Dr. Orkin began by noting that the transfer of hematopoietic stem cells (HSCs), which is the basis for bone-marrow transplantation (BMT), is an established and successful therapeutic modality. In the United States alone, 15,000 transplants are performed each year. To provide a historical context for this approach, BMT was first carried out in dogs in 1956. Mouse embryonic stem (ES) cells were developed in 1982, and human ES cells were discovered in 1997.

Hematopoietic cells reside in the marrow and can give rise to the entire hematopoietic system upon transplantation. In the mouse, transplantation of a single cell is sufficient to produce new tissue. However, blood formation takes place at successive sites (e.g., yolk sac, placenta, bone marrow) during development, and each site has a unique milieu. Moreover, HSCs found at different sites differ in their properties (e.g., cycling versus quiescence). Bone marrow is the site of adult hematopoiesis. Cellular interactions keep stem cells quiescent, although a variety of methods can mobilize cells. These interactions are critical to maintain HSCs in a quiescent state and to permit expansion as needed. The "bone marrow niche" likely represents several niches; cells found in the bone marrow reside in either the vascular or osteoblast niches. Although HSCs circulate at a low frequency and can be harvested from peripheral blood after mobilization, the bone marrow niche has not been recreated ex vivo.

Two types of transplants, autologous and allogeneic, have been successful. Autologous transplants are used for BM replacement after high-dose chemotherapy, whereas allogeneic transplants are used to treat acute leukemias and other chronic blood cancers and syndromes characterized by bone marrow failure. Cells used in autologous transfers are provided by cryopreserved harvests of autologous BM or peripheral blood. For allogeneic transplants, material may be provided by matched family members, unrelated donors, umbilical cord blood, and mismatched family members. Matching a sibling donor decreases the likelihood of graft rejection and graft-versus-host disease (GVHD), and readily available family members tend to be younger and available for subsequent product if needed. Disadvantages of allogeneic strategies include the low incidence of matching sibling donors (fewer than one-third of all cases) and the possibility of graft-versus-leukemia disease in matches that are very close. Hematopoietic transplantation from an unrelated donor is the standard route if matched tissue is unavailable, although many donors lack sufficient high-resolution typing. Moreover, the likelihood of locating an acceptable donor is linked to ethnicity, and some individuals with rare human leukocyte antigen (HLA) types have no acceptable donors available. Umbilical cord blood also contains HSCs in adequate numbers for engraftment, which offers the advantages of a decreased time to donor availability, a greater tolerance for mismatch, and less GVHD. Cord blood can be harvested at the time of delivery and stored for personal use or donated to cord-blood banks. However, cord-blood transplantations often feature an increased incidence of graft failure, an insufficient amount of donor tissue relative to the size of the recipient, and the inability to reassess the donor for additional product. Furthermore, donor cells are naive against infectious agents, and the long-term health issues of the donor are not known. Preimplantation genetic diagnosis (PGD) can be used to create an HLA-matched sibling who can be "disease-free" if mutations are identified. While this procedure is relatively accurate, it is expensive and time-consuming (e.g., it requires nine months to gestate). Clinical challenges that affect the use of HSC transplantations thus include donor availability, prevention of GVHD, therapy for refractory GVHD, immune reconstitution, and amelioration of long-term sequelae.

These challenges accompany issues in the laboratory, including expansion of HSCs ex vivo and the reproducible generation of HSCs from human embryonic stem (ES) or induced pluripotent stem (iPS) cells. Expanding HSCs ex vivo without the loss of stem cell properties will greatly facilitate BM transplantation and improve somatic gene therapy. To date, however, the extent of achieved expansion is quite modest (if any). However, new growth factors (e.g., angiopoietin-like molecules), developmental growth factors, and pharmacologic agents show promise for improving this expansion. Mouse and human ES cells can also differentiate ex vivo into hematopoietic cells, although these products are more typical of embryonic/fetal cells than adult blood cells. To date, there is little evidence that authentic adult HSCs have been produced in vitro, although the addition of specific transcription factors may render this process more viable.

Differentiation is programmed by transcription factors that drive the choice of lineages, and individual transcription factors can be used to drive differentiation toward a specific fate. Dozens factors likely play roles in this process. For the hematopoietic system, colony assays are available for all progenitor cell types, and HSCs and progenitor cells can be isolated prospectively using monoclonal antibodies. However, this process is relatively complex and nuanced. There have also been reports in the literature of HSCs giving rise to muscle, liver, and other tissues upon transplant. This process would be an example of transdifferentiation, or the differentiation of one cell type into a mature tissue outside of its programmed lineage options. Dr. Orkin noted that in the absence of any new data to the contrary, it is highly unlikely that HSCs transdifferentiate into any non-hematopoietic cells or give rise to any cells that do so in the absence of cell fusion. He observed that blood gives rise only to blood under normal conditions. Often, stem cell "plasticity" reports use heterogeneous, uncharacterized cell populations that include HSCs along with many other cell types.

In conclusion, Dr. Orkin noted that BMT is a highly successful modality for replacing the hematopoietic system in patients with a variety of disorders. However, challenges remain in improving the genetic matching of donors and recipients and the preparation and growth of cells ex vivo. Current stem cell research is focused on expansion ex vivo and the generation of HSCs from human ES cells or iPS cells. However, the rich history of HSC research and the development of BMT as a clinical modality serve as paradigms for the stem cell field.

Discussion:

One attendee asked if a donor-recipient match could be too close. Dr. Orkin responded that a perfect match may enhance the graft-versus-leukemia effect, although this may not be an issue for non-malignant disease.

Another participant inquired about the manufacture of cells that are produced ex vivo. Dr. Orkin noted that techniques to harvest peripheral blood or marrow are currently available, but limitations exist with regard to the degree of expansion. However, evidence suggests that growth factors may be manipulated to assist this process. He noted also that the events that occur in the marrow niche are not well understood and that many scientific and engineering challenges remain to be addressed.

Another attendee asked if parathyroid hormone, which has been hypothesized as a stimulus for osteoblasts, has been investigated with regard to HSCs. Dr. Orkin noted that parathyroid hormone research suggests that the factor works only modestly in the mouse. However, he observed that parathyroid hormone and other potential factors could be investigated safely in trials with careful monitoring.

An attendee asked about positive results reported from the use of HSCs in diabetes. Dr. Orkin expressed doubt about these results, stating that it is highly unlikely that HSCs can be traced to a final fate of pancreatic beta-cells. These experiments most likely used a heterogeneous mixture of cells that included cytokines and possibly other stem cell types that could contribute to the reported results.