These intensive regimens ablate the patient’s marrow and if not salvaged with HSC infusion would result in aplasia or severely prolonged cytopenias. High-intensity preparative regimens include chemotherapeutic agents such as busulfan, etoposide, melphalan, thiotepa, cytarabine, carmustine, and are often combined with irradiation, either in the form of TBI, limited field irradiation, or targeted radioimmunotherapy.

Broadly, these regimens are designed to be less toxic than high-intensity regimens, but they represent a spectrum of intensity. The most minimally toxic regimens are designed to provide only sufficient immunosuppression to facilitate engraftment of donor cells. Mixed chimerism between donor and host cells frequently develops for a period following transplantation. For nonmalignant disorders, mixed chimerism may be sufficient to correct the underlying condition, but to treat malignancies, complete donor chimerism is generally desired as donor cells that are not tolerant of host antigens are required to mediate destruction of residual disease.

Two gray (Gy) TBI ± 90 mg/m² of fludarabine, pioneered in Seattle, is the paradigmatic low-intensity regimen for patients undergoing HSCT for malignant diseases [36]. However, for patients undergoing transplantation for nonmalignant conditions, or those who have not received significant prior therapy for their malignancies, additional immunosuppressive therapy may be required to secure engraftment and avoid graft rejection. Modest increases in radiation doses, or added T-cell-specific therapies such as antithymocyte globulin (ATG) or alemtuzumab (anti-CD52 antibody), may be incorporated into these regimens [37, 38].

Additionally, because relapse remains a significant issue following reduced-intensity conditioning (RIC) transplants, several regimens have been developed that are designed to be less toxic than traditional high-intensity regimens, but potent enough in their antimalignant capacity to reduce disease burden sufficiently to allow GVT effects to mediate host cell elimination. Fludarabine combined with modest doses of busulfan, cyclophosphamide, melphalan, or TBI is typically used in those regimens [39–42]. Treosulfan, a busulfan analogue with less liver toxicity and possibly greater antileukemic potential, is currently under investigation as a novel agent in RIC regimens, and initial data suggest lower toxicity [43, 44]. Targeted radioimmunotherapy has also emerged as a strategy to intensify radiation therapy with minimal toxicity. Radiolabeled antibodies targeting receptors expressed on hematopoietic elements allow focal intense radiation exposure to disease with minimal exposure to other organs. Radiolabeled CD20 antibodies have been well established in the treatment of lymphoma, and radiolabeled CD45 antibody shows promising early data in the treatment of myeloid leukemia and myelodysplastic syndrome (MDS) [43, 45]. New strategies to pretarget radiolabeled antibodies directly to malignancy may lead to even greater efficacy of this strategy [46].

As described earlier, autologous transplantation is indicated for several diseases. While autologous HSC in principle should be available for every patient (except possibly heavily pretreated patients or patients with marrow failure states who are unable to mobilize sufficient stem cells for subsequent transplantation), autologous transplantation conveys no GVT effect and there is always concern that some malignant (diseased) cells are returned to the patient with the transplant inoculum. Until more effective gene transfer strategies have been developed, autologous HSC treatments are also not widely available for genetically determined disorders. Given its inherent risks, allogeneic transplantation is generally reserved for diseases not amenable to treatment with auto­logous transplantation. Allogeneic transplantation is virtually always performed with curative intent.

HLA matching is the primary consideration when selecting an allogeneic donor [47]. The Class I molecules HLA-A, B, C and the Class II molecules HLA-DR ± DQ are considered in screening potential donors. In general, an HLA-matched sibling donor is considered the first choice. Because individuals inherit one set of HLA genes from each parent, any sibling has a 25% chance of being matched to the patient, and ∼30% of persons in the United States will have an HLA-matched sibling [48]. In addition, phenotypically-matched related donors are identified for about 1% of patients, and somewhat less than 1% of patients will have an identical (syngeneic) twin donor [49]. Syngeneic donors, however, do not provide a GVT effect.

For patients lacking a matched sibling, an alternative donor must be identified. The typical second choice is an unrelated donor matched at high-resolution (DNA level) typing for at least 8–10 HLA antigens. Due to the efforts of organizations such as the Anthony Nolan Appeal in the UK, the National Marrow Donor Program in the United States, and other organizations worldwide, more than 12 million volunteer donors have been added to registries as potential HSCT donors [50]. Computer searches of those databases very quickly generate an idea as to the probability of finding an unrelated donor. However, the logistics of confirming eligibility and coordinating timing of an unrelated donor frequently take weeks to months, and this delay may be too long to control disease in patients with aggressive malignancies [51]. The probability of finding a “complete” HLA match is currently about 50–60% for Caucasian patients. The probabilities are lower for patients of different ethnic backgrounds [48].

