CHAPTER 26: Hematopoietic Stem Cell Transplantation

The era of hematopoietic stem cell transplantation (HSCT) began in the 1950s, when pioneering studies by E. Donnall Thomas, James Till, and Ernest McCulloch demonstrated the ability of unfractionated bone marrow cells to "rescue" animals from hematopoietic failure induced by otherwise lethal doses of radiation. Such studies had profound implications and consequences for basic research, because they proved the existence of multipotent hematopoietic stem cells and provided a powerful new experimental tool for studying the immune system.

From very early days, it was evident that HSCT also had enormous therapeutic potential but was prone to cause serious and all too frequently fatal complications. HSCT remains as close to a high-wire act as exists in medicine, one in which patients receive potentially lethal doses of chemotherapy and/or radiation. However, as we will discuss, advances in stem cell biology, immunology, and pharmacology have allowed the development of more effective, less toxic HSCT strategies. As a result, HSCT is now being used to treat an increasing number of disorders and a broader spectrum of patients than ever before. This chapter serves as an overview of some of the salient features of this fascinating, rapidly evolving area of hematology.

HSCT has proven to be effective in the following clinical settings:

• Correction of genetic and acquired defects in hematopoietic stem cells (HSC). Until the dream of gene therapy for germline defects is realized, HSCT is often the only hope for those suffering from severe genetic disorders that affect the function of the HSC or its progeny. Replacing the defective HSCs of the patient with HSCs obtained from a normal donor can cure such diseases. HSCT has been used to treat many genetic diseases, including those affecting lymphocytes (eg, severe combined immunodeficiency, X-linked agammaglobulinemia), red cells (eg, severe thalassemia, sickle cell disease), and monocytes/macrophages (eg, Gaucher disease).

• Support for high-dose cancer treatment. The dose-limiting toxicity of radiotherapy and many conventional chemotherapy agents is bone marrow failure due to ablation of HSCs. HSCT overcomes this limitation by providing the patient with healthy HSCs, which can completely reconstitute hematopoiesis and the immune system over a period of weeks to months.

• Generation of graft-versus-tumor effect. When a recipient receives HSCs from another individual who is not an identical twin, the transplanted HSCs give rise to a "new" immune system that may recognize the patient's tumor as non-self and mount an immune response against it. It is now clear that this graft-versus-tumor effect is a very important benefit of HSCT in certain types of cancer, particularly myeloid leukemias.

• Generation of organ graft tolerance. HSCT has sometimes been used to alleviate solid-organ rejection. These types of procedures typically involve the transplantation of HSCs along with a solid organ, such as a kidney or a single lobe of the liver.

• Salvage from genotoxic chemicals or radiation. To date, this indication has applied mainly to rare individuals exposed to nuclear power plant accidents (eg, Chernobyl). It is sobering to consider, however, that the United States government is currently expending large amounts of time and money planning for the consequences of a terrorist attack involving a so-called dirty bomb containing radioisotopes. Such an attack has the potential to cause severe bone marrow damage to hundreds or thousands of people, many of whom may need to undergo HSCT if they are to survive.

There are three types of HSCT. Autologous transplants are performed with the patient's own HSCs. These have the advantage of avoiding immunologic complications but fail to generate the graft-versus-tumor effect and thus produce higher rates of relapse in those being treated for myeloid leukemias. Syngeneic transplants use HSCs from identical twins and have the same advantages and disadvantages as autologous transplants. Allogeneic transplants are done with stem cells obtained from a genetically distinct donor.

The most important determinants of the immunologic consequences of allogeneic HSCT are the class 1 and class 2 major histocompatibility complex (MHC) antigens, also known as human leukocyte antigens (HLA). You will recall that class 1 and class 2 MHC molecules present short peptides derived from intracellular and extracellular antigens, respectively. Although many genes in the MHC locus influence immunity, experience has shown that, for patients undergoing allogeneic HSCT, the most important are the HLA-A, HLA-B, and HLA-C class 1 antigens and the HLA-DR class 2 antigens. The roles of the class 2 antigens HLA-DQ and HLA-DP are less clear. The genes encoding these HLA antigens are closely linked on chromosome 6 and thus are inherited en bloc in what is referred to as a haplotype. An allogeneic HSCT in which the haplotypes of the donor and the recipient are identical is called a matched transplant. HSCT can be performed with one or two mismatched antigens, but the risks of the procedure are substantially higher than with an HLA-identical donor.

