In the 1960s, Till and McCulloch and colleagues provided genetic marking evidence that in mouse bone marrow there are rare cells that can form myeloerythroid colonies in the spleens of irradiated mice, some of which contain cells that can self-renew.^{6 - 7} It was reasonable to propose the existence of HSCs from these experiments,^{8 - 9} although later experiments showed that the colony-forming cells were defined myeloid progenitors.¹⁰ The proposal that HSCs probably exist set the stage for prospective isolation of these cells.

The isolation of HSCs required several conceptual and technical advances, including monoclonal antibodies to cell surface proteins,¹¹ high-speed multiparameter cell sorters,¹² and the establishment of clonogenic assays for all blood cell lineages.^{13 - 17} In a preclinical setting, the cells were shown to regenerate the blood-forming system of lethally irradiated mice.^{1 ,17} These assays eventually led to the prospective isolation of clonogenic mouse¹ and human¹⁸ HSC populations that self-renew and include in their clonal progeny all blood cell types.^{19 - 20} The downstream progenitors in the hematopoietic lineage in mouse and humans have also been isolated.^{3 - 4 ,21 - 23}

The same method was used to isolate a human central nervous system stem cell (hCNS SC) population²⁴ which, at the single-cell level, could give rise to spheres of neural cells that can differentiate to oligodendrocytes, astrocytes, and neurons, as well as self-renewed hCNS SCs.²⁵ Implantation of these cells into the lateral ventricles of newborn immunodeficient mice brains led to site-appropriate seeding of neurogenic zones (subventricular zone of the lateral ventricles and the dentate gyrus of the hippocampus²⁵ ) by these stem cells, where they self-renew for the life of the mouse. Their daughter cells migrate to sites where they differentiate into the neural cell types and the specific tissue architecture specified by the mouse brain regions they occupy.

These results revealed an unexpected aspect of stem cell biology—the cues for self-renewal, site-specific migration, and site-appropriate differentiation and placement into the microarchitecture, at least for brain stem cells, appear largely to be conserved between mouse and humans. This becomes important in studying preclinical capacities of stem cell therapies, for understanding the behavior of brain cancer stem cells, and for neurobiology. Investigators can now study human neuronal cells in the context of the mouse brain, which should be valuable in the fields of neurodevelopment, function, and perhaps neuropathologies. These neural stem cells can be genetically modified, allowing the opportunity to study particular genes in human neural cells in situ for their role in neurological functions.

Stem cell isolation is now about 17 years old, yet only a few tissue stem cells from mouse or humans have been prospectively isolated to date, and really only HSCs have been transplanted in humans to regenerate any tissue. Yet tissue and organ transplants have shown the need, and the feasibility of regenerative medicine, and so the field has a long way to go before we understand and exploit their potential.

Most clinical hematopoietic cell transplants now use mobilized peripheral blood (MPB), which results in rapid (10- to 13-day) engraftment; this is due to increased numbers of HSCs in MPB vs marrow.^{26 - 27} Unfortunately, because the clinical community has accepted the term “stem cell transplants” to include a variety of hematopoietic transplants, only the most savvy oncologist will recognize the difference between unpurified mobilized blood cell transplants and HSC transplants. The following proposed terminology may help clarify this problem: all transplants could be called HCT for hematopoietic cell transplantation. Mobilized peripheral blood could be called MPB; bone marrow, BMT; umbilical cord blood, UCB; CD34⁺-enriched transplants, CD34 HCTs; and true stem cell transplants by their identifying characteristics, eg, CD34+Thy1+HSC. That way the reader can know what was actually done, rather than requiring credentials in stem cell biology.

For treatment of patients with cancer with their own hematopoietic cells following myeloablative chemotherapy, only purified HSCs were free of cancer cells,²⁸ and these were used in several clinical trials.^{29 - 32}