For patients lacking a matched unrelated donor, alternative donor options include mismatched unrelated donors, UCB, and haploidentical donors. UCB, due to its unique immunologic qualities discussed later, has less stringent matching requirements than unrelated donors, and an acceptable cord blood match is therefore more easily attainable. A significant body of rapidly increasing data supports the efficacy of UCB transplantation for patients lacking ­suitably-matched related or unrelated donors. Haploidentical donors—related donors who share only one haplotype (the maternal or paternal #6 chromosome) with the patient—are appealing, because nearly all patients will have a donor who is immediately available (at low search cost). However, the high degree of HLA mismatching inherent to haploidentical transplantation creates significant immunologic challenges, including potential graft rejection, GVHD, and poor immune reconstitution, and haploidentical transplantation at present remains investigational.

Bone marrow was the original source of HSC evaluated for transplantation. Demonstration of the ability to mobilize HSC into the peripheral blood either with high-dose granulocyte colony-stimulating factor (G-CSF) or chemotherapy ± glucocorticoids followed by G-CSF raised the possibility of peripheral blood stem cells (PBSC) as a source for harvesting cells [52–54]. PBSC show a lower retention of Rhodamine 123 (which is typical for noncycling cells), and may have a lower expression of CD38 as compared to bone marrow cells. Unless T cell depleted, PBSC contain significantly greater numbers of T cells than marrow.

An initial concern with PBSC was that they may not be capable of long-term hematopoietic reconstitution. These concerns have been definitively rebuked based on results from autologous transplant patients who have been followed now for more than 15 years and show normal hematopoiesis, as well as observations with allogeneic PBSC transplants, which have resulted in sustained donor cell engraftment [55, 56]. Given the ease of collecting PBSC relative to marrow, the number of PBSC transplants as compared to marrow transplants has increased significantly, and PBSC are currently used for virtually all autologous stem cell transplants in patients with hematologic malignancies. Several issues regarding the relative benefits of PBSC vs. bone marrow, however, are worthy of discussion.

Numerous randomized trials have compared outcomes using either bone marrow or PBSC for patients undergoing allogeneic HSCT following high-intensity conditioning for hematologic malignancies. A meta-analysis reviewing nine trials and 1,288 patients explored the relative benefits of each stem cell source [57]. Engraftment (neutrophils ≥0.5  ×  10⁹/l) was significantly more rapid following PBSC (median 14 vs. 22 days p  <  0.00001). (Of note, G-CSF mobilization of bone marrow prior to bone marrow collection may decrease time to engraftment of cells harvested from the bone marrow [58].) There was no difference between overall rates of acute GVHD between donor sources, but there was a modestly higher incidence of grades III–IV acute GVHD following PBSC transplantations (p  =  0.03). Patients treated with PBSC had a higher incidence of extensive chronic GVHD (p  <  0.00001). Rates of relapse were lower among PBSC patients, and PBSC were associated with improved disease-free survival, particularly among patients with more advanced disease at the time of transplantation. There was no statistically significant difference in overall survival between patients, but the subgroup of patients with advanced disease had improved survival following the use of PBSC. In the aggregate, a recent analysis suggests that for patients with a 1 year relapse risk  >  5%, PBSC are preferable to marrow both in terms of overall and quality-adjusted life expectancy [59].

In patients who are prepared for transplantation with RIC regimens, PBSC have been used virtually exclusively as the source of HSC. This approach has been taken in part because of the greater “engraftment potential” ascribed to those cells and in part related to the larger numbers of cells that are obtainable and the higher T-cell doses [57].

However, in patients undergoing HSCT for nonmalignant disorders, concern about the potential increased risk of GVHD associated with PBSC still favors the use of marrow [57]. Patients with nonmalignant diseases are unlikely to derive any gain from the development of chronic GVHD. Similarly, for patients undergoing haploidentical transplantation, PBSC have been avoided in favor of bone marrow or CD34⁺-selected products, given the high risk of GVHD.