Another important variable in allogeneic transplantation is whether the HSCs are from a related family member or an unrelated donor. In addition to the MHC loci, poorly characterized "minor" histocompatibility complex loci have lesser but still significant immunologic effects. Allogeneic transplants from related donors are generally performed between siblings, who have a greater degree of identity at these minor loci than do unrelated donors. As shown in Figure 26-1, there is a 25% chance that any sibling will be fully matched at the MHC loci. In contrast, it is rare for a patient to have a full HLA match with a parent or a child. Of note, one or more of the minor histocompatibility complex antigens is located on the Y chromosome. As a result, the transplantation of HSCs from female donors into male recipients produces more graft-versus-tumor effect but also is more prone to result in significant graft-versus-host disease (GVHD) (described later).

HSCs from several different sources are used clinically. Most allogeneic transplants from related donors are performed with whole, unfractionated bone marrow or with HSCs harvested from peripheral blood. The latter are typically obtained a few days after administration of a hematopoietic cytokine such as G-CSF (granulocyte colony stimulating factor (Chapter 2)). Allogeneic transplants from unrelated donors may be performed with bone marrow, HSCs harvested from the peripheral blood, or blood obtained from the placenta at birth (so-called umbilical cord blood). Cord blood has the advantage of being available as a byproduct of every child's birth. There is a national initiative to establish public cord blood banks, with an eye toward future use in unrelated allogeneic HSCTs. The principle disadvantage of cord blood is that the number of HSCs in any sample is relatively small, a limitation that can be overcome by pooling of cord blood from two or more donors. Autologous transplantation is generally performed on cancer patients with peripheral blood HSCs obtained at 7-10 days after a dose of chemotherapy with GCSF stimulation. Recovery from the chemotherapy stimulates HSC proliferation and mobilization.

All forms of HSCT require some conditioning of the recipient with myelotoxic therapies. Conditioning is required to kill or mobilize sufficient numbers of host HSCs to open up the stem cell niche in the marrow (Chapter 2), creating space for the HSCs in the graft. Conditioning also may serve to eradicate a host immune system that has run amok (eg, aplastic anemia, Chapter 4) or to destroy tumor cells that may be resistant to standard doses of chemotherapy and radiation (eg, the transformed HSCs that cause chronic myelogenous leukemia, Chapter 20). The precise conditioning regimen used is tailored to the disorder being treated. For nonneoplastic disorders, common regimens consist of chemotherapy with or without anti-thymocyte globulin, an antibody that recognizes and leads to the killing of T cells. For neoplastic disorders, conditioning is usually done with high-dose chemotherapy regimens, sometimes in combination with total body irradiation.

Several days after conditioning, donor HSCs are transfused into the peripheral blood of the recipient. If all goes well, the marrow and the peripheral blood are reconstituted by donor-derived hematopoietic cells and their progeny (Fig. 26-2). The earliest formed elements that appear in the recipient are derived from late myeloid progenitors, followed by a second cohort of cells derived from early myeloid progenitors. However, sustained reconstitution requires engraftment of HSCs, the only marrow cells that possess the key properties of multipotency and self-renewing capacity.

On average, it takes about 3 weeks for the marrow to engraft and begin to produce mature myeloid cells (platelets, red cells, and neutrophils). However, the pace and timing of reconstitution in the recipient varies depending on several factors including the source of the HSCs. In general, unfractionated bone marrow has greater numbers of early progenitors and HSCs than does cord blood, and, as a result, patients receiving unfractionated bone marrow tend to engraft more rapidly. During the critical pre-engraftment period, these patients are profoundly immunosuppressed and are completely dependent on transfusion of red cells to maintain oxygen delivery to tissues and transfusion of platelets to prevent bleeding. Because of agranulocytosis there is a very high risk of severe and potentially fatal bacterial and fungal infections during the first 2 or 3 weeks, prior to the production of adequate numbers of neutrophils from the allograft. Measures aimed at reducing the morbidity and mortality associated with infections in the immediate posttransplant period include housing of patients in special positive-pressure rooms supplied with filtered air to limit exposure to pathogens and aggressive treatment with broad-spectrum antibiotics at the earliest sign of infection.

