also seems to exist in the striatum, thalamus and hypothalamus10.
What is the function of these persistent stem cells in the adult brain? A distinct possibility, so far overlooked, is slow and continuous replacement of neural cells: sphere-forming cells, a population of neural precursors produced at various stages of embryo development, may persist as a ‘leftover supply’ in the brain and other tissues until the necessity arises to proliferate and differentiate11. Very recently, Rietze et al.12 reported that stem cells purified from mouse adult brain appeared capable of generating either neural or non-neural cells. It has long been known that neural stem cells immortalized in culture can be induced to yield both neurons and glial cells13. Clearly, therefore, central nervous system stem cells, preserved in culture, offer a source of replenishment for depleted tissues14. We should also note the use of fetal cells which have undergone initial neural differentiation for treatment of Parkinson's disease. This method, however, requires so much fetal tissue that it is impracticable on a large scale; therefore, for this purpose too, the focus is on stem cells in culture15.
Many classes of haemopoietic stem cells have been identified, capable of giving rise to particular sets of blood cells: among the earliest to be proposed were the so-called CFU-S (colony-forming units spleen), capable of repopulating the entire haemopoietic system in lethally irradiated mice. An analogous class of cells in human beings, including a subset of long-term self-renewing cells, has been shown able to reconstitute haemopoiesis in patients who have received lethal doses of radiation or chemotherapy16,17.
At a less primitive level there are oligopotent stem cells that can produce either a common lymphoid progenitor or a common myeloid precursor, which in turn may give rise to an erythroid or megakaryocytic or granulo-monocytic precursor. Furthermore, so-called colony-forming cells can produce colonies of a restricted type, such as BFU-Es and CFU-Es for cells of the erythroid lineage, CFU-GMs for granulocytes and monocytes, and CFU-MKs for megakaryocytes11. A stem cell acting as a lymphoid progenitor has also been identified19.
The potential of such haemopoietic stem cells has been exploited clinically, not only in blood neoplastic disorders but also in genetic diseases of the immune system20 and now treatment of solid tumours21 (see later).
Adult stem cells can be highly versatile, capable of modulation from one type to another. Neural stem cells, for instance, can transform under selective stimulation into blood cells of various sorts22, and haemopoietic stem cells injected into irradiated hosts have yielded glial cells25 and hepatic oval cells24 among others. Human mesenchymal stem cells have differentiated in culture into adipocytes, chondrocytes and osteocytes25. This wide spectrum of possible differentiation illustrates the impressive potential of early stem cells taken out of their microenvironment. If one wishes to achieve differentiation away from the expected lineage, how crucial is culture and how long must it continue to allow dedifferentiation and a subsequent shift of lineage? Lengthy passage in culture does not seem essential: for instance, marrow stromal cells grown for only five days and injected into the forebrain of neonatal mice were able to differentiate into mature astrocytes, and possibly into neurons as well26.
Haemopoietic stem cells can now be purified to such an extent that single bone-marrow-derived cells are available. Seemingly, one such cell can not only achieve long-term haemopoietic repopulation of an irradiated host but also differentiate into non-haemopoietic elements such as epithelial cells of liver, lung and gastrointestinal tract27.
Transplantation of human haemopoietic stem cells, obtained from various sources, is now established practice for certain disorders28. To obtain sufficient quantities, these cells are usually exposed in vitro to a cocktail of cytokines29.
How frequent are such early clonogenic cells in blood or blood-forming cells? The exact number is still uncertain, but a comparison of bone marrow, peripheral blood and cord blood showed cord blood to be the richest source and peripheral blood the poorest30. Cord blood clonogenic cells expand rapidly in vitro and are already used in treatment of genetic disorders of haemopoiesis and leukaemia31,32.
It seems important to avoid early depletion of the stem cell pool in culture and attention has been focused on the action of negative regulators of haemopoiesis, such as macrophage inflammatory protein 1-alpha (MIP-1-alpha) and tumour necrosis factor alpha33. MIP-1 alpha and leukaemia inhibitory factor protect the repopulating ability of purified haemopoietic stem cells, and preliminary clinical trials to test tolerance to MIP-1-alpha are in progress34.
A new approach is to use a preparative regimen which though not totally myeloablative allows engraftment of transplanted allogeneic cells. This can result in a pronounced graft-versus-tumour effect, and early results are encouraging. 10 of 19 patients with metastatic renal cell carcinoma showed a clear response35.
As regards the central nervous system, a widely used model in animals is the lesioned striatum, in which cells at different stages of maturation have been inserted36. Virtually immortal human nervous stem cell lines are available, and when implanted into the striatum of immuno-deficient mice were detectable in the host parenchyma up to a year after grafting14. Implantation of dopamine-producing cells has been tried not only in mice but also in patients with Parkinson's disease37; transplants of embryo-derived cells as well as adult neural stem cells have also been successfully performed in laboratory models of demyelinating diseases such as multiple sclerosis38,39.
Lately, further evidence of the surprising potential of adult neural stem cells, pointing to a ‘repertoire’ of differentiation very close to that of embryonic cells12,40, has provided further stimulus to work in this area41. Preliminary clinical trials are contemplated on replacement of damaged cartilage or repair of injured tendons. Some groups are trying to grow neural stem cells for transplantation into the brain or spinal cord of patients with severe central nervous system damage14,42,43; modulation by growth factors of cells derived from human microspheres (the early aggregates of neural progenitors) could provide the almost unlimited supply of enriched non-genetically-transformed neurons required for transplantation studies44.
Stem cell transplantation and modulation has potential applications in various other clinical conditions. The turning point has been the recognition and identification of stem cells in several different tissues, with a plasticity that allows transdifferentiation: bone marrow cells, for instance, may be capable of yielding epidermal cells, skeletal muscle cells and even hepatic oval cells (precursors of mature liver cells)24.
When stem cells are introduced in a new ‘niche’, they are believed to undergo a process of reprogramming and differentiation in response to signals promoted from the new micro-envronment45. The nature of such signals, however, remains obscure. It is noteworthy that stem cells from different tissues, such as bone marrow and muscle exposed to a given culture environment, yield cells with similar properties. We already know that areas of muscle regeneration show an influx of progenitor cells from other sources, and in immunodeficient mice transplanted marrow-derived cells can migrate into such areas, differentiate and give rise to muscle fibres47. This suggests a possible means of treatment for degenerative disorders such as muscular dystrophies. Another observation of great interest is that bone marrow stem cells can lead to myocardial regeneration in mice with experimentally induced infarction48; attempts to isolate the responsible cells are in progress.
Another ready source of adult stem cells with high proliferation potential is the skin. Such cells may prove useful not only for repair of skin lesions but also, after reprogramming, for transplantation into other tissues or organs49.
The finding of high plasticity in adult stem cells demands revision of the stem-cell concept50. When the identity of cells capable of repairing damaged tissues is scrutinized, we see that not all stem cells are equivalent; in the central nervous system, for example, most neural stem cells are regionally and temporally restricted, although it is possible to find and isolate rare stem cells which are capable of producing diverse cells types51.
In future, therefore, it will probably be necessary to use stem cells from many different sources, for the repair of damaged tissues and organs. The main limiting factor at present is the poverty of information on signals in host tissue that promote optimal homing and activation of transplanted cells. A promising development, in this context, is a very recent approach whereby a synthetic matrix and controlled-release microparticles are assembled with the progenitor cells before transplantation, thus creating an artificial microenvironment: within such ‘neo-tissue’ each microparticle is designed to supply agents that promote certain aspects of cell function52. This whole area of research is showing a welcome surge of activity53.