Cord blood has emerged as an important source of HSC. Since the first UCB transplant was successfully performed in 1988 on a 5-year-old child with Fanconi anemia using his HLA-identical sibling’s UCB [60], rapid developments in the field have led to a dramatic increase in the use of UCB. Studies demonstrate that the proliferative potential of cord blood stem cells, as determined by the generation of CD34⁺ cells in serum-free medium, is greater than that of adult marrow or PBSC [61]. Cord blood contains proportionally higher numbers of progenitor cells per volume than adult blood and possibly adult marrow [17, 62, 63]. Immunological studies show that cord blood T cells are less immunocompetent than adult cells [60, 63]. Cord blood T cells can be activated via the T-cell receptor, but have a reduced capacity to stimulate [64]. Intracytoplasmatic signaling (following T-cell receptor engagement) in cord blood T cells also differs from that in adult T cells [65]. Cord blood monocytes express lower levels of HLA-DR, CD86, and ICAM-1 than monocytes in adult blood, and produce lower levels of IL-10 and IFN-γ. IL-4 and IL-5 are basically absent [64]. These differences in the immunologic characteristics of UCB may contribute to the tolerability of greater mismatching between donor and patient. Additional advantages of cord blood include the fact that it is a frozen off-the-shelf product that is rapidly available, that there is no danger to the donor, and that the risk of transfer of viruses is minimal. Major challenges to using UCB include the low number of HSC and progenitors in a single cord unit, which can lead to delayed engraftment and graft failure—particularly in adults and large children—as well as possible delays in immune reconstitution.

In an adult, bone marrow is the primary organ of hematopoiesis and the site from which stem cells for transplantation were traditionally harvested [66]. For the purpose of marrow harvest, the donor (or patient) receives anesthesia, and under sterile conditions, multiple small volume aspirates (≈3 ml) of marrow are obtained from both posterior iliac crests [67]. Additional potential aspiration sites are the anterior iliac crests, the sternum, and, particularly in children, the tibia heads. Approximately 10–15 ml/kg donor weight is collected, usually yielding 1–4  ×  10⁸ cells/kg recipient weight (Table 57.4). The concentration of cells is generally higher in children than in adults, and in first harvests than in subsequent aspirations if done within a short time interval after the first procedure [68–70]. The marrow is passed through filters, generally in semiclosed systems, to break up cell aggregates and remove bone particles. If no ABO incompatibility exists and if the marrow is not to be subjected to any in vitro purging procedure or other manipulation, the resulting cell suspension is infused intravenously, generally via an indwelling intravenous catheter (e.g., Hickman line), and cells “home” through the peripheral blood circulation to the marrow cavity to reconstitute the hematopoietic system.

At steady state, HSC circulate at very low concentrations in peripheral blood. Following cytotoxic therapy, however, the frequency of early hematopoietic precursors in the circulation increases dramatically, though transiently, during the recovery phase [52]. The same effect is achieved if cytotoxic agents such as cyclophosphamide, etoposide, or taxol are given for the very purpose of “mobilizing” cells from the marrow into the circulation (Table 57.4) [71]. The administration of recombinant hematopoietic growth factors, such as G-CSF, granulocyte-macrophage colony-stimulating factor (GM-CSF) or c-kit ligand alone or in combination, has similar effects (although presumably via different mechanisms [72]), and the timing of the maximum stem cell yield, around 4–6 days after initiation of treatment with growth factors, is more predictable than with chemotherapy [53, 54]. Furthermore, the absence of cytotoxicity and the associated side effects allows the application of this approach to normal donors and the harvesting of mobilized normal cells from the peripheral circulation. The mechanism involved in dislodging cells from the marrow and forcing them into circulation is incompletely understood. HSC are normally anchored within the marrow environment, presumably through specific receptor interactions with other cells and extracellular matrix. In agreement with that notion, Plerixafor, an inhibitor of the stem cell homing receptor CXCR 4, has been demonstrated to improve mobilization and has recently received approval in the United States [73]. Ongoing studies are investigating optimization of patient selection for the addition of plerixafor during mobilization.