Other complications in the early posttransplant period are directly related to cellular injuries induced by conditioning. Nearly all patients have mucositis, particularly of the soft palate and esophagus, causing swallowing to be painful and difficult. These mucosal lesions are potential sites of entry of pathogenic bacteria. An even more serious complication is veno-occlusive disease, which appears to be caused by damage to endothelial cells lining the venous sinusoids of the liver. Although the precise pathogenesis is unknown, the injury induces fibrin deposition and obstruction of hepatic sinusoids (leading some to suggest that this entity would be better entitled sinusoidal obstruction syndrome). The blockage of blood flow causes liver engorgement and tender hepatomegaly and can lead to hypoxic centrilobular hepatic necrosis and liver failure in severe cases.

Graft failure and graft rejection are other serious early complications of HSCT. The incidence of graft rejection is proportional to the degree of immunologic dissimilarity between the donor and the recipient and thus is relatively high in unrelated allogeneic HSCTs. Graft rejection also is increased by prior immunization of the recipient by transfusions of red cells or platelets. Graft failure may stem from an insufficiency of HSCs in the stem cell preparation, damage to the marrow microenvironment caused by the conditioning regimen, or infection. Patients who fail to engraft must undergo another transplant, typically with HSCs from an unrelated donor, if they are to recover. Understandably, second transplants are associated with a high rate of graft failure and morbidity and mortality related to a second round of conditioning.

Once neutrophils appear, the recipient's risk of bacterial and fungal infections declines dramatically. However, as described in the following section, in allogeneic transplant recipients reconstitution of the lymphoid arm of the immune system is a much more complicated process, one involving a delicate balance between immunosuppression and GVHD, an assault on the host mediated by lymphocytes derived from the graft. Under the best of circumstances, recovery of adaptive immunity in allogeneic HSCT recipients requires many months, and, as a result, these patients are at high risk for viral infections and Epstein-Barr virus (EBV)–driven B-cell tumors for an extended period of time after transplantation.

HSC preparations used in allogeneic transplants contain varying numbers of mature T cells. These cells as well as newly produced lymphocytes derived from donor HSCs have the capacity to "see" host cells as foreign. Curiously, the tissues most prone to GVHD are the skin, the gastrointestinal mucosa, and the bile ducts. The common feature of these tissues is that they line parts of the body that are constantly exposed to bacteria and other pathogens.

The pathogenesis of GVHD is still incompletely understood, but a plausible scenario involves the following series of events (Fig. 26-3). The initiating trigger is believed to be epithelial cell injury caused by a number of factors, including the conditioning regimen, proinflammatory stimuli such as endotoxin, and cytomegalovirus infection. Injured epithelial cells secrete a number of soluble proinflammatory factors that activate resident dendritic cells, tissue macrophages, and endothelial cells, which also may be activated directly by molecules derived from bacteria such as endotoxin. Once a proinflammatory state is established, graft derived T cells specific for recipient histocompatibility antigens are recruited. CD8-positive cytotoxic T cells selectively target the epithelial cells, inducing their death by apoptosis and in the process amplifying the initial stimulus—epithelial cell injury.

Early in the posttransplant period (generally within the first 100, days), GVHD is caused by mature T cells present in the preparation of donor cells. This form of the disease (acute GVHD) can appear quickly and have dramatic clinical manifestations. The incidence of acute GVHD can be reduced by treating the graft with anti-T-cell antibodies prior to infusion into the recipient, but this proves to be a trade-off, because T-cell depletion increases the risk of recurrence of the original tumor as well as the development of new EBV-positive B-cell tumors. Beyond 100 days, GVHD appears to be mainly mediated by both new T cells and B cells that are produced from HSCs or other early progenitors in the graft. This phase of the disease tends to follow a more indolent course (hence the name chronic GVHD) but nevertheless can have devastating consequences.

Acute GVHD presents with rash, diarrhea, and elevated liver function tests. The rash is often diffuse and can be quite striking (Fig. 26-4). The microscopic features are similar in each of the involved tissues, consisting of a sparse infiltrate of lymphocytes associated with apoptosis of epithelial cells (Fig. 26-5). The extent of the epithelial injury varies widely, from a few scattered necrotic cells in mild disease to denuding of surfaces in the most severe cases. Severe acute GVHD often leads to sepsis and a fatal outcome. Chronic GVHD is associated with more subtle forms of tissue injury that lead to fibrosis, atrophy, and loss of function (Fig. 26-6). The skin, gastrointestinal tract, and liver are prominently involved, but other tissues, particularly the lung, also may be affected. Atrophy and fibrosis of the skin may lead to scleroderma-like changes and loss of hair. Fibrosis of mucosal surfaces of the gut may prevent absorption of nutrients and render the gut susceptible to mechanical injury. Fibrosis also may seriously impact pulmonary and hepatic function. In those patients who received total body irradiation as part of the conditioning regimen, fibrosis related to late-stage radiation toxicity also may contribute to organ failure.