In the first human transplantations between HLA-matched siblings, the donor T cells that are present at high levels in bone marrow and MPB (and UCB) recognize the host as foreign and carry out a potentially lethal graft-vs-host disease (GVHD). In mice, rapid and sustained engraftment with pure HSCs could be accomplished without GVHD.^{33 - 36} In allogeneic transplants between major histocompatability complex (MHC)–matched (HLA in humans, H2 in mice) but otherwise genetically distinct pairs, the main requirement for successful engraftment is temporary lymphoablation of T cells.^{33 ,35} Even unirradiated immunodeficient mice can achieve 0.1% to 1% donor cells with HSC transplants (D. Bhattacharya, D. Rossi, D. Bryder, I.L. Weissman, unpublished data, 2005). But when the donor expresses unshared MHC alleles, one must also eliminate host natural killer cells,^{35 ,37} which can kill or reject tissue cells.³⁸ In mice monoclonal antibodies that eliminate host natural killer cells are essential for partial or unmatched HSC or HCT engraftment. Using sublethal irradiation plus anti-T and anti–natural killer antibodies, a “safe” regimen, mice can be transplanted with pure HSCs and be lifelong chimeras without host rejection of the graft or GVHD.^{33 - 35}

Purified hematopoietic progenitors can also be useful in particular circumstances. For example, mice exposed to murine cytomegalovirus, Aspergillus fumigata, or Pseudomonas aeruginosa in the immediate post-HSC transplant period die rapidly because their immune defenses are weak. Cotransplantation of common lymphoid progenitors with HSCs blocks murine cytomegalovirus mortality, even if the common lymphoid progenitors are from fully allogeneic donors.³⁹ Similarly, cotransplantation of common myeloid and granulocyte-monocyte progenitors precludes lethality with Aspergillus or Pseudomonas, again even if the donor is fully allogeneic.⁴⁰

In another example, whole body just-lethal irradiation or exposure to myeloablating chemicals is possible through neglect, intention, and war. The only sure way to be saved from doses of radiation that cause hematopoietic failure, but not irreversible gut damage, is to be transplanted with sufficient numbers of HSCs, requiring some kind of HLA match. However, some progenitors in mice can be radioprotective until the rare surviving host HSCs can regenerate the system.¹⁰

While hematopoietic regeneration is a well-developed field, CNS regeneration is still only experimental. Given that human CNS stem cells engraft and migrate widely in site-appropriate manners in immunodeficient mouse brains (and presumably human brains), several experimental models of neural repair are ongoing. Patients and mice with lysosomal storage diseases, such as Batten disease, undergo both systemic and neural degeneration. Provision of the missing enzymes systemically can result in uptake of the enzymes by diseased cells in the body, but not the brain, and ameliorate systemic disease. Transplantation of hCNSSC–derived populations into mice affected with Batten disease results in amelioration of the neurodegeneration in all parts of the brain.⁴¹

Compression injuries of the spinal cord often result in paralysis following local inflammatory events, and among the pathological hallmarks are areas of cord demyelination.⁴² Transplantation of hCNS SCs/neurospheres into immunodeficient mice that had a controlled crush injury at T9 about 9 days after the injury led to cell engraftment–dependent recovery of hindlimb paralysis and coordination.⁴² The engraftment was mainly oligodendrocytic, resulting in effective remyelination, and sustained presence of the graft was required for sustained recovery.⁴²

In these 2 examples neuroprotection was afforded by injected stem cells and depended on their ability to differentiate, migrate appropriately, and function. It is yet to be determined whether regeneration of neural circuits by replacement with cells derived from CNS SCs can occur, and of course, it is yet to be determined which neurodegenerative diseases sustain neuron loss by direct effects on the neurons vs their nurturing environments.

When HSCs engraft in fully myeloablated/lymphoablated mice, the blood-forming and immune systems are largely, if not completely, derived from the donor. Donor cells that enter the thymus give rise to T cells as well as donor-derived antigen-presenting dendritic cells. The dendritic cells, along with the host thymic medullary epithelial cells, delete developing T cells with reactivity to self-proteins that could be expressed in any tissue or organ from donor or host. Thus these HSC chimeras produce T-cell populations that cannot make immune reactions against donor or host but are capable of making other protective immune reactions for the host. As a result these reconstituted mice are permanently transplantation tolerant of grafts of any cell, tissue, or organ from the HSC donor.³⁷

Tolerance induction for tissue and organ grafts should eventually be followed by cotransplantation of HSC- and tissue/organ-specific stem cells from the same donor source.^{34 ,37} This might be achieved not only from living donors but eventually from classic embryonic stem cell lines or from donor-specific nuclear transfer stem cell lines.^{43 - 45}

These possibilities could usher in the era of regenerative medicine, in which curative intent regeneration of diseased organs and tissues can be achieved with stem cells, rather than chronic support with drug therapies. However, thus far the only cell of choice for these regenerative medicine therapies are the self-renewing tissue stem cells rather than more transient progenitors or actual mature, functional cells.