PBSC (and cells at various stages of differentiation) are harvested from peripheral blood by leukapheresis. The best currently available marker of the cells required for a successful transplant (stem cells; long-term repopulating cells) is the surface glycoprotein, CD34. For autologous procedures the goal is to harvest approximately 3–5  ×  10⁶ CD34⁺ cells/kg recipient weight [74]. Higher cell numbers are targeted if the cells are to be processed in vitro or if cells for a “back-up” infusion or second transplant are desired. The yield of CD34⁺ cells per apheresis (usually 10–16 l over 2–4 h) for autologous transplants varies considerably, largely dependent on the intensity of prior therapy given to the patients. The most reliable indicator of the success of apheresis is the concentration of CD34⁺ cells in peripheral blood [75]. In many patients, one single harvest will suffice; in others, several sessions are required. However, after 2 (or 3) days of apheresis, generally the incremental cell yield of CD34⁺ cells from additional aphereses is small. Large-volume apheresis may be useful in some patients. Usually autologous cells need to be cryopreserved until completion of the patient’s conditioning regimen.

Mobilization of PBSC by G-CSF in healthy allogeneic donors results in more consistent cell yields than in pretreated patients. By and large, 2–20  ×  10⁶ CD34⁺ cells/kg are obtained with a single harvest [55, 56]. As the harvest can be timed such that it coincides with completion of the transplant conditioning regimen in the patient, allogeneic PBSC are usually transplanted fresh. Cryopreservation is possible, however, and has been used, for example, to accommodate donor preferences.

Cord blood is collected either by venipuncture in the third stage of labor while the placenta remains in utero or by gravity drainage of the ligated umbilical vein once the placenta has been delivered [76]. Collected cord blood must be shipped to a bank for further processing and cryopreservation. At the bank, units are screened for size, and appropriate units are processed. Depending on individual bank criteria, minimum size thresholds for further processing are typically approximately 0.9  ×  10⁹ total nucleated cells (TNC) (the minimum size unit currently funded by the Health Resources and Services Administration (HRSA) in the United States). Processing strategies include directly freezing the unit in 10% DMSO vs. various approaches involving red blood cell and plasma depletion. Units may be manually depleted using hydroxy ethyl starch or PrepaCyte-CB or by automated systems including Sepax (Biosafe, Eysins, Switzerland) and AutoExpress (AXP) (Thermogenesis). Though direct freezing results in the lowest loss of total nucleated cells, the large product volumes occupy significant storage space, can cause hemodynamic complications upon infusion in small patients, and result in potentially larger DMSO exposure of patients. RBC and plasma-depleted products reduce these concerns, but some cell loss is inevitable during processing [77].

A variety of strategies is utilized by transplant centers to prepare cryopreserved units for infusion. The thaw and wash technique developed by Rubinstein et al. is most common [78]. More recently, a thaw and dilute procedure has been advocated as a strategy to increase efficiency and potentially improve infused cell doses [79]. Some centers directly thaw and infuse cord blood at the bedside.

As the number of UCB transplants has increased, banking of cord blood has increased as well. There are currently more than 100 active public cord blood banks worldwide with an inventory of greater than 400,000 units [80]. In order to donate to a public cord blood bank, an individual must deliver at a hospital that is affiliated with a public bank. A significant private cord blood banking industry has also developed. For a collection and storage fee, parents may bank their child’s cord blood for future personal use. However, given current indications for cord blood transplantation, uses of privately banked cord blood are extremely limited. Additionally, private banks are not currently regulated, and the quality of units collected is not guaranteed. Though the future potential uses of privately banked cord blood, particularly for regenerative medicine, cannot be predicted, expert opinion does not ­currently encourage private cord blood banking [81].

Fetal liver, if obtained at the optimum time of gestation, provides a rich source of HSC [82]. The concentration of CFU-GM, an indicator for hematopoietic potential, increases with gestational age, most markedly after week 10. However, beginning at about week 14 there is also an increase in cells expressing T-cell markers. Thus, an optimum window for the use of fetal liver cells exists at 10–14 weeks of gestation. The feasibility of fetal liver transplants has been shown in adult primate and canine models as well as in in utero transplants in sheep [21, 83]. While fetal liver cells have been used successfully as intrauterine (fetal) therapy for children with immune defects [82, 84], for reasons of availability, limited numbers, technical problems, and ethical concerns, fetal liver cells have fallen out of favor and are no longer being used clinically.

General complications associated with allogeneic transplantation include graft failure, delayed engraftment, regimen-related toxicity, GVHD, infection, and relapse. Clinical data regarding each of these issues as well as general outcomes with respect to donor source are discussed later and summarized in Table 57.5.