The clinical challenge is to control GVHD while allowing sufficient immune activity to generate a graft-versus-tumor effect (in those being treated for cancer) and foster immune reconstitution. This balance is achieved by careful titration of immunosuppressive agents that inhibit T-cell activation (cyclosporine, tacrolimus, sirolimus), interfere with T-cell proliferation (methotrexate, mycophenolate), or kill T cells (anti-thymocyte globulin, prednisone).

Many clinical studies have now established the therapeutic importance in allogeneic transplants of the so-called graft-versus-tumor effect. This activity stems from recognition and killing of host tumor cells by immune cells in the graft. Because the tumors in which this beneficial effect is most clearly demonstrated are chronic myelogenous leukemia, acute myeloid leukemia, and acute lymphoblastic leukemia, it is often referred to as graft-versus-leukemia effect, or GVL.

GVL and GVHD go hand in hand, because the incidence of leukemia relapse is directly related to the immunologic similarity of the donor and recipient. Thus, relapse in leukemia occurs most frequently in syngeneic transplant recipients and least frequently when donors suffer from some degree of acute and/or chronic GVHD (Fig. 26-7). In diseases like chronic myelogenous leukemia that are particularly sensitive to GVL, the effect may be augmented by infusion of donor leukocytes (as described in Chapter 20). A major question for the field is whether GVL can be teased apart from GVHD. This may be possible if cancer-specific minor histocompatibility complex antigens can generate an effective GVL response.

Historically, the use of allogeneic HSCT has been curtailed by the severe complications caused by myeloablative conditioning regimens and the inability to identify suitable donors for many patients. The latter limitation has been solved in part by the creation of national and international bone marrow registries containing lists of haplotyped donors, a resource that is now being supplemented by cord blood banks. However, even with the most fastidious attention to clinical care, conventional conditioning regimens have unacceptable toxicities for older patients and those with underlying medical conditions.

Based on the insight that much of the treatment benefit of allogeneic HSCT in those with leukemia comes from the GVL effect, recent work has focused on the development of nonmyeloablative conditioning regimens that permit at least partial engraftment of donor cells. In principle, such regimens could greatly expand the number of patients who are eligible for HSCT, including those who are older, who require second transplants, or who have organ dysfunction that precludes conventional conditioning regimens. Recent studies have shown that nonmyeloablative conditioning with chemotherapy is indeed much less toxic, often requiring minimal hospitalization, and can be tolerated by children and adults of all ages as well as those with chronic diseases. Radiation is not required, which may lessen the risk of some late complications of HSCT.

An example of the promise of nonmyeloablative HSCT can be seen in the treatment of an adult woman with sickle cell disease (Fig. 26-8). This less toxic strategy is particularly amenable to treatment of sickle cell patients because, unlike those with leukemia, they can tolerate a minor degree of hematopoietic chimerism. This patient received an allogeneic transplant from a brother with sickle cell trait 2 weeks after an exchange transfusion to lower the burden of SS red cells. The conditioning regimen consisted of nonmyeloablative doses of chemotherapy (fludarabine and busulfan). A second exchange transfusion was done after transplantation to tide the patient over until engraftment was complete. As of 6 months post-HSCT, the levels of Hb A and Hb S in the peripheral blood were about 60% Hb A and about 40% Hb S, the levels expected for an individual with sickle cell trait. With establishment of the graft, the peripheral smear normalized, and hemolysis ceased, lowering the erythropoietic drive and decreasing the marrow cellularity to normal. With appropriate prophylaxis, GVHD was held in abeyance.

With additional experience and a more comprehensive understanding of the immunologic facets of HSCT, it should be possible to use nonmyeloablative HSCT in patients with a wide spectrum of neoplastic and nonneoplastic hematologic conditions. Such studies are currently underway in older adult patients with acute myeloid leukemia and other hematologic conditions that have a poor prognosis with conventional therapies.